Call Search
     

New to Ham Radio?
My Profile

Community
Articles
Forums
News
Reviews
Friends Remembered
Strays
Survey Question

Operating
Contesting
DX Cluster Spots
Propagation

Resources
Calendar
Classifieds
Ham Exams
Ham Links
List Archives
News Articles
Product Reviews
QSL Managers

Site Info
eHam Help (FAQ)
Support the site
The eHam Team
Advertising Info
Vision Statement
About eHam.net

donate to eham
   Home   Help Search  
Pages: Prev 1 [2] 3 Next   Go Down
  Print  
Author Topic: noise from DDS VFO  (Read 10995 times)
AJ4SN
Member

Posts: 69




Ignore
« Reply #15 on: June 13, 2017, 06:15:42 PM »

Egmont,

You are correct, the crystal filter precedes the IF amp. The IF amp is the hybrid-cascode designed by W7ZOI. I was under the impression that it is a decent IF strip. Are there some deficiencies with the design? I'm always interested in improving my receivers.

Stan
Logged
N3QE
Member

Posts: 4882




Ignore
« Reply #16 on: June 13, 2017, 07:31:45 PM »

You are correct, the crystal filter precedes the IF amp. The IF amp is the hybrid-cascode designed by W7ZOI. I was under the impression that it is a decent IF strip. Are there some deficiencies with the design? I'm always interested in improving my receivers.

The hycas is a good clean design.

You are right, the weak link in your setup is the DDS VFO.

I am curious, do small changes like a single step frequency change in the DDS VFO cause a complete change in the buzzes and whines?

If so, you are getting the spurs.

You could help clean it up a lot, by getting, say, a 20MHz output from the DDS VFO, filtering it up there with a 20MHz bandpass, and dividing by 4 followed by a simple low pass to get to 5MHz. This may clean up many (but not all) of the worst spurs you hear. I think a PLL tracking loop could have similar benefits too.
Logged
KB1GMX
Member

Posts: 1496




Ignore
« Reply #17 on: June 14, 2017, 09:57:03 AM »

Likely the noise being detected is not the DDS fault.  DDS even the lowly 9850 has how jitter (unless its reference is junk).
It will have spurs but those tend to be better than 60db down especially int eh under 20mhz realm.  For a 5-5.5mhz VFO it can be very good.

So based on that (and experience using the 9850) if you getting noises and beeps its entirely likely you have oscillation in the
amplifies(s) used to get to 16dbm (Tak3H is a high level DBM using a nominal 17dbm drive).  You may also have  the post mixer  amp doing unstable things as well as the HyCas (the published design is over 100db of gain, do you really need that much?).  Its very easy to combine stable components to get an unstable result if care, shielding, proper grounding, and bypassing are not taken into account.

One topic not handled separately but is part of the text in EMRFD is that shielding is part of any design of a high performing receiver.  That applies to many of the RF circuits to keep them from picking up signals that produce unwanted results.  Keep in mind that a 16dbm signal for the mixer is a 50mW signal and even short leads will easily radiate enough that a stage a few inches or less away will hear it.  If your serious  every stage should be in a box with 50 ohms in and out and those boxers in the bigger box (case) and signals between them that are not DC shielded and bypassed as needed.

Using a high level mixer requires a fair degree of care in the overall design as the leakage products are significant.  Also its alone does not guarantee a good receiver (resistance to overload/intermod) as everything downstream but prior to the selectivity (crystal filter) must be able to handle big signals or bad things result as mentioned.

In this case lacking a lot of info but having used the 9850 DDS blaming it for the "noises" is likely blaming the cow for soggy cereal.  Its likely not the cause but may be a symptom of many combined design or assembly faults.

FYI the DDS has a typical output of 200Mv into a 200 or so ohm load so most DBM require amplification and
should be followed by a band pass or low pass filter as appropriate.

The HyCas amp is a good design but care must be used as wide open gain is 100-120DB, that much gain must be respected as oscillation can be the result.  I've used it and it works well but its input does not properly load a filter without some form of matching network. (L or tuned transformer).

High level mixers (tai-3H or SRA1h) need to be followed with a very robust post amp with the ability to handle large signals as LO leakage alone will be around -30dbm (assume a 45db Lo suppression in the DBM) and can easily be larger.  Overload in this stage will be fatal to good performance. 

Allison/kbqgmx

Logged
AJ4SN
Member

Posts: 69




Ignore
« Reply #18 on: June 14, 2017, 08:24:40 PM »

Allison,

Thanks for your reply. I hadn't thought about how much power 17 dBm represents. 50 mW almost rises to the level of a QRP transmitter! I see what you mean about taking care with the design. I didn't set out to build a receiver with high dynamic range (I'm not nearly technical enough to do that). I just found the TAK-3H mixers on the cheap, and I read an article about the hycas IF. Just a point of clarification, the receiver has worked well with the analog VFO for a several years. I've tried to include a picture of the inside of the receiver with the IF shield off. It should also be on my profile page. You motivated me to update my profile with the latest gear! I'm going to look at the amp that I built for the DDS to see if it might be oscillating, and I'm also going to use the receiver with the LO drive level reduced. I think it should work ok, although there may be more loss in the mixer. I probably have more gain than I need anyway. By the way, I love the expression, ". . . blame the cows for soggy cereal." I think I'm going to use it at work!

73,

Stan

Logged
G4AON
Member

Posts: 1021




Ignore
« Reply #19 on: June 14, 2017, 11:42:07 PM »

Stan, if that open board includes your Hycas IF, it could do with being completely shielded. A Hycas board is sensitive to sub microvolt signals and maybe why you hear some of the DDS noise.

By the way neat assembly!

73 Dave
Logged
G3RZP
Member

Posts: 8123




Ignore
« Reply #20 on: June 15, 2017, 02:11:27 AM »

Quote
Also its alone does not guarantee a good receiver (resistance to overload/intermod) as everything downstream but prior to the selectivity (crystal filter) must be able to handle big signals or bad things result as mentioned.

You need to keep gain down prior to any crystal or mechanical filter, because they produce intermodulation products, and it is relatively easy to have a receiver front end where the IMD performance is dominated by the filter. On occasion, IMD performance can be improved by exchanging the input and output connections, as the filter may well not be symmetrical as far as IMD is concerned. Also, high frequency (e.g. 9MHz) filters are generally worse on IMD than lower frequency e.g. 1.4 MHz filters. Some 455kHz mechanical filter had intercept points as low as -12dBm.
Logged
YL3GND
Member

Posts: 55




Ignore
« Reply #21 on: June 15, 2017, 04:11:57 AM »

Stan, if there is a problem not only associated with spurs, but also noise (lowered sensitivity), than You have equation with 3 multipliers:
1.   DDS noise itself
2.   LO leakage to IF (including noise, depends on mixer construction)
3.   Given level of LO
Apparently the 4th one also counts: IF sensitivity (from the output of mixer), but let’s assume it’s a constant.

This is why I was asking about the circuitry (it’s pity that on eham you can’t put pics directly..) – may be there is no point in putting in +17dBm mixer. Instead if You would’ve been using +7dBm one with a similar LO isolation, You would get the noise 10 dB down.

I mean, if there is some circuitry in the front end that ruins dynamic range more than a mixer, there is no point in particular mixer.

Could You, please, tell (better draw also!) more about your front end circuitry? What is the sensitivity without a preamp? Does preamp exists only for to use “rarely only on 10m band” or..?

Egmont
Logged
YL3GND
Member

Posts: 55




Ignore
« Reply #22 on: June 15, 2017, 04:45:38 AM »

Just acknowledged there are some more valuable thoughts in there – esspecially from Allison KB1GMX, about blaming the cow for soggy cereal. That’s on spot!

About cascode amps (like hycas): I am using one (with 2bjt’s) myself, but only where signals never grown big – in the 80m band ARDF receiver with 6” diameter loop antenna. Where there are stronger signals (even inband ones, after crystal filter) it is counter intuitive to use it: the stronger the signal, the less linear is amplifier due to lowered current by applied AGC.

As I got, Stan, You do not use post mixer amp?

Egmont
Logged
N3QE
Member

Posts: 4882




Ignore
« Reply #23 on: June 16, 2017, 03:19:20 AM »

About cascode amps (like hycas): I am using one (with 2bjt’s) myself, but only where signals never grown big – in the 80m band ARDF receiver with 6” diameter loop antenna. Where there are stronger signals (even inband ones, after crystal filter) it is counter intuitive to use it: the stronger the signal, the less linear is amplifier due to lowered current by applied AGC.

Well then just use 1 or 2 (not 3) stages in the hycas.

A lot of us just habitually put a 20dB or 40dB attenuator on whenever we switch to 80M anyway. I work more countries every year on 80M, than just about anyone else in North America, and I work them all with the attenuator on :-).
Logged
KG7TR
Member

Posts: 1




Ignore
« Reply #24 on: October 05, 2017, 09:05:43 PM »

Everything N3QE stated in his 12 June 17 responses is absolutely true.  I have been experimenting for over a year with a vacuum tube homebrew receiver that started out as a clone of a Collins 75S-3. The original plan was to use Arduino controlled AD9850 DDS modules to replace the first crystal heterodyne oscillator, PTO (VFO), and BFO.  The rest of the circuitry is all vacuum tubes. I used LMH7603 current feedback op amps to get the oscillator signals up to a couple of volts RF to inject into the cathodes of the receiver's mixers, just like the 75S-3.

When I finally got all the circuits built and working results were most disappointing, with birdies and digital sounding spurs all across the spectrum. Some were up to an S3 or S4 level of annoyance. The scope patterns from the AD9850/LMH7603 chips looked like beautiful sine waves, but a little more internet research confirmed my suspicions that there was some low level junk getting into the IF. The first LO/first mixer and PTO/second mixer are followed by two IF amps with a combined gain of at least 80db. The spurs are perhaps only 50-60 db down from the oscillator carriers, so they come blasting right through.  Next I tried si5351 PLL modules, and results were not any better.

These chips are remarkable for what they can do, but they were never designed for sensitive HF receiver applications.  I finally settled on a crystal controlled first oscillator using cheap microprocessor crystals, and an analog tube type VFO.  The BFO does use an Arduino controlled si5351 module followed by an LMH7603 amplifier.  At that point the spurs are too low to make any difference. The radio is working pretty well now. The only birdies are due to harmonics of the VFO, and they are equivalent to less than .1 microvolts at the antenna.

I really wish I had done more research ahead of time, but I did learn a lot in the process.  I echo other's sentiments that these chips might be OK for QRP radios, but not for sensitive receiver applications or high power transmitters.
Logged
G4AON
Member

Posts: 1021




Ignore
« Reply #25 on: October 06, 2017, 02:34:53 AM »

Next I tried si5351 PLL modules, and results were not any better.
<SNIP>
These chips are remarkable for what they can do, but they were never designed for sensitive HF receiver applications.
<SNIP>
I use an Si570 synth in my home brew RX and it's fine. They are also used in the Elecraft KX3 and K3S. The KX2 uses an Si5351. So I doubt they are not suitable for sensitive HF receivers.

I suspect the issue is one of shielding and/or effective decoupling. My home brew RX could do with better shielding but it's adequate. Having a decoupling capacitor only half an inch from a ground point makes quite a big difference when dealing with signals in the +10dBm region, as does assuming a decoupling capacitor must be OK without checking the RF level across it (a 100nF ceramic cap might not be as effective as you think).

On the other point about the Hycas board design, mine works absolutely fine although I've done a couple of tweaks... R6 is shown on the circuit as being 270 Ohm, with that value high ambient temperatures cause the AGC Voltage to drift. Using 220 Ohms stops that drift. There are some adjustments to the overall gain that can be done easily with the damping resistors across the RF chokes, W7ZOI mentions this on his site and details are also on my site: http://www.qsl.net/g4aon/g4aon_rx/

73 Dave
Logged
YL3GND
Member

Posts: 55




Ignore
« Reply #26 on: October 07, 2017, 05:41:44 AM »

Very interesting, Dave! About KX2 using Si5351, have not heard this before. Also I am interested in K3S LO circuitry (contemporary Elecraft is hiding their circuitry) - does it have PLL loop with Si-570, or it goes straight out of the chip?

Egmont
Logged
YL3GND
Member

Posts: 55




Ignore
« Reply #27 on: October 07, 2017, 06:58:18 AM »

About IF amps: As I stated earlier, its basically a bad technical idea (only positive - easy to implement) to steer amplification down by reducing current in amplifier's active device. That's why better designs are using passive attenuators between amplification stages.
I am using one MC1350 on one of my portable rigs myself  Tongue , I would never state it's a great IF amplifier, despite it does what I need to be done in this rig.

N3QE: absolutely wrong - its better to use more such a bad IF amps in a string, as than You don't need to shut them completely down for larger signals.

Recently I came across pirate copy of "EMRFD" book of W7ZOI and others. What he wrote in chapter 6.2 about IF amps is erroneous at times. Like statement in the page 6.17 : "However intercept degrades slower than gain, so IMD products are allways decrising with gain reduction". The numbers given by himself tells us the opposite.

A lot of fairy tails not backed by measurements is circling around.

Egmont 
Logged
G4AON
Member

Posts: 1021




Ignore
« Reply #28 on: October 07, 2017, 07:13:17 AM »

Egmont, the K3S circuits are published on the Elecraft site.
http://www.elecraft.com/

Regarding EMRFD, several circuits aren’t thought through. There is a 2 stage basic amp that even I was able to redesign to have a flat frequency response, plus a few others. Spice modelling works wonders!

73 Dave
Logged
G3RZP
Member

Posts: 8123




Ignore
« Reply #29 on: October 07, 2017, 03:52:12 PM »

Over 40 years ago, I did a design for a 1.4 MHz  IF amplifier using discrete transistors (2N5179 I think) , in which AGC was applied using a PNP transistor in series with the emitter bypass capacitor. By taking the base positive, the PNP transistor was turned off, the negative feedback increased and the IMD performance improved as the gain was reduced. For the life of me, I can't remember why we didn't end up using it in production.

The usual way in ICs was to use a long tailed pair and divert the current from one transistor to the other. An improvement over just reducing the current in a single transistor, but not ideal.
Logged
Pages: Prev 1 [2] 3 Next   Go Up
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.11 | SMF © 2006-2009, Simple Machines LLC Valid XHTML 1.0! Valid CSS!