diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Tubes / Valves (http://www.diyaudio.com/forums/tubes-valves/)
-   -   Strange Grounding Problem - My best guess: ghosts. (http://www.diyaudio.com/forums/tubes-valves/177944-strange-grounding-problem-my-best-guess-ghosts.html)

spice house 26th November 2010 05:49 PM

Strange Grounding Problem - My best guess: ghosts.
 
Longtime reader - I've looked around quite a bit but i cant seem to find this problem.

My grounded grid pre has a ghost infestation all of a sudden. It stopped working for no apparent reason (no part of the system was changed), and i discovered this: For some reason, if i ground the PCB, it stops passing audio.

I had a star grounding centered on the PCB so that the mains earth, - from each rca, and the PCB GND were terminated at one point and it worked great for a year+.

Now, if either the mains earth or the - from the RCAs touch any part of the GND on the pcb, the thing stops passing audio. It becomes silent, though the tubes are still lit.

Since there is no GND to the PCB there is a hum, though the circuit still passes audio. The hum starts when the unit is plugged in and gets worse when it is turned on.

The wierd bit is this is only sometimes! it mysteriously stops, then its ok, then it works for a week, then today, if it's grounded it won't pass signal. Normally if the ground becomes connected it will increase slightly in volume and the hum will disappear.

I have tried: rewiring the whole unit, changing the tubes, eliminating all extraneous components so there is only the PSU and in and out from the pcb connected directly to RCA, plus the grounding changes above.

Is there something I'm missing?

M Gregg 26th November 2010 06:11 PM

Spice,

Can you post a schematic?

Regards
M. Gregg

atmasphere 26th November 2010 06:55 PM

Quote:

Originally Posted by spice house (Post 2377852)
My grounded grid pre has a ghost infestation all of a sudden. It stopped working for no apparent reason (no part of the system was changed), and i discovered this: For some reason, if i ground the PCB, it stops passing audio.

I had a star grounding centered on the PCB so that the mains earth, - from each rca, and the PCB GND were terminated at one point and it worked great for a year+.

Now, if either the mains earth or the - from the RCAs touch any part of the GND on the pcb, the thing stops passing audio. It becomes silent, though the tubes are still lit.

Since there is no GND to the PCB there is a hum, though the circuit still passes audio. The hum starts when the unit is plugged in and gets worse when it is turned on.

The wierd bit is this is only sometimes! it mysteriously stops, then its ok, then it works for a week, then today, if it's grounded it won't pass signal. Normally if the ground becomes connected it will increase slightly in volume and the hum will disappear.

I have tried: rewiring the whole unit, changing the tubes, eliminating all extraneous components so there is only the PSU and in and out from the pcb connected directly to RCA, plus the grounding changes above.

Is there something I'm missing?

This sounds like an RF problem, and it can be related to the grounding scheme.

Here's how you set up the ground:

Set the chassis ground to the middle pin of the IEC power connector- if you have a 3-wire power cord, this is the green wire.

The circuit ground cannot be at chassis ground! Keep it separate and star it to the power supply, which is *not* chassis ground, BTW.

At the signal output, install a resistor about 10 - 100 ohms between the circuit ground and the chassis ground. You could also install this resistor between the power supply ground and the chassis ground.

This will take care of ground loop problems, getting shielding from your chassis and problems with RF (as long as you have your layout ducks in a row). Your house AC wiring, as long as it is to code, will take care of everything else just fine.

There are other ways to execute grounding. They all have their strengths and weaknesses. This is the one that I have found to work the best. You may find that the circuit tends to be a little quieter, blacker backgrounds, that sort of thing.

spice house 26th November 2010 07:48 PM

Quote:

Originally Posted by atmasphere (Post 2377911)
This sounds like an RF problem, and it can be related to the grounding scheme.

Here's how you set up the ground:

Set the chassis ground to the middle pin of the IEC power connector- if you have a 3-wire power cord, this is the green wire.

The circuit ground cannot be at chassis ground! Keep it separate and star it to the power supply, which is *not* chassis ground, BTW.

At the signal output, install a resistor about 10 - 100 ohms between the circuit ground and the chassis ground. You could also install this resistor between the power supply ground and the chassis ground.

This will take care of ground loop problems, getting shielding from your chassis and problems with RF (as long as you have your layout ducks in a row). Your house AC wiring, as long as it is to code, will take care of everything else just fine.

There are other ways to execute grounding. They all have their strengths and weaknesses. This is the one that I have found to work the best. You may find that the circuit tends to be a little quieter, blacker backgrounds, that sort of thing.


Huzzah! a 33 ohm resistor did it. Its dead silent now. My mistake, the actual mains earth is from the 3 prong to chassis, doh. This works because the resistor stops a ground loop... somewhere? oh well either that or magic. Thank you :)

spice house 27th November 2010 05:10 PM

Sadly, the same grounding scheme that worked yesterday does not work today.

Again, now it only works un-grounded, hum and all

http://1.bp.blogspot.com/_u_EoD7hGIP...+schematic.png

spice house 28th November 2010 04:29 PM

any suggestions to find/fix this strange problem? I want to move this unit into my studio, but i cannot until the ghostbusters make their move.

pedroskova 28th November 2010 05:51 PM

I take it that you've measured voltages throughout your circuit when it's working properly and when it is on the fritz...and they're the same?

If your psu is on the pcb and grounded to star ground, you might try grounding from the first psu caps to chassis and not from the star ground connection. Keep your safety (earth) ground as is.

Soonerorlater 28th November 2010 10:24 PM

Intermittent faults are very hard to solve. They are usually caused by changes in temperature. Something might be shorting, something might be going open. Changing grounding schemes might be solving the symptons not the problem. IMHO a thorough quality inspection is in order

spice house 28th November 2010 10:41 PM

Oh great! Thanks for all the comments, I'll probe around to the best of my abilities, hopefully I'll discover something that will translate into "oh replace input cap" or something of that nature

Also, I just noticed the amp is acting quite strange along with the pre... When the the preamp is properly grounded, the way it was originally, and 'not passing audio' like i said, i checked the other inputs of my amp and ALL of them, except where the actual RCAs its plugged in to, have very low music playing, slightly distorted, at about half volume. whaaaat? Lifting the ground on the preamp pcb, this phenomenon goes away.

I hope that made sense, but it must be passing audio if i can hear the music on other (wrong) inputs? very strange! i feel foolish i didn't notice this earlier. I didn't think to use a multimeter on the output of the preamp, doh!

Hopefully this rings a bell for someone, and thank you for your time :)

dmills 28th November 2010 11:25 PM

Methinks the thing is oscillating up in the RF somewhere.

Getting quiet distorted audio in strange places can be diagnostic of this as the following stage demodulates the RF (And grounded grid glass fets can make wonderful RF amplifiers).

One thought, that feedback taken from the output divider will be susceptible to phase shifts caused by cable capacitance and if the thing is only marginally stable that might be sufficient to case it to take off.

Try sticking a modest value resistor in series with the output connector to help isolate the feedback network from the cable, also possibly a few pf of cap placed carefully may help matters.

Regards, Dan.


All times are GMT. The time now is 03:09 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2