neg grid voltage limit

[uneumann]

I never paid much attention to the negative grid voltage limit in the 12ax7 specs. JJ and others spec -50V wrt to the cathode. Since many grids are AC coupled and some big signals can hit the cap, I'm sure I've seen more the -50V on preamp grids. I'm running into this with an OD channel I'm testing and wondering just what happens when the limit is exceeded. Is the failure mode simply arcing from grid to cathode, or is there something else going on? Has anyone ever had experience with this limit causing a failure?

[66 Kicks]

-50V or -55V is the maximum negative DC grid bias. 0V is the maximum positive DC grid bias. Neither of these places any limit on the instantaneous voltage of the signal applied to the grid.

[uneumann]

-50V or -55V is the maximum negative DC grid bias. 0V is the maximum positive DC grid bias. Neither of these places any limit on the instantaneous voltage of the signal applied to the grid.

OK - thanks. So - the spec is for continuous DC - not AC voltages.
That's comforting since I don't expect continuous large negative voltages.

I still do wonder what happens if the spec is exceeded. What does a triode do with -100v Vgk?

I may have to try it sometime just to see.

[SoulFetish]

OK - thanks. So - the spec is for continuous DC - not AC voltages.
That's comforting since I don't expect continuous large negative voltages.

I still do wonder what happens if the spec is exceeded. What does a triode do with -100v Vgk?

I may have to try it sometime just to see.

Do it.. What's the worst that can happen?



Then, report back you're findings, if you please.

[J M Fahey]

Positive voltage can actually reach a few + Volts, only problem is that grid will start passing current, but if source can supply it, so be it.
No damage usually because source driving it in general canīt supply more than a couple mA.
Now if you plug a speaker out into a jack connected straight to a grid (no 68k or whatever series resistor) I expect damage.

As of negative voltage, a previous tube can routinely supply 150V peaks, and maybe up to 250V if waveform is strongly asymmetrical, so far I never heard of a tube damaged by that.

I bet you would need scary voltage, way above 300V to cause a grid to cathode arc; if curious you might experiment with some old worn tube and post results.

[R.G.]

It's an interesting case in imponderables. I think you're right J.M. - the real threshold would be when you get arcing to another electrode. 300V would be a good guess. Probably this varies a lot from tube maker/batch to tube maker/batch.

Fortunately, its easy to prevent damage if it turns out to be a problem. A zener clamp would take care of any issues in some really strange circuit if that existed. As you said, there have been a lot of tubes AC connected to plates with hundreds of volts of swing on them for a lot of decades and no problems have turned up. A clamp would not have audible effects (unless it caused issues in the preceding stage from the current) because the tube has long since cause plate current to cut off.

And the business about driving grids positive is absolutely true. In power tubes, it's what one does to get more power out of the output tubes.
A side effect of driving grids positive is that the tube gets into a conduction region where the plate current / grid voltage starts squashing and it's useful for softer distortion. Providing enough grid drive to push the grid into conduction a bit can be sonically useful if done carefully. Driving a grid from a high impedance source causes abrupt clipping at Vgs ~~ 0 from the inability of the source to drive it positive.

Tubes can and were driven "in clamp" especially in tube logic circuits. The grids were tied to B+, not ground, with high value resistors. It looks screwy, but it amounts to holding the tube constantly on but with the ability to drive it off by negative signals on the grid. It took me some time thinking about this to let it sink in.

[pdf64]

-50V or -55V is the maximum negative DC grid bias. 0V is the maximum positive DC grid bias. Neither of these places any limit on the instantaneous voltage of the signal applied to the grid.

My sketchy understanding is that limiting voltages were 'ac+dc', eg the 180V heater to cathode limit, and only the dissipation was a time averaged parameter; why do you think that the limiting votlages are in regard of the idle / operating point, rather than instantaneous values?
Bear in mind that 'design centre' is the most conservative of the rating systems, the 'absolute max' limit may be somewhat (perhaps significantly) higher.
https://tubedata.altanatubes.com.br/...10/e/ECC83.pdf

[Helmholtz]

I think 66 Kicks is right.

From the 1975 RCA Receiving Tube Manual (12AX7A):

Heater-Cathode-Voltage:
Peak value: +/- 200V max
Average value: 100V max

Grid voltage: (Design maximum ratings)
Negative bias value: 55V
Positive bias value: 0V

According to the general section of the manual instantaneous limits are characterized as peak values.