|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
150.148.218.189
In Reply to: RE: Klyne, Beveridge system posted by AJ on February 26, 2014 at 16:42:00
Good idea. Bill Thalmann actually did that already on my Bev amps. There are four resistors that link the solid state input stage to the tube output stage. Two 1K resistors go respectively to the grids of two of the 4 output tubes (grid-stoppers). The other two are 7.5M and go to the cathodes, I think, of the bottom tubes in the totem pole arrangement. Bill put ferrite beads on the afferent leads (where the resistor leads solder to the input board and before the body of the resistors) of all 4 resistors, early on, in our attempt to quell oscillation. It definitely helped. Then I installed a shield made with TI Shield (Texas Instruments), between the input/driver board and the output stage board, in each amplifier. I really think that what may have happened with the one amp that is giving trouble is that I installed four new output tubes in that amp, which probably brings up the transconductance and enhances the tendency to oscillate. So, something more is now needed. Like I said, the other amp, still on its original tubes, does not oscillate and sounds excellent. The other thing I did that probably does not help was to bypass a 10uF electrolytic that is used as a coupling cap with a 0.1 uF teflon film cap. Possibly the electrolytic was deliberately chosen to limit bandwidth and reduce the tendency to oscillate, until I "tweaked" the circuit and thereby created problems of wide bandwidth. But as a true audiophool, I could not tolerate an electrolytic coupling cap.
Follow Ups:
I know the characteristics of new tubes can cause issue in circuits where old ones existed.
I doubt the 10uf cap was chosen deliberately as a bandwidth limiting etc part. You could try a Solen 10uf in that position it has to be way better than the electrolytic.
If you want to try one, I have a couple that I can send you for the experitment.
This kind of spurious oscillation requires a good scope and injection of signal with lots of time to try grid stoppers, ferite beads and different values for bandwidth.. etc.
Time and frustration like chasing hum and radio frequency noise.
AJ
Thank you for the kind offer, but I've got a surfeit of suitable 10uF film caps I could use. For now, I might try a bipolar lytic in place of the polar one, was going to do that anyway. Don't want to introduce another variable by going to film. I don't know enough about this problem to argue one way or the other, but some knowledgeable guy on TubeDIY did think that the electrolytic was at least in part chosen for its limited bandwidth. Even if that is correct, I've got the 0.1uF teflon bypassing the lytic, so bandwidth is restored, for better or worse.
The maddening thing about trying to cure this problem in Bevs is that you can only model it based on behavior when the amp is connected to the speaker. When the amp is on the bench, disconnected from the speaker, Bill found it to be even more prone to oscillate, if anything. So you have to try a strategy with the whole assembly, using real music input. Then take it apart again to try something else. But keep in mind, I am having the problem ONLY when I play SACDs and ONLY on one channel.
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: