pi Tank Tune & Load Values for 80M

[Naysayer]

Homebrew 4cx300a amp grid driven:
I found a better air variable for my Load cap. Installed and working OK 10-40M but I need to estimate the C values for 80M. New Load cap is another 2 section Johnson, 3kv rated but the smallest section <10pF and both sections equal 205pF. I use a CRL 2500 series, progressively shorting switch to add capacitance for 40 & 80m. I had 40M working fine on the old load cap somewhere around 320pF and it working OK again in that range so that’s my starting point on the Load.

Generally I noted my Tune and Load C values are lower than the values I see on other amps using small Tets. ‘Stray Capacitance from Roller’ causing the lower values is what AI suggested. My Vacuum Tune only goes to 110pF and it’s enough for 40M with pF to spare. Have a latching Kilovac if needed for 80 on the Tune. I will use CRL 6kv ceramics as padders (100, 200 & 300 values on hand).

How to estimate the Capacitance needs for 80M? My best guess is somewhere below 900pF on Load and <250pF on Tune given I have 40 working with <350pF on Load. Is it reasonable to double that for 80 to start?

Pride’s DX-300 uses a 3 section air V with 450pF max each (1 tube) but there’s no minimum value I could find online. It looks at least 40-ish if not higher. On 10 meters my tube pair needs <20pF on Load so adding another tube has me in unknown but lower pF territory.

Am I stuck with trial and error?

 

[brandon7861]

Since none of the tube guys are replying, I'll take a crack at it. I'm probably wrong. Maybe one of the guys here that do this can chime in. Edit: I am NOT a tube guy.

Commence guessing: I assume one tube, AB, 300w, don't know your plate voltage. But if I assume 2000v, that's 1800v peak swing at 90% supply swing, and that's 1273V RMS voltage swing. Ohms law says 1273^2/300=5400ohm for the load resistance.

I don't know what your inductor value is in the output network, but you basically have a pi network with about 5400ohm on one side and 50ohm on the other. Should be simple to work out the caps once you have the inductor value.

 

[brandon7861]

I just did a sanity check using the DX300 schematic as a baseline. Since it uses the same voltage, I assumed the same load (well, i put in 5kohm this time) and 3.5MHz. I see in the A position, the load capacitor has a max of 450p x3 + 160p x2 for a total of 1.67nF max for the load capacitance. On the tune side there is 110p + 165p for a total of 275pF max at the plate tune. I then calculated the total inductance based on the diameter and turns listed in the schematic plus the approximate length of those coils via pictures online. That came to 12uH, and in position A, we have all turns active. Off to simsmith to see if it matches.

yes it does, and we didn't even need the max capacitances available either.

 

[Naysayer]

Thank you Brandon. I’m going to make another try this afternoon if possible.

The online calculators wanted values I did not understand and was unable to provide. Your walk thru showed me how to use and apply that formula.

I’ll report end values but it will take time, it is ungodly hot in NYC so life slows down a bit here. Thanks again for the lesson.

 

[brandon7861]

I think when you use a website like this:

VK1SV Pi tank calculator

you will notice that at 5000ohm to 50ohm, the recommended Q that AI provides (about 10-15) will result in an inductor value that is larger.. Trying to solve with a 12uH inductor and a low Q returns an error for a pi network. It suggests an L match at low Q. Just raise the Q until it puts out an inductance that you can live with. It requires a Q of about 20 at 3.5MHz to use the 12uH inductance I estimated from the amplifier photos online.

It is important to know that a higher Q = higher circulating current, higher capacitor voltages, and the adjustment becomes much more critical. This means don't go nuts with the Q in order to use a smaller inductor.

repeating my disclaimer: I am not a tube guy and I have not done this before. Just going by what I think I gathered after few quick google searches and talking with AI and playing with simsmith.

Edit: seems the guys here actually building kilowatt amps are afraid to chime in. I find that a little odd, but ok.

 

[brandon7861]

not to ramble on, but this paper talks about the formulas for calculating a transistor and a tube Rload, and in this paper, it uses the formula Rload=Vplate/(k*Ip). It then shows that k for class AB is 1.6, and when I punch that in with the max plate current of .25A, we again get 5kohm for Rload. Just another sanity check.