Gary Kemp
My Masterpiece has for some months now, usually right after being switched on, made a loud buzz, that until just recently would slowly fade and disappear after ten seconds or so.  But lately it does not disappear, so I have to act. The sound: 

1. Is not volume dependent.
2. Is equal on both sides. 

I conclude that the problem is with the preamp's power supply.  Does anyone know what I should look for? For those of you knew the man (recently passed away), it was actually assembled by Kelvin (who was an expert) so I'm disinclined to suspect a bad solder joint etc., but of course even the best of us do that sort of thing.  
Quote 0 0
Deleting Attachment...
Gary Kemp
btw of the similar issues discussed on the board, the closest seems to be Andrew T's problem, which turned out to be the delay/relay. 
Quote 0 0
Hi Garry:

You can just pull the relay out, it is a shortening device. If gone you found it. If this was the issue and you use a solid state power amp, first switch on the master piece wait 30 seconds and then switch on your SS power amp, If you have a tube power amp do not worry you can switch them on the same time.

Can you measure voltages? If so please measure, with some measurements we can perhaps pin point better.

Is the buzzing sound low in tone like 60 Hz?




Quote 0 0
Gary Kemp
Thanks very much.  I've disconnected the relay/timing unit, and unfortunately it still does it.  When I turn it on, the loud buzz goes on for about 30 seconds, then it functions normally.  So something's funny with the ps, I assume.  Yes I have a meter; what should I test? I suppose the fault/change must show up in a change in the power supply voltages. 
Quote 0 0
Gary Kemp
I see on the schematic that the ps has an output of 150v. 
Quote 0 0
Gary Kemp
OH god. The problem is not with the ps, but with one side or the other of the preamp circuit,  at least I assume. No such problem occurs when I plug in the GG for the MP (which re-acquaints me with the GG sound; not as warm or dynamic, but so balanced, detailed and solid).  The problem is that one side is silent, after  the hum dies down.  My excuse is that I didn't listen carefully, and didn't twig this because I have a subwoofer which picks up the signal from both sides, so the stereo effect is not as pronounced.  So: any obvious places to look for problems?  
Quote 0 0
First Put volume control to zero, any difference?

then

First Disconnect the subwoofer, it could be possible the cables function as antenna and create buzz or oscillation.

If Buzz persists easy swap the tubes, first the 300B's no change then the driver tube. No change now we need to start measuring.

measure the fillament voltage on the 12au7 it should be 6 V
on the 300 B it should be 5 volt
And yes main DC voltage should be 150 Volt this one can be higher depending your supply voltage.

What is the voltage over the LM317? You could put one probe on the + side of the output capacitor and the ground, should be around 9 to 10 if I recall correctly. I am currently not using this pre amp using another home brew. (perhaps some one can confirm?)

Measure voltage over the low voltage 6800 uF. Should be same between 8 and 9 volt. If not the same let us know. (could be rectifier is not working well or one of the 6800uF caps has a bad solder connection but the voltage regulators still make the voltage on your meter correct (5 or 6 volt) but with quite some interference.)

Lets stop here and see what you get.

O yes a subwoofer does not need to be connected to both channels one is perfect.
Can you measure the resistance between both signal inputs on your cable to your subwoofer. So not between signal and ground but between the signal wire which was connected to the left and right of your pre amp. This could be an important help in finding the issue.
Quote 0 0
Gary Kemp
Again, thanks enormously ..

A bit more evidence has emerged.  One 300B is a kaput; it lights up, but no sound is produced from that side.  But when the tubes are swapped, the unit produces NO sound. So either the tube failed, causing further failure on the side of the circuit, or the other way round. 

I checked the driver tubes; both are fine.   

So there is some fault on one side of the circuit.  I'm just presenting this latest bit of evidence now, in case anyone knows that this is a common fault with a ready remedy, before looking with my meter. 
Quote 0 0
Bruce Rozenblit
It is possible to have a bad socket.  It is also possible that some other failure (lack of negative bias or problem with the LM317) caused the tube to fail early.  Or, you just have a bad tube.
Quote 0 0
Gary Kemp
Thanks Bruce .. just to clarify: a tube has failed, and there is a problem on one side of the circuit.  I'll look under the hood and report back ...
Quote 0 0
Thanks Bruce.

Gary you remember my comment in my last email?

O yes a subwoofer does not need to be connected to both channels one is perfect.
Can you measure the resistance between both signal inputs on your cable to your subwoofer. So not between signal and ground but between the signal wire which was connected to the left and right of your pre amp. This could be an important help in finding the issue. (I should have said here perhaps the reason of the issue)


Well let me explain here a experience I had with using cable making a mono signal from a stereo signal.

I listen in Mono so I connected the out puts of my grounded grid together to go to my mono power amp. (similar to yours how you connect the subwoofer)

When I had this configuration the left channel tries to drive the right channel and vice versa I learned, the lower the output impedance is the worse it gets since the current will be higher. It damaged for example my output tubes in the Grounded Grid over time. ( I noticed music was missing and one channel died)
I was overloadding my output tubes.

Now fast forward.

When I put a resistor from 1K5 in the signal path of both channels before they were joint to one output, all problems were solved. Now the resistance between the left and right  is 3000 Ohms. And that is way higher then the output resistance from the output tube.

So again check the resistance between your two signal wires from your subwoofer. It should not be zero. If so make the modification, it could be that this had effect on your malfunction.

You can check if the LM317 is still correct by inserting the tube you know works. You can measure between the + side on  the output capacitors and the ground, should be around 9 to 10 Volt.

Good luck


 
Quote 0 0
Just received a private message asking why tubes in parallel in the OTL design do not have problems, but when you combine left and right yes you have a problem?

Well;

In the OTL all tubes are driven the same source signal.  In stereo in extreme no signal in one channel and the other channel full signal, now the impedance of the channel without signal is overloading the tube in the channel with a signal.
Quote 0 0
Gary Kemp
Thanks for this. I should have said that in my setup, the MP is connected to the Sub Buddy, which in turn is connected to the Subwoofer (to it, but not from it), and of course the power amps. 

But I went ahead and measured, the (difference between) signal input on one channel to the signal input of the other (on the subwoofer).  It was 0.  Is that significant?  The Sub Buddy has no issues (so far!).  


Robbert wrote:
Thanks Bruce.

Gary you remember my comment in my last email?

O yes a subwoofer does not need to be connected to both channels one is perfect.
Can you measure the resistance between both signal inputs on your cable to your subwoofer. So not between signal and ground but between the signal wire which was connected to the left and right of your pre amp. This could be an important help in finding the issue. (I should have said here perhaps the reason of the issue)


Well let me explain here a experience I had with using cable making a mono signal from a stereo signal.

I listen in Mono so I connected the out puts of my grounded grid together to go to my mono power amp. (similar to yours how you connect the subwoofer)

When I had this configuration the left channel tries to drive the right channel and vice versa I learned, the lower the output impedance is the worse it gets since the current will be higher. It damaged for example my output tubes in the Grounded Grid over time. ( I noticed music was missing and one channel died)
I was overloadding my output tubes.

Now fast forward.

When I put a resistor from 1K5 in the signal path of both channels before they were joint to one output, all problems were solved. Now the resistance between the left and right  is 3000 Ohms. And that is way higher then the output resistance from the output tube.

So again check the resistance between your two signal wires from your subwoofer. It should not be zero. If so make the modification, it could be that this had effect on your malfunction.

You can check if the LM317 is still correct by inserting the tube you know works. You can measure between the + side on  the output capacitors and the ground, should be around 9 to 10 Volt.

Good luck


 
Quote 0 0
Hi Gary:

I am not familiar with the sub buddy.

Having seen and use(d) multiple Bruce designs, I know he designs failure out, fail safe and the design is always rock solid when used as intended.

Hopefully another sub buddy user can reply here.

Quote 0 0
Gary Kemp
So:

The DC voltage to the big tubes 189v (both the same). 
The filament voltage on the tubes is 5v and 6v.  I trust that the difference is not significant. 
The 6800 caps: On the good channel, 9.4v. On the bad channel, 8.0v. Is that significant? 

But I can't get a reading on the LM317. You say: "You could put one probe on the + side of the output capacitor and the ground"; I'm not quite sure what you mean, but the most obvious capacitor is the .01 connected to LM317; no reading there on either side (+ or -), or on the opposite (good) channel. 

Or does this have to be done with the tubes in? 

Thanks as ever ...




Robbert wrote:
Hi Gary:

I am not familiar with the sub buddy.

Having seen and use(d) multiple Bruce designs, I know he designs failure out, fail safe and the design is always rock solid when used as intended.

Hopefully another sub buddy user can reply here.

Quote 0 0
Deleting Attachment...