I started this by trying to test lead free solders to see if there was any difference. Playing led me to interconnects, then to contacts.
If you have any contrary testing you have done at the levels discussed then we might talk. But you and I sometimes read the same papers, look at the same results and draw different conclusions. When I try things out I often do not get results that I was expecting.
You don't think my AP was imagining the extra power line noise? Reasonable and rational examination allowed a possible conclusion of a dirty AC contact, the the other possible explanation is an outside source of interference that stopped at the time I first changed the power cord and it stayed off. So the answer is not absolute. But the knowledge is that when I see the higher level of noise, something is wrong.
BTY the noisy 797 that I just lifted the output from the rest of the buss, now no longer has any output, noisy or not! (After the thunderstorms I went through and rechecked all of the signal levels.) Another bad habit I have is when resuming a play session I repeat the last experiment run to be sure the results are the same. Tells me if I screwed something up or if the equipment has changed. I got that habit after a wasted week during my early college days.
Holm's mentions vibration in a closed contact can let in contamination and form an oxide barrier. I think most folks who do service have seen dirty contacts cause problems.
I did the electrolytic capacitor tests at 0 VDC for a reason, when there is a DC bias the problems don't show up. That supports the anecdotal evidence that capacitor problems showed up worse with bipolar power supplies.
The noise in switches shows up with a very low level signal, but the signal level to create the noise can increase as frequency decreases. There is nothing surprising there to practicing engineers. What this means for audio designers is that they may want to parallel switches to reduce some problems. Of course it is clear that a thick gold plated contact is best.
The literature mentions the problems with low level switching and contact type also thermo electric effects at low levels.
So one of the interesting questions is do crimped end resistors sound different that ones that use a different connection method?
I and others have presented several examples of measurements an order of magnitude lower. I can't follow your interpretation of the pico-voltmeter, the connection of the input wire to the lug is outside the working of the instrument. You can take the input smoothly through zero (+ to -) at the pico-volt level i. e. there is no barrier to overcome on a metal to metal mechanical connection.
I admire your energy for this, but if the listening "verification" remains anecdotal we just go round and round.
The connection of the input lead to the voltmeter is done after cleaning the terminal typically with an alcohol wipe. It doesn't just sit on the shelf, get tarnished and then work properly. The choice of copper is to match the lead wire. That is why it is not gold plated. The over sized connections allow greater clamping force to keep the connection clean. Copper also is quite malleable, allowing the deformation required for a good connection. A disadvantage in a switch as it shortens the switch life. That is why switches use alloys.
(You might try cleaning all of your switches and contacts including power cords and see if you can hear a difference, then talk about anecdotal issues.)
As has been mentioned before in cable comparisons often just removing and reinserting the connectors makes a difference.
The instrument is not a general purpose DVM. It is used by skilled folks.
If you are good at your software you can even design the instrument to measure the impedance curve of the voltage source and pick out high impedance voltage sources in series with lower impedance ones!
Calling out a trick pony and extending it to the general case is weak at best.
The issue remains, do contacts exposed to the air form non-conductive or high resistance films that affect low level signals? I think so, you don't.
Does a high non-ohmic resistance, non-conductive/abrupt conductive or even inherent diode formation in a connection increase distortion?
The nano switch paper even listed a few more issues that we are not even looking at.
Cables and Connectors
Absolutely true in my experience as well. Even in mated pairs where the interfacing metals have been matched and cleaned, over time the connection often will become problematic. Anyone who has had to maintain an audio plant with patch panels has owned the special brushes, cleaning fluid and contact preservative, and has spent WAYYYY too much time learning about the sound problems attributable to old connector pairs. It is why I learned early on to use equipment which used terminal strips instead of XLRs or (yuck) TRS plugs. I have yet to had a problem with audio passing through correctly terminated terminal strips, which can establish a gas-tight, high-pressure termination. Even old 66-blocks, the newer 110-blocks or the superior replacement Krone blocks are much better than connector pairs (when matched to the correct wire type).
Connectors are evil, man; make them stop!!!
Howie
Howard Hoyt
CE - WXYC-FM 89.3
UNC Chapel Hill, NC
www.wxyc.org
1st on the internet
Absolutely true in my experience as well. Even in mated pairs where the interfacing metals have been matched and cleaned, over time the connection often will become problematic. Anyone who has had to maintain an audio plant with patch panels has owned the special brushes, cleaning fluid and contact preservative, and has spent WAYYYY too much time learning about the sound problems attributable to old connector pairs. It is why I learned early on to use equipment which used terminal strips instead of XLRs or (yuck) TRS plugs. I have yet to had a problem with audio passing through correctly terminated terminal strips, which can establish a gas-tight, high-pressure termination. Even old 66-blocks, the newer 110-blocks or the superior replacement Krone blocks are much better than connector pairs (when matched to the correct wire type).
Connectors are evil, man; make them stop!!!
Howie
Howard Hoyt
CE - WXYC-FM 89.3
UNC Chapel Hill, NC
www.wxyc.org
1st on the internet
Hi,
Well, I don't think much, but I can share two more anecdotes...
First, practically all the US/West European gear I encountered in industrial use in east germany in the mid 80's (putting the time of their design usually into the 70's) used mercury wetted reed switches, not dry switch signal relays which even then where common. I presume they had reasons to spend money so freely, so it must have mattered.
Second, when, while blacklisted from real jobs for political reasons, I repaired radios and TV's to make a living we found that taking apart the switches in these devices, which would often cause intermittent and/or distorted audio very quickly (sometimes in stuff that was only weeks of the assembly line), and following a clean with "blue contact oil" (yes, there was such a thing) with covering the switch surfaces and nearly filling the switch with vaseline would dramatically cut recurrence, while laboriously fitting new switches would fix the problem at that instant only for the same unit to come back within a few weeks to month with the same problem.
The same treatment on potentiometers (really old carbon type) was routinely able to not only resurrect pot's with major dead zones but to produce more reliable parts and we used this in preference to fitting new pots, as the new pots would go dead again in no time, while the cleaned and vaseline'd pots staid very reliable. I later also used this on the slider pots in mixing desks, rebuilding them before even fitting...
Of course, this is no evidence for the formation of non-conducting films etc, but clearly something caused those problems... ;-)
Ciao T
Well, I don't think much, but I can share two more anecdotes...
First, practically all the US/West European gear I encountered in industrial use in east germany in the mid 80's (putting the time of their design usually into the 70's) used mercury wetted reed switches, not dry switch signal relays which even then where common. I presume they had reasons to spend money so freely, so it must have mattered.
Second, when, while blacklisted from real jobs for political reasons, I repaired radios and TV's to make a living we found that taking apart the switches in these devices, which would often cause intermittent and/or distorted audio very quickly (sometimes in stuff that was only weeks of the assembly line), and following a clean with "blue contact oil" (yes, there was such a thing) with covering the switch surfaces and nearly filling the switch with vaseline would dramatically cut recurrence, while laboriously fitting new switches would fix the problem at that instant only for the same unit to come back within a few weeks to month with the same problem.
The same treatment on potentiometers (really old carbon type) was routinely able to not only resurrect pot's with major dead zones but to produce more reliable parts and we used this in preference to fitting new pots, as the new pots would go dead again in no time, while the cleaned and vaseline'd pots staid very reliable. I later also used this on the slider pots in mixing desks, rebuilding them before even fitting...
Of course, this is no evidence for the formation of non-conducting films etc, but clearly something caused those problems... ;-)
Ciao T
I have no problem with that, if the mechanical force is insufficient to bring any metal to metal contact there is an open circuit. The asperity of the metal surface, contact force, and the mechanical properties all determine how much metal to metal contact is made. Then there is ballistic or diffusive conduction of electrons where metal meets metal, neither theory has any potential barrier.
A barely made dirty contact that changes resistance with vibration is not the same thing.
Must stop now and save some energy for the feedback article. 😀
Then the issue seems to be one of underlying assumptions. I presume when you set your preamp to CD you leave it there for hours or days or weeks, maybe even longer.
That is just for a single contact how many are in series with the signal?
Do these become contaminated from exposure to air even when closed?
Does that require vibration or just contamination creep?
Does this happen to all mechanical contacts?
I was testing a fairly standard miniature toggle switch (C&K 7201 an unused spare part from a piece of audio equipment) and a carbon volume control.
The issue is not "Does electricity flow through conductive metals", but "Is there a minimum voltage required to start the flow in connectors or other mechanical contacts?"
So the difference is now, "Is contamination always present?"
Is contamination always present in any practical audio interconnects, switches, or other contacts?
Hi Ed, Scott,
Perhaps a simple view on contacts, but since the oxides of silver are conductive, shouldn't that create a low problem contact pair for audio signals? One confusing factor in all this is other forms of deposits. Things like cigarette smoke and cooking oils really cause trouble. Can you tell this is a technician's viewpoint?
Hi Thorsten,
I've also been forced to rebuild various switches and controls used for signal manipulation. These parts were used in both instrumentation and audio equipment. Disassembling, cleaning and rebuilding always restored very good performance as long as you didn't break anything in the process. I used a silicon paste as an oxygen barrier, and oil for any bearings for controls or switches. Like yourself, our observations are that rebuilt controls and switches were more reliable than new parts in some cases. One critical point that many service people may not understand is that for continuously variable controls, the part that becomes noisy are the slip ring (or linear slide) metallic components, and not the resistance element! Flushing these controls full of "cleaner" only serves to wash out the protective lubricants. This leads to permanent damage as the element and shaft bushing both wear out.
One problem that some may not be aware of is the length of time that new parts may spend sitting in storage, and the environmental conditions in the storage area. That means that a "new" part may actually be a year or so old to the factory for manufacturing. The parts (spares) department may receive a quantity of "new" parts for replacement use, then allow those to sit for years on end until ordered by the plucky service technician. Poor devil. The initial or later storage conditions may add several years worth of aging to these parts on top of all that. Some parts need to be rebuilt before they are used for the very first time!
-Chris
Perhaps a simple view on contacts, but since the oxides of silver are conductive, shouldn't that create a low problem contact pair for audio signals? One confusing factor in all this is other forms of deposits. Things like cigarette smoke and cooking oils really cause trouble. Can you tell this is a technician's viewpoint?
Hi Thorsten,
I've also been forced to rebuild various switches and controls used for signal manipulation. These parts were used in both instrumentation and audio equipment. Disassembling, cleaning and rebuilding always restored very good performance as long as you didn't break anything in the process. I used a silicon paste as an oxygen barrier, and oil for any bearings for controls or switches. Like yourself, our observations are that rebuilt controls and switches were more reliable than new parts in some cases. One critical point that many service people may not understand is that for continuously variable controls, the part that becomes noisy are the slip ring (or linear slide) metallic components, and not the resistance element! Flushing these controls full of "cleaner" only serves to wash out the protective lubricants. This leads to permanent damage as the element and shaft bushing both wear out.
One problem that some may not be aware of is the length of time that new parts may spend sitting in storage, and the environmental conditions in the storage area. That means that a "new" part may actually be a year or so old to the factory for manufacturing. The parts (spares) department may receive a quantity of "new" parts for replacement use, then allow those to sit for years on end until ordered by the plucky service technician. Poor devil. The initial or later storage conditions may add several years worth of aging to these parts on top of all that. Some parts need to be rebuilt before they are used for the very first time!
-Chris
Scott,
I don't understand how the plot you post shows no relationship. This is a plot of electrodes placed distance d apart with an air dialectric? Looks like the Ni makes a difference in the breakdown voltage.
I'm sure I'm missing something here and haven't followed this switch contact discussion that closely, but maybe you could fill me in?
I don't understand how the plot you post shows no relationship. This is a plot of electrodes placed distance d apart with an air dialectric? Looks like the Ni makes a difference in the breakdown voltage.
I'm sure I'm missing something here and haven't followed this switch contact discussion that closely, but maybe you could fill me in?
At 1000nm the breakdown is still >200V so I don't see any argument for non-conductive contamination separating a contact so it has a < 1mV barrier. Metal touching metal just conducts. I know the technical references are dense but the jist of the MEMS contact resistance calculation was to figure out the density and size of the actual metal touching metal regions and using standard electron conduction theory to compute it. The gaps and insulating contamination just don't conduct at all. My take on the absense of "micro-diodes" is that if copper oxide is the contaminant the surrounding conducting regions still have virtually no drop so there is no voltage to support rectification.
Hey Thorsten even German women can play golf!
Last edited:
Hi DF96,
The normal application of a thermalcouple heats the wires along equal lengths from the junction. Many lower accuracy K-type probes are intended to be used with the junction in contact with the surface to be measured, just to ensure that both metals are sitting at a similar temperature. Of course, they may also be trying to work with the bad habits generated from thermistor or semiconductor junction use (as a sensing element). Proper industrial probes are designed to bring both wires and junctions to the same temperature along the entire length of the sensor and are calibrated as such.
So, while the junction isn't where the actual voltages are generated, it is acceptable to allow the junction to transfer the measured temperature to both wires equally and along a similar length. And yes, the voltage vs. temperature response is anything but linear.
Well J&J (John & Joshua),
Go. Be happy. Just try not to confuse people who want to go about learning about this stuff for real.
Hi John,
For measuring wire characteristics, you'll probably require larger currents to make any effects easier to measure. Personally, I feel that is a waste of time once you determined that the effects were below your measurement thresholds. Now, measuring circuitry is much easier. This shows that the effects created from circuits and components are magnitudes higher than those created in wire. Since this is the case, wouldn't your time be better spent refining your art as a circuit designer? Otherwise, what you are doing is much like trying to quantify differences in bearing noise and rolling resistance when the tires on the car are flat. Look at the issues having the most effect before worrying about things that don't even show up.
-Chris 🙂
In theory that is true. In practice there may be a thermal gradient across the connection. The act of soldering introduces other variables as these metals alloy (or simply diffuse) into each base wire to varying degrees. Since there is no longer any guarantees about the characteristics of that junction, the practice of soldering or welding using any other material, metal, is not acceptable in the metrology field. Since I was an instrumentation calibration technician in an ISO registered firm that specialized in calibration, I (we - all of us) were routinely tested and observed working.
The normal application of a thermalcouple heats the wires along equal lengths from the junction. Many lower accuracy K-type probes are intended to be used with the junction in contact with the surface to be measured, just to ensure that both metals are sitting at a similar temperature. Of course, they may also be trying to work with the bad habits generated from thermistor or semiconductor junction use (as a sensing element). Proper industrial probes are designed to bring both wires and junctions to the same temperature along the entire length of the sensor and are calibrated as such.
So, while the junction isn't where the actual voltages are generated, it is acceptable to allow the junction to transfer the measured temperature to both wires equally and along a similar length. And yes, the voltage vs. temperature response is anything but linear.
Well J&J (John & Joshua),
Go. Be happy. Just try not to confuse people who want to go about learning about this stuff for real.
Hi John,
For measuring wire characteristics, you'll probably require larger currents to make any effects easier to measure. Personally, I feel that is a waste of time once you determined that the effects were below your measurement thresholds. Now, measuring circuitry is much easier. This shows that the effects created from circuits and components are magnitudes higher than those created in wire. Since this is the case, wouldn't your time be better spent refining your art as a circuit designer? Otherwise, what you are doing is much like trying to quantify differences in bearing noise and rolling resistance when the tires on the car are flat. Look at the issues having the most effect before worrying about things that don't even show up.
-Chris 🙂
Scott, if I understand what you're saying, the issue is that the surface contamination is not homogeneous. There is a roughness on a microscopic scale as well. Where metal pokes past or is not covered by contamination and can touch the other contact, there is no real resistance. Where the contamination insulates, there's far less (or far far less) conductivity. The ratio of these two areas determines how much the contact fouls the sound (in the form of excess noise, sometimes correlated with the music), modulated by vibration, creep, and other mechanical forces.
There is no metal-thing-metal heterojunction with a 1nV (or whatever) deadband, especially in common metals used for electrical contacts.
Is this a correct summation?
If so, the answer is still the same- clean and tighten contacts periodically and try to minimize them.
There is no metal-thing-metal heterojunction with a 1nV (or whatever) deadband, especially in common metals used for electrical contacts.
Is this a correct summation?
If so, the answer is still the same- clean and tighten contacts periodically and try to minimize them.
Hi SY,
No one wants to maintain their equipment though.
-Chris
That seems to do the trick in practice.
No one wants to maintain their equipment though.
-Chris
I accept the need as a given unless I'm willing to give up switching sources, raising and lowering the volume, and hardwire my phono, preamps, crossovers, power amps, and speakers together. With decent quality connectors and wire, it's not too often.
Hi SY,
No argument from me! I was only making an observation from many years of performing said maintenance. 🙂
-Chris
No argument from me! I was only making an observation from many years of performing said maintenance. 🙂
-Chris
scott wurcer said:
Thank you for the explanation on the graph Scott.
In theory metal touching metal conducts, but in practice it doesn't conduct very well. I had this experience with some new 3 inch lithium coin cells I was using as a voltage reference. I stacked them to get the correct voltage, but no dice. Even when loaded with just a cap, the voltage was 3 volts off, without significant pressure applied. Used some silver contact enhancer between them to get the correct voltage.
I don't remember if I tried just cleaning the contact area with Deoxit first, but I may have.
Just other day my dad was complaining that the system was sounding bright after I had switched and then replaced the components back. I told him he was crazy (not exactly those words), because nothing had changed in the system. A week later after lots of complaining and arguing, I noticed that when I'd listen to just the left channel the sound was bright. I investigated and found the woofers weren't working because of a loose connection at the woofer binding posts.
I think metal slammed together in a vice conducts well.
Last edited:
Well the world won't end we still disagree. I don't see the point you are trying to make with that graph. If you go by that then my dirty contact shouldn't conduct until I reach a few hundred volts.
We could discuss two dimensional vs three dimensional pre-arc conditions but that isn't really what I think the issue is.
Last edited:
- Status
- Not open for further replies.