Did you sit on something sharp, or what?
The topic is switching low level signals in the absence of any standing currents. i.e signal switching. NOT telephone like level switching at the end of 1Km of lightning conductor.
Lightning will certainly zap a small reed relay, but if you are designing your MC front end to resist a direct lightning hit .... good luck with that!
No the issue was the article quoted talked about unsealed relays and the films that form preventing them from working. The article did cover some of the methods of preventing those problems.
Now if you want to know why mercury whetted relays are not used for audio, try measuring the dynamic contact distortion.
Also it takes a lot less than a lightning strike to ruin most of them.
One problem with reed relays switching audio is that the relay is encased in a coil. The issue is that any variation in the coil current will induce a voltage in the reed controlled circuit. Not a problem in ATE where you can ignore it, as you only take measurements after a settling time has elapsed. So when you switch it off, before the whetted reed can actually open there will be a voltage surge. Not much but at preamp levels...
There really are relays designed for very low level signals. These typically have quad bifurcated gold plated cross bar contacts. Older types used long levers to activate the contacts to keep the low level signals away from the coil, as not only is the magnetic field an issue, so is the distortion induced by iron near a signal conductor. That of course varies with the circuit impedance.
No it was an actual test of a circuit shown to "Improve" interconnect performance and my observations of what actually happens. You then brought up wire wrap techniques which use a silver plated wire and a gas tight connection. Not quite the same thing. My question is do you read the cites or just comment based on the last 90 words or less?
I have not heard the new benchmark. But want to order one soon.
Then I can compare with the first one I have. Just pointing out the varying thoughts on the subject by other designers. Who to trust? Myself?
This past year I have been on a quest to measure very low distortions, accurately. Now that I can, I wonder what the Alps pot will tell me. I swear by the Cosmic Great One that the sound was degraded when, last year, I put it in series with a <-100dB THD+N line stage. What's up with that?! I suspect oxidation and contacts. Have to take the pot apart after I measure it.
As for activated coils around the reed/contact etal..... good thought.... might be the reason JC/others didn't like what the 'reed relay' did to the sound?; One could design the relay operation such that it had power on it all the time and that would put the reed/contact into an open/unused position (dpdt). Then, only a deactivated coil would let the signal thru its contact.
Thx-RNMarsh
Last edited:
Hi Mike,
Have you been listening to the analogue VC's of the PGA23xx and CS3318 family?
jan
In the mercury relays it is usually "wetting" with a small pool of mercury. They were invented for telephone switching because they offer several orders of magnitude better reliability (or so the hype went).
How do I measure the dynamic contact distortion?
I have a prototype of an IM test set I'm fiddling with that was constructed almost entirely of wirewrap connections. Untouched after many years it seems to be capable of -140 dB IM floor. No signs of any issues with wire wrap connections.
What citations? When I post conflicting well known and experimentally verified facts they are ignored, i.e. Pashen's Law - there are no situations where current can be made to jump a gap at 100mV potentials. If this is observed, I would say first the experiment is flawed in some unknown way.
BTW have you seen Jim Williams article on mercury wetted relay switched potentials to measure settling time of 18bit DAC's? Tek sold these in the 60's as pulse generators, 1ppm in 265nS. AN120-1
I am well familiar with the resistance during release due to the meniscus of the mercury but in use relays are either off or on.
Last edited:
Are you referring to the TEK 109????
They were sooooo neat.. 250 picosecond 100 volt transitions into 50 ohms
jn
Similar deal for mercury whetted choppers (e.g. Potter and Bromfield(sp ?).
The bridge time of 300uS when both NO and NC contacts are shorted became the standard for power transistor power pulse measurements. Run it off the 50/60 Hertz mains and you get the convenient ~~1% duty cycle.
I might continue as to why we think MC cartridge loading to be both useful and desirable.
Exactly how the loading changes the sound at audio frequencies is still a mystery, so far as I know. However, when I started making the Vendetta Research phono stage approximately 30 years ago, I designed in a very high quality 1 turn pot (tantalum film) of 200 ohms in series with a 10 ohm resistor. This way, the cartridge could be loaded with 10-210 ohms, even while the preamp was reproducing music. We found that a 'sweet spot' could be found between 50 and 150 ohms usually, with a few exceptions for a specific type of MC cartridge. Besides my finding, both Dave Wilson and Brian Cheney could find the 'sweet spot' and narrowed it down to 1 ohm resolution or so, so one cartridge might have a 'ss' of 90 ohms, and another might have a 'ss' at 153 ohms, etc.
I also added a summing input setting to bypass the pot entirely and give lower than 10 ohm loading, initially. Back 30 years ago, we did find that some cartridges DID like very low impedance loading, but as the years passed, we found that less loading was optimum, and that even 47K ohm loading was useful for some cartridges.
This is why I want a range from 10 ohms to several thousand ohms (if practical), to be continuously variable with a precision of 1 ohm or so, to allow the customer to find the 'ss' while playing their favorite records and sitting properly in front of their loudspeakers.
It does look to be difficult to do properly, but doable.
Exactly how the loading changes the sound at audio frequencies is still a mystery, so far as I know. However, when I started making the Vendetta Research phono stage approximately 30 years ago, I designed in a very high quality 1 turn pot (tantalum film) of 200 ohms in series with a 10 ohm resistor. This way, the cartridge could be loaded with 10-210 ohms, even while the preamp was reproducing music. We found that a 'sweet spot' could be found between 50 and 150 ohms usually, with a few exceptions for a specific type of MC cartridge. Besides my finding, both Dave Wilson and Brian Cheney could find the 'sweet spot' and narrowed it down to 1 ohm resolution or so, so one cartridge might have a 'ss' of 90 ohms, and another might have a 'ss' at 153 ohms, etc.
I also added a summing input setting to bypass the pot entirely and give lower than 10 ohm loading, initially. Back 30 years ago, we did find that some cartridges DID like very low impedance loading, but as the years passed, we found that less loading was optimum, and that even 47K ohm loading was useful for some cartridges.
This is why I want a range from 10 ohms to several thousand ohms (if practical), to be continuously variable with a precision of 1 ohm or so, to allow the customer to find the 'ss' while playing their favorite records and sitting properly in front of their loudspeakers.
It does look to be difficult to do properly, but doable.
Last edited:
Some Contact history -
Yes, I remember them, too. Mercury contacts prevent arcing. That gave them a long and stable contact life. They broke/disconnected from current cleanly. That lead to the idea that they could be used to make a nice clean edged/corner for square wave or pulse generator with them. Thus, the TEK 109. Noted too that they would only operate it in the upright position or the mercury would move away from the contact. Thus, shipping them caused issues if not shipped "this end Up".
Thx-RNMarsh
Yes, I remember them, too. Mercury contacts prevent arcing. That gave them a long and stable contact life. They broke/disconnected from current cleanly. That lead to the idea that they could be used to make a nice clean edged/corner for square wave or pulse generator with them. Thus, the TEK 109. Noted too that they would only operate it in the upright position or the mercury would move away from the contact. Thus, shipping them caused issues if not shipped "this end Up".
Thx-RNMarsh
I don't recall the manual mentioning shipping orientation nor operational orientation. Sounds reasonable though.
The biggest problem I had was the scopes of the day. It was necessary to use a very dark room and allow time for the eyes to adjust to the dark. Scopes of that time also had triggering issues, what a PITA.
jn
You measure dynamic distortion by superimposing a DC current onto a test signal and comparing that to the plain AC signal.
Did we fail to reach an understanding? Wire wrap works because of the silver plated wire that even oxidized is a good conductor and the square edges bite into the wire to form a gas tight seal. That is different than a simple RCA connection that with scratches from use is not even gas tight for a small contact area.
The cites more than 90 words ago about the Wiki article.
Now do you know the difference between repeatable switching times in a specialty switch and the actual switching time?
Ignoring the sarcasm, I was working towards finding a volume control, many years ago, which didn't have audible artifacts, along the same lines as mikelm. At one stage I looked at the possibility of relay control, but didn't like what I saw, even from a known behaviour POV. Spec sheets, as usual, were completely useless in telling one what was actually important, what the real low level behaviours were - the only solution would have been a suck it and see approach, this is where real world experience is everything.
Was interested in finding a faster solution, to bypass the irritating, audible behaviour of pots - so tried a PGA23xx unit, and was happy with the consistency of behaviour of that - the worst thing is hearing the behaviour change with time, and the history of the unit - disconcerting changes are worse than consistent possibly slightly poorer overall performance ...
If I can find the mercury relays in my stash I'll try that test. How much current?
RE Wirewrap: I must have missed part of the discussion. RCA connectors can be quite marginal. I don't see how the latest fashion with a single point of contact improves the situation, especially wrt: shielding.
Richard
Audio reproduction's "raison d'etre" is to provide a pleasurable experience for the listener - no one else can tell you if it is succeeding or not so I guess you have to trust yourself.
I avoid unsoldered connections in my system like the plague. The line level leads only have RCA plugs on one end, speaker leads similarly so. ( There is a trimmer to zero the servo in the amps right now but the o/p is very heavily filtered. )
I think a good test for the Alps pot would be to insert it in series with the signal as a simple resistor ( not using the wiper ) and compare this with a your favorite brand of resistor, then repeat the test only now passing the signal through the wiper as well as the track. This way you can evaluate the effects of the track and wiper separately both subjectively & objectively.
Thanks for your input
mike
Last edited:
I bought a set of those Eichmann Banana plugs and a set of their Bullet RCA plugs some years ago to test the theory. Waste of money IMO
I would suggest not only avoid, but eliminate. This is a bit like a leaking swimming pool problem - if you still have one leak left, you always lose water. IME there is a significant quantum (sorry!!!) jump in behaviour, quality when the last hole is plugged ...
- Status
- Not open for further replies.