Can I auto-zero a sub-microvolt DC signal with a relay? (null detector)

[Squantor]

Can you give a link to the mentioned HP-forum-post Dieter?

I was actually following that discussion a while ago when I was repairing my HP419A, here is the link:

https://groups.io/g/HP-Agilent-Keysight-equipment/topic/24812737#68881

They moved from yahoo groups to groups.io

Searching for hp419A or the 3400A you can read up on a lot of discussion replacing the photochopper with various other devices. Worth digging through that.

[Echo88]

Thanks Squantor.
Found the following: https://groups.io/g/HP-Agilent-Keysight-equipment/message/68873?p=,,,20,0,0,0::Created,,h11f1+30uV,20,2,0,24812737
But its 30µV instead of 30nV and no mention of H11F1-aging. Maybe i missed the aging-part in some other post?

[dietert1]

It's in a HP419A thread:

https://groups.io/g/HP-Agilent-Keysight-equipment/message/75908?p=,,,20,0,0,0::Created,,H11F1,20,2,0,24814081

Yes, there is a difference between nV and uV. My mistake - difficult to believe that Fluke put something with an offset voltage of several hundred uV into a uV null detector that is supposed to work down to 200 nV. Really, i still don't believe it.

Regards, Dieter

[SilverSolder]

As long as the offset is stable and consistent - it can probably be nulled out?

[dietert1]

I just tested a COTO 6213-1 LED driven mosfet switch we had (old parts from the 90s). The part shows an offset voltage of 0.32 uV. On state resistance is 5.5 Ohms. Off resistance as measured by HP 3456A is 5.35 GOhm.
Another part one can order from RS would be Vishay LH1510 that has a spec: "Switch offset at 5 mA: 0.17 uV typical". Typical on state resistance is 12 Ohm.

That means these opto mosfet switches do not generate "several hundred microvolts". A stable offset voltage of less than a microvolt would be something one may compensate in the input stage of a null meter.

Regards, Dieter

[Kleinstein]

The offset of less than 1 µV would be acceptable for most cases, especially higher impedance use. Beside the photovoltaic, relatively fast part, there may be another slow thermal part, as the LED side will heat up the switch.

However there could still be charge injection from the internal FET. It is no per se clear that charge injection is less than with a normal MOSFET or JFET. It is only the normal capacitance around the switch that is reduced, not that actual redistribution of the gate charge.