HP3458A

[Laval]

Hello everyone.

I just acquired a HP3458A 8.5 digit multimeter. I just got it yesterday and didn't really have time yet to play with it. I plan on analysing and testing this thing as much as I can but it will have to wait a few days. My main concern right now is the NVRAM. This a HP unit, pre Agilent era so it's old. It is most likely equipped with the Dallas chips that have integrated batteries in them. I don't know if they have been changed before, the date code on the chips will give me a hint I guess (I haven't cracked it open yet) but they may have passed their 10 years safe life time. Since the instrument has just been calibrated before chipping, I'm afraid I could lose the calibration data. The DS1220AD-150 NVRAM ICs are readily available at Mouser (they have over 1000 of them) and they are not really expensive. Is there an alternative to these ICs, one that doesn't need batteries. I know there is a model with snaphat batteries that can be replaced but I was wondering if there is a better solution. Also, is it possible to backup the calibration data and upload it back once they have been replaced ?

[Arhigos]

Is there an alternative to these ICs, one that doesn't need batteries

yes. there is FRAM adapters available

Also, is it possible to backup the calibration data and upload it back once they have been replaced ?

Yes. You can backup cal data using this app https://www.eevblog.com/forum/metrology/3458a-logging-via-windows-app-revisited/

[Laval]

Wow, thank you !

[alm]

I suggest you edit the title of the first post from HP3858A to HP 3458A. I figured the HP 3858A might be a HP instrument I'm not familiar with, like a spectrum analyzer, and only clicked on the topic out of curiosity.

I recommend you check xdevs.com for a wealth of information and resources on this meter. For example this article or this article describing the FRAM upgrade.

[Laval]

Yes thank you very much. xdev indeed has very extensive and comprehensive information on that device, it looks like he has repair a few.

[coromonadalix]

and  when done   

you'll need to check  if you don't have the A3 board drifting issue(s), 

it is no longer sold from keysight to personal users,  unless you send your meter there, and for sure you pay lots of $$$ to have it repaired and calibrated ... 1k - 1.2k$

[Laval]

Yes, I'm hoping to avoid the A3 convergence error. This multimeter is very expensive to get calibrated and I don't have good enough instruments to do it. I think you need a Fluke 5720A or better and all the Fluke voltage and resistor references. I think I'll just backup the cal data for now.

In case someone wants to take a look, I posted the calibration report that came with the unit.


Edit: sorry, the forum won't put the pictures in the right order for some reason.

[mendip_discovery]

Did you get a set of results from before they adjusted it?

[Laval]

I asked for that but unfortunately it wasn't available.

[CalibrationGuy]

I've been in the calibration industry for 34+ years. The Fluke 5720A is a great instrument but alone is not capable to really calibrate a 3458A. For uncertainty levels that are really low, necessary if you are chasing the ppms, you need a Standards Lab Calibration from Keysight or the equivalent from Fluke. Also, given the 5720s drift spec., it is important to know when the standard used was calibrated. You can really go down a rabbit hole when chasing ppms. I know because I have 4 3458A meters on my bench calibrated within 6 months but only agreeing within 3.4ppm from lowest to highest meter at the 10volt range. I have an old 732A that reads 10.0000007Vdc with less than 1ppm drift in 18 months.

PS - For anything more than hobby use, you should at least splurge on a calibration done at Keysight, at least once. Good luck!

TomG.

[bdunham7]

I know because I have 4 3458A meters on my bench calibrated within 6 months but only agreeing within 3.4ppm from lowest to highest meter at the 10volt range. I have an old 732A that reads 10.0000007Vdc with less than 1ppm drift in 18 months.

I would be curious to know 1) how far that spread goes from the actual value (presumably from your 732A) and 2) what the spread is immediately after doing ACAL on all of them versus 24 hours after the ACAL.

I'm not sure if you are saying the differences are due to drift or due to differences in the reference standards used to calibrate the meters?

[CalibrationGuy]

@bdunham7: If using my very recently calibrated 3458A (by Keysight Metrology, Standards Lab Calibration, .3ppm uncertainty), the errors are as follows:

1) 3458A#1: +1.8ppm (Keysight ISO17025 calibrated) (cal date 05/25/2023)
2) 3458A#2: +1.6ppm (Keysight ISO17025 calibrated) (cal date 03/29/2024)
3) 3458A#3: +3.3ppm (ISO17025 calibrated, not by Keysight) (cal date 10/30/2023)
4) 3458A#4: +0.0ppm - the reference meter just returned from Keysight Metrology

The reference meter agrees with the 732A within .3ppm (today's reading was even less). That original measurement was taken by the 3458A 18 months ago when the unit returned from Keysight Metrology.

So, those values were taken after ACAL, NPLC300, leads zeroed out. If you wait 24 hours, things get interesting. Sometimes they drift together, sometimes more apart.

My conclusion: The ISO17025 grade calibration is referencing a Fluke 5730A calibrator in all 3 cases while my reference meter was calibrated by comparison to Keysight's JVS. Of course, YMMV.

Almost forgot to mention the obvious which is that the Fluke 5730A calibrated meters agree within 1.7ppm, the two calibrated at Keysight within 0.2ppm!

Hope this was useful to you.

TomG.

[Laval]

I just realized, I probably should have posted this in Test Equipment.

[bdunham7]

Almost forgot to mention the obvious which is that the Fluke 5730A calibrated meters agree within 1.7ppm, the two calibrated at Keysight within 0.2ppm!

Yes, those numbers do tell a story.  Keysight's DMM specs are "relative to calibration standards" as I suppose everything is in a way.  I'm curious to know what the expanded uncertainty (as stated on the cert) of the 10V reference was for the 17025 calibrations that used the 5730A calibrator. 

[CalibrationGuy]

Almost forgot to mention the obvious which is that the Fluke 5730A calibrated meters agree within 1.7ppm, the two calibrated at Keysight within 0.2ppm!

Yes, those numbers do tell a story.  Keysight's DMM specs are "relative to calibration standards" as I suppose everything is in a way.  I'm curious to know what the expanded uncertainty (as stated on the cert) of the 10V reference was for the 17025 calibrations that used the 5730A calibrator.

Their claim is 1.1ppm. Reasonable if you have enough calibration history for the Fluke.

TomG.

[mendip_discovery]

I've been in the calibration industry for 34+ years. The Fluke 5720A is a great instrument but alone is not capable to really calibrate a 3458A. For uncertainty levels that are really low, necessary if you are chasing the ppms, you need a Standards Lab Calibration from Keysight or the equivalent from Fluke. Also, given the 5720s drift spec., it is important to know when the standard used was calibrated. You can really go down a rabbit hole when chasing ppms. I know because I have 4 3458A meters on my bench calibrated within 6 months but only agreeing within 3.4ppm from lowest to highest meter at the 10volt range. I have an old 732A that reads 10.0000007Vdc with less than 1ppm drift in 18 months.

PS - For anything more than hobby use, you should at least splurge on a calibration done at Keysight, at least once. Good luck!

TomG.

I think they used the 8508 to get the uncertainty lower. The 5720A was just used at a source measured by both meters.

[CalibrationGuy]

The guys at the standards calibration lab called me regarding my concerns over the uncertainties across different levels of service and they agree that I should use the meter that was calibrated at the standards calibration lab as the main meter. The 10 volt range on that meter was characterized using direct comparison with their JVS.

I will be sending my Fluke 732s to them soon, so our uncertainties will be on the order of .3ppm or so.

Getting ISO17025 accreditation and maintaining low uncertainties has been a 100K+ effort so far.

TomG.