Protecting G2 from overvoltage - How ?

[kagliostro]

As to clarify why I ask

I and some friends are planning an SE amp that uses a GU-50 as power tube

As many other tubes the GU-50 require on G2 a voltage that is way lower from B+

In our project B+ is around 420V and G2 is around half of it

We can adopt the Hiwatt style PS where two HT winding are rectified and daisy chained



this way if we feed the G2 from the first winding we are not afraid about rises of the voltage that feed G2

But if we use a common TA (without CT), rectify with FWB then use a Gyrator (or Mayda) to eliminate the ripple and we derive the G2 voltage from B+ using a VVR if the VVR Mosfet fails we have the whole B+ on G2 and this will not be a good thing (KABOOOM)

So I'm wondering if the use of a 250V Varistor + an in series Fast Blow Fuse connected between G2 and ground would be a sufficient protection

I've seen that some (EPC B72220-S 252) has a response time of 25 ns

What do you think ?

Franco

[Merlin]

Resistor in series with the MOSFET drain to limit screen current in case of MOSFET failure, long enough for the HT fuse to blow. Or use a cap-coupled rectifier?

[kagliostro]

Thanks Merlin

We have a B+ ~410V @ 110mA and ~230V @ 4mA (at idle) & 30mA (at max) on G2

I don't know if to obtain B+ from a capacitor coupled FWB will be a good idea, despite all 110mA aren't so few for such a configuration

Can you share the source of the picture you posted ? I've all Mr. Blencowe books, so if it comes from there I can read the whole chapter

Thanks

Franco

[astronomicum]

Designing Power Supplies for Valve Amplifiers, Second Addition, Chapter 3.5, Disparate HT Voltages.

Univox used disparate supplies on a number of their models. The 40-50 watt U1236 schematic attached is an example. Also see the U-1221, U-235-PB, and the U-1226. These are not capacitively coupled like the example you provided.

[kagliostro]

That uses a voltage doubler but not full wave

There are also other methods like in some old Philips PA Amps

They use a PT with CT, the half voltage comes from the CT

Franco

[astronomicum]

That uses a voltage doubler but not full wave

Half true. The B+ is full wave, the G2 as you call it is not.

[kagliostro]

We have think to this (Standard PT, no a la Hiwatt)



and the result, if the Mosfet is OK and not shorted is this (near no ripple)



D10 and D11 will act as G2 protection

Franco

[sluckey]

What is the purpose of the M1 circuit? The input is 400V but you have a string of zeners that total 410V. The zeners will never come into play unless the input rises above 410V. So the voltage at the gate will be the same as the input voltage and that will make the output voltage the same as the input voltage.

The only purpose I can see is the output will be clamped to no more than 410V. Is that what you intended?

[kagliostro]

Ciao Steve

The circuit was planned by one friend that is an engineer the LTSpice simulation was draw by one other friend

May be the 400V you read at the exit of the bridge are correct or may be it is a mistake (this thing happened other times, some of the voltages you can read are only a note)

I'll ask him

In the intention M1 is a stabilizer

and now I remember, the intention is to protect the components from over voltages at ignition

Initially the M1 was different and without those zeners

Franco

[sluckey]

May be the 400V you read at the exit of the bridge are correct or may be it is a mistake

The circuit makes perfect sense if the bridge output is greater than 400V.

[kagliostro]

I think the answer is this

... the intention is to protect the components from over voltages at ignition

Because we were discussing about to use 450V e-caps instead of 500V e-caps some time ago, so this new arrangement for M1, I think

Franco