mercury switched input selector

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

For ages I've been meaning to upgrade my passive pre set up with a better switching section. I'm currently using a rotary switch to select inputs- which crackles and cuts out on occasion. I was going to implement a relay system using the rotary switch to engage the relays, but then I got to thinking about mercury switches... I like the fact that they are free of oxidation and will pretty much do nice clean switching forever. On the other hand, I'm a bit worried about electrical resistance. Mercury is not the best conductor in the world. I'm sure my fears are unfounded, but wanted people's opinion on whether or not using a mercury switch would cause any kind of perceivable signal degradation (in the line level range).

Secondly, I'm not sure if using a single switch per channel would be good enough. I'm thinking I ought to use two switches in conjunction so that I could make a spdt to switch in order ground signal inputs not being used (I'm worried about possible bleed though, or noise from un-grounded in's). This would make for a total of four mercury switches per input rather than two. It would also make things a bit more complicated...

As for the actual switching, I'll have some kind of mechanical shaft to turn with flexible wires hooked up to the mercury switches so I can twist a knob to tun things on and off.

I have included a drawing containing both the double mercury switch, and single so this makes a bit more sense.
 

Attachments

  • mercury sw.jpg
    mercury sw.jpg
    253.8 KB · Views: 321
Are the current versions of small signal relays from NEC orPanasonic not good enough for switching line level audio? I have built an input selector with EB2 relays and i am really happy with it. I do not switch the grounds so no problems there. The only thing that you need to implement is a muting circuit after the relay selector if you do not want to hear any switching noise.
 
Gutten tag rrrremus,

Yes, relays would be just fine, I have no troubles with relays. I think it's the simplicity and longevity of mercury switches that interest me. Relays will eventually wear out, plus they need power. I'm also a fan of novel approaches.

I suppose the best way to go about this is to simply build it and see. I was just wondering if anyone knew about any potential problems in passing a signal through a blob of mercury. If that is something that will drop the signal level (due to it's relatively poor conductivity), or dirty things up, I didn't want to go through the trouble of ordering parts and doing a build if there's a fundamental problem with that idea.

Thanks for your response. If the mercury doesn't work out, I'll definitely take the relay route.
 
Resistance of mercury is not a problem. Tests on this forum showed that mud makes a good enough audio interconnect! The only possible snag would be if there is surface corrosion which introduces high impedance or nonlinearity, but as the mercury is sealed this is unlikely. You can do a calculation of a potential divider to convince yourself of this.

Shorting unused inputs is not usually necessary.
 
Are the current versions of small signal relays from NEC orPanasonic not good enough for switching line level audio? I have built an input selector with EB2 relays and i am really happy with it. I do not switch the grounds so no problems there. The only thing that you need to implement is a muting circuit after the relay selector if you do not want to hear any switching noise.

Hi rrrremus,

If you adequately decouple any capacitances BOTH sides of the switch(es) to ground through say 100-220 Kohm resistors you don't normally get any switching noise anyway?
I have done this for years in design with relays and when fixing vintage, with switchbanks and rotary selectors, including H J Leak 60s and 70s equipment, all switch without even the tiniest click in speakers!
 
Last edited:
I'm a bit worried about electrical resistance. Mercury is not the best conductor in the world
If that is something that will drop the signal level (due to it's relatively poor conductivity),
You need to better understand the Physical concept behind it :)

1) Metals do not have intrinsic resistance but resistivity.
https://en.wikipedia.org/wiki/Electrical_resistivity_and_conductivity

Do not get tangled with the Math, just get the basic concept, what it means, and then scroll down to the different metals resistivity table, to compare copper and mercury.

You'll see mercury has ~70% higher resistivity than copper but what should worry you is the actual contact resistance, that's what the current will "see".

2) actual resistance depends on resistivity and contact geometry: contact surface, conductor surface and length.

3) in a solid metal switch, you have a certain contact surface, and contact is not perfect:

a) surface is not perfectly flat, under a microscope you will see mountains and valleys and only mountain peaks on one side will touch mountain peaks on the other, it's very unlikely that just by sheer chance a mountain on one side actually matches a valley on the other improving contact.

This is a copper contact cleaned/polished into visible shiny bright with very fine 800 grit sanding paper:
800x200-s.jpg

and will make "very good" contact .

Flash gold on it, it will protect copper from air oxygen and humidity, but won't change geometry, contact will still be "peak to peak", point is you are not using 100% of the surface visible to the naked eye.

b) on a relay you have light pressure, the springs are light leafs or relay wouldn't switch on-off under the relatively light pull of an electromagnet.

Even on a finger toggled switch you have an internal spring applying contact pressure, which can't be that high or rocker/toggle would be hard to push or shaft hard to rotate.

Contact is much better on a screw held terminal strip , because you can tighten it a lot and even make relatively soft copper "flow" a little improving contact ... but that's intended as a semi permanent connection, not on-off

4) now consider the mercury switch: one side of the contact is solid metal but mercury is liquid, with no pressure but its own weight it will flow and make contact with all of the "hard" contact surface :)

This alone makes for a far better contact, far lower resistance than any comparable solid switch, specially including relays.

There must be a reason for mercury switches being used in all kinds of hairy industrial connections ;)
 
"There must be a reason for mercury switches being used in all kinds of hairy industrial connections "

Since the contacts and Hg are sealed in a vacuum (or inert gas) and there is always a film of mercury between the contacts (due to surface tension) there can be no contact deterioration, even under severe arcing conditions which would weld any dry contact.

The downsides are position sensitivity (not always) and having a poisonous substance in a glass capsule.

and to think or physics teacher used to pour mercury into a tray and we all had great fun poking it about with our fingers!
 
It might be worth reconsidering switching the ground as no switch is always going to be better than any switch.

I'm planning to experiment with mercury wetted relays for my LDR preamp to see if there's any audible difference.

I've got a couple of relays from Pickering mercury wetted, 2 form A, 5v
http://www.pickeringrelay.com/pdfs/98-sensitive-dil-reed-relays.pdf

I've got to build a second LDR preamp first and try it with and without relays. So may take some time.

PS definitely recommend the LDR preamp. It's nothing short of amazing. Thanks George and Rudi.
 
Last edited:
and to think or physics teacher used to pour mercury into a tray and we all had great fun poking it about with our fingers!
Not being water or grease soluble not a good path for contamination .... thanks God.

Now mercury salts are water soluble and do enter easily our bloodstream, also mercury vapour if breathed can get inside our lungs and stay there forever .... it will eventually be absorbed.

But there is a much worse situation, yet it does not seem to have hurt anybody: teeth cavities were filled with gold and mercury amalgam for the last 200 years at least (that's what "gold teeth" actually means) , and stayed inside our mouth, of all places.

Mercury was expected to evaporate or be swallowed ... probably a combination of both.
 
Tooth "amalgam' containing Mercury continues in widespread usage.
Amalgam has proven to be far stronger durable than the best of Dental epoxies.. is why.
And serious research suggests it's actually inert/genuinely unlikely to contaminate, despite being in on'e's mouth.
Ask your dentist ?
 
Last edited:
Wow! Thanks everyone for your replies! The mercury switches are ordered! I will post my results as soon as I have built it.

anjump123, I've ordered enough switches to do it both ways (with and without grounding inputs). Also, I agree that LDR attenuation is a good idea from what I've read about them here. What design is your favorite? I was looking into the Light Speed LDR. I have some of the parts, just got to get around to building it one of these days...

JMFahey, thank you for such a detailed explanation, it's always nice to learn stuff (I'm self taught and rather lazy). You've confirmed my assumption that the amount of surface area contact in a mercury switch makes it superior to other forms of switches. I initially did a quick search comparing the resistivity (thanks for correcting me) of copper and mercury and was concerned due to the drastic difference between them; mercury has 56 times the resistivity of copper, so it had me a bit worried at the time. Seems like a much better idea now.


Cheers to everyone else for your encouragement, tips and suggestions.
 
I'm using the Rudi Ratios latest version of the analog VPre and it sounds great and has the (significant for me) advantage of a remote control.

I'm planning to build a version more like George's original and experiment with stepped attenuator and mercury switches just to see if there's any differences.
 
I think the argument was between those who have tried it and those that just theorised about it.
Once you've heard it, it's a non-issue (provided you've got an amp and source with suitable impedances. The non-zero volume is not a real world issue and the channel balance is liveable with too, provided you've got matched pairs of LDR's but it's not any more difficult than matching transistors IMO.
The resolution can be revealing of poor sources and material, but that's the case with all good equipment.

I don't think I'll ever replace it, just play around with different wiring, connectors and relays. May have to upgrade the DAC, SACD player, Power Amp........
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.