Opamp as tube driver: Protection ideas?

[Franz G]

Hi

Maybe this is the better forum for this amp, than the chip amp forum?

Here, I showed my latest, opamp driven EL84SE amp:

http://www.diyaudio.com/forums/showthread.php?postid=457912#post457912

You realize, that there is no capacitor between the opamp and the tube.

A tube failure could be very fatal for this expensive opamp. Do you have some good ideas, how to protect the chip from the 300VDC from the tube, in case of failure?

Maybe some relais, grounding the opamps output? But how to drive the relais?

Franz

 

[Frank Berry]

You might consider adding a resistor (10k) between the opamp output and the tube grid. Then add a 15 volt zener diode between the opamp output and ground.
If the tube shorts, the resistor will limit the current going back towards the opamp and the zener will clamp the voltage to +15 volts. If you use a very low wattage resistor, the resistor will blow before any damage is done elsewhere.

Just a thought.

 

[moamps]

zener + dus in series

 

[Circlotron]

You might consider adding a resistor (10k) between the opamp output and the tube grid. Then add a 15 volt zener diode between the opamp output and ground.
If the tube shorts, the resistor will limit the current going back towards the opamp and the zener will clamp the voltage to +15 volts. If you use a very low wattage resistor, the resistor will blow before any damage is done elsewhere.

Just a thought.

The opamp output has to swing below ground.
How about a 30v schottky diode from opamp output to each rail and then a resistor to grid as Frank suggests. I say a schottky instead of a normal diode because it's ~0.2v forward drop would not be enough to allow any forward bias of any pn junctions inside the opamp should your tube quit catastrophically.

Would a 10k resistor have any effect on frequency response because of Miller effect?

 

[Frank Berry]

"The opamp output has to swing below ground."

Absolutely correct, sir. Sorry, I should have suggested two zener diodes, series connected, cathode-to-cathode.

"Would a 10k resistor have any effect on frequency response because of Miller effect?"

Probably not. 10k is a low impedance when compared to the impedance of the control grid.

 

[moamps]

THow about a 30v schottky diode from opamp output to each rail and then a resistor to grid as Frank suggests. I say a schottky instead of a normal diode because it's ~0.2v forward drop would not be enough to allow any forward bias of any pn junctions inside the opamp should your tube quit catastrophically.

Hi,
according your idea, maximum back voltage from faulty tube will be Vcc+Vschottky what is IMO too high. Opamp's output impedance is very low and this output will sink all current from tube thru grid resistor. In this moment, voltage at opamp's output can be well under voltage limit Vcc+0,2(Vschottky). Better idea is IMO R1-SCHOTTKY-R2 network, where R1 is defined from (Vcc+Vschottky)/Iopampmax, R2 is +Btube/Ischottkymax.

Regards
Milan