Designing the Ultimate HF Front End
from
Peter Gant, HA5RXZ
on
June 22, 2009
View comments about this article!
Designing the Ultimate HF Front End
Ham radio caters for a wide variety of interests. Some seek DXCC
certificates, others want to paper the walls of the shack with QSL cards
whilst others are dedicated to contests.
My particular interest is HF
receiver front ends because here in Europe there is a greater density of
broadcasting stations and other sources of QRM, this means that separating
the weak signals from the noise can sometimes be a challenge. The various
manufacturers each have their own solution, for example here are the systems
used by the Icom IC-7800, Yaesu FT-dx9000 and the now defunct Hilberling
PT-8000:
Icom IC-7800 Yaesu FT-dx9000 Hilberling PT-8000
Bandpass Filter Preselector Hybrid Amplifier
Preselector Bandpass Filter Preselector
Preamplifier Preamplifier First Mixer
First Mixer First Mixer First IF Amp
Low Pass Filter Roofing Filter Roofing Filter
Roofing Filter First IF Amp Second Mixer
First IF Amp Second Mixer
Second Mixer
All table data has been taken from the manufacturers leaflets but with a
little bit of renaming, for example Yaesu call their preselector a VRF. Look
at the table but in particular look at the ORDER of the various elements, all
three manufacturers adopt different methods for coping with the modern HF
spectrum. All of the engineers who designed these rigs have more knowledge of
RF engineering than me yet they cannot all be right.
My own favorite (although I do not own one) is the architecture of the Icom
IC-7800 and here's why: At the antenna jack there is the full HF spectrum,
this is then reduced to a half an octave using the bandpass filters then an
even narrower bandwidth using the preselector. The preamplifier and first
mixer therefore have to cope with a much narrower spectrum and this reduces
overload. After the first mixer there are the roofing filters THEN the first
IF amplifier which only has to cope with a signal bandwidth which is KHz
wide. This architecture is copied in the IC-7700, if you get something right
then why reinvent the wheel?
So, I think that Icom got it right, but I am of course willing to debate the
issue. What order would you assemble the perfect HF front end and why? Feel
free to contribute front end details of other rigs and to make the lists
similar make your list as far as the second mixer (we can save the debate of
what constitutes a front end until later). Who knows, if we keep the debate
civilized we might learn something.
This article has expired. No more comments may be added.
|
Designing the Ultimate HF Front End
|
|
|
by ZENKI on June 22, 2009
|
Mail this to a friend!
|
Neither of the 3 radios that you mention has the ultimate front end.
The work by PA3AKE has produced the ultimate front end.
http://www.xs4all.nl/~martein/pa3ake/hmode/index.html
If you are going to start dreaming about the ultimate front end, you would be wise to also start dreaming about the ultimate transmitter performance.
The major impediment that prevents us from extracting the maximum performance from our ultimate receivers is the poor state of our transmitter performance specifications.
Until our transmitters IMD, phase noise and keyclicks approach the IMD dynamic range numbers of our receivers at close signal spacing, it would all be a wasted effort.
This is exactly the situation that we have today with radios like the K3 and Orion, great receivers however their below average transmitter performance specifications leaves a lot to be desired. Rob Sherwood has given a presentation on this point.
We need transmitters like the ADAT which have technically perfect and clean transmitters.
http://www.adat.ch/index_e.html
However expecting perfect transmitters from ham manufacturers is just a dream. The only way we will get perfect transmitters is when laws are passed that mandate these specifications under ITU law.
While its nice dreaming about the above radios, all that you are really doing is continuing the receiver performance numbers hyperbole.
BTW the AOR 7070 is due out soon, and according to my spies it is going to blow everything off the planet. A darn shame that it has no matching transmitter with fantastic transmitter specs.
Dream on!
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by G3RZP on June 22, 2009
|
Mail this to a friend!
|
My measurements (published in QEX and NCJ) taken at sunspot max and sunspot min, showed that 100dB of instantaneous dynamic range is about all that is needed - in a place where the noise level is below ITU median quiet rural levels, and 7MHz BC signals can be big. The insidious part of front end performance is phase noise, which may well negate a lot of the attempts at a high spurious free dynamic range front end. Spurs need to be low, too, which can be a problem with DDS. Narrow tuned filters help that, as well as helping to reduce the number of signals giving IMD problems.
Maybe those old designers of tube radios who used three gang tuned circuits actually did know what they were doing!
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by K0BG on June 22, 2009
|
Mail this to a friend!
|
A couple of comments.
The Icom IC-7600 is designed the same way, and once the ARRL, Sherwood Engineering, et. al. test it, I'm sure it will prove to be very close to the 7700 and 7800 in receiver performance (maybe for naught).
I could not agree more with the statement about transmitter IMD levels. The biggest problem is not the transmitters themselves so much, as it is the way they're used. Far too many amateurs, use way too much microphone gain and compression levels, peaky microphones, and other devices to the point that all of the transmitting stages are vastly overloaded. As a result, it is not uncommon for the IMD level to be nearly as high as their fundamental bandpass ones are.
Adding even more injury, they drive their amplifiers so hard, what little overhead is designed into them, is long since flattened out.
The EU and the FCC have rules governing the manufacturers with respect to IMD, yet little if any effort is made to govern on-air activities.
And to paraphrase, until we tackle the self-generated IMD problem, the receivers we have now are more than adequate.
Alan, KØBG
www.k0bg.com
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by WX7G on June 22, 2009
|
Mail this to a friend!
|
While the front end is ever so important the first IF can also make or break receiver performance.
Most modern amateur transceivers (receivers) up convert to the first IF. Because of the high IF frequency (70 MHz) the 'roofing filter' cannot be made with a very narrow bandwidth.
A handful of amateur transceivers, such as the Elecraft K3 down convert to a low IF. This allows a narrow roofing filter. The K3 can be ordered with a 500 Hz roofing filter to provide incredible close-in signal performance. Where this matters most is during CW contesting, especially on 160 meters.
I have not found out if the IC-7800 uses up conversion or down conversion for the first IF. Does anyone know the answer?
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by G3RZP on June 22, 2009
|
Mail this to a friend!
|
|
The EU limitations on transmitter IMD have no requirement closer than 5kHz from the centre of a transmission at HF. Further out, for an HF tx, they are 43dB + 10logP, not exceeding 50dB.
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by W9OY on June 22, 2009
|
Mail this to a friend!
|
On the SDR front we are now developing radios that have wide band image rejection and brick wall filtering on both transmit and receive.
SDR radios are phasing rigs and as such can attain very clean transmit characteristics and receive characteristics. The new programming does its magic in milliseconds as you tune across the bands giving you 100dB image rejection and brick wall transmit and receive filtering. In other words the radio configures itself to best performance on the fly for what ever frequency you choose to operate.
As the quality of DDS and A/D improves these radios will become the only rational choice on a cost vs performance basis alone
73 W9OY
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by AB7E on June 22, 2009
|
Mail this to a friend!
|
I agree with K0BG. While many rigs may have weak transmitter performance, the much greater problem lies with how people use them.
It's pretty obvious to the user when a receiver isn't performing well, but in my opinion, most hams are fairly clueless whether they are transmitting a clean signal or not ... and most hams are reluctant to provide honest feedback that might help the situation. I've heard hams with horrible SSB signals ask how they sound, only for them to be told by the guy on the other end that they sound fine. I've almost never heard someone with 2KHz-wide key clicks receive anything other than a 559 or 579 or 599 report.
My experience is that most hams really appreciate honest feedback. During last January's NAQP SSB a fellow in the midwest with a 59+20db signal was calling CQ but getting rather few callers because his signal was so overdriven it was difficult to understand him. Two hams on frequency commented amongst themselves how badly he sounded but they didn't tell him, so I gave the guy a call and suggested he back down the mic gain and compression. He cranked it down a bit, then some more, and then some more, until finally he sounded pretty good. His signal was literally 15 db weaker (59+5db) but he was far more understandable and immediately started getting replies.
Here's another example. During the recent CQ WPX CW contest, several stations had horrible key clicks, and one large M/M station was so bad on 20m I was convinced he was hot switching the amplifier. So I sent them a carefully-worded email after the contest to describe what I heard. I received a nice reply a few days later saying they had been using two different rigs during the contest, and after checking them out they discovered that one of them (an unmodified Yaesu) was generating bad clicks. They thanked me for the report and said they had already ordered a mod kit from W8JI.
No ... I don't make a point of playing signal quality policeman, but I think if we all gave honestly constructive feedback to others we meet on the air we'd have less trash to complain about.
73,
Dave AB7E
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by KB2FCV on June 22, 2009
|
Mail this to a friend!
|
|
Perhaps a LID filter. So that when there are those 1 or 2 guys that don't understand what 'up' means on the DX station and just keep calling and calling and calling and calling completely wiping out the DX station. That would be awesome.
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by K7LRB on June 22, 2009
|
Mail this to a friend!
|
Dave AB7E),
Very nice post and very true. Probably the biggest problem with being "honest" and forthcoming with people is that, for the most part, people are very "covetuous" of their own "stuff" (ideas, equipment, signal quality, etc.). Therefore, rather than accepting "constructive" criticism, most people get very defensive when they receive what they consider "criticism", constructive or not.
That is probably why most people prefer NOT to say anything rather than provoke someone's ire. This appears to be true of society in general as well as ham radio.
Just look at some of the postings here on eHam. Personally, I think we should exercise at least a modicum of care with grammar and spelling. Sometimes the spelling and grammar is absolutely atrocious, to the point that some posts are literally impossible to "interpret". Yet, "I pity the fool" who makes any comment regarding the apparent lake of literacy. You are labeled "the spelling police", or worse. Two things come to mind: one is that, apparently, this site is checking my spelling, even as I type this; secondly, surely everyone knows a literate person who can post FOR them.
You can apply this to virtually any activity or topic in life in general, again, as well as ham radio. If you mention you own a Ford, there are thousands of people who will gladly jump on you and tell you what a fool you are. Oh, you ride a Honda motorcycle? What an idiot! Oh, you bought a Yaesu? What a moron!!
So the obvious question is why would you subject yourself to such "backlash" and vitriol? Is it really worth it? Sometimes, as you point out, yes it is. My guess is, more often not.
Sometimes people can't understand what you are saying anyway. Give a report of 599C. Don't be surprised if the response is, "Thanks for the great report!" Conversely, send 579X. Hey, there must be something wrong with my signal!
Again Dave, I appreciate your thoughtful, and civil, reply. We should all be so courteous!
73 to all,
de Larry
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by K0BG on June 22, 2009
|
Mail this to a friend!
|
Larry, Dave, et. al.,
First, I apologize to the author, as this response is off the direct subject, albeit related.
I make it a policy to never (!) use compression, not even for that rare, nirvana producing, DX entity. I keep my mic gain on the verge of none, I'm careful about getting too close to others, and I always try to make sure the frequency is clear before I call CQ. I'd like to think I do this as a courtesy to others, even though too few others do likewise.
As self-centered as it may appear, my on-air conduct is part of my personal motto; set the example, don't be one. I don't believe I've heard Dave on the air, but I have heard you, Larry, and it appears you too set a good example. I think that is all any of us can do, and I further believe that we all get our just rewards in the end.
As for the rest, they can hide behind all manner of disguise, berate and belittle me for my responses, and even hate me, if that's their intent. The difference is, their reward is here and now in the pitiful world they live in. Sobeit!
Alan, KØBG
www.k0bg.com
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by WB2WIK on June 22, 2009
|
Mail this to a friend!
|
I've never even seen a Hilberling, let alone use one, so can't comment on that. I have used both the IC-7800 and the FTDX9000 and they're both pretty spectacular.
But "ultimate HF front ends" don't have to be very modern to be good. My old Drake TR-7, c. 1978, uses a half octave preselector followed by a balanced mixer, with no preamplifier stage at all, and is as "bulletproof" as anything I've ever used. It does have a crystal roofing filter also, maybe the first one ever implemented in a piece of ham gear (not sure).
I cannot recall anyone ever overloading the Drake. It's also extremely immune to out-of-band (BC etc) interference, and in the Sherwood Engineering reports has the highest "100 kHz blocking" dynamic range of anything ever tested (at least, that was the case last time I looked), at 146 dB. Not bad for 30 year-old technology.
WB2WIK/6
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by K6AER on June 22, 2009
|
Mail this to a friend!
|
Receiver front ends are not a problem in today’s transceivers. Their dynamic range is well over 100 dB with third order intercept in the plus 30 dBm range. Even the cheapest transceivers are more than sensitive for the limitation is the noise floor coming out of the coax.
Mixer, IF distortion and filter phase distortion is much more of a problem. Even the IC-7800, K3 can suffer from Mixer, IF and filter phase distortion. The last 10-20 dB of performance is generally lost due to a poor antenna systems.
Software defined radios are also changing the traditional mindset on receiver design. Every cell phone in use is a software designed radio.
Sill, I would much rather have a beam up high and a cheep transceiver than a 7800 and a dipole.
|
| |
|
Designing the Ultimate HF Front End
|
|
|
by KC8ZEV on June 22, 2009
|
Mail this to a friend!
|
While many of the high end transceivers perform exceptionally well, my elmer taught me many years ago to put more of your resources (doesn't always mean $ but often times translates into that) into your antenna than your rig.
High end rigs cannot make up for a marginal antenna, period.
Even with modern advances in transceiver design, no circuit has yet to replace a poor antenna.
My two cents, for what it is worth, save your resources for the best antenna you can make work for your location. You might be surprised how that "entry level" HF rig performs.
Don't forget about your feedline too!!
73
KC8ZEV
|
| |
|
Designing the Ultimate HF Front End
|
|
|
by MACKAY3031 on June 22, 2009
|
Mail this to a friend!
|
Things I would want...No phase noise, no passband ripple, a pulse RF noise blanker early in the system that works,very high dynamic range, and a very low noise floor.
The amateur who mentioned cleaning up our transmitted signals is spot on!!
And of course, those that mention the critical importance of the antenna system are also right, however...Improvements in receiver design are always welcome.Spectral noise subtraction and 64 bit A/D converters with excellent dynamic range would be cool too.
DE KI4WCA
|
| |
|
Designing the Ultimate HF Front End
|
|
|
by K4ELO on June 22, 2009
|
Mail this to a friend!
|
A quick glance at http://www.sherweng.com/table.html
will show you a lot.
The K3, Flex 5000, and the Orion II are way ahead of what you have mentioned.
73
Wayne
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by W7ETA on June 22, 2009
|
Mail this to a friend!
|
"What order would you assemble the perfect HF front end and why?"
Can one ponder better answers than antenna and transmitter?
Bob
|
| |
|
Designing the Ultimate HF Front End
|
|
|
by SV9OFO on June 23, 2009
|
Mail this to a friend!
|
I also believe that since ADC's climbed linearly up to the 250 MHz range, there are more to come in SDR developments, like the complete removal of LO's as all conversions can now be in the math size. Greater sampling rates provide better accuracy filtering and an enormous bit depth along with huge linearity allows for even better dynamic ranges. Since evolution on the field keeps pulling performance ratings up the scale, the "perfect" receiver is always the next model to show up.
That, of course, leaves space in the receiver for better spectrum management on the RF side. Lower noise design concepts, including less parts for better performance, new types of semiconductors, hybrid components like electronically tunable RF filters and notches with extreme accuracy, and more.
Maybe new technologies could be evolving on managing spectrum on RF side. There are more to come in this field and I feel that we are witnessing a very interesting phase of transition in receiver performance equal to the introduction of tube supergenerative receivers.
My fears for tomorrow include facts like BPL and other sources of industrial in-band QRM that weren't here before, and that manufacturers will invest more on researching ways to get rid of those interference sources than evolving other characteristics.
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by G3RZP on June 23, 2009
|
Mail this to a friend!
|
The problem with cleaning up transmitters is that it's going to cost. Such an approach as a polar or cartesian loop is quite complex, especially the polar loop for multiple bands. Another similar approach is the CALLUM modulator, which uses two limiting amplifiers, fed in phase quadrature from VCOs with feedback and combined in phase quadrature. That needs very careful adjustment of both phase and amplitude balance and linearity in downconversion in the feedback loops, which is a fairly frightening requirement for multiple bands, although maybe an auto calibration might be possible. Possibly pre distortion in the SDR approach may work, but that requires some form of auto calibration when doing band changes (or even major frequency changes within a band). Pre distortion is probably going to be the most effective from the viewpoint of power consumption.
Reverting to receivers (although this applies to transmitetrs too), the phase noise is likely to be the limiting factor. If we have a +30dBm Intercept Point (IIP3) and a 10dB noise figure, then with a 250Hz CW filter noise bandwidth, we have a 113dB dynamic range. To have an equal phase noise limited dymanic range, the phase noise needs to be -137dBc/Hz with one interferer. For two interferers, we need -140dBc/Hz to get the same performance. Now if the transmitter has its IMD products improved to -60dB, the phase noise requirement for a 3kHz bandwidth is -94dBc/Hz for a 3dB degradation, so it should be -104dBc/Hz. But then we need to also look at the transmitter noise figure. Assume that we have a 1mW level at the base band modulator and a 60dB SNR. We then have at 1kW out, a 0dBm noise signal in a 3kHz bandwidth, or -34dBc/Hz. Now assume that outside the base band, we have 80dB of attenuation, we have -114dBc/Hz, so that allows a total transmitter noise figure of around 10dB, which is pretty difficult - especially with complex high linearity power stages.
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by ZS5WC on June 23, 2009
|
Mail this to a friend!
|
My dream rig would be an HF rig with all signal processing handled by valves-something like my FT-501.
Then add up conversion and roofing filter.
All HT dc supplies must be regulated by mosfets and the frequency control must be controlled by very well filtered VCO's locked to DDS'es..
Now that would be my best performing rig..!
William
ZS4L / ZS5WC
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by SM0AOM on June 23, 2009
|
Mail this to a friend!
|
It appears that the largest improvement potential lies with the transmitters. For decades, this has been pondered on by the radio engineers that are involved in "co-located systems", and proven remedies, such as narrow RF selectivity and low-noise transmitter amplifier chains exist, but are quite expensive from an amateur radio point of view.
I concur with Peter that it is probably the noise performance, phase or broadband, that will limit strong-signal system performance, as the acheivable receiver IMD has reached a point well beyond which is usable with the current transmitters.
SDR-based transmitters may have the potential to create less "close-in" noise than conventional transmitters if properly designed.
It also appears that quite little effort is invested in optimising current amateur transmitter noise performance.
Comparing SM5BSZ:s sideband noise measurements on i.a. the IC-7800 with the best professional HF transmitters shows that the IC-7800 is about 20 dB worse at a 50 kHz spacing, and about 30 dB at larger spacings.
To create a very low-noise amplifier chain, it takes a lot of "tricks" including MOSFET:s in all stages, "noiseless feedback" and not using any resistive or PIN-diode attenuators for level-setting in the signal path. Using these techniques the SRT SSA1020 1 kW HF amplifier obtained a noise floor of - 165 dBc/Hz at 100 kHz spacing, and a noise figure of 6 - 8 dB referred to the input.
The receiver performance attainable in direct-sampling SDR-based designs should be well above any current transmitter, as the "close-in" phase noise problem is much less prevalent in direct-sampling than in mixing designs.
Demonstrations of the Rohde&Schwarz EM510 and the Perseus, where a 50 Hz hum sideband -85 dB from the carrier of a crystal oscillator could be resolved, have convinced me in that matter.
As other have said, the receivers have reached "the end of the road", and further improvements should be directed towards the transmitters.
73/
Karl-Arne
SM0AOM
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by G8UBJ on June 23, 2009
|
Mail this to a friend!
|
|
When all else fails don't for get the attenuator. At the very least it lets you determine who is overloading your front end as opposed to who is overdriving their tx
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by SM0AOM on June 23, 2009
|
Mail this to a friend!
|
The RF attenuator, preferably with 1 dB steps, can indeed be a very effective tool, as it enables the user of an 80 - 90 dB IMD limited dynamic range receiver to place the range just above the ambient noise floor. A much larger dynamic range is seldom usable in practice.
Readers old enough to be familiar with the works of B M Sosin of Marconi and his "receiving failure factor" can recognise the merits of the "attenuator approach".
73/
Karl-Arne
SM0AOM
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by WX7G on June 23, 2009
|
Mail this to a friend!
|
http://www.sherweng.com/table.html
This link provided by K4ELO tells all. Anecdotal evidence, such that the TR4 has a good receiver, is not borne out by the Sherwood data.
A good receiver doesn't just happen. It is carefully engineered from the front end, thru the mixer(s), the LO, and so on to the audio stage.
The Elecraft (receiver) at less than $2000 appears to outperform other amateur transceivers at any price. I would like to see some professional receivers included in the Sherwood testing.
For casual use even a poor receiver works well enough to make many contracts. But get into a dense enviromnet such as the ARRL 160 meter contest and a poor receiver makes for tough going. I've been there. Using a K3 makes all the difference in the world. No more receiver overload and foregoing a QSO because of a nearby signal pumping the AGC.
|
| |
|
Designing the Ultimate HF Front End
|
|
|
by WA0ZZG on June 23, 2009
|
Mail this to a friend!
|
Those that comment about transmitter performance are right. Just do the math. Don't blame your receiver for all the noise you hear. A 100W transmitter with only 25db second order distortion is OK if the signal is weak (say an S5). If the operator now uses the same transmitter to drive a high power amplifier and shows up at 40db over s9 will now have second order at 15db over s9. You receiver will see this within its passband if you listen next to it and the transmitter operator and you will be operating within normal specifications. If the transmitter operator now adds compression and overdrives ALC, second order will be past 20db over S9 on your properly operating receiver.
That is why some radios allow class A mode. To drop second order down to 50db. Now the Second order will only be around an S7. We can now see what the receiver does.
Dave
WA0ZZG
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by K0BG on June 23, 2009
|
Mail this to a friend!
|
It doesn't make much different what you do to the transmitter section; there will always be some clown doing his best to defeat it. Anyone who listens around 14.175-180 already knows this.
We could do away with the built in processors, but then TenTec, Idiom Press, et. al., would just fill in the demand.
No, what we need is better education, and that task will be (is) much more difficult than anyone could imagine. That's why practically no one has tried, and those who have, didn't make a dent!
Alan, KØBG
www.k0bg.com
|
| |
|
Designing the Ultimate HF Front End
|
|
|
by K1DA on June 23, 2009
|
Mail this to a friend!
|
|
Antenna AND location count, but as a practical consideration-- how many MORE contacts would be made if everyone had the ultimate receiver? Few stations outside of the contest and "rare" dx world need 12 grand worth of radio and few need 1.5 kw to talk to the next state on 80 meters. Take a look at 6 meters.. when it opens 10 watts will get you great signal reports, guys have worked dxcc with 100, but the new fad is "converted low band amps" and 1.5 kw of more or less clean RF which will give you some "extended groundwave" when the band is closed, that's all. SO now we chase our tails trying to improve our 6 meter receivers to deal with 1.5 kw signals where 1.4 kw of it is unnecessary. I am sure the "big three" love this for sales but it doesn't make the hobby more affordable or more enjoyable. Being something of a six fan from way back I was pleased to see 6 in the new hf radios-- more people to talk to-- but screaming pileups with their inherent bad attitudes, three people answering at once, and generous splatter are ALREADY trashing a great band and the sunspots are not even back yet. Well, there IS some life on 6 cw.
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by WB2WIK on June 23, 2009
|
Mail this to a friend!
|
>RE: Designing the Ultimate HF Front End Reply
by WX7G on June 23, 2009 Mail this to a friend!
http://www.sherweng.com/table.html
This link provided by K4ELO tells all. Anecdotal evidence, such that the TR4 has a good receiver, is not borne out by the Sherwood data.<
::Did anyone say the TR4 had a good receiver? :-P
I mentioned the TR-7, which has the highest 100 kHz dynamic range ever tested by Sherwood...
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by WX7G on June 23, 2009
|
Mail this to a friend!
|
I meant to say TR7 as one might gather from my reference to the TR7 claim.
Yes the TR7 has high 100 kHz blocking. The Sherwood data is in order of Dynamic Range Narrow Spaced. I consider this to be a more valid measure of receiver performance. As such the TR7 is 28th on the list. Not much of a receiver by the standards of today.
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by N5XM on June 23, 2009
|
Mail this to a friend!
|
|
I'm sure everyone will tell me how stupid I am, and while they may be right, I still want a single conversion rcvr with roofing filters and a digital preselector with rotary control so I can preselect the exact frequency I'm using at any given time. I also want adjustable ACG and slots for xtal filters.
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by KB1GMX on June 23, 2009
|
Mail this to a friend!
|
N5XM:
NO, your not nuts! What you want is what newer radios are
promoting with their roofing filters.
Classic designs that worked and still work had limited gain before the mixer, preselection, selectivity (filters) as near to the antenna as possible and careful
thought given to AGC design.
The problem with modern radios is there is image rejection filters so they can do DC to daylight but
some commercial station 200khz up on 40M kills them.
The desire for wide tuning range and instant bad
switching are compromizes that balance against low
noise and rejection of adjacent signals.
However all of that helps you none at all whe the band is populated with lids running AKM (All Knobs at Maximum) making whole sections of bands unusable.
Allison
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by WB2WIK on June 23, 2009
|
Mail this to a friend!
|
>RE: Designing the Ultimate HF Front End Reply
by WX7G on June 23, 2009 Mail this to a friend!
I meant to say TR7 as one might gather from my reference to the TR7 claim.
Yes the TR7 has high 100 kHz blocking. The Sherwood data is in order of Dynamic Range Narrow Spaced. I consider this to be a more valid measure of receiver performance. As such the TR7 is 28th on the list. Not much of a receiver by the standards of today.<
::Its 100 kHz BDR is a very, very important figure, one which is important to me and anyone living in an RF crowded environment, because it's one of a small handful of receivers that can content with the 1.2 million Watt SW BC station which is up the street from me (KVOH on Simi Peak). The K2 couldn't, and neither can most other receivers.
However to improve the close in BDR, all one needs to do is replace four $5 crystals in the roofing filter, to change the roofing from 8 kHz (standard) to 4 kHz (modification). I've done that, and now its 20 kHz BDR is as good as a K3.
Fairly cheap mod, takes about five minutes.
WB2WIK/6
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by WB2WIK on June 23, 2009
|
Mail this to a friend!
|
>RE: Designing the Ultimate HF Front End Reply
by WX7G on June 23, 2009 Mail this to a friend!
I meant to say TR7 as one might gather from my reference to the TR7 claim.
Yes the TR7 has high 100 kHz blocking. The Sherwood data is in order of Dynamic Range Narrow Spaced. I consider this to be a more valid measure of receiver performance. As such the TR7 is 28th on the list. Not much of a receiver by the standards of today.<
::Its 100 kHz BDR is a very, very important figure, one which is important to me and anyone living in an RF crowded environment, because it's one of a small handful of receivers that can content with the 1.2 million Watt SW BC station which is up the street from me (KVOH on Simi Peak). The K2 couldn't, and neither can most other receivers.
However to improve the close in BDR, all one needs to do is replace four $5 crystals in the roofing filter, to change the roofing from 8 kHz (standard) to 4 kHz (modification). I've done that, and now its 20 kHz BDR is as good as a K3.
Fairly cheap mod, takes about five minutes.
WB2WIK/6
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by WB2WIK on June 23, 2009
|
Mail this to a friend!
|
>RE: Designing the Ultimate HF Front End Reply
by WX7G on June 23, 2009 Mail this to a friend!
I meant to say TR7 as one might gather from my reference to the TR7 claim.
Yes the TR7 has high 100 kHz blocking. The Sherwood data is in order of Dynamic Range Narrow Spaced. I consider this to be a more valid measure of receiver performance. As such the TR7 is 28th on the list. Not much of a receiver by the standards of today.<
::Its 100 kHz BDR is a very, very important figure, one which is important to me and anyone living in an RF crowded environment, because it's one of a small handful of receivers that can content with the 1.2 million Watt SW BC station which is up the street from me (KVOH on Simi Peak). The K2 couldn't, and neither can most other receivers.
However to improve the close in BDR, all one needs to do is replace four $5 crystals in the roofing filter, to change the roofing from 8 kHz (standard) to 4 kHz (modification). I've done that, and now its 20 kHz BDR is as good as a K3.
Fairly cheap mod, takes about five minutes.
WB2WIK/6
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by N2EY on June 23, 2009
|
Mail this to a friend!
|
To N5XM:
What you want is an Elecraft K3.
At least that's the closest thing I know to your requirements.
73 de Jim, N2EY
|
| |
|
Designing the Ultimate HF Front End
|
|
|
by VK3DWZ on June 23, 2009
|
Mail this to a friend!
|
The best "front-end" I have come across is the one in the Collins R-390(A) series. Using valves (what else!) and tuneable band-pass filters these wirelesses exhibit a very strong, and hearty front-end indeed. Indeed, this front-end was rated by Sherwood Engineering as A+. And with careful gain distribution throughout the rest of the receiver, the '390(A) is probably one of the best receivers ever.
Back in the 1950's, those gentlemen at Collins sure knew how to design wireless receivers. Even the professsionals agree that valves are best -- I once worked in a Receiving Station, and we found the transistor receivers virtually useless. The transistor receivers exhibited loads of ovedrloading, and intermod., whilst the valve recivers -- Racal RA.17L's -- were squeeky clean.
We use a "Collins" R-390A here and have never had problems with intermodulation, overloading, double-spotting, etc.
|
| |
|
Designing the Ultimate HF Front End
|
|
|
by W9GMT on June 23, 2009
|
Mail this to a friend!
|
I appreciate the discussion here and I've learned a lot. The ultimate front end for me starts way before the electronics. Please give me old fashioned knobs to control functions. It starts with the Human-radio interface. Menus are nice, but things that I might want to change from time to time for general operation of the rig should be available at the first menu level, or better yet, as knobs or switches on the front panel. The settings of those knobs or switches should have positive feedback about how they are set. For example, whenever a switch or knob for some function is touched, a section of the screen could announce the function of that control, and where it is currently set. Tone, color, numbers, visual analog or bar graph or rotational position, or meter face could pop up as an indicator. And no more of those little knobs, please, like the power output knob on my 756pro3. I would like to have a switch for attenuating the power from full power to just a few watts to tune my antenna or a knob to turn down the power when I switch from SSB to digital modes.
Lets consider displays, controls that provide audio, visual and tactile feedback with enough room for stubby fingers and large enough for those with poor eyesight. Not a radio designed for people with handicaps, but one that is universal in design.
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by G3RZP on June 24, 2009
|
Mail this to a friend!
|
|
By 'Blocking', do you mean 'gain compression', or SNR degradation caused by phase noise? Both phenomena are often called blocking, because they both lead to reception failure, but they are distinctly different causes of failure. It is no use having good gain compression if the phase noise is inadequate.
|
| |
|
Don't forget what's after the Front End
|
|
|
by N9DG on June 24, 2009
|
Mail this to a friend!
|
K6AER: "Mixer, IF distortion and filter phase distortion is much more of a problem. Even the IC-7800, K3 can suffer from Mixer, IF and filter phase distortion."
And don't forget the FT-2/9K's with these kinds of issues as well..
It's fine and good to say that narrow selectivity upfront in the signal path "protects" everything else down stream. But you do have to ask why are the downstream circuits so susceptible distortion to begin with? The current marketing focus is to push and promote roofing filters, but somewhere along the way the broader understanding of the underlying reasons for their need has gotten lost. Now "roofing filter" is a buzword check off item for many when they go radio shopping. They don't actually look at how well they perform (hint: down-conversion radios are far better performers in this measure).
There is something to be said for keeping the amount of "stuff" in the signal path to an absolute minimum. The less stuff there is, then the less opportunities for that stuff to have distortion problems. The direct sampling and QSD based RX's embody this philosophy. As such they turn in the best close-in, not just @ 2 kHz, but *less* than 2 KHz IMD DR numbers. Roofing filter radios always fall apart in this regard once the multiple signals are inside their roofing filters, no matter how narrow they are. And the up-conversion radio struggle to get good 2-5 kHz numbers. Some even can't even get good IMD DR performance at 2 kHz. I think a majority of the ops in NA they run into problems with close spaced IMD DR long before wide spaced BDR problems. The sad part there are many, many, ops who don't even realize it when they are having close-in RX IMD DR problems. They just think it is "dirty" TX's on the frequencies next to them. Are there dirty TX's out there? ... Sure…. But there are not as many as their up-conversion radios may make them think that there is..
Then there is the whole issue of in-band audio IMD. That is a subject that I think has been ignored for far too long. Again the "less stuff in the signal path" radios consistently outperform multi-conversion superhets, and certainly the up-conversion radios in this measure. Just look at the QST expanded test reports that have the in-band IMD plots..
|
| |
|
RE: Don't forget what's after the Front End
|
|
|
by SM0AOM on June 25, 2009
|
Mail this to a friend!
|
I have no argument in principle with N9NG:s reasoning,
but I feel that performance figures have to be seen in their proper perspective.
A quoted 2 kHz spacing dynamic range in a receiver of 90 - 100 dB takes no account of the adjacent channel suppression of the interfering transmitters, but assumes perfect transmitters.
One may identify two major cases in amateur radio operations, one is the SSB and the other the CW case.
For the SSB case,field data from major manufacturers of professional ground HF systems indicate that an adjacent channel suppression of a properly operated SSB transmitter is in the order of - 50 dB for the first adjacent channel (+/- 3 kHz) and - 70 dB for the second (+/- 6 kHz).
Very little actual field data can be found about amateur transmitters, but figures used for ITU/ICAO spectrum planning in aeronautical HF (quoted in ARINC 719) of - 30 dB for the first and - 40 dB for the second adjacent channel may be representative.
The 2 kHz IMD performance of the receiver becomes insignificant if it is an order of magnitude better than the transmitter adjacent channel suppression, which is the case for the vast majority of the receivers in the Sherwood table.
In the CW case, we also have to consider the keying sidebands of the interfering signals. If their averaged energies are larger than the IMD products, any receiver IMD will be masked by the sidebands.
In very grave cases of key-clicks, peak-acting AGC may reduce the receiver sensitivity even at large
frequency offsets.
I have made comparisons between professional and amateur radio equipment in these matters, and the differences of 500 Hz adjacent channel suppression (used in ITU spectrum planning) were in the order of 25 - 30 dB at a keying rate of 30 WPM.
These comparisions were made between Collins HF-80, SRT and Telefunken systems, which are using Gaussian keying envelopes according to ITU-R SM.328 and the IC-735 / IC-706, which by no means are among the worst "key-click offenders" in amateur gear.
It also appears that the in-band IMD of current SSB receivers is far better than the in-band IMD of the transmitters. Speech clipping and some forms of compression may create an in-band transmitter IMD in the order of - 20 dB.
Finally, the extremely good close-in IMD and blocking data that direct sampling SDR receivers can provide is a challenge for the future to the transmitter designers.
73/
Karl-Arne
SM0AOM
|
| |
|
RE: Don't forget what's after the Front End
|
|
|
by G3RZP on June 25, 2009
|
Mail this to a friend!
|
Karl-Arne,
Are those adjacent channel suppression figures inclusive of IMD products? if so, 50dB is pretty aggressive.
|
| |
|
RE: Don't forget what's after the Front End
|
|
|
by SM0AOM on June 25, 2009
|
Mail this to a friend!
|
The values for ground system adjacent channel suppression come from Rockwell-Collins information regarding the HF-80 and Spectrum 2000 systems.
Some of the spectrum plots are referenced in "Single-Sideband Systems and Circuits"
They are overall time-averaged spectra using speech modulation with quite long integration time, which may represent a best-case situation where the IM power is averaged out in some extent. It is certainly a much more benign test than two-tone modulation.
I have tried to re-create this system set-up in connection with customer acceptance measurements of both the SRT SSA1020 and the Collins PA-2250, using shaped noise (CCITT weighting filter) as modulation and a shorter averaging time. My results were in the -45 to -50 dB suppression range for the first SSB adjacent channel.
It also appears to be in line with the results obtained with the best tube-equipped transmitters; the
Collins 208U-10A manual quotes a specification of > 43 dB using the NPR method of measurement.
Also, an old spectrum plot that was in the CCIR reports and in ITU-RR Appendix 17A showed an first adjacent channel suppression in the order of 50 dB using averaged speech modulation.
When ISB telephony and VFT telegraph transmissions were in vogue, adjacent channel suppression was a critical parameter for transmitter designers and operators. Careful adjustment of a transmitter chain could net a long-time value of > 40 dB using multi-tone modulation,even with solid-state transmitters.
73/
Karl-Arne
SM0AOM
|
| |
|
RE: Don't forget what's after the Front End
|
|
|
by G3RZP on June 26, 2009
|
Mail this to a friend!
|
Karl-Arne,
Any idea how these measurements related to a 'normal' 2 tone IMD test?
Back in the days of big valve HF tx's, it was usual to aim for -36dB relative to tone for ISB: how this relates to averaged suppression on voice, I don't know.
|
| |
|
RE: Don't forget what's after the Front End
|
|
|
by ZENKI on June 26, 2009
|
Mail this to a friend!
|
Hi Karle Arne
Can you describe how these professional IMD tests are done and what equipment is used? Unfortunately there is very little information in the public domain on how one goes about doing these tests and how you interpret the data.
The "notched noise test" is mentioned in SSB Systems and circuits however no details are given on how to do this test.
It would be great if either the ARRL or the RSGB would start using these testing methods for product reviews, it certainly would give a more realistic picture on how ham transmitters are performing in the real world.
Zenki
|
| |
|
RE: Don't forget what's after the Front End
|
|
|
by SM0AOM on June 26, 2009
|
Mail this to a friend!
|
The two-tone test acceptance limits for the PA-2250 were if I remember correctly -32 dB IM3 and -38 dB IM5 relative to one tone.
Actual equipment was better in this respect with
about 3 - 5 dB.
Equipment used for measuring these values were a Racal 9083 two-tone generator, an SRT TD90 exciter which measured > 55 dB IM3 and a Marconi TF2370 spectrum analyser connected to the back end of a power attenuator.
For the noise modulation test an FM broadcast receiver was tuned to an unused channel and the audio fed through a CCITT telephony weighting filter to the exciter input. The intention was to try to emulate the average spectral characteristics of voice.
The drive level from the exciter was adjusted so that noise peaks reached the rated PEP as viewed on an oscilloscope.
NPR tests on SSB eqiupment are usually done by connecting a noise source in parallel to all 4 inputs on an ISB exciter, and adjusting the levels so that
all channels show the same average power output on a spectrum analyser.
The noise input to one channel "in the middle" is then removed, and the decrease of noise level for that channel is noted. The difference in level of the "Quiet Channel" is then the adjacent channel or IMD suppression expressed as "Noise Power Ratio" or NPR.
This practice comes from carrier telephony, and is a very sensitive indication of proper operation of a whole chain of signal processing.
If I have had access to a 4-channel ISB exciter when making these measurements in the mid-90's, I would have been able to compare all three methods.
In a way, progress has been "backwards" in the transmitter adjacent channel performance aspect.
The Telefunken ISB transmitters that the PA-2250'sreplaced could measure -42 dB IM3 and -48 dB IM5 when all RF feedback and envelope feedback circuits were correctly adjusted on the measurement frequency.
73/
Karl-Arne
SM0AOM
|
| |
|
RE: Don't forget what's after the Front End
|
|
|
by SM0AOM on June 26, 2009
|
Mail this to a friend!
|
Addition to above:
A quite readable overview of NPR testing can be found at
http://www.commsdesign.com/design_corner/showArticle.jhtml?articleID=200900037
which also covers NPR techniques in modern multi-carrier mobile telephony systems.
73/
Karl-Arne
SM0AOM
|
| |
|
RE: Don't forget what's after the Front End
|
|
|
by G3RZP on June 26, 2009
|
Mail this to a friend!
|
Actually getting back to the thread......
The Hilberling was, according to reports, abandoned because of the difficulty of meeting EU standards for commercially available amateur radio equipment.
Now they did not need to use a Test House, provided thay had an adequate QA procedure. This is laid down in the Radio & Telecomunications Termnals Equipment Directive. (R&TTE). The levels at that time (they are about to change) were not excessive, and in line with FCC Part 97 requirements. The requirements were purely spurious emissions from transmitters and receivers (the latter at -57dBm below 1GHz and -47dBm above 1 GHz): the standards have just (last week!) been changed and approved to go out on Public Enquiry and now align with the Radio Regulations. They are 7dB tighter than FCC part 97 at HF, purely because FCC managed to implement the ITU requirements incorrectly. The other requirements were in EMC immunity and the amount of rubbish put back on the mains and the effects of brown outs and mains failures. (This latter is important - one UK test house passed a set of paging transmitters. In service, when the mains failed, they came up on 145.000MHz, and, in some cases, the resultant SWR caused the PA to blow up!)
Additionally, the 'Big Three' as well as TenTec have manged to meet the standards without a problem. So one wonders exactly what the real reason was.....
|
| |
|
RE: Don't forget what's after the Front End
|
|
|
by WB2WIK on June 26, 2009
|
Mail this to a friend!
|
Re: Hilberling it may have been inadequate funding to continue the project, or poor marketing feedback after "showing" the rig at Dayton (I think that's where they showed it -- I wasn't there). I "heard" the "display" rig was not connected and was under plexiglas or something so people couldn't touch it.
That's not a good way to garner a following.
I don't believe it would be difficult to meet the EU standard, lots of stuff does.
|
| |
|
RE: Don't forget what's after the Front End
|
|
|
by ZENKI on June 26, 2009
|
Mail this to a friend!
|
ETSI EN300 states that for marine HF transmitters the spurious out of band emissions should be the following.
1.5khz to 4.5 khz 31db below PEP
4.5 to 7.5 khz 38db below PEP
7.5 to 12.5 khz 43 db without exceeding 50mw
These figures are for a 100 watt transmitter.
The NTIA states
1.75 khz to < 5.25 khz 28 db below PEP
5.25 khz to < 8.25 khz 38db below PEP
>8.75 43 + 10(Px)
I cant easily summarise ITU SM-329, however the ITU has a very tight IMD spectrum mask for out of band emissions(OOB) for HF equipment. Certainly no amateur transceiver would meet would meet present ITU standards, they are that poor.
There is a considerable amount of transmitter data for IMD tests done on DRM transmitters for various bandwidths. In every case the best 2 tone IMD test numbers produced good IMD figures for various wider DRM bandwidths. For amateur service it would probably be a good idea to use the old FCC commercial SSB IMD standard for transmitters. That standard had a very tough IMD mask that covered everything on HF.
This old FCC standard simply stated that 3rd order IMD had to be 30db below one of 2 tones(36db below PEP), 5th order 36db below one of 2 tones((42db below pep)) and the most critical of all the 11 the order, 43+10 log (average power). In the era of super stations this figure should be set at 80db(90db below PEP) and this would cover everything from QRP to QRO. There is little point setting the standard only for transmitter power only when we know most people use amplifiers. If this did become the standard for HF amateur equipment you would only need laws for LIDS!
As usual we will get arguments about the costs of making ham equipment meet these standards. I don't see that as issue when we have radios like the IC7800 that costs $10,000 dollars that cant even meet any of these basic IMD standards, whose fooling who? Hams will be the only losers in the long run towing the transceivers manufacturers line that HF transmitters are good enough when they are clearly are not up to the standard required.
Zenki
|
| |
|
RE: Don't forget what's after the Front End
|
|
|
by SM0AOM on June 27, 2009
|
Mail this to a friend!
|
To summarise the above, it appears once again that transmitter adjacent channel suppression the system performance on the amateur bands.
I share the sentiments that tighter limits should be introduced, but there is a problem spelled "deregulation" in many countries. In some countries, there are no references to ITU-RR (or EU regulations) to emission limits within the amateur radio bands.
A radio amateur there can transmit with a very "lousy" signal without risking any enforcement, as long as the modulation and IM sidebands stay > 20 dB down at the band edges (and, there are no sub-bands in these countries).
To return to the original discussion, it appears that Hilberling put a lot of faith in their wideband amplifier by putting it before any front-end selectivity. Narrow-band preselectors or bandpass filtering before any amplification makes life very much easier for the "downstream system".
The performance of their wideband amplifier was quoted as an output IP of + 50 dBm, which is in line with what you find in the more expensive antenna multicouplers.
An amateur with either many "super-station" neighbours or using large wide-band antenna arrays (such as log-periodics or rhombics) may notice the differences in IM2 and IM3 performance when using a narrow preselector before the receiver.
I am however not entirely convinced of the real needs for very high IP receivers for current HF operations.
Measurements made in the mid-90's at two large receiving sites (using TCI 12 dBi log-periodic antennas beamed due south)failed to detect any significant reductions of the effective receiver sensitivities by the IM power generated by the Rohde&Schwarz NV-14 wide-band active multicouplers (2 dB gain, +35 dBm IP3), even at night-time and at frequencies near the 6,1 and 7,1 MHz broadcast bands.
It would be most interesting if more recent field data of actual received noise floors and signal levels could be collected and presented in a form like the tables by Peter in QEX 1/2009.
The "power wars" between the broadcasters seem to have become part of history, as the numbers of extremely strong signals have diminished. On the other hand, we may have to worry of the far-off sideband noise from HF DRM transmitters.
73/
Karl-Arne
SM0AOM
|
| |
|
RE: Don't forget what's after the Front End
|
|
|
by G3RZP on June 27, 2009
|
Mail this to a friend!
|
SM329 does have limits for amateur equipment. Outside 250% of the necessary bandwidth (which is an arguable number, but the 99% power bandwidth is usually used), the limits are -43dB +10 logP, not exceeding 50dB below 30MHz and not exceeding 70dB above 30MHz. The European implementation (ERC REC 74-01) specifies minimum necessary bandwidths in order to prevent problems at UHF and microwaves with phase noise making the limits impossible to meet. I went to the CEPT SE21 meetings and got those 'easements to practicality' through, and I should be presenting a proposal to ITU SG WP1A in September on behalf of IARU to adopt those easements in SM329.
The current ETSI Standard for commercially available amateur radio equipment, EN301 783 has different requirements. This is because the limits were derived from EN300 684, which was an EMC Standard, produced before ITU TG1/3 introduced limits for the amateur service into SM329. However, last week, at the ETSI TC_ERM (Technical Committee, Emc and Radio Matters) a new version, prEN301 783 was approved to go forward for Public Enquiry, and that does have requirements aligning with ERC REC 74-01.
|
| |
|
RE: Don't forget what's after the Front End
|
|
|
by G3RZP on June 27, 2009
|
Mail this to a friend!
|
Karl-Arne said
>It would be most interesting if more recent field data of actual received noise floors and signal levels could be collected and presented in a form like the tables by Peter in QEX 1/2009. <
I thionk you mean 2001 or 2002
For the update, see NCJ in about March or April last year! Also my lecture at the RSGB HF Convention last October.
If you want the slides for the lecture or the article, send me your email address.
|
| |
|
RE: Don't forget what's after the Front End
|
|
|
by SM0AOM on June 27, 2009
|
Mail this to a friend!
|
Well, I managed to find two tables comparing the
7 MHz signal level and noise floor situations in 2002 and 2007 in the "Letters to the Editor" column of 1/2009 QEX.
I have no argument with the SM329 limits, but national rules may be so vague that radio amateurs cannot claim any protection against "lousy signals" within the band limits. If someone contacts the Swedish authorities and complains of "splatter" or key clicks, he is met with a "shrug of the shoulders". Only if licensed users outside the amateur radio bands complain, there may be enforcement.
On a side view, tighter spectral limits may turn out as a two-edged sword.
There are discussions in Sweden regarding further deregulation of amateur radio, and one of the paths that the legislators may follow could be tighter spectrum limits paired with lower allowed output power and type-accepted equipment.
"Good-bye, home construction and modified surplus"
73/
Karl-Arne
SM0AOM
|
| |
|
RE: Don't forget what's after the Front End
|
|
|
by G3RZP on June 27, 2009
|
Mail this to a friend!
|
|
The choice of 2.5 times the necessary bandwidth was done to keep the 3rd and 5th order IMD out of it. There have been howls of protest from the broadcasters that under the European EMC Directive, home brew amateur radio and amateur radio kits were excluded. Incidentally, other electronic kits aren't, but that is a can of worms that the European Commission don't want to touch ! Some people in EU administrations aren't too happy about not Type Approving amateur home brew equipment, but since it would be impossible to enforce, and under the R&TTE, you can self certify, it's not likely to happen, provided amateurs tend to go for good house keeping. Especially with the current trend to no enforcement. Good housekeeping is things like using mains filters, for instance, to keep rubbish being fed back up the mains, and not using half wave rectification, which puts DC on the mains. Keeping harmonics under 50dB - not difficult. In fact, it's a licence requirement in Eire, as it is in the US (the US is -43dB because the FCC got it wrong, but never mind, it's the thought that counts...)
|
| |
|
Designing the Ultimate HF Front End
|
|
|
by W3ULS on June 27, 2009
|
Mail this to a friend!
|
It's great to read a thread with such knowledegable people:G3RZP, et al. Some ramblings:
The RSGB mag has carried a Peter Hart review of the new Icom 7600 in which was found that phase noise (as I recall) actually was worse at wider kHz spacings. Anyway, whether my memory serves or not, there was some strange anomaly in the lab test results that raised a question as to whether it was a problem with the tested ICOM sample or was something more serious.
The K-3 as I recall (again depending on memory) was released with much ballyhoo about close-in IMD3 being plus 100 db. I believe one of the owners said that achieving the 100 number was a design goal. Given the fact that such performance is invisible (per previous comments) in the real world of hamming, the decision was strictly a marketing gimmick. Adding insult to injury, the transmit IMD3 initially was inferior, and the problem was addressed at least partially in subsequent iterations of the K-3, as I understand it.
I miss Tom Rauch on this thread. He has opined (memory again) that close-in IMD3 above mid 80s is unnecessary, and even that is only under severe CW contest environments. When I was listening to CW contets in my former QTH, a Ten-Tec Jupiter and a Yaesu FT-950 provided all the selectivity I needed. I kept a Ten-Tec OMNI VI on standby just in case.
73 to all
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by G3RZP on June 28, 2009
|
Mail this to a friend!
|
I don't always agree with W8JI, but
>He has opined (memory again) that close-in IMD3 above mid 80s is unnecessary, and even that is only under severe CW contest environments.<
ia a sentiment that I do heartily agree with. Mainly because the close in phase noise will be a limit before the IMD is. It can also be very hard to get really good in band IMD from crystal filters, especially the higher frequency ones - because of Hooke's law. Early mechanical filters were notoriously poor in this respect, although the Rockwell-Collins ones from about 1985 or 86 onwards were (still are? - haven't used one for years) very good.
I'm not totally convinced about the overall superiority of DDS, even when combined with a good PLL. The spur levels can be a problem, and there was a paper by a guy called, if I recall correctly, Samuelli some 20 odd years ago on the mathematics of spur levels in DDS systems. Fractional N PLL systems are another approach, and although the original ones weren't very good on phase noise, the use of sigma-delta techniques can allow very good performance at low power. So, of course, can a 6C4 with an input of about 1.5 to 2 watts, a high Q coil on a ceramic former and good air spaced variable capacitor.....
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by SM0AOM on June 28, 2009
|
Mail this to a friend!
|
Valid point about the IMD performance of high-frequency crystal filters.
Narrow bandwidths at VHF IF frequencies result in extremely high loaded Q's, and as Q can be defined as reactive energy/active energy, it results in extreme mechanical stresses in the quartz material at higher input levels. From correspondence with crystal filter manufacturers in the 80's, I recall that an IP3 specification of +35 dBm for a VHF crystal filter was considered as "tough".
One way of circumventing this is to split the signal path after the first mixer into two or more, as Telefunken/Deutsche Aerospace did in the E1800 receiver.
Finally, a lot can be said in favour of the "KISS" philosophy. I am still intrigued about the G3PDM receiver (RadCom 1971) where (single-tone) adjacent channel suppressions in the order of 140 dB (!) were claimed. This design used a varicap tuned push-pull (Kalitron) E88CC triode oscillator, and it would probably be possible to do even better by using "Huff and Puff" stabilisation of the oscillator.
73/
Karl-Arne
SM0AOM
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by WA2JJH on June 28, 2009
|
Mail this to a friend!
|
The SDR-direct A/D converter should yeild the best front end.
However, I agree with WIK(Steve). If you have not used one, you will see this rig sounds fantastic.
If you replace the vari-L front end DBM with a Minicircuits mixer, the TR-7 rocks even more.
We do not need .18uv sensitivity! (.4-6uv is fine) Thats why the TR-7 is almost silent, then the DX pops right in.
I guess direct A/D detection makes SDR rule with better BDR, no filter blow by, and no cheap audio amp and speaker found in many analog rigs.
Of course the nifty factor of changing your rig by software is a wunner too.
I hope one SDR rig will use open source code. Be real cool to tylor the front end with Basic or Fortran.
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by W3ULS on June 28, 2009
|
Mail this to a friend!
|
". . . and no cheap audio amp and speaker found in many analog rigs."
That's why I have kept my JRC JST-245. Very high quality audio is present in both receive and transmit modes. Even though I care little for phone, I keep it for SSB contests and because for my money it was the best analog transceiver produced in the 90s.
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by G3RZP on June 29, 2009
|
Mail this to a friend!
|
One does wonder about a claim of 140dB. It would need an average phase noise of around -174dBc/Hz in the adjacent channel.
We had an interesting discussion in the Homebrew forum about mixers and the noise figure of the 7360. I need to make some measurements.
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by SM0AOM on June 29, 2009
|
Mail this to a friend!
|
Regarding the really ultimate HF front end, it would of cause contain a superconducting preselector.
I long thought that this was some kind of "high-tech joke", but to my surprise they were for real.
In the mid 1960's superconducting filter research was done at the Naval Research Laboratories, and was published in 1967.
I found a reference in an early 70's Swedish electronics journal about this work, and an online summary was found at the IEEE website:
"Superconducting Ultrahigh Q Tunable RF Preselector
Arams, Frank R.; Fradkin, Joseph; Kornfeld, David; Sard, Eugene W.; Siegel, Kenneth
Electromagnetic Compatibility, IEEE Transactions on
Volume 9, Issue 3, Dec. 1967 Page(s):110 - 123
Digital Object Identifier 10.1109/TEMC.1967.4307478
Summary:The application of superconductivity to the problem of reducing radio frequency interference (RFI) is described. A lowloss, extremely narrowband superconducting filter was developed that is tunable from 6.3 to 21 MHz. This filter has an unloaded Q of 350 000 to 600 000 over the tuning band, bandwidth adjustable from 60 Hz to 50 kHz, calibrated external controls for frequency tuning and bandwidth adjustment, and good resettability and stability. An optimum circuit design provides minimum bandwidth variation with frequency tuning and minimum resonant frequency variation with changes in bandwidth. Performance tests on high-frequency receivers showed reductions in intermodulation interference susceptibility of 40 dB and improvements in FSK Teletype signals of 23 dB."
I for sure would not mind having a tunable multipole filter with bandwidths in the order of 50 Hz upwards in front of an HF receiver or as a "postselector" after an exciter, especially when the tuning range includes 7 MHz.
73/
Karl-Arne
SM0AOM
|
| |
|
Designing the Ultimate HF Front End
|
|
|
by WA2JJH on June 29, 2009
|
Mail this to a friend!
|
Karl..your knowledge of SDR and physics are outstanding.
But liquid nitrogen cooling to almost eliminate scott noise(thermal noise produced by all Componants) IS OVER KILL :)
Heck, I did an artical called......What do you want in your new $10,000 radio. Most ehsm responders said,,,,,,,WE WANT $6000 in rebates and change. :)
I wrote that piece of literary trash 5 years ago,
Back then it was either a crummy SDR-wannabe called the Pegasus.
The other extreme was the 10K Icom 7800.
Many hams complained that a A/D front end would still get clobbered by a LID running 1500w OUT. I agree.
Perhaps the new rigs should have a 600W 50V quad mosfet final. The final would be controlled by software
modable ALC. No more twin 3-500z overdriven to hell and back.
I do see great innovations. I perdict an all digital rig to be a real performer. Most of these almost all digital spec programmable rigs can be sold as a kit for under $2000.
-
I build over clocked PC's. The payoff is that all the expensive ASus MOHERBOARDS have many clock options.
The overclock chips can be software adjusted in 100hz increments. One clock can be tuned from 150mhz-4ghz+
I can see SDR rigs using these clocks.
See what I am leading TO? The computer provids L.O.s and other freq pumps for the rig too.
Gee, wish you posted here more often. Psdhaps some hams may want to put the JA rigs to shame.
The younger hams get hooked by a combo Ham rig/video game. Bet those computer whiz kids can program better than the adult that wrote the code for the Yeasu FT-950.
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by WA2JJH on June 29, 2009
|
Mail this to a friend!
|
W3ULS...I agree. Your 1990 vintage JRS is like my Kenwood TS-950SDX(not yuasu FT-950)
Kenwood used to be smart. The last revision of the TS-950SDX. 4 or more 8 pole filters in ths 1st and 2nd IFs in the main RX did the heavy work.
The used DSP in the 3erd IF snd demod/audio out.
The DSP, DDS, distorionless RF speech proc. and DSP to boot.
The audio out chip was ultra high grade. It would deliver over 3W of audio into a 4 ohm load.
The rigs today deliver under 2W with 10% THD.
That is into an 9 ohm speaker to make it even worse.
With only 8db of compression, and rhe DSP voice eq set for my voice, the pre and after RF filters built for the special speech proc,,,,,,,,,,,I would hold my own with people using Amps. The TX funals output a clean 150W.
I will go SDR, However, my Drake TR-7 and Harris rigs will be in my grave when I SK
73
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by SM0AOM on June 29, 2009
|
Mail this to a friend!
|
Well, thanks for nice feedback :-),
The main reason for using superconductivity in the preselector (which used a lead alloy as material) was to eliminate all resistive losses in the filter circuits, thus enabling the extremely high loaded Q's quoted.
This was accomplished by liquid helium cooling, a by no means simple feat.
Also, as a side effect, the thermal noise in the filter circuits will be almost non-existent, but for practical HF applications this does not make much of difference.
The main point is that if it is possible to put a narrow filter at the signal frequency in front of all signal processing, the "downstream receiver" does not have to be very "high-grade".
I have not read the whole original paper, but it seems reasonable that one of the receivers that the filter was put ahead of was the R-390A. The quoted improvement of 40 dB in IM suppression over an already good receiver speaks for itself.
Finally a spelling correction to my previous post,
"of course" is of course spelled "of course".
The phone rang while I was typing.
73/
Karl-Arne
SM0AOM
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by W3ULS on June 29, 2009
|
Mail this to a friend!
|
WA2JJH:
Your description of the Kenwood is an exact replica of my experience with the JST-245. I found that with 150 watts of clean audio, a simple antenna and a good location, I could break into DX and contest pileups almost at will (not all, certainly). I attribute the success to the absolutely clean SSB generated by the JST, plus a true 150 watts. Also, for some reason, the standard Heil mike worked very well with it—something that Dave Heil told me he had heard from other JST owners.
This was in the days of better sunspot numbers. Today, with DX concentrating on 20 and below, I'm sure I would not have the same kind of luck.
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by G3RZP on June 29, 2009
|
Mail this to a friend!
|
|
There was an interesting article in RadCom back in January 2008 on SDR and IMD. It was shown that the current concept of input intercept was not applicable, as the intercept point was a variable depending upon the input signal levels. This means that a graph of intermod levels against input level is needed to properly evaluate the IMD of an SDR rx.
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by WA2JJH on June 29, 2009
|
Mail this to a friend!
|
liquid helium......yowza! I think liq. helium is the closest to absolute 0.
Confess....did you dip apples and rubber balls into the liquid He?
After a few seconds of o degrees K, the apple and rubber ball will shatter like glass if dropped.
The ultra shock was when I freezed a cockroach with the warmer liquid nitrogen. The Bug looked dead for 20 minutes. After thst, the roach just walked away!!!!!!!
Enough of my goofy undergrad experiments. However, we were breeding vestigial winged fruit flys.
My Proffessor told me the experiment was over.
A fruit fly lives for 4 days. We took a 500W UV lamp, and exposed them for 2 hours. LD-100/24H was the result. They all dropped dead in less thn 24 hours.
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by SM0AOM on June 29, 2009
|
Mail this to a friend!
|
I never came into closer contact with liquid helium.
It was quite expensive, and all equipment (MASERS, Josephson mixers, paramps etc.) at the Onsala Space Observatory that used liquid He used some form of closed circuit system to keep track of the precious liquid.
And, a lowly EE student in the 70's was not allowed to get into closer range of such equipment when doing laboratory assignments in Space Communications and Microwave Engineering.
Those were done using much more "robust" equipment such as surplus parabolic dishes, transistor preamps and Collins 51J-4:s in sun-noise, antenna diagram and interferometer measurement setups.
Liquid nitrogen, on the other hand, was ubiquitous around the University, and if some remuneration such as fresh buns at coffee for the institute staff could be provided, some could usally be "liberated" for use in various amateur radio related experiments such as noise figure measurements and cooled GaAsFET-preamps.
GaAs-technology has changed weak signal VHF/UHF forever, but this is worth a thread of its own.
73/
Karl-Arne
SM0AOM
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by WA2JJH on June 29, 2009
|
Mail this to a friend!
|
W3ULS....yup, I used rack up DX 10X easier back 25 years ago.
A few hams eluded to the fact of filthy transmittters. Postulating all the linear amp use has made SDE a must.
Yes, the rig makers should aim for the commercial standard.
I do think if you want a front end for all cndx, SDR is the way.
Say you had a QTH with lots of EMI AROUND. Simply change a few lines of code. Emulate anything from a nuvistor pre-amp to the pre-ampless front ends used in just about every commercial HF rig.
I can only see kits and homebrew make a comeback for
less money.
We will spend our ham bux on algorytyhms. The chips are falling in price.
Sdr is too expensive now FOR THE NEW HAM. Standrds for open source code will be hard to sell to the rig makers.
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by SM0AOM on June 30, 2009
|
Mail this to a friend!
|
Peter, a most interesting discussion about the 7360 in the Homebrew forum.
I have seen data elsewhere which is in line with the findings that this tube (valve) may be somewhat overrated.
The overall selectivity plot for the G3PDM receiver (page 387 of June 1971 RadCom) that shows a response -140 dB down 3 kHz outside the SSB filter centre frequency remains an intriguing question seen in the light of modern knowledge of oscillator noise.
Push-pull tube (valve) oscillators can very good in the far-off noise aspect, I have used the HP3200B VHF oscillator for mixing exciter outputs into the range of a measuring receiver having a VHF crystal filter in front, and it appears that its phase noise performance at +/- 100 kHz is not very much worse than a crystal oscillator.
Perhaps an hybrid solution using tubes (valves) in the high-level parts and semiconductor technology in the low-level and signal-processing parts is an interesting alternative for both receivers and transmitters. With increasing HF ambient noise floors, the noise figure of an 807 as receiver RF amplifier may not be a problem any longer.
One could build onto the ideas from the W1KLK receiver from QST 1972. The extraordinary performance claimed for this design may however be seen in perspective to the quite bad strong-signal characteristics prevalent in most other amateur receivers of that era.
73/
Karl-Arne
SM0AOM
73/
Karl-Arne
SM0AOM
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by G3RZP on June 30, 2009
|
Mail this to a friend!
|
Karl-Arne,
Just how many filter poles do you need to get 140dB rejection even 6kHz away from the nominal carrier frequency? The G3PDM rx only used 6 crytsals at 1.62 MHz. It did have a sharp Af filter for CW.
As you say, doubtful figures by today's standards.
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by SM0AOM on July 1, 2009
|
Mail this to a friend!
|
It would probably take at least 10 poles at 1.62 MHz, and filter "blow-by" would most likely limit the ultimate attenuation to 100 dB or less.
In defence of G3PDM, the SSB selectivity was half done by the crystal filter and half by an elliptic low-pass AF filter.
Still, it appears that the values may be somewhat exaggerated.
Using the quietness of the vacation period here in Sweden, I made some measurements last evening of the single-signal dynamic range on an E1800 receiver. Outside the first IF filter, (about +/- 6 kHz) the dynamic range is limited by synthesizer phase noise and gain compression of the first mixer (difficult to say which is dominant).
A + 5 dBm signal at 30 kHz spacing was needed to reduce an 18 dB SINAD CW wanted signal in 300 Hz
BW by 3 dB.
Inside the first IF bandwidth, things changed considerably.
The first IF amplifier and the second mixer enter compression at - 20 dBm or so (110 dB over the noise floor) which rapidly reduces wanted signal SINAD, and the exceptional phase noise performance is apparently masked by this gain compression.
The bottom line is that in order to gain any practical advantage of very low close-in noise synthesizers, the narrowest possible first IF becomes a necessity.
This may speak against up-conversion designs due to reasons discussed earlier.
73/
Karl-Arne
SM0AOM
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by G3RZP on July 1, 2009
|
Mail this to a friend!
|
The Mark 2 version had the elliptic function filter - that's the one in the 5th edition of the RSGB Radio Communication Handbook. The Mark 1 version in the 4th edition had 2 poles of RC filtering and a valved AF output stage. There's quite a lot of differences between the two versions once you're away from the front end and synthesiser.
As you say, any more than 100dB is dubious - that always made me laugh about the adverts for Atlas radio with the 120dB stop band - as if you could use it even if you could get it actually in the radio.
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by WA2JJH on July 1, 2009
|
Mail this to a friend!
|
The ATLAS had a passive front end. Just about every mil spec radio has no RX preamp. Simply band filters connected directly to a double balanced mixer. Any extra gain can be made up in the first IF.
Passive front end with up conversion, and 3KC 1st IF filter that would not allow one to get FM RX were a few of the TR-7's secrets.
The K2 is single conversion. There are about 10 monolithic xtal filters. 6 are cascaded. This saves money on the 8 pole filters found in many analog rigs.
The big kicker is that the K2's frond end bandpass filters are tuned to cover only the HAM Bands.
Either use varacter tuned frond ends or give up general coverage for superior front ends. That applies to SDR or any analog rig.
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by WA2JJH on July 1, 2009
|
Mail this to a friend!
|
|
Karl, do you agree that if Hams could deal with Ham band only RX, a 1st passive IF that would be wide enough only for SSB/CW(or data)....Good front ends become an easy task?
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by G3RZP on July 1, 2009
|
Mail this to a friend!
|
I do not like varactor tuned front ends. Never really got them to work satisfactorily in tactical military radios (30 - 88MHz) at the input signal levels we wanted, although that was in the mid 1970's. A better approach would be digital capacitors - capacitors switched in a 1 - 2 - 4 - 8 - 16 sequence. Even there, I'd look very carefully before using PIN diodes.
Restricting the rx to ham bands only makes a lot of sense: the front end filters can be designed properly, although you would then end up with a lot more coils than you might like from a commercial viewpoint. Keeping the IF down at 1.4 or 1.6 MHz also reduces the IMD problem in crystal filters.
Or go SDR - but I haven't yet seen published results that tell the whole story. For example, doing IMD at various levels and plotting the IMD product level against input level with sufficient granularity to produce a graph. Doing selectivity tests by measuring the input level at various offsets to cause a SINAD degradation of 3 dB from a 10dB SINAD. Doing selectivity tests on a narrow (250Hz) filter with offsets of a few hundred Hz - to catch the direct conversion approach followed by a low pass filter followed by an ADC. This doesn't mean they aren't around - I just haven't seen them in forms that don't still leave questions.
And at the end of the day, we get back to the fact that most hams aren't in environments where they CAN use more than 95 to 100dB of instantaneous dynamic range!
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by WA2JJH on July 1, 2009
|
Mail this to a friend!
|
Peter, thank you for the info. All the mil radio's I have owned, never had varactor tuning.
My favorite commercial rig I own is the Harris RF-3200. It has a real VFO knob. It is actually opto-isolated knob contolling a VCO.
Has a special CPU controled speech proc. nd ALC. So the 150W output would seem like more. spectual output was fsr cleaner than ANY ham rig.
The ARRL used to use them as CW exciters for their CW practice sessions.
The TX filtering is excetional.For TX and RX far more filter networks are used. This rig is commercial/mil.
For 1980's technology, it is exceptional.
Paid $1500 for it on ebarf. Worth every penny.
You are correct. I did not do enough research to make my blanket statement. Trying to compare it's front with commercial/mil SDR is a hard task.
Important thing is not to believe the hype of the $6000++ Ham rig ads.
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by G3RZP on July 2, 2009
|
Mail this to a friend!
|
WA2JJH said:
>Important thing is not to believe the hype of the $6000++ Ham rig ads.<
How very, very true! I suppose it's always been that way. I saw a copy of an article in R8 magazine from 1936, reviewing an RCA marine rx, which complained about the meaningless IRE standards on sensitivity (they only quoted how much input for how much output, not SNR). So not much in hype terms has changed over the years.
Interestingly, sensitivity numbers have changed. Historically, Europe used microvolts EMF or dBmicrovolts EMF: the US used microvolts PD, which is 6dB better on a spec sheet. Then the radar/ECM people came along, and we have dBm, which is just as much of a fallacy as microvolts PD. It's the power that would be delivered to the receiver if the receiver had a 50 ohm input impedance. If you're designing receivers to be used froma non 50 ohm source (such as MF/LF rx's), it's pretty meaningless. For 2MHz maritime stuff, we used 10 ohms in series with 250pF as the source impedance, and that makes a lot of difference compared with 50 ohms!
|
| |
|
RE: Designing the Ultimate HF Front End
|
|
|
by WA2JJH on July 2, 2009
|
Mail this to a friend!
|
Do not believe the hype or the sequal. One of the more technical OM's believed the Yeasue;s claim of TS-2000.
The ad said of the same proud linage of the FT-9000DX.
The $2000 radio was nothing but trouble. Even after 3 software upgrades. The AGC hsd one bloddy set of lungs!!! It breathed. He tyried ever tweak-a tweak!
Funny...It is supposed to be a software controlled radio!!!
He finally got it to work by addingn a mil spec stande alone audio DSP, and putting the audio through his stereo!
I asked if had a second proc to sell. He said, if you have to ask how much....you cannot afford it! hi-hi
Actually it was some rare prototype he was involved with.
Back to front ends. The K2 also uses a grounded base bipolar transistor as it's pre-amp. It is there just to get unity gain after going through the "Ham Band" only lossier filters.
I do love my Kenwood TS-950SDX. The Kenwood of old, always had great audio TX. On RX...filter blow by was a problem with steong sigtnals around. The TS-930 had the pre-amps wired wrong!
I see a tight roofing filter is offered foe the TS-950sdx. Heck the radio comes stock with a total of 6, 8 pole filters. Perhaps one more wont hurt!
|
| |
|
RE: Designing the Ultimate HF Front End TNX!!!!!
|
|
|
by WA2JJH on July 2, 2009
|
Mail this to a friend!
|
I must say this was a nice technical exchange. TNX Fer artical. Tnx to all those "across the pond"
Lots of grest info. No Bickering!!! :)
Hate to be the last one to coment. MORE PLSE!!!!!
73 DE MIKE WA2JJH
"
|
| |
|
RE: Designing the Ultimate HF Front End TNX!!!!!
|
|
|
by ZENKI on July 2, 2009
|
Mail this to a friend!
|
"I must say this was a nice technical exchange. TNX Fer artical. Tnx to all those "across the pond"
Lots of grest info. No Bickering!!! :) "
All we need now is a well defined set of minimum specifications for both the receiver and transmitter. Maybe we could have 2 sets of specifications, the bare minimum and a set of specifications that defines state of the art practice and technology.
Zenki
|
| |
|
RE: Designing the Ultimate HF Front End TNX!!!!!
|
|
|
by G3RZP on July 3, 2009
|
Mail this to a friend!
|
Now if we clean up our transmitters, what then? I suspect that once you get passed the -50 or -60dB IMD levels, other things will come in to play. like IMD in connectors, joints in antennas etc. Probably find the Steppir with its sliding contacts produces some. Bear in mind the problems at cell sites, and how they have gone over to DIN 7-16 connectors, rather than N types and so on. there comes a point where it's just not economic to go further. Like all the environmental reductions.
Years ago, Plessey Microwave Materials here in the UK sold 'anti-intermodulation grease'. I kid you not. It was a very expensive conductive grease, formulated to withstand sea water, and was used on bits of warship rigging and so on to avoid IMD from the 'rusty bolt effect'. Now Plessey MM doesn't exist, I don't know what the Navy do - there's not enough of them these days to make it worthwhile, and they don't have enough money to go to sea that often. I guess that if we could interest a few CBers in such a product, a fortune could be made.....
Meanwhile, I'll chug along with my old FT102. It's quite adequate in terms of RF performance, even if it doesn't have lots of memories and bells and whistles. It is repairable, which is more than can be said of a lot of modern rigs with surface mount, ball grid array ASICs and lead free solder growing whiskers.
|
| |
|
RE: Designing the Ultimate HF Front End TNX!!!!!
|
|
|
by SM0AOM on July 3, 2009
|
Mail this to a friend!
|
I think that we still have some way to go until
"passive IM" limits the HF performance for amateur operations.
The "topside" or "rigging" noise that the anti-intermodulation grease was a remedy against can be one of the limiting factors when operating several 1 kW HF transmitters onboard a naval vessel, but it appears from recent studies [Steve Hubbard: "A high dynamic range VLF to HF active receive whip antenna", Proc. HF04] that transmitter noise sidebands and IM in the receiver and antenna input circuits still are the main limiting factors.
My firm belief in the matter is that making transmitter performance AND operator knowledge better is the way to go. There is an annoying trend among some HF operators to turn "every knob to the right" and use excessive power even for local contacts.
This was the subject some years ago of a discussion at the SSA web-forum, where this was posted: "a lot of new operators have as their single goal to purchase an HF rig, a compressor microphone and a kilowatt linear..." to which I responded: "when this category also invests in an antenna, we will have an even worse problem." Which netted me a few nasty e-mails and phone calls, but also some positive feedback.
Following commercial practice, the elimination of the "mic-gain" control may be one of the steps towards cleaner signals. Too many operators have no understanding of the results of overdriving an SSB transmitter in terms of adjacent channel emissions.
Also, returning the ALC to the purpose it once was intended for, avoiding overdrive of the PA stage, may also make for less "splatter". The current amateur rig practice using ALC both for increasing average power, limiting and power control can be very contraproductive, and it escapes me why TGC is not used more for level setting and power control in amateur gear. The proper use of TGC would eliminate a lot of the "ALC modulation" sidebands that plague some transmitters.
73/
Karl-Arne
SM0AOM
|
| |
|
RE: Designing the Ultimate HF Front End TNX!!!!!
|
|
|
by G3RZP on July 3, 2009
|
Mail this to a friend!
|
Karl-Arne,
The trouble with ALC is that people expect it to do things it really shouldn't be used for. Now part of this may also be due to the fact that manufacturers look on it as the total gain control, many operators do not understand how the tx works and what the various control loops or controls do. Long before I had an SSB rig (or was even in the industry!) the RSGB Bulletin carried enough info to understand basically what an SSB tx did, and how it did it. (Although I never really understood Third Method!). I suspect the 'turn the knobs to the right brigade' have no understanding of what happens: maybe that's why manufacturers simplify the rig in terms of controls. But there's no reason why power can't be set by a set of switched attenuator resistors in an automatic tune up sequence, rather than using ALC with its inevitable overshoot. The Marconi MST 30kW tx of 1964 used a servo driven pot in an RF attenuator for fine setting of output power. ALC should be the thing that catches the occasional peak exceeding the ratings, not operating all the time.
Maybe we're too old fashioned!
|
| |
|
RE: Designing the Ultimate HF Front End TNX!!!!!
|
|
|
by SM0AOM on July 3, 2009
|
Mail this to a friend!
|
It appears that dealing with professional HF practices
makes you somewhat "spoiled" and critical.
The servo-controlled pot for level setting in the Marconi or Telefunken "self-tuning transmitters", or the photo-electric attenuator in the SRT gear did its duty admirably, and I see no reason why a similar approach should not be used in amateur radio gear.
For production cost reasons it could be realised as a software controlled attenuator
somewhere in the TX signal path before the pre-driver stage, which is set by a software routine at the first transmission after a frequency change.
With this approach, the ALC could have a quite limited range, and be there to avoid "hitting the ceiling".
Using this approach, and VOGAD instead of the "mic gain"-control, amateur SSB could more often be so nice sounding as a newly adjusted Collins 618T or a Marconi H1141 ISB ground transmitter.
73/
Karl-Arne
SM0AOM
|
| |
|
Designing the Ultimate HF Front End
|
|
|
by KD7RDZI2 on July 3, 2009
|
Mail this to a friend!
|
In my view there are several ways to get a good receiver able to cope strong signals such as BCI.
I will comment on some receivers I have and the one that I decided to build myself.
Drake R8 (built in 1994): passband filters, RF ampli or attenuator, 1st mixer, if ampli, 45Mhz filter, 2nd mixer with image rejection, LC filters at 50Khz.
I like this architecture. Using wideband antennas rarely overloads. The receiver does handle very well strong signals. However the if filter at 45Mhz is wide enough to receive narrow FM, i.e. 12Khz. The first IF filter is 12Khz wide. During contests with very close strong and weak signals the weak ones tend to disappear. Using the AGC OFF the R8 does a little better.
Drake 2C (built in 1967): preselector, preampli, 1st mixer with if at 3.5-4Mhz, 2nd mixer 455khz, 3rd mixer IF at 50Hz, LC filters.
There is no roofing filter but the circuits are tuned so if you look at the spectrum there is some filtering already after the first mixer. The receiver does not overload and has good ability to handle out of the band strong signals such as broadcasters. During contests I would find that the performance to hear weak signals close to the strong ones could be far better.
Using AGC OFF and using an external mixer instead of the 2nd at 455khz or in place of the 3rd at 50Khz converting it at about 10-12Khz and connecting it to the PC for IF-DSP filtering and demodulation (using DREAM or WINRAD) I think I found a DX machine that might be usable for contests as well.
Cheap portables such as the RP2100 (built in 2006) or the Degen DE1103 (built in 2004) have low pass filter, an up conversion at about 55Mhz, a filter and a second IF at 450 455 Khz. They don't have passband filters, and because of this, they tend to overload. Using a very good preselector and using an external mixer similarly to the mod I did for the 2-C, I do not notice a significant difference between these cheap portables and the R8 and the drake 2-C. I have the impression however that during contests the signals are clearer on the modified Drake 2-C. In my view there is no much difference using a 70$ dollars receiver with a good preselector and a 1000$ receiver.
The architecture of my homemade HF RX is very simple and I believe is very low noise:
- good preselector Seifert PSE61, with 16db preampli and attenuator from 6 to 40 db (the attenuator is never needed);
- 1st 50 ohm passive mixer with IF at 10.7Mhz; Low noise DDS
- 1:8 impedance transformer
- Roofing filter
- 8:1 impedance transformer
- 20db 50 ohm ampli
- 2nd 50ohm passive mixer with IF at 12Khz. Another DDS is used
- 1:100 audio transformer
- Soundcard with DREAM or WInrad Software for demodulation and IF-DSP filtering. The AGC is done by DREAM and WINRAD.
Performance is great in all cases. Even using cheap portables outstanding results can be obtained. Personally I think that the modified drake 2-C and my homemade RX are the solutions that I prefer. Contrary to most modern receivers, they have one thing in common: they have a downconversion... just like the Elecrafts and the Orion...
73s,
KD7RDZI2
|
| |
|
RE: Designing the Ultimate HF Front End TNX!!!!!
|
|
|
by GM3SEK on July 4, 2009
|
Mail this to a friend!
|
G3RZP wrote:
"The trouble with ALC is that people expect it to do things it really shouldn't be used for. Now part of this may also be due to the fact that manufacturers look on it as the total gain control"
Yes - many rigs use the ALC loop to implement the front-panel RF PWR control. This requires too much gain in the ALC loop, leading to overshoots, clicks and splatter.
Measurements by SM5BSZ have shown that the receiver front-ends of most transceivers have improved to the point where our transmitters are now BY FAR the dominant cause of interference involving strong signals.
http://www.sm5bsz.com/dynrange/dubus205/dubus205.htm
As SM5BSZ, Peter and Karl-Arne all agree, the peak output power should be controlled by a combination of speech processing and a preset level of TX gain. The role of ALC should be only to handle small overshoots due to re-peaking, so only about 3dB of ALC control range is required.
Peter said: "Maybe we're too old fashioned!" but we are only repeating what the Collins engineers knew, understood and published 30 years ago. It's the commercial transceiver designers who still haven't caught up.
Transmitter IMD in amateur transceivers has actually grown worse over the years, because the design teams have been focusing too much on receiver performance... and so too have the equipment reviewers, who have accepted poorer and poorer IMD figures without comment.
Peter added:
"many operators do not understand how the tx works and what the various control loops or controls do."
Quite so... but designers now have the potential to do something about this, by programming the TX to calibrate itself and store the correct gain settings for each band and mode. In other words, they can now build some Artificial Intelligence behind the control panel, if the real thing is lacking out front!
There are also sneakier methods. Our club had one of those "all controls fully clockwise" people - a nice guy, and a good contest operator, but he couldn't be trusted to leave things alone. So we disconnected the RF PWR and MIC GAIN controls, and replaced them by trimpots hidden inside the rig. Sure enough, the next morning we found both controls turned way high... and he'd never even noticed.
73 from Ian GM3SEK
|
| |
|
RE: Designing the Ultimate HF Front End TNX!!!!!
|
|
|
by G3RZP on July 4, 2009
|
Mail this to a friend!
|
I suspect that in the final analysis, assuming that high order products are proportionally lower than lower order ones, about 50 to 60dB on PEP for 3rd order is probably all you can practically use. It's not common, in my experience, to find amateur signals 60dB above noise on HF, and there is an averaging effect comes in on voice, too.
VOGADs have a problem if there's too much range on them - stop talking and the kids yelling next door, the dogs barking and the neighbours mowing the lawn, plus the blower for the amplifier, all drive the tx to full PEP!
|
| |
|
RE: Designing the Ultimate HF Front End TNX!!!!!
|
|
|
by QRPNEW on July 4, 2009
|
Mail this to a friend!
|
The only hope for an intelligent transmitter with perfect specifications will come from one of the SDR makers like Perseus, FLex or Phil Covington. The current players including Tentec, Elecraft and most of the Japanese manufacturers are just not interested in the subject of perfect transmitter design. They all believe that they are delivering perfect radios! If Yaesu or Icom think that they have perfect radios with receivers with such substandard IMD dynamic range numbers, what chance do you think you have of convincing them that their transmitter design techniques are flawed?
Considering that something like the Perseus costs $1500 dollars and it has some of the best
receiver specifications that one can buy, I dont believe it would be that hard building a separate SDR transmitter for much more money. In reality a perfect transceiver and transmitter should not cost more than $5000 dollars. Only SDR designs will deliver the kind
of transmitter performance that is required. The days of super white elephant radios like the Yaesu FT9000 are well and truly over.
Flexradios new QRP radio the Flex1500 would have been the ideal platform for a perfect transmitter station. However this radio does not look like that it will have an ideal set of transmitter specifications, it will just be another cheap QRP radio thats new. The flex would have been good candidate to use with the Perseus receiver. It would have been an easy task to build a ultra-linear external PA for this Flex design assuming its specifications were ideal. Why would you bother when you know that the transmitters IMD performance and ALC control system is far from ideal?
Why is it so hard for ham manufacturers to build great IMD performance at low QRP power?
There are so many great low power 100 to 500mw commercial transmitters that could have been used
as a model for this ideal transmitter, I am thinking of transmitters like the Rockwell HF8014A. The military manufacturers had the right idea building ultra linear low power drivers. These drivers which were easy to keep clean and filter out spurious products at low power keeping the costs low long before they reached the final PA.
The current manufacturers have to much vested interest in their design techniques to throw them out and start with a fresh professional approach. ADAT appears to have done this, however much of their work has not been reviewed, so who knows if its pie in the sky marketing hype that may never reach most of us?
Anyone seen a review for the ADAT radio?
|
| |
|
RE: Designing the Ultimate HF Front End TNX!!!!!
|
|
|
by QRPNEW on July 4, 2009
|
Mail this to a friend!
|
The only hope for an intelligent transmitter with perfect specifications will come from one of the SDR makers like Perseus, FLex or Phil Covington. The current players including Tentec, Elecraft and most of the Japanese manufacturers are just not interested in the subject of perfect transmitter design. They all believe that they are delivering perfect radios! If Yaesu or Icom think that they have perfect radios with receivers with such substandard IMD dynamic range numbers, what chance do you think you have of convincing them that their transmitter design techniques are flawed?
Considering that something like the Perseus costs $1500 dollars and it has some of the best
receiver specifications that one can buy, I dont believe it would be that hard building a separate SDR transmitter for much more money. In reality a perfect transceiver and transmitter should not cost more than $5000 dollars. Only SDR designs will deliver the kind
of transmitter performance that is required. The days of super white elephant radios like the Yaesu FT9000 are well and truly over.
Flexradios new QRP radio the Flex1500 would have been the ideal platform for a perfect transmitter station. However this radio does not look like that it will have an ideal set of transmitter specifications, it will just be another cheap QRP radio thats new. The flex would have been good candidate to use with the Perseus receiver. It would have been an easy task to build a ultra-linear external PA for this Flex design assuming its specifications were ideal. Why would you bother when you know that the transmitters IMD performance and ALC control system is far from ideal?
Why is it so hard for ham manufacturers to build great IMD performance at low QRP power?
There are so many great low power 100 to 500mw commercial transmitters that could have been used
as a model for this ideal transmitter, I am thinking of transmitters like the Rockwell HF8014A. The military manufacturers had the right idea building ultra linear low power drivers. These drivers which were easy to keep clean and filter out spurious products at low power keeping the costs low long before they reached the final PA.
The current manufacturers have to much vested interest in their design techniques to throw them out and start with a fresh professional approach. ADAT appears to have done this, however much of their work has not been reviewed, so who knows if its pie in the sky marketing hype that may never reach most of us?
Anyone seen a review for the ADAT radio?
|
| |
|
RE: Designing the Ultimate HF Front End TNX!!!!!
|
|
|
by G3RZP on July 6, 2009
|
Mail this to a friend!
|
|
Very good IMD at 100 or 200 or even 500mW is one thing. AT 1000 or 1500, it's another. It would need either a Polar or Cartesian Loop, and even they have close in noise problems at the levels of performance we're talking about. Maybe an active adaptive predistortion system might cope, but I don't immediately see how to do that.
|
| |
|
RE: Designing the Ultimate HF Front End TNX!!!!!
|
|
|
by W8AAZ on July 7, 2009
|
Mail this to a friend!
|
|
Lovely discussion, but what will this ultimate perfection cost me? Well we will wait and see, perhaps some day this technology will trickle down somehow into affordable HF rigs for hams like me, but meantime the bands are rotten and the QRM is poppin'
|
| |
|
RE: Designing the Ultimate HF Front End TNX!!!!!
|
|
|
by G3RZP on July 8, 2009
|
Mail this to a friend!
|
|
Unless some method of adaptive pre-distortion can be produced, I don't see the cost of the highly linear PA stages dropping at all.
|
| |
|
Email Subscription
You are not subscribed to discussions on this article.
Subscribe!
My Subscriptions
Subscriptions Help
Related News & Articles
HF Mobile Noise Abatement
Making CW More Popular... A Proposal:
Which HF Transceivers are 'Perceived' to be Superior:
Ham Radio's Unique Selling Point?
Which HF+ Transceiver is Best for Emergency Use?
Other Radios Articles
Cool Radio
Boatanchor AM on a Budget
Keep Your Cool!
SDR: Ham Radio of the Future?
Classic Rigs
|
