classic p-p 812A homebrew issues

Discussion in 'Amplitude Modulation' started by KC2ZFA, Aug 29, 2021.

ad: L-HROutlet
ad: l-rl
ad: L-MFJ
ad: Left-3
ad: abrind-2
ad: Left-2
ad: l-BCInc
ad: L-Geochron
  1. KC2ZFA

    KC2ZFA Ham Member QRZ Page

    hello all, this summer I sorted out my modulator (older thread here), put together a dds vfo using the W9BHI board, sorted out my 6ag7-807 mopa exciter for 160, 80 and 40, and recently got to the testing stage of my long in construction final stage using all classic parts and a pair of 812A’s in push-pull.

    Here are my issues (some may not be issues per se, newbie here):

    1) with a 3 turn swinging link on the output I could only load the PA up to 150 mA at 1300V and saw 72-75% efficiencies on both 80 and 40. So,

    2) I then went to a 4 turn link and now I can load up to 280 mA at 1285V with 72 mA on the grids (-125V, all grid leak). Great, but I’m seeing 285 watts on the dummy load for a 79% efficiency…Can this be real ? the dummy load shows a 1.3 SWR.

    3) Now the whammy: I let go the PTT, no excitation, so that -65V cutoff bias goes to the bottom of the grid-leak resistor (a relay switches the bottom to the filament return on xmit). I switch on the HV and momentarily (~1 sec) I see 250 mA on the plate which cuts down to zero after the momentarily…there’s no grid current and no output while this happens. Is this normal ????

    Any help with 3), in particular, will be much appreciated.


    Attached Files:

    WA3JVJ and N3RYB like this.
  2. W7TFO

    W7TFO Premium Subscriber QRZ Page

    Off the cuff, I'd say the bias switching is going somewhere other than where it belongs.

  3. N2DTS

    N2DTS Ham Member QRZ Page

    Why not have a diode supply the fixed bias all the time and let grid leak to the rest?
    WA3JVJ and WZ5Q like this.
  4. K5UJ

    K5UJ Ham Member QRZ Page

    What's the value of your grid leak R? It's wire wound right, with a bypass mica cap to ground on the grid side of it? How are you measuring your grid bias? I'd just drive the grids with enough power to develop the needed bias across the leak resistor, making sure you don't exceed the tube's maximum grid current rating. The data sheet should say how much bias v. is needed to get into class C. It should also give you the peak RF drive grid voltage. With this you can figure out how much grid leak resistance you need to drive the grids to a certain amount of grid current and get the needed bias v. without drawing too much grid current. Remember that the data sheet will be for one tube and you have two, so your grid current will be double and leak resistance value 1/2 what it would be for a single tube. You can measure the bias under drive directly with a small pie wound choke on the grid, resistor, bypass cap junction and a VTVM on the other side of the choke.

    I've never run 812s before so there may be some need for fixed bias I'm not aware of. I'd think that if you key the driver and final plate v. at the same time, the 812s should be okay.

    812s are a classic triode RF tube. Congratulations on a great design, although I'm kind of partial to vertical mounting, hi. Love that link coupling and p.p. hardware.
  5. KC2ZFA

    KC2ZFA Ham Member QRZ Page

    Dennis may have a point…”I came through and I shall return” :)

    ‘UJ: the grid leak is 1.75 kΩ to account for the two tubes. And for your viewing pleasure a pic of the linkage to swing the link.

    ‘DTS: I may go your way, let me first see if I made a wiring mistake.

    It almost seems like the tubes have to momentarily think about cutting themselves off when they’re sitting there with -65 on their grids and I suddenly hit them with the HV.

    Attached Files:

  6. N2DTS

    N2DTS Ham Member QRZ Page

    Maybe some sort of feeback as things are not shielded.
    The output coils are right above the exciter coil!
  7. KC2ZFA

    KC2ZFA Ham Member QRZ Page

    It's the 1930's in my basement :)

    And, it turned out Dennis was partly right...the cutoff bias was going where it needed to go but it was doing so
    via a cold-solder it seems, as I hit them with the HV (no excitation) the tubes had to work to pull the cutoff
    bias through the cold point. Now the plate mA stays put to zero when I put the HV on them. I intend to keep the HV
    on the modulator and the finals during receive.

    And here's two short clips at two different power levels on 40M (the clips have a few secs of dead air then show the
    power/SWR, the plate mA, the grid mA, and the HV on the plates):

    Not long before you have to suffer my presence on the air :p
    AH7I likes this.
  8. W9BRD

    W9BRD Ham Member QRZ Page

    Be seeing you!
  9. W9JSW

    W9JSW Ham Member QRZ Page

    This is indeed good progress! A very nostalgic design as well. PP 812 modded by 811s will be a head turner.
  10. K4KYV

    K4KYV Premium Subscriber Volunteer Moderator QRZ Page

    This is less a problem with triodes, since they are less power sensitive compared to screen grid tubes. Even so, when unshielded input and output coils are in close proximity, it's a good idea to mount them so that their axes are perpendicular to each other. I recall when I first built the exciter/driver stages in the HF-300 rig, I got each low-level stage working properly, one at at time, before proceeding to the next stage. I would fire up the preceding stage at reduced power, tune the tank circuit to resonance, and then temporarily mount the following tuned circuit so that the coil could be moved around. Then that next tuned circuit was resonated, and using an rf sniffer (a simple diode rectifier and milliammeter), connected to a temporary link wound over the main coil, the main coil was moved around and the position where a null in the detected rf was noted, and permanently located as near to that position as possible. Those input and buffer stages use screen grid tubes, but the whole driver string was stable enough, but even more so after I had completed the rig and used it for a couple of years, when I replaced the somewhat squirrelly 807 (beam tetrode) with an 802 (true pentode).

    I wanted the rig to work on CW as well as AM. The 807 had an annoying back-wave, which I was able to cure only by neutralising the stage. When the 802 was installed as a replacement, it too had to be neutralised to get rid of the back-wave. I suspect external capacitive coupling, not the internal capacitance of the tube, was causing the feed-through. Nevertheless, the whole thing has functioned for decades with no trace of back-wave, instability or parasitics; the resonant dip of each of the tuned circuits is velvet smooth.
    N2EY, KC2ZFA and WZ5Q like this.

Share This Page