Grid Stopper

[lndm]

If they are needed. For me this is the big question as it leads to how to find out.

So far, EC8010 mentioned the 10kHz square wave. Is it necessary to vary this 10kHz at all to catch a high Q oscillation? How about levels?

Miles Prower mentioned a xmitter that I see as potentially intermittent.

Can you avoid:- pressing an apparently stable amp into service only to have it oscillate later?

[Tubes4e4]

Hi Hans,

BTW the grid is not a true spiral in most tubes, it is made like a spiral but the turns are short circuited at 2 places along each turn [...]

You obviously got a point here

Tom

[tubelab.com]

Can you avoid:- pressing an apparently stable amp into service only to have it oscillate later?

Yes, this is very possible. If the user swaps tubes and the new tube is different than the tube that the amp was originally tested with (aren't they ALL somewhat different). It can also happen if the operating conditions are different from your test conditions (temperature, supply voltage, load, output level).

I try to test my amps with as many different tubes as I can find. If I find that a stopper is needed, I try all possible tubes, pick the worst one, find a suitable value for a stopper, and double it. If a higher value still leaves plenty of room in the high frequency department, I will go higher. On a guitar amp I will often add a small (10 to 100pF) silver mica cap from grid to ground after the stopper.

How do I test? I run the amp with sine, square, and triangle waves of varying frequency AND amplitude. If you are testing an input circuit that is connected to a volume control, try it over the entire range. I have found that the worst spots are at the two extremes, and exactly in the middle. If you are dealing with a tone control stack, you must try as many combinations as possible. If you are dealing with an output stage, try it with a deliberate mismatch. Speakers are not a constant impedance, so your amp will see a varying load. I test all amps with at least a 2:1 overload and a 1:2 underload. In otherwords I run an 8 ohm amplifier at 4 ohms and at 16 ohms. I also test my amps under a varying supply voltage using a Variac. I will usually monitor the output with one scope channel, and the driver tube plate with the other scope channel. Avoid attaching a scope probe to any grid, this can add enough load to stop an oscillation. Often an amp will oscillate breifly as it enters or leaves clipping, so I drive just into clipping with a triangle wave, and vary the frequency.

Is this extreme? You may think so but I have had amplifiers come back oscillating even after they have passed these tests. They have been guitar amps, which have a large amount of gain in a small package, that is purposefully overdriven, and often operated with poor electrical power, and 2 or 2 extra speakers connected (usually in parallel). That is what my extreme testing is trying to duplicate. It is often impossible to duplicate the users environment in a lab.