[Amps] Novel Screen Grid Regulator
4cx250b at muohio.edu
Sat Sep 29 13:25:19 EDT 2012
Thanks for your comments, Paul, Carel, and Carl. I think a zener
across the + input is a good suggestion, Paul. I hadn't worried about
protecting the op amp because the 50:1 voltage divider would reduce the
voltage spike at the input to 8V, assuming the TVS did its job. According to
the data sheet, the TVS acts in a nS or so and fails in a shorted mode, so
there shouldn't be a problem with it. Still..... one can't be too cautious.
And Carl, I can't find any info on the IXCP0M90S. You sure you typed
the part numbeer correctly? And Carel, that's an interesting circuit, which
is similar to the Alpha circuit (except for the polarity and use of bipolar
devices). Thanks for flagging it.
Now I'm weighing a design decision for the amp as to the best way to
key it up. Normally, in standby one leaves on the plate voltage and screen
voltage and applies cutoff bias (for the GU-74B about -150V). When the amp
is keyed, the bias is reduced to about -50V,to get the proper resting
An alternate way is to keep -50V operating bias on the tube at all
times, and key up the amplifier by keying the screen voltage. I see two
benefits to the second approach. The first is that the bias circuit is
simplified, since one only needs a single -50V bias supply. The second
benefit is that the approach eliminates possible turn-off problems for the
amp. In my design, when the amp goes off-line a vacuum relay instantly
interrupts the HV. If I use the first approach, I'd need another very fast
way to kill the screen voltage at turnoff, because it would otherwise decay
slowly compared to the HV, thus potentially damaging the tubes.
The main problem with the second approach (keying the screen
voltage) is maintaining flashover protection. I'd like to use a fast solid
state relay to key the screen voltage. Previously I've used the Clare
PLA-140, which handles +/- 400V at 250mA and comes in a 6pin DIP, and I
like it a lot, but it would never stand up to a flashover current pulse.
There are similar issues with implementing a screen current trip circuit. So
anyway, I'm still mulling over possibilities....
> Very interesting circuit since the MOSFETs function like a high-power,
> active voltage divider. I did note that the non-inverting input to U1a
> be susceptible to the effects of both the normal screen voltage and
> a flash from the screen. However, provided that Z1 acts quickly enough
> R10 remains intact, then concern is probably minimized. Still, it may
> sense to Zener-clamp the non-inverting op-amp input to a level just above
> the value found in normal operation. Then, if adding Zener-clamping, I
> think it would be a good practice to add series R at the non-inverting
> of 1K. The input Z to the non-inverting input of the LM358 is quite high,
> allowing for some additional series R protection.
> Paul, W9AC
> I made (and used for my current 160mtrs tetrode th289 amp) almost the same
> for grid supply. But will work also good for positive screen supply (use
> and reverse diodes).
> 73's Carel.
> Ive been using the IXCP0M90S, good to 900V at 1-100ma. Keeps those
> screen grids safe from over current or collateral damage when there is an
> internal failure.
> It does keep 8122's happy in my case.
More information about the Amps