Print

[ZS5WC]

 :)I want to convert an old Yeasu FL-1000 HF linear from 4 6JS6C to 4 QQE06-40 VHF tubes.
Each tube contains two tetrodes , so each tube has two anodes.
Each anode stray capacitance is 2.6 PF , so for 4 tubes with 2 sections in parallel is 20.8pF.
The Anode voltage is 850V, Anode current at full load will be around 800mA.
So using Rp(dynamic) =850/(2.1*0.8A)=around 500 Ohm.
Using this formula:https://people.physics.anu.edu.au/~dxt103/calculators/pi_tank.php
and using loaded Q of 12, it gives me values for the PI network: C1=76pF, C2=233pF, and L is 0.17uH.?
Considering Cp is 20.8pF, assuming layout etc.(Cs) will be another 20pF, leaves us 36pF on the Plate tuning Cap?.

Now, If I reduce the dynamic plate impedance, I assume by lowering HT, C1 in the PI increases.
Do you think my summation is correct?
If I lower HT, PEP output will be lower?.
73 de William
ZS4L / ZS5WC

[AA7IS]

Of course PEP will drop a bit.
I'm wondering how much of the actual amp are you using at that point? the box, the HV and filament xformer, the tube sockets? maybe a meter?

I don't think sticking multi-element tubes will be a good idea for 50 Mhz so much capacitance between all those elements that will be very hard to calculate..
I've been around ham radio since the 1960's and can't recall seeing anyone running four tubes on 6 meters, of course anything can be done but a two tube solution would be so much better, with that tiny 850 volts you don't really have any option to use a much better tube.
I think I would try to find a xformer that will fit into the box that will give better then 1200 volts then you can use a lot of different tube pairs to get 500 watts or better.
850 volts is just so low........

[KD6VXI]

Voltage doubler and a single or pair of Gi7b tubes.

Done.

1kw input all day long.
--Shane
WP2ASS / ex KD6VXI

[ZS5WC]

 :)Thanks  I think the GI-7B will work but only single tube, as supply having been doubled again will be too weak.
850V is entirely suitable for these VHF tubes, in fact hv is over the ratings , but you can run them off about 1200v at a push.
I have the tubes and bases, my worry or question is the stray capacitance for 4 tubes, if I will have enough reserve so to speak to tune the plates.
I have seen a similar design implemented with 3 tubes. 350w PEP out which is a good start on 6m.
73.

[G3RZP]

In plate modulated AM, those devices (aka as 5894 in the US) are rated for 600volts of HT in ICAS, which is 1200 volts peak on the anodes. But you would really need a regulated screen grid supply and a higher voltage bias supply.

[ZS5WC]

 :)100% will use screen regulated supply.
And perhaps trip circuit.
MY question still is: using the the PI network calcs on line, if the use a lower dynamic anode impedance, i.e. you lower plate volts and increase anode current by adding perhaps another tube, the value of C1 increases.
C1 being: (Plate stray capacitance, chassis stray capacitance and VC1 variable capacitor)
So in effect you increase the plate stray capacitance, but because C1 increases by an larger margin you can ignore the extra plate capacitance.
Does this seem right?.
de William, ZS4L

[VE7RF]

If you are using the superb  GM3SEK  PI  spread sheet,  you will  notice that you can insert any value for stray uh  between the anode and the main C1 tune cap.  By inserting a tiny amount of stray uh (typ fractional amount), this extra inductance + the tube's stray C forms a step down L network.   That will transform the plate load Z down to a much lower value..... low enough that a practical PI network can  then be designed and built, using a reasonable tank Q.

That trick is used all the time on the bigger metal triodes and tetrodes on upper HF and 6M bands.

Typ, you don't want the value of the extra coil to be larger than the main tank coil.

But it only works on tubes that don't require a parasitic suppressor.

I would be using loaded B+ / 1.8 X plate current to arrive at the plate load Z.
On the GM3SEK PI spreadsheet, there is various options to derive the plate load Z.

If a suppressor is not being used, just enter zero for  the parasitic components.

If you have an LCR meter, you can measure the anode to chassis  stray C, so you know exactly what you are dealing with.

[ZS5WC]

 :)Thanks for your valuable input!.

Will look at the GM3SEK spread sheet too.
Yes , these tubes have parasitic suppressors but also have internal neutralization.
Does that mean I cannot add series L?. worth a try if needed anyway.

Many Thanks,
William

[VE7RF]

:)Thanks for your valuable input!.

Will look at the GM3SEK spread sheet too.
Yes , these tubes have parasitic suppressors but also have internal neutralization.
Does that mean I cannot add series L?. worth a try if needed anyway.

Many Thanks,
William

The GM3SEK  PI  / PI-L spread sheet is the gold standard.   Everybody's else's  version does not factor in stray L and stray C.

When they talk abt loaded Q, the old method that Eimac and the ARRL used was based soley on the C1 portion.
IE: old books said to use a Q of 10.  In reality, that equates  to an overall network Q of 12.....and an input Q ( C1 end) of 10...and an output Q ( C2 end) of 2.    Input + output Q = total network Q.

You want the total network Q as low as possible.   Or you will cook the tank coil.  On the spread sheet, only the overall network Q can be tweaked, then it spits out both the input + output Q.

The lowest u can go, is = the sq root of the plate load Z / output Z (50 ohms).    IE: say a 4-1000 with a plate load Z of 5 k ohms.   5000/50=100.   sq root of 100 = 10.   Q can't be lowere than 10.   You will see this effect when using the spreadsheet. ( it will blank out with rows of ### signs.

I also use the spreadsheet to caluclate PI tuned input networks. Enter 50 ohms where the plate load X goes..and 50 ohms for the output.  Problem is, the spreadsheet will not handle 2 x equal X's.   The...'fix' is to change just one of em slightly.
IE:  50 ohms on input...and 49.999 ohms on output, problem solved.

Notice on tuned inputs, like 50 ohms to 50 ohms, with an overall network Q of 4, the input + output Q will both be 2.

Jim  VE7RF