Oh dear, another drifty 3458A

[perdrix]

I'm currently running the Service note 18 drift test, and am already way out of bounds after a few days



The error bars on the CAL72 values are set to 3.4E-5 percent of the reading which I think is the 0.34 ppm magic number.

The meter was powered on after a long period in storage on the day before the first reading in the graph.

Is this a definite case of "you need a new A3 board", or is there some hope that it can be fixed or will stabilise after a few more days?

Thanks
Dave

[TiN]

I'd leave meter to run a week or two after storage and then do the test. You may have got just long storage stress.

[tggzzz]

Keep measuring before spending money

My Solartron 7081 has an interesting maintenance procedure after being in storage: wrap it in insulation and keep it powered for 24 hours to raise the internal temperature.

I used a blanket, and the internal temperature rose to, IIRC, around 40C. It was noticeably better thereafter.

I make no comment about the relevance or otherwise to your 3458.

[perdrix]

Doh!  It's nowhere near as bad as I initially thought - the limits are 0.43ppm per DAY, not 0.34 per week which is what I had misremembered!

I've since updated the spreadsheet to set the errors bars at 4.3E-5 percent.   

The only period in which the limit was exceeded was the first one where the delta was 0.9ppm.

The next 24 hour drift was 0.13ppm.

So fingers crossed!

Dave

[MisterDiodes]

It's probably fine, and you'll probably see it settle down a bit more.  Let it run for -at least- a few weeks after a long power-off storage period before gathering any serious data.

[Pipelie]

fingers crossed!

[mimmus78]

My 3458a despite having a jumpy reference with a horrible temperature coefficient A9 has a very good "resistor switch/chip".

23/12/2016   35.20   997.876849
09/01/2018   35.20   997.876575

0.3ppm per year seems to qualify as "good enough" also for the 0.34ppm per week (but also month or year)

[TheSteve]

My 3458a despite having a jumpy reference with a horrible temperature coefficient A9 has a very good "resistor switch/chip".

23/12/2016   35.20   997.876849
09/01/2018   35.20   997.876575

0.3ppm per year seems to qualify as "good enough" also for the 0.34ppm per week (but also month or year)

This looks impressive - has it been powered on for the entire time(it is only legit if powered on).

[mimmus78]

My 3458a despite having a jumpy reference with a horrible temperature coefficient A9 has a very good "resistor switch/chip".

23/12/2016   35.20   997.876849
09/01/2018   35.20   997.876575

0.3ppm per year seems to qualify as "good enough" also for the 0.34ppm per week (but also month or year)

This looks impressive - has it been powered on for the entire time(it is only legit if powered on).

No not 24/24, let say 20% ... anyway still impressive. It has the new analog board.



Inviato dal mio ONEPLUS A5010 utilizzando Tapatalk