Does 2A3 require a grid stopper?

[Ty_Bower]

Does the 2A3 absolutely require a grid stopper, or can it be considered optional? If I had a 2A3 amp that had none, should I add them? Does the presence or absence of feedback in the amp have any bearing on the question? If I were to add them, what value should I use? Would 1/2 watt, 100 ohm carbon composition be sufficient?

[kevinkr]

I use 1K grid stoppers with 45/2A3/300B mounted right at the socket as a matter of course.

Whether or not you actually need the grid stopper with a low transconductance type like the 2A3 is a matter of debate and the quality of your layout. (It's not debatable with higher transconductance types - they will oscillate even with very good layout.)

In many instances if the distance between the driver and output tube is small there will be no evidence of oscillation, more than a couple of inches of wire though between the grid and what's driving it, and a grid stopper becomes a very good idea.

I think it is good design practice particularly if you do not own a high performance oscilloscope that would allow you to sniff for RF oscillation. (With a suitable scope at some point in the verification process you should use a sniffer loop to avoid loading high impedance nodes which might prevent them from oscillating exactly when you are looking for oscillation.)

By habit I now use grid stoppers pretty universally, saving a lot of aggrevation later. I use a fair number of high transconductance/RF types in my designs. (Believe me I learned the hard way.)