[Amps] low Q output matching

[peter chadwick]

Carel,
When the amplifier is running in any class other than A, (except in one case which I'll deal with later), the tank circuit 'rings' with each current pulse, and 'fills in' the voltage swing at the anode. So if you just had a wideband transformer, you'd only be putting current in for around half a cycle - maybe for 220 degrees in Class AB. Now the exception for that is if you run in push-pull, as the solid state amplifiers do, and as audio amplifiers used to do with valves. 
So there comes a point as Q is reduced where distortion will start to rise. Also, the power output will start to drop: exactly where that point is, I don't know, but my feeling is that going below a working Q of around 5 will probably show this up.
Untuned valve amplifiers in Class AB use push-pull, distributed type amplifiers: for just one  band, that's unnecessary. They are also very inefficient in terms of power in for power out. The other problem with a transformer is the shunting effect of the output capacitance of the valves: this will make the anode load reactive and mean the valve is seeing an elliptical load line, which in turn produces some phase distortion: that phase distortion can lead to degraded IMD. 
73
Peter G3RZP

========================================
Message Received: Nov 12 2012, 02:40 PM
From: "Carel, pc5m" 
To: amps at contesting.com
Cc: 
Subject: Re: [Amps] low Q output matching

Hi Peter (and Gerald)

Basically my aim is to get the simplest output match circuit possible
(simplest = lowest value of C's, because high valued, high current, high
voltage variable C's are difficult to get and/or expensive). I don't mind
even 0 dB of harmonic suppression.... because that is easily solved
afterwards with fixed valued C's/L's lowpass filter. My question is
basically are there any other parameters I am overlooking (beside the afore
mentioned harmonic suppression) if I go for this ?. In essence, could I just
a high power 3kW broadband 50 to 1700 Ohm RF transformer and the tube would
be happy ?
Gr, Carel.

-----Original Message-----
From: Amps [mailto:amps-bounces at contesting.com] On Behalf Of peter chadwick
Sent: zondag 11 november 2012 9:32
To: amps at contesting.com
Subject: Re: [Amps] low Q output matching

What is the value of the lowest Q?
What harmonic suppression are you aiming for? It should be 43dB + 10 log P,
without exceeding 50dB. If there's a tuner on the antenna, that may well
mean you don't need the low pass filter.
73
Peter G3RZP





========================================
Message Received: Nov 11 2012, 07:46 AM
From: "Carel, pc5m" 
To: "'AMPS'" 
Cc: 
Subject: [Amps] low Q output matching

Hi, I was wondering what else, besides bad harmonic suppression, will happen
when I design my (single band) output matching network (PI-L) with low Q
(lowest Q needed to get the match done) ?

When using low Q design I will get
- much smaller (variable) capacitors, especially for the bigger load C, so
easier to get components
- broader no-tune range
-smaller RF stress on the components

As an example for a 160mtrs amp, with Rload of 1700Ohm and intermediate
impedance around 330Ohm:
With Q of 10:
CLoad =1200pF
L-l=10uH
Ctune=293pF
Lpi=33uH



With lowest Q:

CLoad=400pF
L-l=10uH
Ctune=132pF
Lpi=33uH



When you plot both matching networks in a Smith diagram you see immediate
the difference. Low Q is much more straightforward.
To get the needed harmonic suppression a lowpass filter with fixed
components is easy to construct (first harmonic at 3,8MHz needs 25dB and
second one 20dB to get the same values as higher Q design). But maybe I have
overlooked something . Thanks for your ideas 73's Carel, pc5m





_______________________________________________
Amps mailing list
Amps at contesting.com
http://lists.contesting.com/mailman/listinfo/amps
_______________________________________________
Amps mailing list
Amps at contesting.com
http://lists.contesting.com/mailman/listinfo/amps

_______________________________________________
Amps mailing list
Amps at contesting.com
http://lists.contesting.com/mailman/listinfo/amps