I would like to get any feedback or experience info on whether Deoxit
would help sufficiently to improve mechanical switch contact performance
used in very low (nV) level circuits. I posted this question in the
tekscopes group a while back, but with no response. I was surprised that
there was no info from such a wide-ranging group, but maybe here in
volt-nuts, with narrow focus, it will stand a better chance. Here is
approximately what I wrote, slightly edited, for consideration.

Then:

"One of my projects involves boxing up some special transformers and a
switching system, for isolating and dividing down AC signals by large
factors (E-3 to E-6), down as far as the 1 nV RMS region. A fair number
of switching elements are needed to route and select various transformer
output taps to send to the single output connection. There are two
transformers to cover two frequency bands 10 Hz-1 kHz, and 1 kHz-100
kHz, and each has four outputs for the decade dividing. The switching
elements need to have as low an on resistance (<< 1 ohm) as possible,
and function well at extremely low levels. Since this is an AC system,
I'm not worried about Seebeck effects, just the low level contacting
ability. The primary sides of the transformers are at more normal
levels, so the switches there are not critical.

I have all sorts of nice low level regular and Hg reed relays that would
do the job, but since the output side is to be isolated, to minimize
interference, it would be much better to do all the switching passively
with mechanical switches. This would avoid needing electricity, and the
proximity problems of having relay coils and capacitance in the low
level environment, and power supply and line noise and ground loops (no
power cord), or messing with battery power. Especially, there would be
no power transformer emissions to worry about.

I have lots of mechanical switches of all sorts. I'd like to go with a
rotary wafer type to select the transformer taps and frequency ranges. I
have mostly standard Ag plated type contact ones and some Au ones. I can
usually build a switch from pieces for almost any arrangement.

Another, but quite complicated option is to use low level reeds actuated
by a mechanically driven magnet, but doing that would be a big project
in itself, so scratch that, unless a simple way pops up. Another way is
to make heavy analog switches with big MOSFETs, but that's a lot of
parts, a battery, and issues with capacitance and crosstalk. So, good
old mechanical switching seems the best way to go - if only it can
actually be done.

The big question then, is how to get a regular, environmentally exposed
contact to be usable in the nV region. Wetting the contacts with Deoxit
is the only practical thing I can picture so far. There is another class
of contact "protectors" based on synthetic oils, but I think I need
chemical action too, so these may not work - I'll have to investigate
that too.

So anyway, does anyone have experience or knowledge of how Deoxit would
behave for very low level contacts? It may boil down to a "try it and
see" scenario, but it would be nice to have some info in advance."

Now:

I have opted for the "try it and see" scenario, and have built it up so
far using mechanical rotary wafer switches. The low-level section will
use all gold plated contacts, with Deoxit G100 applied. Also to be
included, is a "contact cleaning" mode, where some higher DC (couple
volts and mA) can be temporarily applied to all contacts from a built-in
Li cell circuit. Since I don't have enough of the right kinds of Au
plated contact wafer parts available, the higher level input (primary)
side will include some regular Ag plated contact switching too. These
will be chemically cleaned first, then have Deoxit D100 applied.

I have it mostly mechanically built, almost ready to begin wiring up and
experimenting. This will be ongoing for a while, and I would appreciate
if anyone here does have experience with such low level applications
with exposed mechanical contacts. I know we usually talk about DC here,
while this is an AC application, but the same principles apply.

Ed

Have you considered using HP 44421A multiplexer boards from a 3497A? They are cheap, though terminal blocks can be hard to come by. They are high spec reed relays and if you get the thermocouple compensated terminal blocks you can adjust for thermal EMF.

Reg

 On Saturday, June 6, 2020, 02:15:47 PM CDT, ed breya <eb@telight.com> wrote:  

I would like to get any feedback or experience info on whether Deoxit
would help sufficiently to improve mechanical switch contact performance
used in very low (nV) level circuits. I posted this question in the
tekscopes group a while back, but with no response. I was surprised that
there was no info from such a wide-ranging group, but maybe here in
volt-nuts, with narrow focus, it will stand a better chance. Here is
approximately what I wrote, slightly edited, for consideration.

Then:

"One of my projects involves boxing up some special transformers and a
switching system, for isolating and dividing down AC signals by large
factors (E-3 to E-6), down as far as the 1 nV RMS region. A fair number
of switching elements are needed to route and select various transformer
output taps to send to the single output connection. There are two
transformers to cover two frequency bands 10 Hz-1 kHz, and 1 kHz-100
kHz, and each has four outputs for the decade dividing. The switching
elements need to have as low an on resistance (<< 1 ohm) as possible,
and function well at extremely low levels. Since this is an AC system,
I'm not worried about Seebeck effects, just the low level contacting
ability. The primary sides of the transformers are at more normal
levels, so the switches there are not critical.

I have all sorts of nice low level regular and Hg reed relays that would
do the job, but since the output side is to be isolated, to minimize
interference, it would be much better to do all the switching passively
with mechanical switches. This would avoid needing electricity, and the
proximity problems of having relay coils and capacitance in the low
level environment, and power supply and line noise and ground loops (no
power cord), or messing with battery power. Especially, there would be
no power transformer emissions to worry about.

I have lots of mechanical switches of all sorts. I'd like to go with a
rotary wafer type to select the transformer taps and frequency ranges. I
have mostly standard Ag plated type contact ones and some Au ones. I can
usually build a switch from pieces for almost any arrangement.

Another, but quite complicated option is to use low level reeds actuated
by a mechanically driven magnet, but doing that would be a big project
in itself, so scratch that, unless a simple way pops up. Another way is
to make heavy analog switches with big MOSFETs, but that's a lot of
parts, a battery, and issues with capacitance and crosstalk. So, good
old mechanical switching seems the best way to go - if only it can
actually be done.

The big question then, is how to get a regular, environmentally exposed
contact to be usable in the nV region. Wetting the contacts with Deoxit
is the only practical thing I can picture so far. There is another class
of contact "protectors" based on synthetic oils, but I think I need
chemical action too, so these may not work - I'll have to investigate
that too.

So anyway, does anyone have experience or knowledge of how Deoxit would
behave for very low level contacts? It may boil down to a "try it and
see" scenario, but it would be nice to have some info in advance."

Now:

I have opted for the "try it and see" scenario, and have built it up so
far using mechanical rotary wafer switches. The low-level section will
use all gold plated contacts, with Deoxit G100 applied. Also to be
included, is a "contact cleaning" mode, where some higher DC (couple
volts and mA) can be temporarily applied to all contacts from a built-in
Li cell circuit. Since I don't have enough of the right kinds of Au
plated contact wafer parts available, the higher level input (primary)
side will include some regular Ag plated contact switching too. These
will be chemically cleaned first, then have Deoxit D100 applied.

I have it mostly mechanically built, almost ready to begin wiring up and
experimenting. This will be ongoing for a while, and I would appreciate
if anyone here does have experience with such low level applications
with exposed mechanical contacts. I know we usually talk about DC here,
while this is an AC application, but the same principles apply.

Ed

volt-nuts mailing list -- volt-nuts@lists.febo.com
To unsubscribe, go to http://lists.febo.com/mailman/listinfo/volt-nuts_lists.febo.com
and follow the instructions there.

Thanks guys, for the responses already - I should have asked here in the
first place. Here's a consolidated answer:

Reg, I could use a low level scanner if it was for a general or one-time
measurement, but this is to be a compact, self-contained AC divider, for
isolated low level signal generation for experimenting with lock-in
analyzer setups. The transformer modules to be used came from a Fluke
5200A attenuator board. I initially planned on only the low and middle
band transformers, for 10 Hz to 100 kHz with good flatness, which covers
the range of all my LIAs. But then I figured I'll include the high band
(up to 1 MHz) too, using the last, little transformer. Since it's AC,
I'm not worried about thermal issues.

David, thanks for the link. I have seen and have the Keithley low level
measurements bible somewhere, but I think this is the newest edition
I've seen. I have copied the pdf into several places now.

Todd, great find! It looks like the 262 used good old copper-copper
rubbing (to cut the oxides) contacts for low level and low thermal
characteristics. Also note the lowest, final dividing resistor network
is all-copper resistance and connections. I've seen this before in old
Keithley gear. My model 148 nV meters use similar in the feedback
divider, and also heavy lug Cu-Cu switches in the front end ranging. I
see the reference to the Cramolin oils in the 262 manual - these were
indeed the forerunners of what became Deoxit. It's a long, complicated,
and kind of interesting story - I studied it a few years ago.

So anyway, I'm feeling better about going with my mechanical switches
with Au and Ag contacts and Deoxit combo. If worse comes to worse, I
guess I could scrape off the plating and go Cu all the way, but I don't
think that will be needed.

Ed

One thing I should mention too is that because it's a wide bandwidth AC
system, with a wide signal range, there are issues with stray
capacitance in many places, so the internal connections and routing have
to be carefully made, and as compact as possible.

Ed