741's are not a good choice for the control loop, they may be internally
compensated so easy to use, but the result is a shunt regulator with poor
dynamic regulation, and that is what you dont need.
Static regulation is easy, to sub mV variation if you need. Dynamic
regulation to keep that pesky G2 voltage stable when keying CW or SSB
contesting, is more difficult to do, and do well.
>From (vague) memory, Ian 'SEK (and some others) used the now less than
common 709 uncompensated op-amp, and did the overall compensation
externaly taking into account the shunt regulator device itself. Result?
Excelent dynamic regulation, if a little phreaky to setup if the layout
involves long wire runs. I've since then built G1 and G2 regulators into
the RF deck as that is where they are needed, plus the settings are
specific to the tube(s) used.
I've also experimented with 800V/15A power MOSFET's, from APT for shunt
reg use in the past, with quite good results. The only issue with them,
is the Gate input and "Miller" capacitances. But, with a low Z gate
drive (bipolar push-pull) things started to become good. Guess if you
wanted to, you could Cascode them with a smaller power FET in the lower
half. I've not tried that however.
I also used the tube G1 voltage as the reference for the G2 regulator.
OK, it's a bit more of a faff to set things up, but if a tube goes
G1-Cathode short, G2 then collapses as a result too, preventing any PSU
woes. (Important, if your plate supply is 'more than capable' of what
you need!) I first used this in a multi type Tube tester at work I
adapted, it now handles Y681's, 4CX600, 250, 350 series, 8122's, and
other similar things..
For a real RF amp, the G1 regulator output could be a nice low Z source
too, so you could drive the tube into AB2 if needed, with no operating
point wobbles. That also has the benefit if the tube goes G1-G2 short,
the G1 shunt reg then just dumps the G2 regulator source current,
shutting off the now failed tube. (Constant current reg, again based
arround an APT power FET, one bipolar and some passives.)
Another thing I did, while messing about with these things. Was to setup
another Power FET so that it was switched by the 'B Gate' output from my
'sope (a Tek 465) That in turn switched in/out another resistor that was
placed across the regulator output, so I could play with the dynamic
regulation, in real time. THEN! You find out just how good (or bad)
your active shunt regulator realy is, without p'ing off the occupants of
20m!
I also do a similar thing at work, with the broadband distributed amps,
to rapidly blank/unblank individual tubes (18 or more in DC parallel!)
when tuning the things after a rebuild, so as to make sure they all huff
and puff together, also (with an inductive coupler) to check the
individual plate currents are all about the same (ish!)
Regards to All.
Dave G0WBX.
On 2 Jan 2012 at 12:00, amps-request@contesting.com wrote:
> From: Mike Tubby <mike@tubby.org>
> Subject: Re: [Amps] [rfamplifiers] 4CX5000/10000 Screen power supply
> help needed
>
> Hsu,
>
> You can make a good shunt screen stabiliser using an op-amp and a
> tetrode, this was done many times for 4CX250B, 4CX1000A etc. before
> the Tetrode Boards from Ian G3SEK...
>
> See here (second half of article):
>
> http://www.agder.net/la8ak/d23.htm
>
> and here:
>
> http://www.qsl.net/g3uvr/4cx250/G4AJW-psu.pdf
>
> for the 4CX10000D you would need a bigger tube ;-) The BBC
> long-wave transmitter (198KHz) used to use a 4CX1500A as the shunt
> screen stabilizer for the 250KW tubes made by Thompson ESF.
>
> I'm just not sure which tube you should select for the job... perhaps
> someone on the list can make recommendations...?
>
>
> Mike G8TIC
_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps
|