calibrating a calibrator...on the cheap?

[willb]

I recently bought a working Fluke 335A DC standard for 250$ shipped to my door. I own several meters, and felt like it was a decent price and it would be fun to play around with. I've got it today and started playing with it. It does work, but I want to verify it's accuracy. I borrowed a freshly calibrated (June 2013) HP 3455A meter from work to verify my calibrator. Our meters get sent out to a highly reputable cal lab twice a year. I don't own anything of that class, my best meters are my HP 3478A, Fluke 45 and 87-IV.

I was wondering if it would be a good idea to use the calibrated 3455A to use as a transfer standard to adjust my 335A? Or should I just leave it as is? Is the 3455A good enough to use as a transfer standard? This is just for hobby use, so I don't really want to fork out more for calibration than what I paid for the unit.

I took the following picture after both units being powered up for 30 min.

[Wim13]

The Fuke has a acc. of 0.002%
The Hp has a acc of 0.002% also, and 0.005% for 90 days...

That gives uncertainty of at least - or + 0.007% of 10 volts is 0.7 milliVolt

on your display you have 0.65 milliVolt off, thats within the range, can not get closer i think,
because you dont have more resolution.

[PA4TIM]

My 332 needs about 2 hours to warm up.

[eurofox]

My 332 needs about 2 hours to warm up.

Holy Moly this is the alibaba place of instruments 

[PA4TIM]

My 332 and some other calibration gear  O0

[branadic]

Damn, this is like "Who has the longest d*ck?"
You guys don't have kids, do you? What the hell do you need all the "wastage of energy" for?

[M. András]

Damn, this is like "Who has the longest d*ck?"
You guys don't have kids, do you? What the hell do you need all the "wastage of energy" for?

cold climate aka room heater

[TopLoser]

Last year I was offered a Fluke 7500A and 7525A. Looked almost new, I had no idea what they were worth and offered £1000 for the pair. Sold them the next day for 10 times that amount! I like calibration equipment now!

[orin]

I think the simple answer is that you don't stand a chance adjusting the 335A using a 3455A. 

Even the 10V range calibration is looking for +/- 10uV difference from a 10V standard (actually, it's +/- 1uV from a standard cell but who uses standard cells any more?).  That's one count on the 3455A's 10V range.

As for the linearity adjustments, they require setting to at worst +/-3 uV and the 3455A on its own certainly can't do that.  [I recently adjusted a 343A which has looser requirements and even using a 34461A's averaging feature to eke out an extra digit of resolution, it was a near impossible task.  It did end up better than it started out, as far as I could tell using the 34461A anyway.]  Probably doable with a 3458A...

BTW, you should do your 10V check on the 10V range with the dials set to 10.000000.   As Wim13 pointed out, the Fluke accuracy is 0.002% which is 2mV on the 100V range.  Your result would be in spec. given the 0.005% of range + 3 count uncertainty of the 3455A (90 day spec).  Also, make sure the 335A isn't current limiting - the reading shouldn't change if you turn the current limit knob up.

So give it a few hours of warmup and do the measurement on the 335A's 10V range.

[PA4TIM]

Even the 10V range calibration is looking for +/- 10uV difference from a 10V standard (actually, it's +/- 1uV from a standard cell but who uses standard cells any more?). 

I do, On top of that 19"rack you see a brown cabinet. I have it from the first owner including all papers until its retirement. After that it was not used the last 30 years and stood all that time unpowered on the same spot of that same cal lab. It is transported with much care to my lab. Restored with some support from Guildline.

This is what is inside:

[Mechatrommer]

You guys don't have kids a wife, do you?

grammar corrected, thanks!

[willb]

I think the simple answer is that you don't stand a chance adjusting the 335A using a 3455A. 

Even the 10V range calibration is looking for +/- 10uV difference from a 10V standard (actually, it's +/- 1uV from a standard cell but who uses standard cells any more?).  That's one count on the 3455A's 10V range.

As for the linearity adjustments, they require setting to at worst +/-3 uV and the 3455A on its own certainly can't do that.  [I recently adjusted a 343A which has looser requirements and even using a 34461A's averaging feature to eke out an extra digit of resolution, it was a near impossible task.  It did end up better than it started out, as far as I could tell using the 34461A anyway.]  Probably doable with a 3458A...

BTW, you should do your 10V check on the 10V range with the dials set to 10.000000.   As Wim13 pointed out, the Fluke accuracy is 0.002% which is 2mV on the 100V range.  Your result would be in spec. given the 0.005% of range + 3 count uncertainty of the 3455A (90 day spec).  Also, make sure the 335A isn't current limiting - the reading shouldn't change if you turn the current limit knob up.

So give it a few hours of warmup and do the measurement on the 335A's 10V range.

Thanks everyone for the replies!

I did have the 335A set to 10.00000 on the 100v range with the 3455A on the 10V range, High Res mode and AutoCal enabled, if that makes any difference. I didn't realize that the first dial goes to 10, hence why I was using the 100v range on the 335A. I feel silly now. I guess i was too excited to play being new to me. The unit was not current limiting, I did vary the current limit knob with no change. I left everything running for 3 hours, and it seemed to have stabilized to 9.99919 on the 100 range. I will repeat the test tomorrow with the proper settings and ranges and post here the results.

[willb]

I'm also looking for a user manual and service manual. Google isn't turning anything up.

edit : got it.

[orin]

Even the 10V range calibration is looking for +/- 10uV difference from a 10V standard (actually, it's +/- 1uV from a standard cell but who uses standard cells any more?). 

I do, On top of that 19"rack you see a brown cabinet. I have it from the first owner including all papers until its retirement. After that it was not used the last 30 years and stood all that time unpowered on the same spot of that same cal lab. It is transported with much care to my lab. Restored with some support from Guildline.

I should have known better than make that comment around here

Orin.

[staxquad]

I'm also looking for a user manual and service manual. Google isn't turning anything up.

edit : got it.

The Guildline had 1ppm accuracy specs, wonder why they gave that up? 

Does your workplace have a calibration certificate with the calibration values for the HP 3455A?  That would be a more accurate transfer than just a meter within calibration.

for others:
http://assets.fluke.com/manuals/335A_D__imeng0000.pdf

Was watching that unit you purchased.  Might have gone for it if I hadn't purchased my voltage source.   Mine doesn't have high voltages or current.  

I got my voltage source to check my 5 1/2 digit and lower meters.  I'm satisfied it does that, so no need to send the unit to a calibration lab to validate all 7 digits,  6 of the digits seem to be close enough. 

the voltage source, 2 hp 3478As and the seller's 34401A seem to be on the same page

[ve7xen]

GR "Division of Chronetics"

Nice.

[willb]

I don't think I can get the cal paperwork for the meter. I'm going to check today, but I doubt it.

I watched the unit for a while on eBay, it was listed at 275. Made an offer at 200 and got it. I'm going to run it again today after work for several hours. I had it operating for an hour or so yesterday on the 10v range with the meter reading 9.99954v. so far it seems in spec, so I'm pleased for the price.

[Dr. Frank]

.. Our meters get sent out to a highly reputable cal lab twice a year...

Oh, congratulations for that beautiful instrument..

Simply wait another 3 months, then use a REALLY fresh calibrated 3455 with +/-20ppm uncertainty to check your 335A, dial set always to X.00000V in 10V range.

I don't expect a difference to your measurement now, because such old 3455 DMM normally are much more stable than the formal specification indicates.

Range calibration of 100V, 1000V may also be possible with a 3455, but to a limited uncertainty only. (i.e. worse than spec).

Linearity alignment is difficult with a 3455A, as it lacks the Secondary Specification (Transfer stability and Linearity).
A 34401A would be suited better, but not good either, i.e. 2ppm linearity.

This is a wonderful design project, to build a precise chain (divider) of 10 identical resistors, including a Wheatstone Bridge for linearity alignment. <1ppm of linearity is required.
Can be built easily from 12 precision wire wound 10k resistors and a bunch of trimmers.

Rem.: You also need a stable 10,0000V source to feed this divider and to produce a linear 1,2, 4,6,8,10V sequence.
See manual of the 335D version, which is much more instructive regarding calibration.

I've done that successfully exactly for that purpose, see my box here, in the range calibration configuration
(description in "Ultra Precision LTZ1000" thread)

Frank

[MetraCollector]

Damn, this is like "Who has the longest d*ck?"
You guys don't have kids, do you? What the hell do you need all the "wastage of energy" for?

Don't be so curious !

[willb]

I was too busy to look into the cal paperwork today, but I did have it running for 5 hours when I got home. I think I'm going to leave it as is for now, as it does seem within spec, on the 10V range anyway.

[Dr. Frank]

I was too busy to look into the cal paperwork today, but I did have it running for 5 hours when I got home. I think I'm going to leave it as is for now, as it does seem within spec, on the 10V range anyway.

Spec. states 20ppm initial / 24h uncertainty, so it might be trimmed to exactly 10.0000V reading.
What about the other two ranges? Did you try do measure linearity?

It's within the 90d spec, but depending on the reference used inside, it can do better.

What version do you have, concerning reference assembly, and concerning Chopper?
Is it the mechanical chopper, or the FET chopper already?

Have you opened the box already???

Frank

[staxquad]

I was too busy to look into the cal paperwork today, but I did have it running for 5 hours when I got home. I think I'm going to leave it as is for now, as it does seem within spec, on the 10V range anyway.

and from the same seller? (good stuff!)....
you will never have adequate self esteem if you don't get one of these:
http://www.ebay.com/itm/290967150462?ssPageName=STRK:MEWAX:IT&_trksid=p3984.m1438.l2649

who knows if the cells are any good anymore?

your unit from the same seller? 
http://www.ebay.com/itm/Fluke-335A-DC-Voltage-Standard-Differential-Voltmeter-/261261437395?pt=LH_DefaultDomain_2&hash=item3cd46529d3



Made in Canada to boot! Too bad it ain't a Guildline T12

Twelve Weston standard cells inside a temperature stabilized oven. Each mercury cell has separate output terminals. Cells are built according the process first devised by Edward Weston in 1892.
Nominal output voltage of each cell is 1,01830 V at 20 degrees centigrade. Cells are individually calibrated with an accuracy of about 1 p.p.m.

more accurate and more stability than a 3458A.  The best a 3458A can do is 4ppm/yr

At any room temperature, other than 20°C, the output voltage of the Weston standard cell could be computed according to the formula:

V(t) = 1,01830 – 0,0000406 ( t – 20 ) – 0,00000095 ( t – 20 )² + 0,0000001 ( t – 20 )³

where t was the actual room temperature.

This means that the absolute voltage differences among the twelve cells were not higher than one microvolt, also inclusive of the measurement errors. Not bad, indeed, and I would like to know who, today, can achieve the same results, even with the most complex and spectacular solutions! I have no way to verify its calibration, but a simple check, with a digital multimeter at the room temperature, returned 1,019 V on all the 12 cells.

http://www.edelpro.com/Guildline%209152.htm

[PA4TIM]

Nice, but you miss the top cap over the terminals. The picture you show is a 4 cell 9152/T4 cabinet. Do you have a 12 cell version  too ?
http://www.pa4tim.nl/?p=2141 This is my 9152T/4 with pictures of the inside made when I changed foam.

[staxquad]

Nice, but you miss the top cap over the terminals. The picture you show is a 4 cell 9152/T4 cabinet. Do you have a 12 cell version  too ?
http://www.pa4tim.nl/?p=2141 This is my 9152T/4 with pictures of the inside made when I changed foam.

the 4 cell is for sale and I am quoting from a website that has a 12 cell (I need to take a cold shower now)

but, being in Canada, maybe I can stumble on one one day