Which connectors for Resistor Box

[derZockerMP]

Hello, I would like to build a Resistor Box with standard values 100 Ohm, 1k, 10k, 100k, 1M with precision Resistors from Vishay. The Tolerance schould be around 0,01% and 0,2ppm drift per Degree. The housing sould be a standard metal thing. The critical question is the right choice of the connector. Can i choose standard 4mm banana jacks like these?

https://www.conrad.de/de/p/staeubli-slb4-f6-3-n-x-sicherheits-laborbuchse-buchse-einbau-braun-1-st-1678343.html

Or sould I use BNC?
Any recommandations?

Thanks for help and I hope i will be welcome as a new member by the voltnuts

[Gyro]

If  you're only going as high as 1M, the shrouded 4mm sockets you linked should be absolutely fine. You probably (definitely) want to consider 4 terminal Kelvin connection for the 100R and 1k though to eliminate wire and connector resistance..

[Alex Nikitin]

Unfortunately, Stäubli seems to be reluctant to name the isolation material used in these sockets. It could be important for larger resistances - I've measured some Nylon sockets to have about 3.3 Gohm leakage resistance to the mounting panel, which might create a 0.03% error for 1 Mohm resistor.

Cheers

Alex

P.S. As Gyro noted, you would need a 4W connection for 10k and below, if you care about ppm level of precision, however for 100k and above a (quality) BNC is a good choice.

[guenthert]

If  you're only going as high as 1M, the shrouded 4mm sockets you linked should be absolutely fine. You probably (definitely) want to consider 4 terminal Kelvin connection for the 100R and 1k though to eliminate wire and connector resistance..

   You will want to use 4W resistance measurements for low value resistors to eliminate the (ill specified) resistance of the test leads, but for that you don't necessarily need four terminals.  If you look e.g. at the wiring of the ESI SR104 standard, you'll see that internally the I and P terminals are connected using a fairly short conductor and a much longer then leads to the resistance element(s).  That's fine, as the internal wiring isn't changing and those conductors are included in the calibrated value.  It's a matter of convenience hence.

   TiN is fond of pure copper terminals it seems.  I think those are generally too cumbersome (delicate and absolutely requiring cleaning before use) and belong to specialized use cases.  Low thermal EMF terminals for the Potential terminals (if there are separate I and P terminals) would be nice, but I think many resistance standards get away without (many better DMMs can separately measure and subtract thermal EMF during resistance measurement and bridges usually offer to reverse the driving current).


EDIT: "convenience" above isn't quite cutting it.  If there are exactly two positions for I and P conductors to attach, then it doesn't matter how it's done.  If however there's a continuum of positions, then minuscule variations will yield varying results.  That's clearly an issue for high precision measurements or low value resistances.

[derZockerMP]

Thanks for the answers. I'll try the standard banana jacks. Before that, I'll try to measure the isolation resistance as best as I can with my K2400. It's not perfect, but it should do the job.

[EC8010]

Neither BNCs nor 4mm terminals are perfect. But BNCs are good for higher impedances because of their intrinsic shielding. I have converted my 8 decade resistance box to BNC connections and made a 100k to 100M 3 decade box with BNC connections because BNCs are quick to plug up. It allows me to quickly insert a variable resistance and know that I haven't introduced hum. Unfortunately, when you use BNCs and DMMs, you introduce thermocouple errors in the BNC-4mm adapters. Prema recommended wrapping a towel round the adapters to reduce the effect. The ultimate is beryllium copper screw terminals, but they need to be cleaned each time. The cheap silver-plated brass non-stackable 4mm plugs seem to be better than stackable 4mm plugs. Each time, it seems to be a compromise between convenience and quick plugging versus thermocouple errors.

[mzzj]

This could be one of cases where UHF connector makes sense

[EC8010]

Could be. "Coaxial Electrical Circuits for Interference-Free Measurements" suggests using MUSAs. The MUSA connector is a coaxial video connector developed by the UK Post Office to allow video signals to be plugged on patch bays with the same ease as audio. Television broadcasters used MUSAs a lot for baseband video. However, a panel MUSA connector is quite deep and they were quite expensive, so I'm not going to be changing all my kit from BNCs to MUSAs.

[Irilia]

I think Gold plated is the way to go, it doesn't need cleaning but offer the best connection you can ask for the price

[coppercone2]

that BPO connector looks nice. the machines gone postal

that sounds like something you use for detecting muons