classic p-p 812A homebrew issues

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

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

    KC2ZFA Ham Member QRZ Page

    John, the t/r setup puts the ground of the grid leak resistor, 1773 Ω, onto a little bias ps which I’ve set at -65V. In receive the tubes are cut off and no grid current flows.

    the two HV supplies each have 56kΩ, 200W, bleeders. Some heat there yes but the cabinet has a fan to bring in cool air under the floor end of its back door, there’s an inch+ of a gap and there are louvers on both sides for exhaust.
  2. KC2ZFA

    KC2ZFA Ham Member QRZ Page

    I excite this thing to 72 mA grid current for 128V bias, all derived from the resistor, with 1300 plate volts.

    the ancient scrolls say -115 bias for telephony at 1250V plate. And the max grid bias is -200.
  3. W9JSW

    W9JSW Ham Member QRZ Page

    Thanks for the explanation. Makes sense. Have you published your schematic?
  4. K5UJ

    K5UJ Ham Member QRZ Page

    Agree. With a choke input of at least 10 Hy., I don't think step start is needed. But I put a thermistor in series to the primary of a modulator supply transformer because its 32 mfd filter capacitance was blowing the line fuse. That ended the fuse blowing.
    Those are RCA designed older triodes. My comment pertained to zero bias triodes, such as those developed by Eimac in the '50s and '60s that do not need a protective bias supply.
  5. K4KYV

    K4KYV Premium Subscriber Volunteer Moderator QRZ Page

    There is very little information to be found on how to accurately determine grid dissipation, and it's not a simple straight-forward calculation of DC grid current × DC bias voltage. With grid leak bias, the power dissipated in the grid leak resistor is calculated by multiplying Ig × DC bias voltage, but how does that relate to power dissipated in the grid? Furthermore, what if a near-ideal fixed voltage source, with negligible internal resistance, is used to supply the bias voltage? The same power is dissipated in the grid at a given Ig and DC grid bias regardless of how bias is achieved, but with the ideal fixed bias supply, there is no bias resistor to dissipate power, yet power is still dissipated in the grid. As we well know, with a conventional class-C amplifier using grid leak bias, a certain amount of RF power from the driver stage is needed to fully drive the grid. How much of that power is used up simply heating the grid leak resistor, and how much is actually dissipated in the grid?

    Most of the information found on this subject is oriented primarily to class AB2 and class B linear amplifiers, not class-C, even though this is an important consideration with FM transmitters as well as AM (and CW). For example, W8JI touches on this subject in his website, and this at least partially explains why the driver stage nearly always must deliver considerably more output than what the tube chart lists as the RF grid driving power requirement for the tube.

    To add another factor to the equation, manufacturers' data do not always include the maximum safe grid dissipation for a given tube.

    Just as anode dissipation can not be determined by simply multiplying Ip times Ep in a working amplifier, control grid dissipation can not be determined by simply multiplying voltage by current in a sub-class 2 amplifier (i.e. AB2). Beware of any article or author who tells you RF grid voltage can simply be multiplied by average grid current to determine grid dissipation.

    Grid dissipation, like anode dissipation, is a time-integrated function of instantaneous dissipation throughout the RF cycle. Computer models have made the difficult task of calculating grid current simple and accurate. Short of that you need to do an actual complex analysis of the grid system, such as in a
    Chaffee analysis.

    In grid driven stages, assuming total power applied at the grid is accurately known, grid dissipation is determined by deducting bias power from RF grid power. This is very similar to the short-form determination of anode dissipation from anode power and RF power output. This does not easily work in cathode-driven PA's, because drive power is mixed with output power.

  6. W9JSW

    W9JSW Ham Member QRZ Page


    Are you on the air yet?

  7. WD5JKO

    WD5JKO Ham Member QRZ Page

    Hi all,

    I was just reading through the thread, and enjoying the comments, and the eye candy photos.

    My only comment relates to my experience with a triode based P-P RF Final. I had a 500 watt DC input P-P final with 8005 type tubes. I cannot say enough good things about those, with one big exception, today they are scarce, and extremely expensive. I tried 572B's, and was able to raise the B+ up to run 1 KW DC input. After beefing up the modulator from two 808's to four 808's, I had 600 watts audio into a resistor load (2250v B+ on the modulator), but I could not modulate those 572B's to 100% no matter what I tried. I had 2500v B+ on the RF stage, and rheostat variable grid leak. I ended up punting, and tried 805's after a socket change. The difference was night and day. The 805's modulated so nicely to 100% and beyond. More grid leak, and then more RF drive, and even better. I ended up running that rig for several years with 2500v on the 805's (1Kv above rating). I got 1KW DC input, and 850 watts RF out on 75M. No color in the plates unless I whistled in the Mic.

    Since the plate curves on an 811 are the same as on a 572B, I find it hard to believe those running 811's in the RF final can get good modulation that approaches 100% with good linearity.

  8. N2DTS

    N2DTS Ham Member QRZ Page

    You can if you put higher voltage on the modulator then the RF deck and have the right mod transformer match.

    It seems a lot harder to get over 100% modulation with less voltage on the modulator or even the same supply/voltage
    as the RF final.

    I think you can run 811's at 2000 volts on the plates, unsure about 2500, the most I ever tried was 1750 with some bias.
    572b's are rated much higher then 811's and you could likely run 2500 to 3000 volts with some bias.
    530 watts out at 2000 volts for the 572b, should do a KW input rig easy at 2500 volts.

    In my case, I ran a common robust supply and switching between 812a, 811a, and 572B tubes gave no
    difference at all that I could detect.
    Neutralization might have been different, but otherwise no change.
    WA3JVJ likes this.
  9. WA3JVJ

    WA3JVJ Ham Member QRZ Page

    Good post, I like the pictures. Your rig is similar to one of my Globe King 400 decks, where I use a pair of 812’s in push pull with about 1400 vdc on the plates. I would like to amplify one of N2DTS’s comments; i.e. fixed protective bias that is diode coupled to the grid leak resistor. The diode acts much much faster than the mechanical relay and should prevent that short period of the amp running away, I use such a circuit.

    With link coupling I would like to recommend a tuning capacitor, (it can be low voltage), in series with the output link. It is series connected between one end of the link and ground, assuming that you are using coaxial output. This will improve coupling between the plate tank and the antenna. I am using 65% of a variable capacitor of about 570 pf or so and a three turn link to a 50 ohm antenna circuit on 40 meters.
  10. KA4KOE

    KA4KOE Ham Member QRZ Page

    I didn't know that BHI had a digital VFO board/kit?

Share This Page