SOLARTRON 7150 Replace mains filter now ? and calibration

[chickenHeadKnob]

I love this analogy, i think i may use it when i email them tomorrow

Couldn't you just quietly, on official looking letterhead, just invoice them for the rental of your solartron "transfer standard", which they used to calibrate their calibrator ?  .  just to see what happens.

[dimlow]

LOL! I liked that

[mzzj]

Ah, thank you! I will change the spread sheet and see if it makes me Happier , But i think what its come down to is they have calibrated a 20ppm meter with a 30ppm Calibrator, Am i right in thinking that ?

I think you are comparing eggs with apples. Your solartron is specified for 20ppm FOR 24 hours. More relevant specification is the 2-year spec: 100ppm + 10 counts.

Gosh, I will always get headache from these "traceable" calibration certificates. At least your calibration is so called "with data"

ANSI Z540-1 Compliant calibration = someone puts on sticker to your meter
calibration with data = someone puts sticker to your meter and gives you some results
traceable calibration with data =someone puts sticker to your meter and thinks that results are traceable to national standards
17025 Accredited Calibration = you might actually get something trustworthy 

[jucole]

I would ask the cal lab to state the ppm for the 202mV range of their cal meter on the day,  then state the calculated ppm of an "ideal" 7150 on the 200mV range.  If the TUR is worse than 4:1, I would ask for my money back!

[dimlow]

ok, going to give then a call

[robrenz]

I think you are comparing eggs with apples. Your solartron is specified for 20ppm FOR 24 hours. More relevant specification is the 2-year spec: 100ppm + 10 counts.

Not so.  IMO If they say they can calibrate that specific meter, the lab should be able to calibrate and adjust to the 24hr values with at least a 4:1 TUR on a 6.5 digit meter.

[jucole]

I emailed the makers of the actual cal meter this morning who replied: "Model 3010A (8ppm) would provide the necessary capabilities appropriate for your Solarton 7150"

I then asked about their suggested TUR ratio, they replied:

"The Solartron 7150 specifications are :
70ppm in 5 1/2 digit mode (Best DCV spec)
35ppm in 6 1/2 digit mode (Best DCV spec)
So @ 8ppm accuracy (best DCV accuracy) the 3010A offers 4:1 ratio


I worked out the 7150 to be 22 ppm -  but I didn't factor-in anything else, so 35ppm sounds perhaps right.

The moral of the story is, if you have to measure apples with oranges - just don't forget to juice them first! 

[ecat]

On the other hand, their UKAS accreditation ( from bottom of http://www.alpha-electronics.com/StaticPage/256/ElectricalTestEquipmentUKASCalibration.html , pdf link on http://www.ukas.org/calibration/lab_detail.asp?lab_id=1443&vMenuOption=3 ) states they are only good for 32ppm on DC volts (2V to 20V @ 26ppm). Also, as we now know their unit was calibrated in March 2012 (see dimlow's test cirtificate attached to Reply #17) we have to assume the 12 month  figure of 30 + 3 applies ( 30 + 3 = 33, where do UKAS get the 32 from ? I guess 11 months does not equal 12 months)

So, yes, I think they should have said something before accepting the job, but they are not hiding anything as the information is there for anyone to read. The problem, as always, is knowing the questions to ask in advance .

[robrenz]

I emailed the makers of the actual cal meter this morning who replied: "Model 3010A (8ppm) would provide the necessary capabilities appropriate for your Solarton 7150"

I then asked about their suggested TUR ratio, they replied:

"The Solartron 7150 specifications are :
70ppm in 5 1/2 digit mode (Best DCV spec)
35ppm in 6 1/2 digit mode (Best DCV spec)
So @ 8ppm accuracy (best DCV accuracy) the 3010A offers 4:1 ratio


I worked out the 7150 to be 22 ppm -  but I didn't factor-in anything else, so 35ppm sounds perhaps right.

The moral of the story is, if you have to measure apples with oranges - just don't forget to juice them first! 

AFAIK Transmille shows thier specs for the calibrator accuracy only. you have to add the uncertainty of the standards the calibrator was calibrated to to their values. All thier specs say (Accuracy relative to calibration Standards).

[mzzj]

I think you are comparing eggs with apples. Your solartron is specified for 20ppm FOR 24 hours. More relevant specification is the 2-year spec: 100ppm + 10 counts.

Not so.  IMO If they say they can calibrate that specific meter, the lab should be able to calibrate and adjust to the 24hr values with at least a 4:1 TUR on a 6.5 digit meter.

Wrong. If I send our  3458A 8½ digit meter with high stability option to any calibration lab I am pretty sure NONE of them are going to (or are able to) calibrate it with 24hr values with TUR 4:1 

[mzzj]

On the other hand, their UKAS accreditation ( from bottom of http://www.alpha-electronics.com/StaticPage/256/ElectricalTestEquipmentUKASCalibration.html , pdf link on http://www.ukas.org/calibration/lab_detail.asp?lab_id=1443&vMenuOption=3 ) states they are only good for 32ppm on DC volts (2V to 20V @ 26ppm). Also, as we now know their unit was calibrated in March 2012 (see dimlow's test cirtificate attached to Reply #17) we have to assume the 12 month  figure of 30 + 3 applies ( 30 + 3 = 33, where do UKAS get the 32 from ? I guess 11 months does not equal 12 months)

So, yes, I think they should have said something before accepting the job, but they are not hiding anything as the information is there for anyone to read. The problem, as always, is knowing the questions to ask in advance .

For UKAS accredited uncertainties you have to actually calculate pretty much everything. And if your calibrator has good history(many years of calibration history with much lower drift than manufacturer specifies) you can also use better stability estimate than manufacturer's spec sheet.

Dimlow's (thread starter) calibration certificate  is NOT UKAS accredited calibration so Alpha Electronics is free to put any sort of uncertainties to their certificate.
UKAS-accredited  ISO/IEC 17025 calibration certificate would look quite different and probably cost 5-10 times more. 




 

[robrenz]

Wrong. If I send our  3458A 8½ digit meter with high stability option to any calibration lab I am pretty sure NONE of them are going to (or are able to) calibrate it with 24hr values with TUR 4:1 

You are correct about the 3458A the TUR starts to shrink at that level. 
I am correct when I specifically said 6.5 digit meter. 

The calibration of my Fluke 8846A:
1.0 V scale at 1.0 V   Measurement 0.999999    Uncertainty 3.1 µV       24 hour spec 24 µV
greater than 4:1

[mzzj]

You are correct about the 3458A the TUR starts to shrink at that level. 
I am correct when I specifically said 6.5 digit meter. 

The calibration of my Fluke 8846A:
1.0 V scale at 1.0 V   Measurement 0.999999    Uncertainty 3.1 µV       24 hour spec 24 µV
greater than 4:1

"Shrink" to put it slightly.
I would say it's diminished. For example our local National Measurent Laboratory can calibrate 3458A with 2 ppm uncertainty and manufacturer 24h specification is 0.6ppm
What is that, 0.6:1 TUR?  And I had to pay approx 1500 usd for this level of calibration
Agilent UK cal lab seem to be able  to do 0.8ppm according to their accreditation scope BMC (best  measurement capability)
My guess would be that is already against their josephson-standard. Must be bloody expensive, ie several thousand dollars.

My point? I guess my point was that not that much point looking at 24h stabilities.

[dimlow]

I contacted Alpha electronics today and spoke to the guy that calibrated the meter. He said he was not aware of the accuracy of the meter and calibrated it to the ukas spec set out by the ukas manager of the company as they calibrate all meters that way. He was just simply following his procedure. He again offered my money back. I think i will accept, but before i do i will talk to the ukas manager on Monday, he had last week off. They do apparently have something that can calibrate the meter. He mentioned a 1710A or 710A or something like that. I tried looking it up but could not find any details, so don't know what this instrument is.

How exactly do you calculate the ppm of a meter? I see above the jucole contacted Transmile and got a value of 35 ppm for the solartron, but says he worked it out at 22 ppm. I thought it was 20ppm. I want to be sure about this before a speak to the ukas manager. ( Transmile is about 40 mins from here, they offered to calibrate and adjust the meter for £150 )

This calibration business has taken far to much of my time now, i was stupid to think it was as simple as just taking the meter to be calibrated and the company would sort it all out for me. Then hand me back a fully calibrated meter. I now have no confidence in the calibration and im £65 + VAT down. All i really know is that the meter works. Lets hope the ukas manager can put my mind at rest.

[mzzj]

This calibration business has taken far to much of my time now, i was stupid to think it was as simple as just taking the meter to be calibrated and the company would sort it all out for me. Then hand me back a fully calibrated meter. I now have no confidence in the calibration and im £65 + VAT down. All i really know is that the meter works. Lets hope the ukas manager can put my mind at rest.

Be happy with your current calibration certificate and stop worrying nonsense. 
Are you planning to use your meter for something super-accurate within first 24 hours or what? 

Just my 2 cents as a Calibration Engineer who spends far too much time with these things. 

[jpb]

I've been following this thread with interest as I'm in a similar position. I've just bought a ADCMT 7451A 5 1/2 digit meter on e-bay and would like to get it calibrated and adjusted if need be as a standard but it seems this calibration business can be very expensive involve shipping the instrument and may very well not include any adjustment at all.

Looking at different websites (eg RS calibration services) there is no clear statement as to whether or not they will adjust the instrument or whether this counts as a 'repair' and cost extra.
Do they calibrate it to the 24 hr spec, the 90 day spec or the 1 year spec?

What is the proper spec anyway - eg on most DC Volts ranges it is listed at 0.0020% + 2 digits which should be 20ppm + something like 2/300,000 if the reference is near the top of the 320,000 count display which I guess is around 27 ppm for the 24 hour spec (+/- 1C) but for 90 days it is 60 ppm + 3 which is around 70 ppm (+/- 5C), the yearly spec is up at 140ppm+3 or say 150ppm.

I too thought it would be very simple - especially as the meter is closed box calibration just requiring the pressing of buttons to adjust the readings rather than any trimming of pots etc.

Given the yearly spec is outside 0.01% perhaps a DIY approach with something like the Geller labs voltage source and a home made voltage divider is a better route.

[mzzj]

I've been following this thread with interest as I'm in a similar position. I've just bought a ADCMT 7451A 5 1/2 digit meter on e-bay and would like to get it calibrated and adjusted if need be as a standard but it seems this calibration business can be very expensive involve shipping the instrument and may very well not include any adjustment at all.

Looking at different websites (eg RS calibration services) there is no clear statement as to whether or not they will adjust the instrument or whether this counts as a 'repair' and cost extra.
Do they calibrate it to the 24 hr spec, the 90 day spec or the 1 year spec?

What is the proper spec anyway - eg on most DC Volts ranges it is listed at 0.0020% + 2 digits which should be 20ppm + something like 2/300,000 if the reference is near the top of the 320,000 count display which I guess is around 27 ppm for the 24 hour spec (+/- 1C) but for 90 days it is 60 ppm + 3 which is around 70 ppm (+/- 5C), the yearly spec is up at 140ppm+3 or say 150ppm.

I too thought it would be very simple - especially as the meter is closed box calibration just requiring the pressing of buttons to adjust the readings rather than any trimming of pots etc.

Just to mess your head more: There is no universal standard how meter manufacturers specify meter accuracy. Some use k=2 confidence interval (95% probability) for their specifications,  on the other hand for example Agilent uses k=4 or 99,994 confidence level for 34401A specifications...
Manufacturer X may specify 20ppm/year with k=2 confidence. Means that there is 5% probability that meter  X doesn't meet the 20ppm specification.
30 ppm Agilent specification for 34401A is on the other hand with k=4  confidence level and it means that there is 0.006% probability that Agilent 34401A doesnt meet the 30ppm specification.
30ppm from Agilent is better than 20ppm from manufacturer X 


I am going to say this for last time: Using 24 hour specs for TUR calculation is nonsense unless you are planning to calibrate your meter every 24 hours!

[jpb]

I am going to say this for last time: Using 24 hour specs for TUR calculation is nonsense unless you are planning to calibrate your meter every 24 hours!

I agree when it is a case of checking the meter (i.e. not adjusting it if it is in spec) but if the meter is to be adjusted as a starting point perhaps after a new purchase or if the calibration has been lost then the 24 hour spec seems to be a good target for the adjustment. I would guess that it represents the tightest region that you can realistically expect to be able to adjust the instrument into.

[jucole]

I am going to say this for last time: Using 24 hour specs for TUR calculation is nonsense unless you are planning to calibrate your meter every 24 hours!

I'm confused; why would calibrating your 7150 meter with one 4x better require you to calibrate your meter every 24 hours?

[jucole]

I've been following this thread with interest as I'm in a similar position. I've just bought a ADCMT 7451A 5 1/2 digit meter on e-bay and would like to get it calibrated and adjusted if need be as a standard but it seems this calibration business can be very expensive involve shipping the instrument and may very well not include any adjustment at all.
Looking at different websites (eg RS calibration services) there is no clear statement as to whether or not they will adjust the instrument or whether this counts as a 'repair' and cost extra.

You might be lucky enough to find one close to where you live then you could perhaps take it in yourself to save on shipping cost.

Regarding adjustment, you have to read the service manual to understand how it is done for any particular meter;  with the 7150 in say the 200mV dc range, you apply 199.999 v, then read back the value, then tell the meter what the REAL value is;  the meter then stores the result;   The meter now knows the offsets and will apply them thereafter to hopefully give you the correct result;  So you could say the "adjustments" are already done by the calibration software running over GPIB.

Tell them that if the meter needs any additional work (repair) then to let you know before proceeding.

Do they calibrate it to the 24 hr spec, the 90 day spec or the 1 year spec?

For hobby use;  just ask them to calibrate it to as close to the "as new" as they can;   If it's done properly when you compare the result to the manufacturers spec. it will tell you the real age of the meter!

What is the proper spec anyway - eg on most DC Volts ranges it is listed at 0.0020% + 2 digits which should be 20ppm + something like 2/300,000 if the reference is near the top of the 320,000 count display which I guess is around 27 ppm for the 24 hour spec (+/- 1C) but for 90 days it is 60 ppm + 3 which is around 70 ppm (+/- 5C), the yearly spec is up at 140ppm+3 or say 150ppm.

The spec you describe is an indication by the manufacturer of the "best accuracy" after you have bought it;  and how you might expect your "best accuracy" to change over time.

[dimlow]

Looks like i didn't give the forum an update on this. I hate unresolved posts, you read all the way to the end and there is no final solution.

I spoke to Alpha electronics and they gave me my money back. That was ok but i still have an un-calibrated meter, or at least a calibration that i cannot be certain of. As mentioned above Transmile state that the meter should be calibrated with uncertainties of 8 ppm not 20 ppm as Alpha electronics did.

Over at Mjlortons Forum i detail a phone conversation i have with Calibrate.co.uk Martins Sponsor. When i phoned i didn't get the expected answers to some questions i had about Ukas and ISO 17025. Later in the thread the MD of Calibrate.co.uk replied and also said they were getting in the 8 ppm Calibrator and i could have my meter calibrated on this. I have now since spoken to the MD, Barry and will be sending the meter in for calibration with them. Within 1 hour of arranging the calibration i was presented with  a UPS label via email. All i have to do now is pack up the meter and call UPS to collect it when I'm ready.

Will let you know how it goes.

[ober]

Does anyone have dimlow's great set of images saved?
Picasa is no more.