DIY Receiver for AM

Discussion in 'Technical' started by BBurns, Sep 14, 2014.

  1. BBurns

    BBurns Member

    Hello All,

    I'm looking for a do-it-myself article for building a ham-band AM-CW receiver. I built one of Byron Goodman's 80 and 40 meter double conversion designs back in high school radio shop in 1955-56, and never could get it working properly. The shop teacher only knew regens, and there was no test equipment for superhets. An un-closed gestalt, or something like that you know.

    So, it's in my bucket to build a receiver for CW and AM with good enough performance to actually be useable on the air. I'm not thinking of trying to duplicate the convenience of a commercially made receiver of the period. I'm content to use plug-in coils, and ganged tuning capacitors aren't necessary---more knobs to twist?---more fun. Besides, I still have the HQ-150 that Carl restored for me.

    Caveat. I'm a confirmed "glowbugger", and this is all about nostalgia anyway, so I'm only interested in tube based designs.

    In looking through all the old literature for a couple of years it seems that regens and simple superhets with regenerative I.F.'s are good at CW, but AM?--not so good. When you get up to the sophistication of the "Junior Misers Dream" the bloody thing has a product detector, and is specialized for CW and SSB!

    So, I'm really interested in any suggestions that any of you might have.


    Brian Burns Ex: W5BRO, K6UCD
  2. W5HRO

    W5HRO Administrator

    Sounds like you just want something better than a regen, but still keeping in simple.

    The issue would be in making it single or double conversion. If you want to keep in simple then do what RCA did with the BC-348 and use a 915kc IF. That way you can eliminate going double conversion with it.

    For IF alignment just build one of the simple wobbulators or find a cheap sweep generator. Then all you need is a scope. You could probably find an old junker BC-348 for the 915kc IF cans, etc.

    P.S. The hombrew forum is for projects already in the works or already completed. It's more of a show-and-tell type forum.
  3. BBurns

    BBurns Member

    Hello Brian,

    There is a lot of praise for the RAK and RAL Navy regens as communication receivers. I’m trying to stir up some interest over in the regenerx group in coming up with a homebrew regen that draws on the qualities of selectivity and strong signal handling of those receiver designs.

    This is a brand new project so not much has happened yet. The only bits of received wisdom so far are two RF stages, and mechanically very sturdy construction. The two RF stages are more for isolation of other stages than for amplification.

    The regen gets around the superhet problems of images, birdies, and finding I.F. transformers. It’s presumably somewhat easier to build, and hasn’t the alignment difficulties. So we’ll see what happens. I’m good at the mechanical construction part, but am entirely dependent on others for the circuitry.

    I’ve got an old Hickok sweep generator, and a scope, plus some other test equipment.


  4. W5HRO

    W5HRO Administrator

    The main problem with regens as everyone knows is the lack of AVC control, or more to the point reducing the RF gain when strong signals are received.

    We have discussed this in the past and I still say the best way to do it would be to use a single RF amp stage at the input like a 6SJ7 with a control line coming back to its screen supply to reduce that screen voltage when strong signals are received. The trick would be figuring out a way to produce the control voltage, but I honestly don't think it would be too hard to accomplish. It would require adding control tubes or some solid state circuitry to control that voltage. One tube would work fine though like in a pass transistor type configuration maybe. The controlled screen supply would just need to be an independent supply and not derived from the plate. And I say control the screen and not the control grid because tubes like the 6SK7 designed for AVC control wont have good enough sensitivity to make a decent RF amp stage connected to the antenna. An antenna connected to a high-pass filter to an 6SJ7 would though. AVC tubes were primarily designed to handle large input voltage swings from other tubes and not to amplify tiny signal voltages

    P.S. I still say a single conversion receiver using a 915kc IF strip with a Pullen mixer would be easier to build and make work though.
  5. KM1H

    KM1H Guest

    The RAK/RAL were not made for AM, the RAK audio filter is way too narrow and the one in the RAL sounds worse than space shuttle audio. They are excellent for CW and are even being used on digital at the 600M experimenters band.

    Quasi AVC is built in with the audio limiter which also does a better job on noise spikes than the usual diode NL in most comms radios.

  6. BBurns

    BBurns Member

    Hello Brian and Carl,

    As for AVC in a regen, I ran across an article in QST for Feb. 1940 entitled "A Regenerative Preselector With Output Metering Bridge" by H.O. Talen, W9PYQ. What caught my eye was his mentioning its better AM reception due to what appears to be some AGC action that allows advancing the regeneration control closer to oscillation, when used in a regen. I'm attaching a .pdf of the article, enlarged somewhat for easier reading, if you print it out.

    Right Carl about the RAL and others being too selective for AM. What intrigues me about that series is their reported ability to handle strong signals. I'm wondering if those two RF stages, with their associated tuned circuits, provide some up-front selectivity, or whether they were just there to completely keep the oscillating detector from radiating. Perhaps both?



    Attached Files:

  7. BBurns

    BBurns Member

    Hello again,

    Attached is an earlier article by Talen which gives a more thorough explanation of the AGC and noise limiting use of the circuit with a TRF regen.



    Attached Files:

  8. KM1H

    KM1H Guest

    I can attest to the RAL's signal handling abilities Brian, I used one aboard ship for almost 2 1/2 years with a roughly 150' longwire about 40' away from a parallel TX antenna running normal ships traffic at 500W out. Frequencies were in the 6 and 9mc range and I didnt even know they were operating when I was on 40M.

    The front end tuning is very sharp which attests to RCA's design of very high Q coils. Even on 40M off the coast of EU and AF in the Atlantic and Med there was no overloading from the SWBC flamethrowers and lots of QSO's were made with the ships emergency use TCS-13; its RX was useless. I sometimes used the RBC which was often superior to the RAL but only had limited access to it.

    For a bit more flexibility I built a Knight T-150A and operated 15-40M CW and AM. Not the greatest rig but I chose it for size, about 6db more RF, and light weight. Key clicks were horrendous and I was banned from using it on CW until that got fixed.

  9. WD5JKO

    WD5JKO Member

  10. BBurns

    BBurns Member

    Hello Carl,

    "The front end tuning is very sharp which attests to RCA's design of very high Q coils."

    So would it be fair to say that the RAL's strong-signal handling ability was due to its two RF stages with their high Q tuned circuits?

    I'm really interested in whether a regen can be built to be a "useable on the air" communications receiver in these latter days.

    Received wisdom is that air core coils are the highest Q---is that right?

    Those flamenco guitars are nearing completion. I'll trade you straight across for one of your Buicks (;->)...

    Thanks a bunch,

  11. BBurns

    BBurns Member

    Hello Jim,

    I've not followed the whole thread, but I've watched that video. I suspect that radio works well enough for SWLing, but for use as a ham communications receiver on the air I suspect that it will suffer from the usual regen problems of pulling, and overloading. Vlad is certainly to be complimented for his work at getting the most out of a low parts count design.

    I'm willing to build a more complicated design with the hope that I will actually enjoy using it on the air. I'm also aware that such a receiver has to be built "bully for strong" in order to be reasonably stable.


  12. KM1H

    KM1H Guest

    The pair of 6D6's are loafing and there is some reason RCA chose remote cutoff tubes instead of 6C6's; I havent dug into the circuit for several years.

    Define useable. The RAL is the ultimate commercial regen for CW IMO. My National SW-58 aint half bad when I use a DSP audio filter for CW but for AM the PP 45's and 10" speaker sound just fine.
    Your not going to get "real" performance from minimalist designs either.

    For a mid 30's design they didnt have much choice. Today ferrites offer higher Q up to lower HF or so depending upon the wire size, # turns, density, etc. of conventional coils. Ive used powdered iron toroids to replace superhet 30's era coils on the highest range that became lossy due to moisture, chemical reactions, or whatever. Even image rejection improved.

    My Brian retired August 1 from the USAF and has moved back to Germany from Tbilisi where he starts a new job today. I dont think he will have much guitar time available for awhile!

    Im probably going to start thinning the vehicle herd again soon, the 38 Roadmaster will be included; Ive received a couple of very nice offers this year whenever it goes out on a drive.
    The 37 Roadmaster with a 1969 455 Buick driveline will stick around awhile; its more fun to drive.

  13. W5HRO

    W5HRO Administrator

    Front in tuning (selectivity) is very important in ALL receiver designs. The "Q" of the coils isn't that big of a problem today if you use enough poles. What you want is a good high-pass filter (HPF) between the antenna and first RF amp.

    Anyway, it's really easy to do today even with tube receiver projects. Below is an example of a very small 8-pole filter.

    Here because it's 50-ohms to 1Meg (grid resistor) you could just use fixed .001uF caps all the way across as long as you use the right tunable inductors (coils). You can get the coils from Coilcraft and mount everything on those small PC boards Radio Shack still sells. They have several different sizes too and all you need are 4 metal standoffs to mount them underneath the chassis next to the RF amp tube. The first peak is right at 1.8Mc and look at the attenuation below that :eek:


    If you are going the regen route then I definitely would throw in an RF amp stage with a HPF like this to keep from sucking everything in below 1.8Mc down to DC.

  14. W5HRO

    W5HRO Administrator

    Actually 10-poles works even better. Those Coilcraft coils are so small that adding a couple more wouldn't even be an issue.


    Let's see you get this kind of filtering/attenuation using 65-year old or older technology :icon_shh:
  15. BBurns

    BBurns Member

    Hello Brian,

    That high pass filter looks like a really good idea! A real "noise getter-outer" (a technical term). I'm a believer in doing all you can to add to selectivity and noise reduction before you do any amplifying.


  16. BBurns

    BBurns Member

    Hello Carl,

    "Define useable. The RAL is the ultimate commercial regen for CW IMO. My National SW-58 aint half bad when I use a DSP audio filter for CW but for AM the PP 45's and 10" speaker sound just fine. Your not going to get "real" performance from minimalist designs either."

    Useable for me means enough strong signal handling ability to be able to keep a QSO going in spite of one of the big boys coming out to play next door. I guess what I'm trying to figure out is what things make the RAL able to do that so well. It certainly had to be a design objective of the RCA engineers, knowing that there was going to be a lot of RF activity in close proximity aboard ship. Selectivity can always be added in the audio section, and sensitivity shouldn't be a problem.

    I'm really not interested in minimalist design. I did that with tin cans and wire from an old auto voltage regulator when I was a kid. If getting good communications level performance out of a regen requires more tubes and more parts, no problem.

    "For a mid 30's design they didnt have much choice. Today ferrites offer higher Q up to lower HF or so depending upon the wire size, # turns, density, etc. of conventional coils. Ive used powdered iron toroids to replace superhet 30's era coils on the highest range that became lossy due to moisture, chemical reactions, or whatever. Even image rejection improved."

    Would "Kits and Parts" be the best source of data on toroids?

    "Im probably going to start thinning the vehicle herd again soon, the 38 Roadmaster will be included;"

    I probably mentioned in one of our phone conversations that I bought a 38 Century Coupe for $50 when I was 15. I new when I sold it that it would be valuable some day, but there was no way I could afford to keep it at that age---I just figured out, that was 59 years ago!


  17. W5HRO

    W5HRO Administrator

    There's the other topic High-Pass Filters where you can spend the money on an overpriced unit like the one I'm using on my NC-303 or you can just build your own.

    The neat thing about regens is that you can get narrow pass-band selectivity without using IF's and filters, but it still doesn't address the issue of rejecting interference from BC stations below 1.8Mc or other noise coming from down lower. When I lived in Tulsa back in the day when I was the most active on the bands I argued with others local hams there because they never understood the importance of having selectivity at the very beginning of the first RF amplifier stage. They would sit there and tell me I had my terminology all wrong and that selectivity only happened via the IF stages ::)

    Anyway, they were clueless as to what I was even talking about and they failed to grasp that a high-pass filter is also selectivity and having both will produce even higher RF energy through the IF stages. If you start out by amplifying everything from DC on up as with "dynamic receivers" you will lose a lot without even realizing it. Unfortunately receiver designs started going that route because when it came to multi-band operation it was easier to just do it in the IF stages instead. It was a design trade-off because of the amount of room it would require and complexity of needing more band-switching. Back then coils and components in general were large so you can imagine what it would take to switch all of the extra LC circuits around. Just imagine an HRO receiver with all of the plug-in coils built-in. It would have been huge!
  18. KM1H

    KM1H Guest

    I guess it matters where you live as far as BCB problems are concerned.

    Ive no problems here with the 1933 SW-58 on 80 and up. I dont have the BC coils but there are 7 1-10KW stations LOS from here on the hilltop.

    In 1954 I built a 5 tube AC/DC regen set from a Popular Mechanics article using Radio Row in NYC and the local town dump for parts; yes it had a tuned RF stage. Learned CW on it from W1AW on 80M and a few months as a Novice on 80 and 40. Never heard any BC crud surrounded by 250W-5KW stations out on Long Island.

  19. BBurns

    BBurns Member

    Hello Carl and Brian,

    There is not much BCB AM activity around where I live, but I was thinking that high pass filter would tend to cut down on whatever noise, atmospheric or man made, that was below 160.


    Brian B.
  20. KM1H

    KM1H Guest

    Try it again with real world numbers. Those coils have low Q as well as high distributed C, the plot wont be as pretty.

    The USN was concerned about multiple shipboard transmitters a megacycle or more away causing receiver blockage. They ran full QSK that way.
    Us hams have to worry about signals in the sets passband. While I enjoy my RAL its not for contest weekends!
    Without AGC audio selectivity works well and a strong signal off to the side isnt pumping the output.

    The easy way to get performance these days is front end bandpass filters followed by roofing crystal filters at the up conversion to VHF as is done with all GC coverage transceivers...just cut some diodes and you can TX 1.5 to 30mHz also.

    Most decent tube sets have good out of band selectivity since they had tracking (or manually tuned) preselectors. The R-390A is an excellent example of top IM performance with just a single RF amp. The 455kc IF sets with 2 tracking RF stages gave decent image rejection even on 10M.

    I use a modified RME DB-22A for marginal single RF amps sets such as the Hammy HQ series. Use just a single 6GM6 RF amp at V-1, existing gain control, a 100-270pf mica cap between pins 1-5, and with 3 tracking tuned circuits and voila, lots of sensitivity when wanted and no images. It helps the dual RF Hallis also since they have poor upper band sensitivity and less than stellar image rejection, even the SX-28 and SX-42.

    The earlier DB-20 (later production with the 6K7's) is also a candidate for modification using one of the hotter octal tubes such as the 717A mushroom tube, or a 6GM6 on an adapter, moving the gain control to V-1 and reducing the B+ a bit. You could even get carried away using a 6EH7 on an adapter.

    There are many tube options in all cases since there is no AGC used and sharp cutoff tubes with even higher Gm can be used.

    Your HQ-150 has the 6GM6 mod but it still suffers from images on the higher frequencies as do all brands and models of that genre.

  21. W5HRO

    W5HRO Administrator

    Those are real world numbers and it was done on the Spice simulator. They are low "Q" coils and if you use enough of them (enough poles) it works fine and that was my point. It's not necessary to always use big high "Q" coils. I've built similar ones in the past using Coilcraft's and if you do it right, minimal loss as well. Maybe only a dB or so when sweeping them with an IFR spectrum analyzer. On the low bands you won't even notice it. You can even add one BiPolar NPN transistor at the end of it if you want to overcome any loss and give it a little gain to drive the tube which can help on the higher bands like 10-meters, but it's not necessary. It's only an option.
  22. KM1H

    KM1H Guest

    Adding a preamp at the end of a lossy filter does nothing for the cumulative noise figure increase from each pole. At the higher frequencies it could be detrimental; use the preamp to drive the filter at unity gain if the receiver is already sensitive enough....otherwise include a gain control. Since no antenna is 50 Ohms J0 the filter response will vary, use the preamp to drive 3-6dB pad to stabilize the filters impedance.

    My own at work lab experiences with Coilcraft is that they are not suitable for anything even remotely critical in the signal path. Spice is only as good as the data entered provided by the manufacturer; that leaves Coilcraft out in the overly optimistic end of reality IMO. It doesnt take much work to wind 2uH coils on a decent size toroid.

    When using a Beverage or other low noise negative gain antenna any added loss will add to the NF which WILL affect hearing those weak ones. While this is an AM forum I do like digging into the noise on CW in a 200-250 Hz bandwidth and even use those antennas on AM :icon_shh:
  23. W5HRO

    W5HRO Administrator

    The problem is Carl I've done it before. I've laid out small PC boards for the coilcraft coils with an NPN at the end of the filter to the RF amp tube and it works fine. The trick is laying the board out properly and don't use a noisy regulator. Also properly bypass the regulator's input and output leads to ground. That's usually where the noise comes from. Just changing to a different regulator or changing the bypassing will greatly effect the SNR.

    Anyway, on HF from 30Mc on down most of that stuff isn't even and issue, except for the possible regulator noise. Even the solder mask on the boards wont really effect the capacitance and tuning like it will up in the VHF/UHF ranges and above. You can just use one of the small Radio Shack boards and it will work fine. You can also use slightly smaller caps and then increase the uH value of the coils a bit, but I wouldn't go any higher than about 4 to 5uH max. It's just harder finding slightly smaller value precision caps.

    A good 8 to 10-pole HPF using those small coils will work fine. And if you want more gain say like up on 15 and 10-meters then just add the NPN transistor between the HPF and the tube. It just takes a little to overcome the interelectrode capacitance between the grid and the plate of the1st RF amp tube. Most tube receivers don't hear shit on those higher bands because of it and the lossy stock filtering. Just bring up the slight bit of loss the HPF will induce then bump it up only about a dBm or two higher. And you don't need a JFET, MOSFET or a GASFET because we are talking about HF and the impedance wont be an issue either. Using an NPN will work fine. Most tube input circuits are between 500K to 1Meg so its not going to load down the NPN's output. You can actually make a tube receiver hear better than 1uV on 10-meters if you do it right. I've actually got them better than -110dBm before to almost -115dBm.
  24. KM1H

    KM1H Guest

    The problem Brian is that Ive been doing this since possibly before you were born starting in 1964 when I had run of the National R&D lab to experiment on my own time and with improving my 75A4, building VHF/UHF converters for TX and RX and then transverters.
    I got a real education in the late 70's when I discovered the noisy regulators in my Drake R-4C's and then the TS-930 while working to improve the phase noise.
    In the 70's to mid 80's I was Tempest certified and worked on keeping circuit signals inside the circuit for CIA, NSA, and DOD projects for Sanders Associates and then Wang Labs. Learned a lot about bypassing, ferrites, shielding and PC board design. Working in a screen room was enlightening and I later used that knowledge to build DX and contest stations that set records as well as did not prompt any TVI/RFI calls.
    In later years I did R&D as high as 300 GHz for commercial and DOD projects.

    I disagree and component noise is important at any frequency including audio. Poor PC boards and cheap resistors affected the PLL in the TS-940.

    I'll repeat: you dont want to place a preamp after a lossy filter if the receiver has poor sensitivity to start with. A 2N5109 in a Norton circuit will give you 10-11 db gain, a 1.5dB NF, and IP3 in the 30's.

    Most any decent tube radio from the mid 50's and later has better than 1 uV sensitivity for a 10dB SNR on 10M, even on AM, out of the box. The stock NC-300 measured <.5uV (-113dBm) and a stock rebuilt HRO-60 came in at .8uv (-109dBm). A simple 6GM6 swap in the 300 brought it down to .25uV (-119dBm)and a 6GM6 in V1 and a 6BY6 in V3 plus a few component changes did the same in a HRO-60 which was still using mid 40's tubes into the 60's.

    Every improvement peeled off another layer of readable signals on 15 to 10M and a frontend HPF was never needed. I cant think of any reason to use one here on any tube or SS radio that I actually use and that includes the 160/80/75M parallel inverted V's with a single coax feed at 180'.

    Tubes that were marginal on 10M were long gone by the mid 50's when TV IF's went to the 40mc+ range and FM took off before SS caught up. Unfortunately some ham manufacturers never did catch up before going under.

    Using MDS as 3dB above noise add -6dBm to the above.

    In a typical 2400 Hz SSB or sharp AM filter the thermal noise floor is at -140dBm with the average SS quality rig coming in around -130dBm on SSB and a CW filter can get it right down to the floor. My very highly modified 75A4 and the past pair of R4C's did that as did a stock R-390A which had minimal losses in the front end. I use my R-390A as the benchmark standard.
  25. W5HRO

    W5HRO Administrator

    For anyone interested below is what I was talking about despite what some might think and it will fit on one of those small boards.

    I've built similar ones in the past and they work fine. This one could be powered by regulated +150V from inside of the receiver. Just use a zener from the HV supply line. I picked a simple 2N3904 NPN switching transistor because they are good for almost 40V at 200mA continuous and up to 100Mc. It was one already in the program's library so I used it for the example. Just search around for the best low noise transistor to use.

    Below are 500kc to 50Mc results of after the amp (OUT), before the amp (TP1), and then the circuit itself. If you want better sensitivity on the higher bands like 15 and 10-meters then this is what you have to do. The tubes themselves and the old over-sized lossy components wont cut it. The HPF will also cut way back on the noise and BC interference.

    The whole point is that RF energy is lost and wasted when the receiver is trying to amplify everything from DC on up. If you remove the bottom end by rejecting it and selecting everything above then that wasted energy can be amplified by the amplifiers instead and put through the IF. The end result is higher detected IF levels at lower RF input levels.

    P.S. Good receiver sensitivity levels are one thing, but throw intelligence on it (modulation) and those numbers will often crumble. The IF signal itself going to the detector without modulation needs to be at least 20dBm above the noise floor. Anything less than that wont cut it.