Why do dirty switch contacts cause harmonic distortion?

[mkiijam]

Title says it all ( I think)

I have heard and read that the culprit is Silver Sulfide, not Silver Oxide. But that still raises the question (and perhaps begs the question, too) for me of what causes the harmonic content when the audio goes through this compound / impurity?

 

[Philbo King]

It depends on the specific type of dirt and corrosion present. Both can act as intermittent semiconductors.

 

[k brown]

I have a gut hunch that the distortion caused by poor contacts would not be harmonic, but some other kind.

Thoughts?

 

[mkiijam]

Well, I am watching my distortion analyzer and my spectrum analyzer and the fundamental drops as is expected, but even / odd spikes creep up. i.e. I can clearly see harmonic content.

 

[OneRoomStudio]

From OSHA:

“Generally, the first step in semiconductor device fabrication involves the oxidation of the wafer surface in order to grow a thin layer of silicon dioxide (SiO2).”

I think @Philbo King is right. The oxidation/dirt results in a crude semiconducution. Depending on where in the circuit the switch is (i.e., shunting signal to ground), it can act similarly to a diode and clip the signal (resulting in harmonic distortion).

 

[mkiijam]

Well Silver Sulfide is a known semiconductor for sure. Decreases resistance with heat. I've heard that's the real issue with bad switches contacts?

 

[Brian Roth]

After decades or working with switches and relays in desks and tape machines, there is sort of a "spectrum" of switching contacts, as well as connectors.

1. Wonderful, fresh gold plated contacts.
3. No audio passes at all.

I skipped #2....that is everything in between. Sometimes, the failing contacts cause very subtle audible problems. At worst, the audio literally "cuts in and out".

My 2 cents. Use switches and relays with gold contacts for audio. If possible, enclose the contacts in sealed packages.

Bri

 

[Brian Roth]

One further thought for folks dealing with decades-old equipment. I clearly recall a discussion (more than a few years) with a "main" design engineer from the Ampex audio recorder division. Ampex recorders were a premium brand...like Studer, etc.

Larry told me Ampex designed their products for reliable life for ten years. Reason? Customers buying their products had a 10-year amortization tax plan.

Amazing that so much of the equipment from all those companies in that era still operate decades past "the expiration date on the milk bottle".

When Men Were Men....lol!

Bri

 

[Mr. Moscode]

After decades or working with switches and relays in desks and tape machines, there is sort of a "spectrum" of switching contacts, as well as connectors.

1. Wonderful, fresh gold plated contacts.
3. No audio passes at all.

I skipped #2....that is everything in between. Sometimes, the failing contacts cause very subtle audible problems. At worst, the audio literally "cuts in and out".

My 2 cents. Use switches and relays with gold contacts for audio. If possible, enclose the contacts in sealed packages.

Bri

Yup.

The oxidation prevents low level information from getting thru until it gets so bad then it becomes intermittent.

If I'm not replacing the switch I clean sometimes with 2000 grit emery or al oxide paper, rinse with some CRC and follow with DeOxit. Make sure the CRC cleaner is safe for plastics.

I have also replaced pushbutton switches and RCA jacks with gold plated. No more problems.

Good luck.

 

[sodderboy]

Over the decades, there is all sorts of chemistry happening with the original lube, heat, internal off-gassing, external smoke, dust, skin, etc. Add to that the cleaning products used over the same decades. The issue with spray cleaning them is that the lube removed has to be replaced in some way, both for the actuator and the contacts.
I have been maintaining a few Dualitys and a 9K for almost two decades and have found the Caig ProGold to be the best cleaner/conditioner in one over time. I was just mapping the 9K yesterday, and many of the switches cleaned in the past stay clean. Especially the ones that get used regularly.

Biggest best thing for switches is to be used regularly.

Others are a few pushes away from usable. When I started cleaning switches I would have to clean 4-7 switches per module, but yesterday's mapping was 1-2. And hardly any harmonics component on the bad switches which was out of the ordinary. Just fuzz and crunch. Some just need another Pro injection.
I was never a fan of ultrasonic dipping, but it certainly worked well with SSL 4K's. It was that once you stripped all the factory lube out, you had to keep with a regular dipping/lubing schedule or it would get noisy real quick. I think it is a disaster on 80 series Neves.
I would never use emory cloth on an audio switch contact, and I can't get to the contacts in a 6PDT e-switch, ITT Schadow, or Neve rotary anyway.
Mike

 

[Mr. Moscode]

"SODDERBOY: I would never use emory cloth on an audio switch contact, and I can't get to the contacts in a 6PDT e-switch, ITT Schadow, or Neve rotary anyway."
YMMV

2000 or 3000 grit is more of a polish and removes silver tarnish without chemicals and I don't over do it. I've cleaned hundreds of ITT Schadows over the years, no callbacks. Another mild abrasive is a pencil eraser which I've used on the older larger rotary switches in receivers and test equipment on contacts I could get to. It buys a few years.

I also cleaned many TV set turret tuners with a pencil eraser in the olden golden days. No callbacks.

 

[Philbo King]

Yeah, pencil erasers work. Just remember to use a bit of deoxit or similar to flush away the rubber residue.

 

[RCAguy]

If memory serves, oxidized switch or connector contacts act like imperfect diodes, partially squaring half the audio waveform, thus adding odd-order harmonic distortion.

 

[sodderboy]

I'm down with emory from silver jewelry and metalwork.
If I have a B67, how am I getting into those dirty winkies to burnish the pins with emory? Or a MCI with them soldered to a PCB? It is a mechanically sealed switch. And the wipers are ridged, the "self-cleaning" ha-ha. Do they get burnished as well? Interest piqued!
Mike

 

[sodderboy]

I've cleaned hundreds of ITT Schadows over the years, no callbacks.

Please describe how. I've dealt with ITT in MCI and Studer mostly. You disembowel from the front? How?
Mike

 

[Disco Volante]

I've cleaned a good many of those old push-button switches in (mainly) hifi-gear with a soak in mild citric acid, then flush with water, ipa and blow out with clean dry air. Citric acid only removes the black or green oxides, does not attack the metal. Not a single callback. This also works if you can't dismantle the switches, but of course you have to be thorough, and excercise those switches while they're soaking.
We use a similar method for cleaning up bronze-castings in my day-job..
Just got a ultrasound machine which will hopefully make a breeze of all that cleaning;-)

Cheers, V

 

[Philbo King]

So IPA is prefered over lager & stout?
Just kidding, I know it's isopropyl...

 

[Cqwet Dbdfte]

The first semiconductor device was a copper-copper oxide rectifier.
Silver oxide is not as bad but can be made a P-type semiconductor. Silver sulphide is worse in this respect.
If you can avoid mechanical contact in favor of silicon switches (with known parameters) you may have an advantage. If not (hermetically sealed) reed relays with Rhodium plated contacts would be better. Mercury wetted parts are getting harder to find but may be good.
Gold is soft and usually very thin, wears off easily exposing brass whatever under.
Open to air and dust contacts for critical low level signalling is asking for trouble.

 

[living sounds]

If not (hermetically sealed) reed relays with Rhodium plated contacts would be better.

Reed relays are what is used in my console mostly. Lot's of them per channel. After more than 30 years, many started to either cause distortion or fail completely. I finally replaced every single one of them. Not cheap, but thankfully they were all socketed.

 

[Cqwet Dbdfte]

Coils or contact surfaces failed? Sockets can get wonky too and need replacing.
Latching relays won't need much coil power