DMM relay lifetime

[LaurentR]

I was plotting an I/V curve using the 34461A by programmatically switching between Voltage and Current modes and realized that each switch took a bunch of relay clicks.

Looking up the topic, I found an NI page pegging a relay lifetime to about 10^5 cycles and showing how their switching products tracked relay actuations (and the counters are software accessible).

Looking at my little experiment, I was drawing about 1000 points, so that's a few 1000s relay actuations, a non-negligible fraction of  10^5.

I didn't find any spec in the DMM's manual. Are these specs commonly mentioned?

Are programs that trigger a lot of relay switching (like my current vs. voltage plotting) basically not recommended?

 

[Paul Moir]

10^5 is shockingly low for a relay lifespan.  Would this happen to be contact lifespan at full current?

Normally you operate high cycle count relays at much less than their rated current, since this will greatly increase their lifespan and allow you to reach their mechanical lifespan.  For example, a typical Omron ice cube relay is rated at 50 million mechanical operations.  In my experience, that's a little conservative in normal applications.

[LaurentR]

10^5 is shockingly low for a relay lifespan.  Would this happen to be contact lifespan at full current?

I got the number from this NI page.

[Paul Moir]

Oh, I'm sorry.  You said NI and somehow I got a relay manufacturer vaguely stuck in my head.  For some reason I was thinking of TE. 

[LaurentR]

I got an answer from Keysight. They do not spec relay actuation or guarantee them.

However, they pointed me to a relay spec.
There is an electrical spec (5.10^5, in line with the NI number) and a mechanical spec of 10^7 or 10^8 depending on the relay type.

They claim they are careful to remove the load when a relay actuation is necessary, so it should be closer to the mechanical spec then the electrical spec (which assumes full rated load).

This matches Paul's comment well:

10^5 is shockingly low for a relay lifespan.  Would this happen to be contact lifespan at full current?

Normally you operate high cycle count relays at much less than their rated current, since this will greatly increase their lifespan and allow you to reach their mechanical lifespan.  For example, a typical Omron ice cube relay is rated at 50 million mechanical operations.  In my experience, that's a little conservative in normal applications.

[dfmischler]

I guess the obvious answer is that you need twice as much equipment so you don't need to tell the meters to switch modes or ranges 

[LaurentR]

I guess the obvious answer is that you need twice as much equipment so you don't need to tell the meters to switch modes or ranges 

Yes yes. I actually bought a used 34401A for that purpose but was hoping to do 2 voltages measurements, but doing 1 or 2 current measurements (especially the supply) at the same time would be great 

[radioFlash]

You could set the meter to a manual range fixed range if you can accept the lower accuracy for the lower part of the range.

[LaurentR]

You could set the meter to a manual range fixed range if you can accept the lower accuracy for the lower part of the range.

Indeed.
It turns out that the higher ranges between current and voltage measurements switch to/fro without any relay being involved, but I believe that's only for 100V and 1A/3A/10A. I haven't found a working combination (i.e. without relay activation) at lower ranges. With 6 digits of precision and if (absolute) accuracy is not required, that can certainly work for mV/mA measurements.