Modified Heising or Shunt-Fed?

Discussion in 'Amateur Radio' started by W5HRO, May 30, 2017.

  1. W5HRO

    W5HRO Administrator

    This topic has been around before, but I have not seen anyone bring it up in a very long while. Years back there were a few guys experimenting around with the old shunt-fed method in their Johnson and other transmitters. Basically, it’s the old method of removing the DC load from the secondary of your push-pull modulation transformer when modulating a class C RF amp. This is also often referred to as the Modified Heising method, but incorrectly I think.

    The old shunt-fed method is below...


    I was looking my McMartin BC transmitter out in the garage and it has the big 50H reactor choke. That transmitter uses the choke in a similar method shown below which I think is probably a much better configuration than the first one above and I think is a bit closer to what really could be considered Modified Heising, but using the push-pull audio amp.


    Anyway, both configurations are almost identical and I was thinking of trying the 50H reactor choke in my HB transmitter just to see what would happen and if the fidelity would improve even more. The fidelity was always fantastic with that transmitter anyway, but it might be fun to just throw it in there temporally and see. I think there is a spot where I could hang it with some strong cord or rope. Any thoughts or suggestions?
  2. W5HRO

    W5HRO Administrator

    And then here's another one they called parallel-fed which is just another way of saying shunt-fed, but this method completely removes the DC from ever touching secondary winding period by using two 4uF coupling caps. I think I actually like this method best because it provides better protection when switching the B+ on and off to both the modulator and RF amp. This would work best when you have separate B+ supplies for each side like my HB transmitter does.

  3. WD5JKO

    WD5JKO Member

    I am in the process of restoring a Mosley CM-1 receiver. This is a small, somewhat rare vintage receiver that uses five 6AW8's for all functions. I have been collecting some small audio chokes (like a 30H 50ma choke, and two 50H 3 ma chokes). For kicks, I tried the bigger 50ma choke with the audio output tube along with a 4.7UF polypropylene capacitor. The idea was to use a "parafeed" arrangement to take the DC off the primary of the audio output transformer. It worked fine, but I could see no improvement whatsoever. I did notice that the lows suffered slightly based upon more of tilt on square wave when dropping below 1 Khz. Perhaps a bigger series cap would have helped. Went back to stock...not worth the trouble.

    Tried to upload a 10 sec .mp4 file; not allowed.

  4. W5HRO

    W5HRO Administrator

    Yeah, 4.7uF was probably just way too small for a 4-ohm load (speaker). From what I remember the CM-1's had really good audio into a 4-ohm speaker but only at 1/2 watt. I doubt with that low of a step-down load and power it would make much of a difference no matter what you did anyway. That's what I discovered with the D104 buffer circuit. With only a 300-ohm load it takes a much bigger coupling cap to get the lows up all of the way to where you can actually hear them.

    With plate modulated class C transmitters though the load Z is usually a few thousand ohms so the cap values don’t have to be so large and I think you will notice a slight fidelity improvement, but it just depends on how much of an improvement. You did bring up a good point though about the cap value. I think it just depends on the secondary load Z and how much internal capacitance the choke itself has.

    I also think McMartins design is probably the most practical as it eliminates the need for any coupling caps the other designs require. It just uses the one to ground at the bottom end. I’ve heard from others too that out of all of the old AM BC transmitters the McMartins had about the best sounding audio when compared to many of the other transmitters and their reactor choke arraignment more than likely contributed to it.

    Oh, and that's the other thing, power supply chokes and audio reactor chokes are not really the same thing. Power supply chokes will never work or react as well at audio frequencies. Most people use them because that's all they can find and its better than using nothing at all, but finding and using real audio reactor chokes instead will work much better.
  5. WD5JKO

    WD5JKO Member

    On my choke, a few tests.... (10X scope probe!)
    My Beckman 27XT says > 20H, and 400 ohms DCR, and 150pf ( for Cap -leads tied together, and measured to case).
    Made a setup with audio generator, scope, and series resistor 10K
    Feeding the choke from a 50 ohm source thru a 10K resistor... Other side of choke grounded, as is the case.
    The .5x or -6db point across the choke is 70hz...calculates out to be 22.7H at 10K XL
    Run up the frequency flat till a resonant dip at 100Khz, and that goes away if I float the metal case.
    Resonance tests:
    .47uf parallel cap, 38hz peak -> L = 37H
    .1uf parallel cap, 100hz -> L = 25H
    .01uf parallel cap, 380Hz -> L=17H
    Excited to give that dude a try..
    Plate Choke.png
    plate choke2.png
  6. W5HRO

    W5HRO Administrator

    I picked up the two chokes below a couple of years ago back when I was looking for one to enhance the 4-400's screen modulation, but I wound up just using a standard 10H Hammond choke and then force modulating the screen with the capacitive divider instead. In the end I just needed a choke to help isolate the supply line from the screen itself. These are probably just supply filter chokes though. I could have sworn one of them was an audio reactor.


    Also, some of the Gates and Collins transmitters used the same reactor configuration as the McMartins. The BC-1H Kelly posted here in the technical forum for example also has it. I cannot remember what the value is in that one though.

    Basically you take your RFPA loads ohms times 8 to get the needed H value. So, if your RFPA load ohms is 6K then 6 X 8 = 48H. However, the 8 comes from 8H for every 1000 ohms of your RFPA load so the 8 is used when working with every 1000 ohms.

    In your CM-1 receiver project your load ohms is only 4 ohms, right? so the 8 needs to be converted. 8 / 1000 = .008 so .008 X 4 = .032H or 32mH. Did I do all of that right? If I did then I don’t see how adding a 50H or 30H audio reactor to the output of that receiver would ever make any difference. It would make sense because the math, if correct, comes out to only 32mH not 32H. You must be using a different configuration.
  7. W5HRO

    W5HRO Administrator

    I don't know if anyone tries or still does this to any of the smaller transmitters like the Johnson's, Heathkit's and etc. but for those the choke wouldn't need to be a very large H or at least not like the 50H used in BC transmitters.

    Say you have a 600V plate supply running at 300mA of plate current. That would be 2K. So, 2 X 8 = 16H so a 16H to 20H 300mA to 400mA choke would work great if you can find one. You could also series two 8H or 10H chokes together providing the current capability is still the same and make sure the phase is right when you connect them together. Two of those C-18A's in my last post in series for example would probably work great.

    There were a few people who did that with the old Johnson Valiant’s years back since the modulation transformer was always an issue, but the problem with that transmitter is it has 2 tubes modulating 3 tubes so it will always be under-modulated doing it. There was one guy who separated out and stepped up the plate supply to make it higher for the 2 modulator tubes so it would work. He divided everything off using a second rectification circuit using a bridge for the modulator side, but I don’t remember how well it ever worked. I no longer have the drawing for it though.

    You can also just hook it up using the accessory socket on the back of the Johnson's so the choke is outboard. You can just put it in a separate shielded box that plugs into the back of the transmitter. If you really want to make those transmitters Hi-Fi then that's what you need to do.
  8. W5HRO

    W5HRO Administrator

    I was just in the garage this morning looking over the McMartin's 50H audio reactor and that thing is too big to try and temporally hang in the HB rig. It's a 750mA choke for one which really wouldn't work very well given the fact my plate current in the HB transmitter only runs at 330mA. It would more than likely degrade the Hi-Fidelity. The HB's modulation transformer was custom wound so it's ratio is right dead smack on the money anyway. I’d better leave well enough alone.
  9. KE7TRP

    KE7TRP Member

    my 304TL X 4-400 is similar to the diagram above. It has a separate supply for the modulators and transmitter and is connected that way..
  10. K4TQF

    K4TQF Member

    Let's call it "Barton" modulation.. Raymond Heising's method was class A and used no transformer. The mod transformer & class B modulation were Loy Barton's ideas. He first discussed it in his 1925 master's thesis while at University of Arkansas. When he returned to UA in 1928 he commissioned the construction of world's first modulation transformer. It wasn't until RCA hired him in 1930 that he actually applied for the patent under the name "Radio signaling system". His method was used in WLW's 500KW fire breather and most future RCA made AM transmitters.

    Heising says as much and gives Barton credit in the attached paper.
  11. WQ5Q

    WQ5Q Member

    So, after you do all the audio mods to a Viking II supposedly the only thing left to "improve" is the mod trans. It might be worth using a mod reactor to take the DC off the mod trans to see if there's an improvement. I may try this when I get the current Viking II mods completed.
  12. WQ5Q

    WQ5Q Member

    My Gates BC1H1 is being reassembled. There are two large chokes. The one I think is the mod reactor (L47) is labelled 10H at 1A. That seems low. I can't find anything in the manual.
  13. W5HRO

    W5HRO Administrator

    That cannot be the mod reactor, 10H is too low. That's probably a supply choke. The mod reactor should be at least 30H minimum. Hopefully you were given all the correct iron that went with the transmitter and nothing was ever swapped out with smaller values.

    It all depends though on what your RFPA load ohms is, but 2500V @ 670mA for class C phone for example comes out to 29.84H so it's 30H minimum. 3.73 X 8 = 29.84H (30H).
  14. WQ5Q

    WQ5Q Member

    I agree...I think the other big choke is the mod reactor and this one is the HV power supply choke. 10H would be about right. I looked again at all the schematics I could find for Gates 1kW transmitters and none show the value of the mod reactor. Regardless my broadcast engineer/ham buddy that I got the rig from will know and he also has the equipment to check. He's going to come over when I've got it back together and help me fire it up on the original frequency into its internal dummy load.

    But I need an L meter here so I can check these things. Would be handy to take to swap meets too. Actually I have one but it doesn't read that high. It's more for RF circuits. I always take it to swap meets.

    This one from eBay is less than $40 and reads up to 200H! I wonder how accurate it is?


    Ok, I checked dc resistance...the choke I thought was the mod reactor was only 57 ohms. The other was 136 ohms so makes more sense that it is the mod reactor.