The other missing part is specifying the regime under which this is applicable, but it does make a nice marketing word soup for snake oil merchants, since the dermatologist or fund manager who is lapping this up will not understand what is meant by "disordered systems," "tunnel junctions," or "Fermi energy." Context is vital in physics and to be avoided at all costs in high end audio marketing.
I think you've done a fine job of fertilizing the weeds.
The cited text is at the page 30, second half of the page.
The limitations of low signal propagation through mechanical contact is nothing new, every manufacturer of rotary switches or relays specify the low limit of operable currents. Silver-silver contacts usually insure 10...100uA currents. For 1uA one needs rhodium, etc. This does not mean, that lower currents will not pass through, but they will flow in a sparkling manner. I'll try to support these values by references. 1mV signal through 10kOhms input impedance produces 0,1uA, though.
"You elitist!
I don’t need a PhD in physics for to understand what a disordered audio system is and what it needs to be purchased and installed in order to turn it back into an ordered audio system." (said the fraud manager)
I have faith in the laws of natural balance.
There will always be people who are willing to energize the reasoning faculty of their brains.
Thank you Vladimir
George
I have checked out all datasheets available to me (relays and switches), and there is no better contact in the world than 0,1mV-1uA. This level is achieved in rotary switches with contact rings made of bulk noble metals (no copper with deposited layers).
That means, that if one ever would attempt to send a control signal with resulting current less than 1uA, there is big probability that the intended action will not be effected.
In musical signals, constituents of below 1uA, will be averaged and strongly distorted.
That means, that if one ever would attempt to send a control signal with resulting current less than 1uA, there is big probability that the intended action will not be effected.
In musical signals, constituents of below 1uA, will be averaged and strongly distorted.
Last edited:
It's pretty easy to check for yourself, assuming you have some basic test gear (signal source, spectrum analyzer).
Those quoted values are for currents that will help wipe off oxides at make and break, it will only slowly cause contact resistance to rise if used regularly below those levels, a one time use above those levels cleans the contact again, that effect regularly bit people who put electronic cd systems onto old fashioned contact breakers as the current into the main transistor driving the ignition coil was always too low to cut through the oil film that slowly built up.
I was put off using relays, because the manufacturers weren't game to say that they would operate reliably below certain signal levels - this smacks of sticking one's finger in the wind engineering ... unless lots of handwaving makes all the difference - some of the time ...
I've found relevant paper, in russian, published in
L.Safonov, A.Safonov "Technologies in Electronic Industry", No2, 2009
Ïðÿìîóãîëüíûå ýëåêòðè÷åñêèå ñîåäèíèòåëè. Ðàáîòà ýëåêòðè÷åñêèõ ñîåäèíèòåëåé â öåïÿõ ñ ìèêðîòîêàìè è ìèêðîíàïðÿæåíèÿìè
What is clear from the paper, that many contact issues appear in uA and sub-mV range, non-conductive films of 10-100 angstroms grow and are destroyed by currents. For signals above 1V we usually do not have problems becuase of fritting, growing dielectric films are destroyed by discharges. However, below 50mV fritting becomes impossible and contact becomes unstable.
There is also frequency effect (picture eclosed), for 10uV AC signals of various frquencies contact resistance varies with frequency, rising sharply at 20kHz. This is because current density increases in near-surface non-metallic (quasi-conductive) region.
L.Safonov, A.Safonov "Technologies in Electronic Industry", No2, 2009
Ïðÿìîóãîëüíûå ýëåêòðè÷åñêèå ñîåäèíèòåëè. Ðàáîòà ýëåêòðè÷åñêèõ ñîåäèíèòåëåé â öåïÿõ ñ ìèêðîòîêàìè è ìèêðîíàïðÿæåíèÿìè
What is clear from the paper, that many contact issues appear in uA and sub-mV range, non-conductive films of 10-100 angstroms grow and are destroyed by currents. For signals above 1V we usually do not have problems becuase of fritting, growing dielectric films are destroyed by discharges. However, below 50mV fritting becomes impossible and contact becomes unstable.
There is also frequency effect (picture eclosed), for 10uV AC signals of various frquencies contact resistance varies with frequency, rising sharply at 20kHz. This is because current density increases in near-surface non-metallic (quasi-conductive) region.
Attachments
So my low output moving coil cartridge, generating microvolt signals which pass through cartridge pin connectors and at least 2 RCA connectors, is doomed to failure.
When I think about a typical audio system I am wondering where all these switches and contact are that we are talking about that are running at such low current or voltage that we are worrying about? Perhaps a switch in an independent DAC or perhaps a low level switch selecting difference sources but how many places are these switches? Why not just turn up the unit to a high output temporarily to clean these contact points if that is the issue? Isn't the average output of a cdp these days in the 2v range at the output connector and between a preamp and amp wouldn't that be similar? If you are one who no longer listens to a phono stage why even include a preamp in the chain, just go direct from cdp to amp removing all those possible switches and contact points, of course you would have no eq function but so many seem to hate that idea here anyway. What is the point of talking about 1ua of current unless we are only talking about very low level signal in a dynamically wide band music source at a very quit passage and then your average power would be higher than this? Is this purely a theoretical conversation?
ps. I wasn't thinking of the low output of a mm or mc phono stage, then I can see the low currents but these would normally be a few connectors and not necessarily a switch position, How do you deal with that on a normal basis besides disconnecting and reconnecting the contact points once in awhile and using a contact cleaner?
ps. I wasn't thinking of the low output of a mm or mc phono stage, then I can see the low currents but these would normally be a few connectors and not necessarily a switch position, How do you deal with that on a normal basis besides disconnecting and reconnecting the contact points once in awhile and using a contact cleaner?
Last edited:
IMO, all conductivity theories are far from being able to descrive sound effects, but we could take their conclusions as a kind of warning. I remember very well, when I excluded relay-based input selector from signal path in SE 300B tube amp, sound definitely got better.
I think, based on these numbers an electrometer could never work. Neither a picoammeter. Both exist in high accuracy versions using connectors and switches on the inputs. This reads to 10 fA: Model 6485 5-1/2 digit Picoammeter with 10fA Resolution and this reads to nanovolts: Model 2182A Nanovoltmeter Both have detachable input cables. And this has a 200 TOhm input impedance Model 6514 Programmable Electrometer .
This is not to say that special techniques shouldn't be used at these levels. Keithley has a very good book on making these difficult measurements accurately. Welcome to Keithley?s Knowledge Center - Keithley Instruments Inc. They have books on both switching and low level measurements. Free downloads.
This is not to say that special techniques shouldn't be used at these levels. Keithley has a very good book on making these difficult measurements accurately. Welcome to Keithley?s Knowledge Center - Keithley Instruments Inc. They have books on both switching and low level measurements. Free downloads.
Thanks for the references Demian. It just seems that practical applications would just cause so many problems that I have a hard time with all this about sonic problems due to switches and connections. We have been listening to audio systems for 50 years with so many switches and interconnections and I just have a hard time with associating all these problems to those switches. Perhaps I just haven't been shown the difference if the switches had been removed from the chain, but I just think about all the switches, sliders and interconnects in record production and PA applications and it would seem that this would really amplify the problem by a factor due to the number of these problems.
The definite book:
Electric Contacts
Theory and Application
By Ragnar Holm 4th edition (1967)
http://www.amazon.com/Electric-Contacts-Application-Ragnar-Holm/dp/3540038752
George
Electric Contacts
Theory and Application
By Ragnar Holm 4th edition (1967)
http://www.amazon.com/Electric-Contacts-Application-Ragnar-Holm/dp/3540038752
George
Demian you beat me to it, I might add this pico-voltmeter which my customers verify in performance.
EM D.C. PICOVOLTMETER MODEL P12 Specification.
I might add this relay stuff based on contacts exposed to ambient goes out the window with vacuum or noble gas sealed relays. 50nv resolution at picoamp currents is an everyday thing. We test chopper amps for micro-volt offsets and femto-amp input currents with several cheap COTO relays in series with each input, and that's after 10e7 operations.
Try this, http://www.cotorelay.com/datasheets/Coto Technology 3500 Reed Relay.pdf
They work fine at 50nV and sub-picoamp currents by experience at the 10e7 uses level.
Last edited:
- Status
- Not open for further replies.