ad: ProAudio-1

Home-brewing a Plate modulated Push-Pull 810 deck

Discussion in 'Amplitude Modulation' started by W2BTK, Aug 29, 2019.

ad: L-HROutlet
ad: l-rl
ad: L-MFJ
ad: Left-2
ad: Subscribe
ad: Left-3
  1. K4KYV

    K4KYV Subscriber QRZ Page

    I use a magnet on my hardware and never use anything steel, even if nickel plated. Not only for RF, but anything that carries current. I once had a steel nut that someone had put on one of the terminals of a rectifier filament transformer get hot enough to burn my finger. It caused the filament voltage to drop enough that the 866As flashed over on occasion.
    Last edited: Aug 31, 2019
  2. W2BTK

    W2BTK Ham Member QRZ Page

    Holy crap. I didn't know about that stuff. I think most of the RF paths are bolted down with zinc-plated hardware..... You know. Stuff I got for free. I am going to have to keep an eye on it or buy stuff. I'm torn.
  3. W2BTK

    W2BTK Ham Member QRZ Page

    Also Don- What do you think about the lack of center taps on the driver link and output coils? Do you use a safety Bias on your rig, and how much grid resistor are you using? I'm trying to compare...
  4. K5UJ

    K5UJ Ham Member QRZ Page

    At some point in the past, none of us knew it.
    With RF what happens is that ferrous metal carrying RF current vibrates at frequency on a microscopic level. This generates heat. Heat causes resistance to rise (and steel isn't the best conductor anyway), and nuts can work loose over time. What about steel towers used as antennas? Well, obviously anything that massive with that much surface area will not present the same sort of problem.
    Usually what's suspect are nuts, bolts and washers. I usually check homebrew I've purchased, and kits others have built because a lot of guys don't know this, or don't think it matters. It's kind of hard to find nickel hardware, copper costs a fortune, aluminum isn't very strong, so what's left is brass. A lot of hardware stores have brass fasteners and washers but they've become kind of pricey. Going forward start maintaining a stash of sheet metal. Don't throw anything out. Even small scraps of brass, copper or aluminum can come in handy and whenever you cannabalize anything, save the RF hardware and everything else except old leaky caps. Look for brass and copper hardware at hamfests.
    K4KYV, W2BTK, AG5CK and 1 other person like this.
  5. K4KYV

    K4KYV Subscriber QRZ Page

    No need for a grounded midtap on the input or output links, nor to connect one side of the link to ground. I let mine float, leaving no pathway to ground for common mode current. I use grid leak bias plus enough fixed protective bias to keep the tubes well past cut off when there is no rf drive present. Diodes are connected in the protective bias circuit so that the grid leak bias developed with grid current, overcomes the fixed bias voltage and reverse biases the diodes, effectively disconnecting the fixed bias supply, allowing the tubes to function with 100% grid leak bias as if the fixed bias supply weren't there. In one of my homebrew rigs I use broadcast receiver type rectifier tubes as diodes, and in the other, silicon diodes.

    With plate modulated class-C finals, modulation linearity is better with 100% grid leak bias than it is with fixed bias or with a combination of fixed and grid leak bias. I bias my finals to 3 or 4 times cut-off, not just twice cutoff as the handbooks recommend. This requires more driving power but it reduces the conduction angle (shortens the grid drive pulses) and increases efficiency and modulation linearity. Just be sure not to exceed the rated grid dissipation of the tube.
    W2BTK, WZ5Q, N6YW and 1 other person like this.
  6. W2BTK

    W2BTK Ham Member QRZ Page

    I need to look into something like what you have there with the diodes, that is very cool. As of now the fixed safety bias is just always on, which I guess will work for initial tests. Thanks for the response, it's very helpful!
  7. W2BTK

    W2BTK Ham Member QRZ Page

    I think the MB-150 grid tank is ready to go and all sorted out. Here's a shot of it.

    Attached Files:

    W2VW likes this.
  8. W2BTK

    W2BTK Ham Member QRZ Page

    Well, I took a three month hiatus from doing anything with this, mainly due to work. However, the past couple days I have been able to work on it and made progress. You're basically looking at a finished transmitter and power supply at this point. Not much left to do. There are still a few things to be done including the entire 'finishing touches' process but I figured I would show what I have so far. The plug in tank coils are not behaving well. The HDVL for 40 meters is really close for the 3885 khz area. Might be able to do something with that. Otherwise, I will probably have to roll my own for the rest of the bands.I'll follow up with more when there is more.

    Attached Files:

    AG5CK, N1BCG, W2NBC and 1 other person like this.
  9. N2DTS

    N2DTS Ham Member QRZ Page

    A while ago I built a nice push pull triode rig:
    I was lucky enough to be able to get some good KW coil sets at the time.

    Later on (15 years?) I wanted to build another 812 rig modulated by 811's, a nice combo, and tried to make coils.
    Total mess, the thing did not tune up well at all and looked worse so I gave up on it.

    The rig above worked wonderfully, 15 kv vacuum caps out of paging transmitters for neutralization, they worked fantastic, the only rig
    that I ever got to totally neutralize...
    A nice 350 watts of carrier with as much audio as carrier, adjustable power supply using relays to select 1200 to 1700 volts.
    No fans or blowers.

    Good luck with your 810's!
  10. K4KYV

    K4KYV Subscriber QRZ Page

    In a push-pull final, one of the main concerns is balance. Each tube should draw the same plate current as the other. The only way to determine this is to use separate grid and plate current meters, one for each tube, which is the way I did my HF-300 rig. When I first built that rig I used 810s in the final, but the problem is that they take a lot of driving power to run in class-C plate modulation service. I replaced them with type 8000s, which are practically identical to the 810, except for the mu of the tube, but they take only about 75% the driving power. The grid drive and bias requirements are different, so I had to replace the grid leak resistors (in my circuit, a separate one for each tube) and re-neutralise; otherwise they performed the same. But eventually, as my 8000s aged, I couldn't get them to balance any more, so I replaced them with type HF-300s which I have been using for years. The characteristics of the HF-300s are very close to that of 8000s, practically the same grid current and bias voltage, and about the same driving power. I used the same grid leak resistors, and run them at the same grid current as the 8000s. I had to re-neutralise and replace the plate caps, since the HF-300s have a jumbo size plate cap and the glass envelope is almost 2" taller.

    Back to the balance problem; originally I swapped around tubes in my collection until I found a pair that ran close to the same plate current for each tube. Sometimes the grid currents would be off a little, but I pay more attention to plate current balance than to grid current balance; if the grid currents are way off balance something is wrong with one of the tubes, but if both tubes are good, the grid currents will match "closely enough for gov't work", so I don't worry about it too much. OTOH, if the plate currents are substantially out of balance you lose one of the advantages of push-pull, which is cancellation of 2nd harmonic output, plus you can't run full power to the pair without overpowering one of the tubes.

    A few years ago I came up with a solution to the balance problem, and now I can balance the tubes perfectly, assuming they are not too far off to begin with. I use a dual differential air variable capacitor, about 12 pf per section and the same plate spacing as the final tank capacitor. Mine was originally the neutralising capacitor removed from a parted-out military BC-339 transmitter, which used a pair of 833As in push-pull in the final, CW and RTTY only. It was made by Cardwell, 12 pf per section, with each section isolated from the other both stator and rotor. I turned one of the sections around 180°, so that it is at minimum capacitance when the other is at maximum. This small dual variable is wired in parallel with the main split-stator plate tank capacitor, which is 240/240 pf @ 7000 volts to cover 160-20m using plug-in coils. Adjusting the small air variable brings the tubes into balance if they are slightly off. I have found the balance to vary slightly from band to band, probably to slight non-linearity in the main tank capacitor sections.

    If the Cardwell neutralising capacitor is not available (I occasionally see them at Dayton and other large hamfests), a pair of circa 15 pf capacitors with the same or greater plate spacing as the final tank cap can be used, and usually these can be ganged together to work with one control knob. Someone with good machine tools, or even skilful with hand tools, could probably fabricate a pair of usable capacitors from scrap.

    Nice looking job, but one problem I see is that you used those crappy ceramic plate cap connectors that were effectively designed to pull the plate caps off tubes when you try to remove them if they have been in place for any length of time (speaking from experience). Good 812s are too hard to come by, to risk destroying them with those things!

Share This Page