Help diagnosing noise

[dsavitsk]

I have a headphone amp with a single 5842 per side that is transformer coupled to some headphones. I am getting a noise on one side that sounds like a static sound -- like a connection not being made. It sometimes turns into a squeel for a bit, and will then go away. Sometimes it goes away for hours, sometimes only for a few seconds. Also, the sound is clearly mostly on one side, but sometimes it seems to move to being on both sides, which says ground issue to me. It is the sort of thing I would typically associate with a poorly connected jack, but that doesn't seem to tbe the case.

I tried swapping the tubes and the noise stayes on the same side. I also resoldered all of the joints. I figured it was a dirty socket, but wiggling the tube doesn't seem to do anything.

The tube socket is in a pcb and has 4.7K smd grid stoppers on all 4 of the grid pins. B+ is about 160V and the tube is biased with a single red (2V) LED.

Any suggestion of what to look for -- unfortuatley I don't have a scope, but this is coming through OPTs and is clearly audible, so it is not ultrasonic.

Any suggestions where to look?

[SY]

Get a scope- it sounds like you have an oscillation. Tracing that down without the right tools will be an exercise in frustration.

[Giaime]

It happened to me, the same description of the problem, when I didn't care about heater - cathode insulation voltage. I don't think it's your case, but since you didn't post a schematic it could be

[kevinkr]

Try swapping the 5842 from one side to the other and see if the noise follows the tube.

I have run into similar problems with other tubes as well as the 5842 and in most cases and for reasons I don't purport to understand elevating the filaments above the cathode voltage eliminated the problem. In this application 10 - 15V above the cathodes ought to be enough. The filament circuit needs a good ac ground as well. (read as shunt cap to your star ground)

Another thing that may help is a very small resistor, say 10 ohms in series with the plate and right at the socket.

Also I never bother with all of the other grid connections for audio use, one is sufficient. In VHF applications using all of the grid connections is important due to internal lead inductances that become relevant at those frequencies - for audio applications this is just ungainly..
Make sure those resistors are none inductive!

[dsavitsk]

I fiddled with the headphone jack a bit last night, and the noise went away. So, thinking that it was just a bad jack, I replaced it and of course the noise is still there. But, after a bit of warming up, it is so intermittant so as to be non-intrusive enough, that I keep using these tubes. It is a very nice sounding tube.

Originally posted by SY
Get a scope- it sounds like you have an oscillation. Tracing that down without the right tools will be an exercise in frustration.

I don't think I'd even know what to do with a scope if I had one. Any suggestions on how to learn to use one to track down problems like this? I have used one before -- years ago in a modern physics lab, and I can get access to one, but I'd like to go into the process at least having a clue what I am doing.


[QUOTE]Originally posted by kevinkr
Try swapping the 5842 from one side to the other and see if the noise follows the tube.[quote]

Still there

I have run into similar problems with other tubes as well as the 5842 and in most cases and for reasons I don't purport to understand elevating the filaments above the cathode voltage eliminated the problem. In this application 10 - 15V above the cathodes ought to be enough.

Presumably I'll just use a pair of resistors to form a voltage divider -- about what size should they be? That is, how much current should be flowing?

The filament circuit needs a good ac ground as well. (read as shunt cap to your star ground

Does this mean a cap from the 15V point to ground? Any particular size cap?

Make sure those resistors are none inductive!

They are 1206 size metal film -- I think Morgan Jones suggested using these.

[SY]

If you've got a copy of "Building Valve Amplifiers," there's some very useful material on troubleshooting with a scope. If you don't have that book, it's one you ought to. Somewhere on the Tektronix site, there's a very nice pdf manual ("XYZs of Oscilloscopes") on oscilloscope basics. Search around for it.

That said, the heater idea sounds very reasonable. Yes, run a voltage divider to float the heaters about 10-30V above the cathode and bypassed to ground. Again, "Building Valve Amplifiers" details this.

[dsavitsk]

Originally posted by SY
If you've got a copy of "Building Valve Amplifiers," there's some very useful material on troubleshooting with a scope. If you don't have that book, it's one you ought to. Somewhere on the Tektronix site, there's a very nice pdf manual ("XYZs of Oscilloscopes") on oscilloscope basics. Search around for it.

That said, the heater idea sounds very reasonable. Yes, run a voltage divider to float the heaters about 10-30V above the cathode and bypassed to ground. Again, "Building Valve Amplifiers" details this.

Hey, it worked. I used a 220K and a 47K resistor and a 100uF/50V cap to do the bypass. B+ is about 165V or so.

I'll still pick up a copy of the other Morgan Jones book, though.

Thanks for all of the suggestions.

[kevinkr]

Hi Doug,
Sorry I did not get back to you last night, but I did note that SY gave you the additional information you needed to get the job done.

The values you chose are just fine. Glad to hear it worked, thought it would..