[Amps] The GU84B and GU78B tubes / some brief history

[Ian White GM3SEK]

Paul Christensen wrote:

>
>If you could list the top few important parameters to low distortion 
>tetrode sign, what are they?  I think most of us know the importance of 
>well-regulated screen voltage.  Other top factors?
>

First of all, do everything that you would for a triode: give it a 
reasonably stable heater voltage, and an anode voltage that remains 
reasonably stable at maximum current demand.

Both tetrodes and triodes require a stable G1 bias voltage (relative to 
the cathode) but this is done in different ways, each with its own 
distinctive set of advantages and pitfalls. GG triodes create negative 
grid bias by taking the cathode positive with respect to ground, using 
the familiar zener diode or "electronic zener" to create a constant 
voltage drop. This is very simple because no external bias supply is 
required, but the challenge is to maintain a constant voltage drop when 
the cathode current is varying strongly.

Tetrodes do it differently, but they always require a separate negative 
bias supply for G1, so that adds some complication. Voltage 
stabilization is much easier because very little current is involved, 
but it's important that the G1 bias voltage remains constant if the tube 
is driven into small amounts of either positive or even negative grid 
current. If the grid voltage varies when current flows, that will 
produce a sharp kink in the transfer characteristic, possibly leading to 
severe high-order IMD.

Another feature of the G1 supply is that most tetrodes have a very small 
grid dissipation, so they require some kind of protection against 
excessive grid current. In the normal setup for class AB1, my boards 
will generate an ALC or warning signal at about 100uA of grid current, 
and will trip the whole amplifier at about 3mA (which is well below the 
threshold of physical damage to the grid).

As for the screen supply, the obvious point is that tetrodes do require 
another separate power supply, with good voltage regulation for both 
positive and negative screen currents. Protection and fault recovery 
adds yet another layer of complication. Screen supplies are covered at 
much greater length in my QEX article: 
<http://www.ifwtech.co.uk/g3sek/boards/tetrode/tetrode-3.htm>

And then you have to think how to switch between RX and TX. If full QSK 
is not required, I strongly favor switching the screen to ground (in the 
typical DC-grounded cathode setup). This has the advantages of 
definitively cutting off the anode current in the RX condition, leaving 
the screen grid safely grounded and the screen supply safely 
disconnected. Protection in the RX condition is very important because, 
when you think about it, this is how an amateur amplifier actually 
spends most of its operating lifetime!

The problem is how to switch these relatively high screen voltages 
RELIABLY at full QSK rates. MOSFET switches have been tried, but are 
destroyed by the first HV arc that comes along... so that means 
mechanical relays, which have their own different set of problems 
including a limited mechanical lifetime. The best scheme I've been able 
to come up with is to use electronic grid-block keying for QSK while 
*also* switching the screen relay with a short hang-time to avoid it 
switching on individual dots and dashes. This gives the advantages of 
both switching methods but requires some careful sequencing.

To summarize, tetrodes require much more support circuitry than triodes 
do... at least, they do if you're intending to get everything right.


Well, that was the short(er) version, and before the second cup of 
morning coffee I'm sure there are other things I left out...


-- 

73 from Ian GM3SEK
http://www.ifwtech.co.uk/g3sek