|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
50.0.128.73
In Reply to: RE: I think I fixed the problem with my Beveridge OTL DD amplifier posted by Lew on October 13, 2020 at 11:50:29
Lew,
I applaud your sense of Caution!
When I worked on these amp with Roger in the late '80s-early '90s,
we would typically set the unit up, turn it on to get some measurements, then turn it off and walk away for a day (at least) before coming back to it.
Roger would let me know what needed to be changed/repaired/replaced, and after waiting - I would get to it-
Yes, there are two amps within the enclosure - the signal amp - ss to bring the signal up a bit - before feeding it to the HV supply (the AC signal ran on top of the DC supply to the diaphragm)...
The big issues with the Tube amp - the HV amp for the stators & diaphragm was the rectifier, and then the carbon comp resistors.
I would replace all of those components...
If you want to play with the signal amp - go for it- but be careful not to increase the gain.
Happy Listening
Follow Ups:
I already mentioned this to 6bq5 via private email, but I thank him for pointing me to check the CC shunt resistors in the "good" amplifier. What I found when first I paid attention to the DC voltage readings was that ALL of the CC resistors were either about 20% above their nominal resistance value or had completely gone short circuit, meaning the meter could read their value as "OL". One other read about 1.1M ohms. This was causing aberrant current sharing across the PS capacitors. (The Beveridge PS has to develop 3200V in two 1600V sections. It uses 8 X 450V electrolytic capacitors in series, between ground and the 3200V output, in order to do that. Hence the necessity for the shunt resistors to promote current sharing.) I noticed that some of the capacitors were not seeing their fair share of the voltage, and others were seeing way more than their 450V limit. Further, as a result the voltage was drifting under my observation. So, if I had just put this amplifier back into service, it would probably have failed catastrophically at some point, soon. In fairness, the CC resistors are ~40 years old, as my amplifiers are dated "1979". I tossed them, and I tossed the filter capacitors too, although they had been changed about 4-5 years ago after I bought the system. Only about 4 out of 8 were grossly defective, but I don't trust the remaining 4 to hold up. It was folly to have changed the caps without also changing the shunt resistors, but live and learn. Now I will rebuild with new parts.
Edits: 10/18/20
although I have always thought of you as a knowledgable guy, I didn't realize until now that your hands on experience with these amplifiers was in the presence of RAM and with access to his ideas and judgement.In the amplifier that was oscillating (or at least was very noisy,whether or not it was truly oscillating), I found that ALL of the eight 470K ohm carbon comp resistors that are used to shunt or as bleeders, were "bad", in that the measured R was either "OL" or > 1M on my Fluke. Knowing that it is the fate of CC resistors to increase in value over time, unless they are made by Allen Bradley apparently, I first replaced them with Roederstein metal film types. To make matters worse, and because the 2W Roedersteins are physically large, I mounted them on the "bottom" side of the output board, the side with the tracings, facing the input board. That turned out to be a real no-no. The oscillation got worse. These amplifiers do not like any parts on the side between the output and input boards, and there is apparently a good reason why RAM or HB used CC types in the first place. I eventually had to go back to CC types, and they had to be mounted as per original. (CC resistors in "no-man's land" were also not tolerated.) One of many lessons I learned. I have to think that the shunt resistors were way over value for quite a while before I discovered the problem, yet most of that time the amplifier played music, albeit it was always much more unstable than its mate. I've got the "good" amplifier also on my workbench, so I can compare the two during the process of quieting down the noisy one; I need to check the CCs on that one too. In contrast, the 1K CC resistors that seem to be grid-stoppers are perfectly OK even after 40 years.
Five years ago, after I purchased this system, Bill Thalmann in Springfield, VA, worked on the amps to do some things I didn't feel skilled enough to do, given the voltages involved, like replace all the tube sockets. At that time, Bill installed all new HV rectifiers, so we are OK there. Except, during my early trouble-shooting, I de-soldered one end of one of the rectifiers in order to check it. When I tried to re-install it, I boogered the PCB tracing for the lead that goes to ground (cathode I think). I first got around that by soldering the end of the rectifier to the tracing but in a way that it was not secured to the board. But it looked just fine. Three or so months later, after feeling a sense of defeat over and over again, and in desperation really, I decided to do a better job of securing the rectifier to the board, for no good reason. THAT turned out to be the final step to making the amplifier quiet. Although the solder joint I first created, direct to the tracing, looked totally OK, it must have been "cold".
Edits: 10/14/20
You mentioned that the cc resistors would not work on the bottom side of the board either- so the bit about the metal films may well be inconclusive. There are newer metal film types now that are about 1/4 the size that you would expect: a 1 watt version of these is half the size of a 1/2W cc resistor! They are also lower noise in our experience.
The fact that the amp is this easily upset suggests to me that it operates right at the edge of its phase margin- a tad bit too much feedback. Due to the kind of load it has to drive, opening up the bandwidth with newer parts seems unlikely to help out.
Well, the CC resistors as shunts mounted on the bottom side of the PCB did not cause as much noise as did the metal films mounted in the same position, but the noise was further reduced when I moved the CCs to the top side of the PCB, where they were originally. As you can well imagine, my goal is to enjoy the Beveridge speakers, so I did not dwell on the question of whether metal film resistors could be made to work, or not. I had both types on hand, so I went with the most optimal choice. True though, the schematic calls for 470K/1W. I bought Roederstein 475K/2W from Michael Percy Audio, just because he is reliable and fast; they turned out to be fatter than the OEM CC types in 1W rating. Thus they really had to be mounted on the bottom side of the board because no room on the top side. There are plenty of carbon film types that would fit nicely on the top side too, like the KOA Speer from Mouser. But I had CCs, and I knew they would work, at least without adding to my problems.
Edits: 10/15/20
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: