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 4 5 [6] 7 8 9 10 11 ... 18 Next   Go Down
  Print  
Author Topic: Designing & Building a High-Peformance Subminiature-Tube Regenerative Receiver  (Read 177345 times)
VE3LYX
Member

Posts: 814




Ignore
« Reply #75 on: May 14, 2015, 07:07:17 PM »

not there for gain  (never hurts) qte
Actually it does. The absolute No 1 reason many (100s of) regens dont work or perform properly is too much signal either by an RF amp or too much antenna. There is a concept the a regen is a step down from modern rxs in its ability to hear. Actually it will give even the best of them a run for the money. It is however a difficult radio for some to run and I have met folks who simply cannot operate one.  
However the other comment building it to see what happens I wholeheartedly endorse. Famous engine builder Smokey Yunick once said "One good experiment beats 1000 expert opinions." And that applys here as well. Go for it. Be honest with your findings (don't lie to yourself) and have fun. Don't be afraid to retrace your steps should you find yourself down a path of no potential.
don
« Last Edit: May 14, 2015, 07:26:27 PM by VE3LYX » Logged
KB1GMX
Member

Posts: 1829




Ignore
« Reply #76 on: May 14, 2015, 07:53:34 PM »

Yep, after years of building them I still do because of that bang for the buck.

The key factor for RF amp is allowing the regen to be loose (barely) coupled
to the RF source.  Then the net gain is small but isolation is a useful attribute
as then the antenna size or tuning is then not a factor.   Bad design or failed
experiment does not mean its wrong.

The other critical factor is still the fact that we already have enough RFI and a regen
oscillating or on the verge will radiate.  Granted its low power but around here there
are people within less than a mile of me and a milliwatt does carry even on a short
piece of wire.

The other way I like to play with them is running the detector at say 1700-1800khz
fixed a simple effective IF and use a tunable converter ahead of it.  You get the
selectivity advantages or regen and stability while being able to tune upper HF.

Also Smokey would have loved CADD and modeling, alas they were in the future,
but he could experiment now.

In the end, after more than 50 years of playing with them, when I need a receiver
I do think about it and then remember why it was interesting but not what the task
called for.  I still remember trying to get WABC NY 770khz 40 miles away with
WGSM 740khz less than a mile away.  No regen was good enough to sort that out
and I tried all I came across. I would repeat that process with transistors starting
around '65. The first successful radio that could was a regen at roughly 455khz
with a converter before it and a audio amp after (6sa7 and  6SN7).

Three of those 1ad4s could easily be a superhet (self oscillating converter,
regen if, audio amp).  Its the next step in receivers.   Look at what AA1TJ has done.

With regens there are many dead ends and many pockets where they shine. 
Its appeal is why after nearly century we still use them both commercially and
experimentally.  There are places where they fit though most of those are technically
superregens.  They were the building block leading to modern receivers and
often a part of them.  Have fun and experiment, but do not limit yourself.


Allison
Logged
KB1WSY
Member

Posts: 1309




Ignore
« Reply #77 on: May 15, 2015, 03:18:49 AM »

Three of those 1ad4s could easily be a superhet (self oscillating converter, regen if, audio amp).  Its the next step in receivers.   Look at what AA1TJ has done.

With regens there are many dead ends and many pockets where they shine. Its appeal is why after nearly century we still use them both commercially and experimentally.  There are places where they fit though most of those are technically superregens.  They were the building block leading to modern receivers and often a part of them.  Have fun and experiment, but do not limit yourself.

My next receiver will be a superhet. If anything they are easier to understand, because each stage has a single function!

I have the parts for the ARRL RS-3 (regenerative superhet). I also have the parts for the W6TNS "Novice Q5er" converter along with a beat-up old BC-453 to restore. Or I might branch out and build an entirely self-designed superhet with subminiature tubes ... based on previous designs of course and with a lot of help from you kind folks, if you can bear it. I've already learned an enormous amount of tube theory, just building the regenerative set. This is despite having built a fair number of things in the past; those were all either kits, or entire ARRL projects, so no "designing" was involved.

The first order of business however is to build the best regenerative receiver I can build.

What I am finding now is that I have a much better "feel" for the purpose of individual components, and an appreciation of appropriate component values -- while still committing numerous "clangers," some of them dangerous! Where I need a lot more knowledge is in some of the basic RF theory: impedance, reactance, and so forth. (I have read quite a lot about these things but it seems to go in one ear and out the other.) Theory and practice need to go hand in hand.

The other way I like to play with them is running the detector at say 1700-1800khz fixed a simple effective IF and use a tunable converter ahead of it.  You get the selectivity advantages or regen and stability while being able to tune upper HF.

In your opinion, are pure regens less suitable for upper HF? The first regen that I built (the ARRL 3-transistor set from 1968) works fine on 20m, although this did require a fair amount of optimization of the coil and capacitors (when I built it exactly as described by ARRL, I could not get it to oscillate on 20m, or alternatively it would motorboard wildly; and antenna loading was tricky).

When the basic regen circuit has been settled and tested, I will be adding a 20m coil (the antenna I'm installing this summer, in the limited space available, is a 40m/20m trap dipole). The plan is to use 6-pin coil forms in the detector, enabling all of the frequency-setting capacitors to be mounted inside the coil form. The RF stage will use 4- or 5-pin coils. What I will basically end up with is an old-fashioned "TRF" with a single RF stage and a regenerative detector. Once I've got all that working, I'm hoping to improve the audio side, adding more amplification, a gain control, and an audio filter for CW. I might even build a 100kHz crystal calibrator (some kind of routine calibration tool is pretty essential with a regen, especially a multiband one). Once all that is done and the design has been finalized, I'll build a "proper" version of this radio, using good construction techniques and in a more compact footprint with a nice tuning dial and controls.

(Also, at some time in the future, it would be cool to build a "minimalist" miniature set for portable use, probably just two tubes and squeezed into a small case. It would use lower-power, low-voltage 2E32s, which would require only a single AA battery for filaments and a couple of tiny A23s for the B+ or perhaps just a couple of larger 9V batteries. The main challenge is miniaturizing the tuned circuit ... but that's a subject for the other thread.)

By the way, based on my limited testing so far, the two-tube version of my regen (with no RF stage) out-performs my ARRL "3-transistor" regen on 40m. In particular, it has even greater sensitivity and smoother regeneration. However the comparison isn't really fair because an enormous amount of optimization and tweaking has been done on the tube set; in comparison, my ARRL build has only minor alterations compared to the published design.

I can also see a big advantage to using a regen as a fixed IF. A regen requires a lot of skill to optimize for each frequency, so once you have it "stable" (well, as stable as a regen can be!) it makes a lot of sense to leave it alone and use a converter to do the tuning!

not there for gain  (never hurts) qte
Actually it does. The absolute No 1 reason many (100s of) regens dont work or perform properly is too much signal either by an RF amp or too much antenna. There is a concept the a regen is a step down from modern rxs in its ability to hear. Actually it will give even the best of them a run for the money. It is however a difficult radio for some to run and I have met folks who simply cannot operate one.  
However the other comment building it to see what happens I wholeheartedly endorse. Famous engine builder Smokey Yunick once said "One good experiment beats 1000 expert opinions." And that applys here as well. Go for it. Be honest with your findings (don't lie to yourself) and have fun. Don't be afraid to retrace your steps should you find yourself down a path of no potential.
don

I agree about "too much signal." That's why it's useful to have various methods of attenuation available.

I've been retracing steps frequently ... sometimes it's frustrating because you get something working really well, then you try to "tweak" it and it gets worse, and then you can never quite get back to where you were before!!!

The cool thing about using these subminiature tubes with their own leads, and low voltages, is the ease of experimentation. One of these days I will smash one of them with a hammer so that I can admire the industrial construction feat (these tubes are still plentiful).

73 de Martin, KB1WSY
« Last Edit: May 15, 2015, 03:53:16 AM by KB1WSY » Logged
KB1GMX
Member

Posts: 1829




Ignore
« Reply #78 on: May 15, 2015, 05:39:51 AM »

I agree tubes seem to outperform bipolar trasistors and usually fets.  I believe
the transistor problems is way too much gain.  I've tried the three transistor
and it was not satisfactory. 

My feeling in regens is that HF selectivity is not adequate.  If the regen
is oscillating its a direct conversion receiver with the inherent limitation no
rejection of the opposing sideband.   

The RS3 is a fun radio.

Like you said if you break the problem of reception into a modular one then
optimizing becomes easier.

Allison
Logged
W1BR
Member

Posts: 4196




Ignore
« Reply #79 on: May 15, 2015, 05:43:53 AM »

[q
Edited to add a quick question: I dithered about the polarity for the 1µF electrolytic on the RF Gain wiper. I've now "fixed" it so that the positive plate goes to ground ... which I think must be correct since the bias is negative?

MDM

73 de Martin, KB1WSY


Also remember that the filament polarity affects the tube bias. 

Pete
Logged
KB1WSY
Member

Posts: 1309




Ignore
« Reply #80 on: May 15, 2015, 07:04:30 AM »

Today I used my lovely new LCR meter to get to the bottom of the "screwy tuned circuit" issue. I completely dismantled the tuned circuit and did the following tests. Note that when using the leads on the meter, I started out by zeroing out their capacitance (using the CAL button).

--The fixed mica capacitors, in isolation, all measured within their tolerance (i.e. plus or minus 5%).

--The variable capacitor had a maximum of 51.1pf and a minimum of 7.1pF. I had assumed 50 and 5 -- good enough.

--When I wired the fixed capacitors and the variable into a network, in isolation from the radio, it had a maximum capacitance of 53.0pF and a minimum of 50.0pF. My calculations had yielded 51.7 and 48.5. Again, good enough, and well within the tolerance of the components.

--I tested the tank coil, after disconnecting it from the radio. It is mounted on an Amphenol polystyrene form. It has 18 turns, a diameter of 1.28" including the diameter of the 20AWG wire, and a length of 0.75". It measured at 10.9µH with the (disconnected) 3-turn tickler mounted (close-wound, directly under the tank). After removing the tickler, the tank coil measured at a slightly lower 10.5µH (it was fun to get direct evidence of the effects of mutual inductance). According to Wheeler's equation, the coil should be 10.0µH. So again, "good enough" (the difference is probably partly caused by the inductance of the leads).

--Edited to add: The meter also shows that the tank coil has a Q of "0.55" -- I presume that is a ratio and that this would conventionally be rendered as "55".

Finally, I tested the capacitance between the input of the grid-leak detector and ground. At this point, the tuned circuit is dismantled, so there is nothing between that point and ground. The meter showed a capacitance of 20.6pF and I find this extremely interesting because it accounts for the screwiness I've been discussing: it agrees almost exactly with my calculation that there was something like 23pF 21pF of stray capacitance messing things up. But is the testing methodology valid? Surely the capacitance of the circuit is different when it is "powered down" and there are no voltages on the tube?

I believe the next step is to rebuild the detector stage entirely, using a more "linear" layout from left to right. I'll eliminate the "criss-cross" wiring and run a 14AWG copper-wire ground bus down the middle of the circuit (in the current build, my grounding points are haphazard). Maybe I'll finally discover a wiring error, despite checking my work several times already. I am determined to get to the bottom of this -- unless one of you experts thinks that a 20pF "input capacitance" for the grid leak is normal?

One possibility that just occurred to me is a defect in the tube, causing a large increase in inter-electrode capacitance. I took it straight from its NIB packaging and don't own a tube tester (and anyway, few tube testers are configured to test these little tubes).

Hmmm ... a fancy and expensive LCR meter just to confirm what I knew already! (Still: You can never have too many pieces of test equipment.)



73 de Martin, KB1WSY
« Last Edit: May 15, 2015, 07:16:43 AM by KB1WSY » Logged
KB1WSY
Member

Posts: 1309




Ignore
« Reply #81 on: May 15, 2015, 02:41:25 PM »

So I rebuilt the detector from scratch, but with the same circuit. This time I built it from left to right, without any arbitrary "wires crossing over" and generally with much shorter signal paths. I also used a "new" 1AD4 tube from my stock, just in case the previous tube was defective.

But when I powered it up, with the aforesaid 10.5µH coil and about 50pF of capacitance, it was more than 700KHz off (it was tuning around 6200KHz instead of around 6950KHz). This is almost as "bad" as in the previous build.

So what I've done is remove 16pF of shunt capacitance, so that there is now about 34pF across the coil. In that configuration, it is tuning the 40m band, starting at around 7000KHz.

The set is working nicely (as before) but I am quite disappointed not to have been able to identify/correct the stray capacitance. In the new configuration, when I measured the (powered-down) capacitance at the input of the grid leak (prior to connecting the tuned circuit) it was about 12pF.

It looks like I may have to write off this issue as one of the "mysteries of the universe" and just get on with building the RF stage now. Unless anyone has any other suggestions....

73 de Martin, KB1WSY
Logged
KB1WSY
Member

Posts: 1309




Ignore
« Reply #82 on: May 15, 2015, 04:55:00 PM »

On the positive side of the ledger:

I splurged and bought two low-current Hammond filter chokes, 60-Henry each.

I installed them as the plate loads in both the detector (replacing the small audio transformer primary that I was using as a load) and the audio amplifier (replacing an old 7H filter choke).

Substantial improvement!

In the detector, regeneration begins earlier and is spread over a larger range. There is substantially more control, especially over strong signals. Thankyou KB1GMX (at least I think it's him who suggested using a large-value choke in this application).

In the audio amplifier, there is substantially more gain (thank you G3RZP, who first suggested using a choke load in this stage).

73 de Martin, KB1WSY

« Last Edit: May 15, 2015, 05:07:22 PM by KB1WSY » Logged
KB1GMX
Member

Posts: 1829




Ignore
« Reply #83 on: May 15, 2015, 07:25:23 PM »

Martin,

Just passing on info I got back when from other old hands.  All of the best regens
tube or transistor had a choke or transformer to get high sensitivity though there were
a few that used active devices to create the same effect.

The reason it works is the plate resistance of tube is very high so the higher the
load resistance (or impedance) the greater the available gain.  the differnce is at 1khz
the choke looks like 370,000 ohm load (impedance)  without the DC resistance
where the resistor was 10K.  The plate resistance of the tube is about 500K so you
get most of the signal rather than loosing it in the tube.  To visualize it just think of
a AC voltage generator with 500,000 ohm resistor feeding the 60H coke VS 10 K resistor.
If the generator puts out 1V the choke has .425V across it, where the 10K resistor has
.0196V.  That is about 21.6 times more signal.

Since your into the experiment... you can break up a regen to a point in to modular sections...
Start with a gridleak detector than add another tube running as a Q-multiplier.
You can get circuits out of the Handbook.  That's  case for a externally excited regen (for AM)
and its run to oscillating for CW or SSB then its really a product detector or cirect conversion
receiver.


Allison
Logged
KB1WSY
Member

Posts: 1309




Ignore
« Reply #84 on: May 16, 2015, 03:49:43 AM »

More progress. Until now, I've been connecting the antenna directly to the "hot" side of the tuned circuit, via a "gimmick capacitor" (two wires twisted together, a couple of turns). I did this because in the previous build of this radio, I could not get separate antenna-coil coupling to work.

This morning, following yesterday's total rebuild and the addition of a choke load to the detector plate circuit, I had another go. I removed the gimmick capacitor, then added a 2.5-turn antenna coil about half an inch above the tank coil. I connected the bottom end of the coil to ground, and the top of the coil (the end farthest from the tank coil) to the antenna.

Success! Reception is good, and if I'm not mistaken it is a bit less susceptible to overload than it was before -- perhaps not surprising now that the detector is no longer directly connected to a 20-meter-long longwire antenna.

The other good news is that after making the above changes, I needed to add back to the tuned circuit about 6pF of additional fixed capacitance in order to bring the bottom of the frequency range down to about 7000KHz. The capacitance of the tuned circuit at the bottom of its frequency range is now around 41pf, which is "only" about 8pF "wrong" for my 10.5µH coil (instead of the previous figure of about 14pF).

I'm glad to say goodbye to the "gimmick capacitor" which always struck me as, well, a gimmick ... and a source of various "issues" including a narrowing in the tuning range.

I will now proceed with building the RF stage. I will leave the above-mentioned, loosely coupled "antenna coil" in place and will re-purpose it as the "plate coupling coil" from the RF stage, and see what happens. It may need to be adjusted, but it's a good start.

The reason it works is the plate resistance of tube is very high so the higher the load resistance (or impedance) the greater the available gain....[snip]

Thank you for the detailed explanation. The various types of load, and the methods of interstage coupling (resistor, choke, transformer) are explained in the literature but often rather opaquely: they tell you the merits and de-merits of each alternative, but often don't go into the "why."

I'm surprised that choke coupling wasn't more common. I suspect it was, in part, because high-value inductors were large, heavy and expensive (and they still are!).

Since your into the experiment... you can break up a regen to a point in to modular sections... Start with a gridleak detector than add another tube running as a Q-multiplier. You can get circuits out of the Handbook.  That's  case for a externally excited regen (for AM) and its run to oscillating for CW or SSB then its really a product detector or direct conversion receiver.

There you go again, giving me ideas ... this project is getting more Baroque by the day! For the moment I'll keep it narrow -- building the RF stage and optimizing the "basic radio" -- before branching out and adding more stages.

73 de Martin, KB1WSY
« Last Edit: May 16, 2015, 04:07:36 AM by KB1WSY » Logged
KB1WSY
Member

Posts: 1309




Ignore
« Reply #85 on: May 16, 2015, 05:40:01 PM »

Major surprises today.

I added the RF stage, and grid-biased it following G3RZP's instructions.

Huge improvement in the receiver's performance!

A reminder: this is all in the 40m band.

(1) Stability is now very good (it was already "good" but there is a still a major improvement now). No hand capacity effects (unless you put your hand "inside" the set near the detector coil). You can put your hand on the antenna wire and shake it, without audible effect. You can copy CW with no tone variation and SSB demodulation is great -- you can listen for several minutes without touching the dial.

(2) There actually is an improvement in usable RF gain (=somewhat improved sensitivity).

(3) The RF gain pot give good control over strong signals, and it has absolutely no very little "de-tuning" effect unlike the detector/regen controls. In fact it is so useful that I have changed the bias battery. It was originally 4 NiMh AA cells (about 4.8V) but I've replaced it with a Li-Ion "9V" cell (actually 7.4V) to give even more control over strong signals. (Edited to add: This modification is still under assessment. It adds only a small degree of additional attenuation.)

(4) In a change to the previously posted schematic, I made the RF stage tunable. By "peaking" the RF stage, substantial additional gain is available (if desired).Edited to add: I am not sure whether it is worth modifying the receiver to install a two-gang tuning capacitor and attempting to implement accurate "tracking." Additional bands will be added later -- 20m and 80m -- which is one reason to consider implementing "tracking," despite the difficulty of implementation.

(5) I am not totally sure about this, but there may also be a moderate improvement in selectivity.

By the way, the coupling in both the RF stage (antenna coil) and the detector (tank coil) is very loose. The antenna coil is a couple of turns about 1/2" 1" above the 18-turn RF-stage tank coil. The plate coupling coil between the RF stage and the 18-turn detector tank is also a couple of turns, 1/2" above. Is it possible that this, on its own, is improving selectivity?

Edited to add: I think I've figured out how the improvement in selectivity happens. With the gain-controllable RF stage, it is now possible to set regeneration to the threshold of oscillation (= most sensitive, and most selective, point on the dial) even when copying strong signals. This only works because, with the RF gain control, you can now attenuate the strong signals that otherwise, would be intolerably distorted or strong enough to cause "blocking." So: At the very least there is an improvement in selectivity when copying strong (or moderately strong) signals, and that is noticeable on a night like tonight with heavy CW contest activity in the lower part of 40m.

I will post an updated schematic once the current "tweaking" phase is over.

This set now has six controls (leaving out the power switches) from left to right:
--RF Peaking capacitor (small vernier dial).
--RF Gain pot.
--Main Tuning capacitor (large reduction drive/dial).
--Screen Regeneration pot.
--Detector Plate Voltage pot.
--Regeneration Throttle capacitor.
... and it doesn't even have an AF Gain control yet.

All the controls are useful, amazingly enough. The challenge is to select the right permutation for the signal at hand ... it has always been said that more skill is required to operate a regen, compared to a modern superhet.



Famous engine builder Smokey Yunick once said "One good experiment beats 1000 expert opinions." And that applies here as well. Go for it. Be honest with your findings (don't lie to yourself) and have fun. Don't be afraid to retrace your steps should you find yourself down a path of no potential.
don

Don, as I said, this substantial improvement in performance was a total surprise. My purpose in adding the RF stage was more a case of "hygiene." Big surprise.

73 de Martin, KB1WSY
« Last Edit: May 16, 2015, 07:44:13 PM by KB1WSY » Logged
KB1WSY
Member

Posts: 1309




Ignore
« Reply #86 on: May 17, 2015, 04:50:06 AM »

In the end, no "tweaking" was required. The new RF stage worked without needing modification. I tried using a higher-voltage grid bias battery by the improvement in attenuation wasn't substantial so I've kept the original 4.8V supply.

The set now broadly conforms with Charles Kitchin's design principles for of a "high performance" regenerative receiver (http://www.arrl.org/files/file/Technology/tis/info/pdf/9811qex026.pdf) except that I've used a tuned RF stage instead of his broadband amplifier.

In the future:
--Aditional stage(s) of audio amplification, allowing an AF gain control and a small speaker (or modern low-impedance 'phones) as a listening option.
--An audio filter. It looks like Kitchin uses passive filtering (switching capacitors into the audio line) but I want something more sophisticated, namely an active filter with adjustable "boost" frequency and passband (very useful, plus I get to learn about filter theory). There are several circuits in old ARRL and RSGB publications -- the Audiofil, the Selectojet, and others. Seems like there are at least two approaches: audio "tuned circuits" on the one hand, and phase-canceling circuits on the other (am I right about this?). It strikes me that there are at least two advantages to an audio filter in a regenerative set: (1) allowing discrimination between closely spaced signals, especially CW ones; and (2) filtering out the regeneration "rush noise" when operating on the high-sensitivity threshold of regeneration and copying weak signals.
--Adding several more bands: the 20m and 40m 80m ham bands and a couple of SWL bands (the latter, because the audio quality of AM reception is excellent if you greatly lower the plate voltage, and I'm a BBC World Service addict). This will be done by winding new coils and changing the values in the capacitor network (eventually, the capacitors will be mounted inside the coil forms).
--However KB1GMX has a good point in advising the use of a frequency changer instead (in effect creating a superhet with a regenerative tunable IF). I have a 3.5MHz crystal which could be the basis for an 80m converter (tuning "backwards," subtractively from 7MHz to reach 3.5MHz) and a 20m converter (using the 7MHz harmonic of the crystal, adding to 7MHz to reach 14MHz). Or I could follow his suggestion and "retune" the regenerative receiver down to a lower "IF" (1.75MHz or 3.5MHz) which presumably would give greater selectivity while still being high enough to avoid images on 20m. I guess the next question would be whether to convert the current RF stage to be the mixer of the frequency changer; or instead, to build an "outboard" frequency changer to place in front of the RF stage.
--Edited to add: Silly me! I actually have a crystal for exactly 7000MHz. It's an FT243 transmitting crystal in my collection, but I'm a General and thus barred from using that frequency (and probably wouldn't risk being so close to the band edge, even if I were an Extra). So that's a perfect building block for a very simple 20m converter, for starters.
--A calibration system. This will probably just be a one-tube 100KHz oscillator. I will add a new, small-value variable "Calibration" capacitor to set the band edges.
--Transmit-receive circuitry, probably just a small reed relay.

When this process of continual experimentation/improvement is "finished" I will build a "proper" receiver in a nice handmade wooden case, in a much smaller form factor and with better layout and short, logical signal paths. Ideally I will design and make my own PCBs for this, although maybe I'll just stick to point-to-point wiring for the first iteration. A PCB build would be much more compact, apart from the controls, which are numerous and take up a lot of room (necessitating a substantial front panel).

Here's the latest schematic (this is exactly the current build):



73 de Martin, KB1WSY
« Last Edit: May 17, 2015, 06:02:22 AM by KB1WSY » Logged
KB1WSY
Member

Posts: 1309




Ignore
« Reply #87 on: May 17, 2015, 06:55:11 AM »

I've devoted a bit more thought to the idea of building a 7MHz oscillator (I have a crystal) as the basis for a 20m converter. In addition to 40m, I'm mostly interested in 20m -- when I start transmitting, it will be 40m, with 20m following soon after -- after I have built my second transmitter.

Looking over the circuits for converters, they often follow a 3-stage topology: RF amp, oscillator, mixer.

Since I already have an RF amp I presume it would be logical to put the mixer in between the existing RF amp and the detector. Alternatively I could convert the existing RF amp into a mixer, but then I would lose the advantages of having a good, gain-controllable RF stage (or would I?). If I leave the existing RF amp in place, I will still need to adapt it to cover 20m as well as 40m -- either with a new coil/capacitor combo, or with a large-value variable capacitor swing to capture both bands on one coil.

So my thinking is: use a strapped 1AD4 as a triode Xtal oscillator, and then inject that signal into the receiver with a new 1AD4 pentode mixer stage just ahead of the detector, for use when receiving 20m. Add circuitry to "bypass" the mixer when using the set for 40m reception.

The next challenge is figuring out how to do a Hartley oscillator with a directly heated triode. I'm working on it ... will build and test the oscillator first (it can also serve as a cool band-edge calibrator!).

73 de Martin, KB1WSY
« Last Edit: May 17, 2015, 07:01:48 AM by KB1WSY » Logged
KB1WSY
Member

Posts: 1309




Ignore
« Reply #88 on: May 17, 2015, 08:04:38 AM »

For Peter (G3RZP). Three weeks ago, in a different thread, you and I discussed the problem that the voltage on the AF stage grid was varying when the regeneration control setting was changed. This was an indication that RF was leaking into the audio stage.

Specifically, these were the grid voltages at the end of April:
--Regen control at maximum resistance (50K) and no regeneration: -0.4V.
--Regen control in controlled oscillation (about 20K): -0.8V.
--Regen control shorted and oscillation a loud squeal: -3.0V.

Today I checked again, after the total rebuild of the detector stage and the installation of a large-value choke load in the audio stage. The new grid voltage values are as follows.
--Regen control at any setting, except the maximum setting where "squeal" sets in: a rock-steady -0.34V.
--Regen control shorted and oscillation a loud squeal: -1.2V.

I am assuming that there is nothing unusual about the spike in voltage while "squealing" because that's a natural result of extreme *audio* (not RF) feedback. So I think it is safe to say that the old problem of RF reaching the AF grid is fixed. I did pay more careful attention to bypassing during the rebuild, including installing a thick 14AWG tinned copper wire as a ground bus alongside the terminal strip, and keeping capacitor leads shorter.

73 de Martin, KB1WSY


« Last Edit: May 17, 2015, 09:17:03 AM by KB1WSY » Logged
KB1WSY
Member

Posts: 1309




Ignore
« Reply #89 on: May 17, 2015, 09:08:47 AM »

So I've been reading all about Colpitts oscillators in QSTs from the 1920s -- from a series called "How Our Tube Circuits work" by Robert S. Kruse. Great stuff. I also looked at the standard circuits in "Understanding Amateur Radio" (ARRL 1963).

Trouble is that right around the time that crystal oscillators became widespread, so did indirectly heated cathode tubes. So I couldn't find a *directly* heated Xtal oscillator circuit.

However I've noticed that in a number of applications, a "cathode" current was developed from a filament tube by tapping the middle of a capacitive divider across the filament. So I've improvised and here's the circuit. Wish me luck, I'm building it, unless someone jumps in with a sharp rap across the wrist:



73 de Martin, KB1WSY

Logged
Pages: Prev 1 2 3 4 5 [6] 7 8 9 10 11 ... 18 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!