Help designing device to drain voltage?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Forgive me for not knowing if this is the right forum, or if this is already covered somewhere.

I have an apparently unique problem in my system, that has been robust to changes in everything from the main breaker panel through the installation of a dedicated line, to the use of a d-i-y common mode choke and dc-blocker, and through replacement of power cords, components, interconnects, speaker cables, and speakers. After four years of on-and-off diagnostics, the problem seems to be that something is causing a DC voltage to build up between the components, and can only be dissipated by severing the interconnects, after which everything works fine for a while until it builds up again.

A next-door neighbor is finishing a graduate degree in electrical engineering, and he proposes skipping the messy business of trying to explain how this could be possible (clearly it shouldn't be), and going straight to the fix, by which he proposes to tie the chassis of each component to a capacitor, and thence to earth, with gates installed on either side of the capacitor to prevent the chassis of any piece from ever being tied directly to earth: When the capacitor is charging, the switch between the capacitor and earth would be open, and when the capacitor was "full," that switch would close at the same moment that the one between it and the chassis would open. Then, when the capacitor was discharged, the process would reverse itself.


Has anything like this been tried in these fora before? I'm presuming that there would need to be resistors in the line somewhere to slow the dissipation of the capacitor, but beyond that I'm not sure how much my EE-friend would have to know (in particular, whether he'd need access to the schematics on all my components), before we could proceed safely.

Thanks in advance for your help. Sorry if this is a newbie question.

Dave.
 
I believe that the inter-web routed your question to the wrong planet, because electricity just doesn't work like that on our planet Earth!

For example, on our planet a small battery will have a DC voltage across it's terminals created by chemical means. To discharge this small battery you could use a jumper wire to short circuit the terminals. Placing a capacitor across the battery terminals will not discharge the battery.
 
The problem is in my communication of the proposal, not the planet-of-origin for the electrical properties in question.

Specifically, the capacitor wouldn't be inserted between the components -- which is where it would have to be for the battery analogy to be germane. Instead it would be inserted between a component (?each component?) and earth, and its only purpose would be to store whatever pernicious energy is being shorted exactly as you describe, since it can't be shorted directly to earth without tying chassis directly to earth. Likewise, the energy building up between the components can't be shorted by a drain wire directly between the two chassis, a la the battery analogy, because that would create a ground loop.

The EE guy across the street proposed to get around this problem by creating a setup that works like a lock in a canal, wherein the chassis would never be tied directly to earth because one or the other of the two doors at either end of the lock would always be closed. In this analogy, the capacitor is only acting as the "tub" of the lock, filling with charge from the upstream side while the downstream switch is closed, and then discharging the rest of the way to earth while the upstream switch is closed.
 
A next-door neighbor is finishing a graduate degree in electrical engineering, and he proposes skipping the messy business of trying to explain how this could be possible (clearly it shouldn't be), and going straight to the fix, by which he proposes to tie the chassis of each component to a capacitor, and thence to earth, with gates installed on either side of the capacitor to prevent the chassis of any piece from ever being tied directly to earth: When the capacitor is charging, the switch between the capacitor and earth would be open, and when the capacitor was "full," that switch would close at the same moment that the one between it and the chassis would open. Then, when the capacitor was discharged, the process would reverse itself.
I'm sorry, i just don't see how this would work. I'm pretty sure i understand what you are trying to get accross, but a capacitor won't charge unless both ends are tied accross something that is developing a voltage (IE the equipment & earth). With one leg tied to the equipment, even through a switch, the other end of the cap must be connected to something. If it isn't part of a circuit it'll do nothing at all :D

E2A:- if i was you i'd try attaching some high value resistors from the equipment chassis in question directly to earth. It appears whatever the problem is takes days to manifest itself, so try something like 10 megohm or greater resistors - i'd suggest 100 megohm & see what happens
 
Last edited:
Okay, now we're getting somewhere. Perhaps what the EE-kid across the street is proposing is to complete the circuit back to (the other?) chassis while the cap is charging, and then the switch mechanism would interrupt both of those connections, and at the same time close a path to ground?
 
Administrator
Joined 2007
Paid Member
What stops working ? How long does it take to stop working ? What are the symptoms as you listen to it ?
All sounds very bizarre to me. The interconnects will tie all the equipment together anyway...
From what you describe it could be anything from aliens to dry joint/cracked print that disconnecting the interconnects "fixes"
 
a) Most soon to be EE's have access to multi-meters and should know how to use them. Maybe he could measure these mysterious voltages.
b) Many audio equipment chassis's are already connected to ground. Only units with rated insulation are not.
c) What happens to the sound when this voltage builds up?
d) Are you sure that it's an EE he's studying for?
 
Perhaps you could tell us what the functional symptom are as opposed to relaying a diagnosis and solution.

Forget the voltage. What actually happens in practise? You turn on the system, play something, then some kind of fault occurs which is alleviated by disconnecting the elements of the system. What happens?

w
 
Thanks for asking, Speedskater & Wakibaki -- I definitely need to start at the beginning but I'm not very articulate in these fora and I didn't want to inundate you with needless details.

What happens is that over hours or, in some cases, days, there accumulates a dry, reedy overmodulation -- not unlike the high-midrange and low-tweeter distortion that one would associate with damaged crossovers -- that gets progressively worse in spite of powering everything down, and can be "reset" to perfect sound by disconnecting the rear aprons.

The problem is robust to a one-item-at-a-time change of everything in the entire system, and has changed in its sonic characteristics so little that, after four years of fighting this, I can barely tell the difference between amps -- and even speakers! -- compared with the difference between listening with this problem and without it.

I'm convinced it's not listener fatigue (or tinitus, or something else personal) for two reasons: First, because non-audiophile friends have commented unbidden on the difference, with the problem vs. without, and second because multiple performance problems only tangentially related to sound quality have failed to reproduce themselves on anyone's test bench.

Here are but four examples:

1) A Parasound A23/P3 combo that would develop a staticky whooshing sound in one channel and eventually drop that channel to zero -- never happened for the local tech, or at Parasound HQ, either one.

2) A Naim Nait 5i whose input selector would periodically freeze and not accept input either through the front panel or the remote until the AC power was disconnected and reestablished. Sent to NAIM USA, they had it for two weeks -- including taking it home to play with, at one point, and never could make it happen.

3) An Arcam FMJ CD-23 that, when playing with IC's connected, gets so hot to the touch, both at the top chassis cover just above the transformer *and* at the interconnect sockets, that you jerk-back your hand. Doesn't happen when powered-up but not connected, and doesn't happen nearly as bad when powered-up and connected and not sending a signal down the IC's. Sent to two out-of-town service specialists, neither of whom could ever make it happen.

4) A series of three separate Sony DVD-players, all of which would periodically issue a loud, scratchy "bang" down the signal path, and none of which have ever done so in any of the houses of the three friends to which I gave them.

...The reason for the OP's emphasis on dcv between components is that this is the only measurable aspect of this thing I've ever been able to replicate with any predictability: When I was using conventional interconnects (coaxial design where the signal return was also the shield), I could set my multimeter to dcv and hold one lead to the shell of the rca connectors at the amp end, and the other lead to the amp chassis, and I would get a voltage that would get progressively bigger as the problem got audibly worse, and would not reset to zero as a result of powering down.

Whether this is the cause or a symptom is not clear to me. Neither is the science behind how this could ever happen, since the rca interconnects themselves are connected directly to chassis at the point of contact, are they not?

Anyway, that's where I'm at. Sorry this one got long.
 
Administrator
Joined 2007
Paid Member
Certainly weird...
I'm no expert on the mains distribution system in other parts of the world, so perhaps you could clarify some points.
The mains supply is the only common link in all this... all the symptoms, many pieces of kit etc.
I take it you are on 120 volts or so ?
Is all the equipment connected to the same socket in the house ?
What I am getting at is, is it possible there is an external distribution issue.

Is your 120vac mains from a "centre tapped" transformer (that gives the option of 240vac for large items of powerconsumption somewhere in the street ?)
Could there be an issue there.
A floating/poor/damaged "ground" or center tap.

What your friend is suggesting seems more like trying to "fiddle" the issue rather than find the real cause.
 
My understanding is that all US power supplies are from 220V centre tapped supplies, effectively supplying 2 phases 180 degrees from each other. My knowledge of this arrangement is rather limited, though, because nobody else is crazy enough to do it that way, so EE students in Australia don't have to worry about it.

If you're finding that significant DC voltages are developing between grounded chassis points, you have a serious (and possibly dangerous) electrical wiring fault.

I would suggest turning off your power at the distribution board, and then testing the resistance between the chassis' of some nearby equipment. It should be very low, say < 10 ohms.

As an alternative diagnosis, try a test with all connected equipment (source, preamp, and amp) connected to a powerboard, so everything comes from just one wall outlet. If the problem persists, things start to get very puzzling indeed.
 
Mooly & TheSeekerr,

Yes the US uses a 240V center tapped transformer to provide dual (split-phase) 120V (& 240V) to most homes. A very few homes receive 1 phase of 3 phase industrial power. I was surprised to read, how often the neutral/ground leg goes bad causing the two 120V legs to divide the 240V input unequally. Like maybe 60V on one leg and 180V on the other.

But if this is really a DC problem, maybe the units are all DC coupled and a servo circuit has gone defective.
 
Mooly:

Yes, 120 volts, though when I test it with my multimeter it wobbles a little each time, and around a different centerline each time. One day it will wobble +/- 0.4v around 121.5, and the next day it will wobble +/- 0.6v around 119.7, etc. Some people have suspected that this could be the problem, but others have said that the transformers inside modern audio equipment is designed to deal with far bigger fluctuations than this.

I don't know what "center tapped" means, but my dryer is 220 and runs off a circuit in the same breaker-box as the rest of the house, so presumably this is what you mean and the answer should therefore be "yes, my home's wiring is 'center tapped'."

The Seekerr:

All equipment is plugged into the same outlet on a dedicated AC line. A single cable runs from that outlet to a set of common mode chokes wired in parallel and each terminated in its own outlet, to which each piece of equipment is thence connected. For a while I was also using an isolation transformer between the common mode chokes and the outlet, but my local tech thinks that the isolation transformer isn't earthed (though I'm not convinced that just because he doesn't get the yellow light where he's supposed to on his tester, that this is automatically proof that an isolation transformer isn't earthed, since it is, after all, supposed to be isolate things).

Last night I tried removing all the power filtration from the setup and connected all four pieces of equipment directly to the four outlets on the dedicated ac-line, and got no improvement (I also didn't seem to get any degradation, either, though this finding was somewhat less conclusive).

Speedskater:

As my OP clearly demonstrated, I am at -- or perhaps beyond -- the ragged limit of my involuntary knowledge of any of this and, accordingly, I don't know what test(s) would distinguish a problem of "dc-coupling" from an ac problem. If you hum a few bars, I can fake it with my multimeter and my thick rubber gloves. Just tell me what to test, and I'll test it.

If it's interesting: I've had similar results when trying the system with all other breakers turned off, and I've had similar results regardless of whether the TV is on or plugged in or neither.

It is *possible* that some of this is down to cable dressing, too, since there seems to be some "noise" in repeated trials of my experiment about disconnecting and reconnecting everything -- specifically, the last two times I've tried it, it hasn't worked.

At this point I'm willing to consider *all* possible explanations, but with the caveat that no replacement of electronics, speakers, or any of the three types of cabling have thusfar made the tiniest difference.

I have not tested resistance between the gear yet -- great idea. I'm assuming I do this with everything powered down but fully connected, yes?
 
Administrator
Joined 2007
Paid Member
In post number 9 you have a whole list of equipment that does strange things. To me that says there is an external issue.
Mains fluctuations are normal and nothing to be concerned about normally. What your meter isn't telling you is whether there are large spikes in those fluctuations (a faulty center tap etc) as it just gives an average value.
Transformers in equipment just give an output based the turns ratio, and pass on the input (primary) to the output (secondary)... they don't "smooth out or deal with fluctuations"... they just pass them on. However, any equipment such as CD player or DVD players etc will have fully stabilised supplies making them imune to normal fluctuations. Power amps and so on don't normally have that.
It's anything with "logic" circuits etc that needs a stable supply that has regulators.

Can you rig an extension lead up from another circuit (socket) in the house and try it all runnibg on that ?
 
Yes, I can do that -- but with one big caveat: The dedicated line into which the entire kit is currently connected is the only one in my house that is earthed with the modern, third-pin system. All the other branches in the house are of the old-fashioned, positive-and-neutral-only design.
 
earthing and fault bleeds

Uh I have recent experience with strange return design causing incorrect behaivoir. I built an audio oscillator, the two transistor design from the GE transistor manual seventh edition, to exercise the inputs of my preamps with a stable waveform so I could track where the inbalance in channels was coming from on my several preamp devices. The circuit was driven from a 9v power supply (6dc transformer winding) isolated from mains earth. The circuit was producing a 4.5V PP waveform about a 6vDC mean. I ran the waveform through 180K resistors to two pots to ground, a 25K pot for 1v nominal waveform for CD-Radio type inputs, and a 500 ohm pot for the magnetic phono input. I used this with a 2-1 phono cable to drive the two inputs of my 1960 PAS2 preamp, and ordered some new caps to replace the paper one coupling caps. No problem. The bottom of the potentiometers was tied to -6vDC, thence to the ring of the phono plugs. I then tried this device to exercise the inputs of my 1992 "disco mixer" which has the rings of the phono inputs floating about 9v on an 18v DC power supply, and the tips and rings of the phono cables connected through resistors and filtering capacitors (not in series with the inputs) to the plus and minus inputs of op amps. The original 4558 op amps produced a distorted waveform at the right frequency with way too much time spent at the positive rail. The 33078 op amps I replaced them with produced DC.??? Everything got better when I put .047 uf capacitors in series with the tips on the oscillator. The op amps couldn't handle the DC offset between the tip and ring of the input jacks. The tube amp handled it fine.
I would suggest you put 1 or 5 mohm resistors between both tips and rings of your input and output connectors (or 2 and 3 of xlr connectors) to the minus supply. Then I would put another 1 to 5 mohm resistor from minus supply to safety earth of the power supply on one component, either the preamp or the power amp, but not more than one (to prevent ground loops) . I would also put a Metal oxide spark arrestor between PS minus and earth on each component. These are usually blue, have a UL, VDE, CSA or other safety logo on them, plus usually a voltage including the mains they are designed for. I salvage my spark arrestors from dead PC power supplies, these seem to blow up about yearly. Do what you want, but in my experience static builds up between components due to imbalences in power supply and signal circuits. Pro audio equipment connect the chassis together with the pin ones of the XLR connectors, and if hum results, they buy a "DI unit" or a "transformer coupler" to break the ground loop caused by each piece of equipment having a safety ground to the main and a signal ground going to the next piece of equipment.
 
Here are but four examples:

1) A Parasound A23/P3 combo that would develop a staticky whooshing sound in one channel and eventually drop that channel to zero -- never happened for the local tech, or at Parasound HQ, either one.

2) A Naim Nait 5i whose input selector would periodically freeze and not accept input either through the front panel or the remote until the AC power was disconnected and reestablished. Sent to NAIM USA, they had it for two weeks -- including taking it home to play with, at one point, and never could make it happen.

3) An Arcam FMJ CD-23 that, when playing with IC's connected, gets so hot to the touch, both at the top chassis cover just above the transformer *and* at the interconnect sockets, that you jerk-back your hand. Doesn't happen when powered-up but not connected, and doesn't happen nearly as bad when powered-up and connected and not sending a signal down the IC's. Sent to two out-of-town service specialists, neither of whom could ever make it happen.

4) A series of three separate Sony DVD-players, all of which would periodically issue a loud, scratchy "bang" down the signal path, and none of which have ever done so in any of the houses of the three friends to which I gave them.

OK. These diverse phenomena are not attributable to a single cause within the wiring of your house.

Have you considered exorcism?

Sell up quietly and move elsewhere. Never mention the subject again.

w
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.