Setting up a Metrology lab

[opa627bm]

Hi all,
We have about 400+ 34470A, Keithley2460 and 2461 on site and recently we had bad experience with transcat.
After browsing through the website, I think 34470A was a bad choice and we will start moving 34470 to the cooperate loan program and switch to DMM7510.
We have talked to Fluke cal about the possible setup and they recommended 5730A calibrator + 10A Amplifier , their latest 8.5 digit DMM and metcal software bundle.
They will offer to write the script for us for the SMUs as they dont have it in their library.
Is there anything else we might need in order to have a running lab just for those equipment? 
Regards,
Li

[Conrad Hoffman]

Do you need any sort of certification? Maybe start here- https://a2la.org/accreditation/calibration/#:~:text=A2LA%20is%20the%20largest%20and,virtually%20every%20field%20of%20measurement.

Probably not going to be cheap! Very few people find in-house calibration practical these days.

[opa627bm]

Hi Sir,
We dont , the problem is our annual calibration budget is ~200K+ and the fluke calibration set is only 150K and SW + Cal is about 10K per year, so if we can get it going , we can easily justify the fluke setup.
(we had some equipment that have cal password changed and they never asked about it and somehow it has a cal sticker on it)
what we think is to have a calibration lab available to engineers so they can always use it to ensure their equipment is in spec.
Regards,
Li

[mendip_discovery]

Hi Sir,
We dont , the problem is our annual calibration budget is ~200K+ and the fluke calibration set is only 150K and SW + Cal is about 10K per year, so if we can get it going , we can easily justify the fluke setup.
(we had some equipment that have cal password changed and they never asked about it and somehow it has a cal sticker on it)
what we think is to have a calibration lab available to engineers so they can always use it to ensure their equipment is in spec.
Regards,
Li

Because calibration is not adjustment. They don't need the code if they don't have to adjust it.

Having an internal lab isn't too bad and many mech firms will cal their own kit internally to save money.

Some of the challenges you will face is around convincing customers etc that any conflicts of interest are handled correctly. For example you need to test meter but a manager goes over you as they don't want it taken out as it is needed all the time and as soon as you take it away they can't produce products, how do you mitigate that problem etc. What is to stop the scared cal engineer from just making up a cert and giving the shouty manager a label.

Record before and after adjustment readings. Adjust calibration frequencies accordingly.

Keep an externally calibrated meter in the met lab for checking the Fluke unit once a month as if that drifts out of spec then you could be certifying lots of stuff.

I would say get it up to ISO 9001 for starters then work towards 17025 even if you don't get accredited for it just use the sections as a guide.

[HighVoltage]

Is there anything else we might need in order to have a running lab just for those equipment? 

Moving from the 34470A to the DMM7510 seems to be a very good step forward.

Depending on what you use the instruments for, you might also want a 10k and 1 Ohm resistor standard for your lab. And to check your 5730, it would be good to have a 732C.

[jonpaul]

why not use the Keysight cal service ?

j

[Conrad Hoffman]

If you don't need certs, get a few standards and maybe a calibrator and have the engineers do a reality check periodically. Just send out the stuff that actually has a problem. You'll find modern equipment rarely goes out of spec and has to be adjusted.

[opa627bm]

why not use the Keysight cal service ?

j

To be honest, I do not like working with them, nor trusting them anymore, since we had few bad experience with their customer support team. (one of the VNA we ordered they promised to have the Impedance measurement software package as default but it wasn't and they just went radio silence)

Also, their build quality has shifted a lot from Agilent era, we had 30 unopend 34470A that was purchased back in 2019 and we opened them up this year then sent to transcat to cal, 1 have a broken front panel USB, 1 have a very noisy reading in the last 2 digits when calibration short was installed. (during incoming IQC, transcat passed them all) and we haven't gone through them all yet.

Their customer service is slow and not really helpful ( we are not a small company, part of FAANG)
Lastly, their animosity towards the home lab community made me decide to move away from them as much as possible. 

[Kosmic]

You should check page 525 of the 34470a service manual. It gives you an idea of what you would need for calibration. Other than the 5730A calibrator + 10A Amplifier, I think you need to find something for frequency and capacitance.

[RYcal]

Just gonna throw my 2 cents in here. I use Fluke products and their software in my accredited lab. It works well and I know their support team are world class people. But you have to know how to make work your way.

I have a nifty little MET/CAL procedure for comparing both the calibrator and the DMM against each other and throwing it into an excel template for analysis. I use this as means to track drift etc. in my lab.

One other thing you may be missing is training in all aspects of calibration including the software. I lost count of how many times I have been called upon to help out people running MET/TEAM and MET/CAL after the agents have taken your money and washed their hands of you. This is what lets Fluke down in my eyes. The people selling their products don't actually know how to use it themselves.

Let me know if you go down this track and need support outside of Fluke and their horrendous pricing hahaha.

Cheers

[Henrik_V]

Li,
to run your own calibration lab, I miss another refrerence(s) to do regular self validations, one is at least needed, two are better to know what is wrong . 

Most important: Talk to your assesor/ your acceditation organisation , even if they are not allowed to assit you, they are usually willing to give hints   A lot of documentation wil be needed.  But usually no 17025 acceditation (while you be close to it )

[mendip_discovery]

Something else that I have noticed when I visit customers who used to do calibration internally.

Get it in writing that the person/team who are doing the calibration will be given the time to do it. Too often I have seen it fall to bits because managers feel they can pass on other jobs to the calibration person/team until it becomes something they do when they have time in between other jobs. Worth noting in the risk assessment as well.


With the kit you will have the opportunity to pull some of the meters used on high risk tasks and do random spot checks on a few important ranges/points. Of course log this and put it on a spreadsheet.

[DavidKo]

From my experience (we are the small firm to calibrate electronic measurement devices by ourselves in contrast to mechanical measurement devices), when we sent the multimeter to external lab for calibration (hard to say what they really do) you do not see it for some period of time. In case that you have 400+ devices you must still have some devices on calibration, maybe you will not need so many of them in case that you can do calibration/adjustment in house (only the shipping will take at least 2 business days). You can have still some multimeters turned on in calibration lab (it is better for them; or just turn some on when you know that some multimeter is coming for calibration) and when the multimeter comes for calibration, you can immediately perform it on one from turned on devices and send it back with fresh calibration (another time saving, since the calibration will last longer compared to return from cal lab). Even if we speak about saving 1 week, than in the amount of devices you have, it is a lot of time. This is the way how you can get better support from top management, since they always hear on savings/speeding things up.

[mendip_discovery]

As DavidKo suggests you no doubt have a certain amount of spares. Make good use of these as you could pull a meter in the morning and take it in the lab, calibrate and take it back later that day. This is good for the meter as it wont have cooled much in the transport but no so good for stabilising in a temperature controlled environment. I would suggest you use a spare, swap it out, take the one to be calibrated, power it up and leave in the lab and calibrate the next day.

At work we receive the unit, book it in, quick visual check for damage. Then it goes into the temperarure controlled lab for at least 24hrs. On the day of calibration it is powered up for at least 30mins before calibration, depending on the type of unit. Its then calibrated and if it needs adjustment that is done. If repairs are needed we contact the customer etc. Then some sanity checks are done on the cert (another person looks at it for smelling pistakes). Then it gets sent back to the customer. We quote 5 working days partly as engineers can be onsite and workloads can get quite manic at times.

[Conrad Hoffman]

+1 on giving things time to stabilize. IMO, fast calibration = bad calibration.

[alm]

Indeed, spare meters could help with this. If a meter needs calibration, pull a recently calibrated spare and put it in its spot, while you can take your time calibrating the meter without holding up production.

To me it seems that the big challenges aren't so much in buying the box from Fluke, but more in training and staff/time.

How about the lab itself? Do you have a lab with good enough temperature and optionally humidity control?

[opa627bm]

Hi all,
thanks for all the suggestions.
We are going to have an on side domo next week with their calibrator.
also I have requested them to add the 732C and 10K reference standard in the quote.
the quote also includes a 3 day on site met.cal tutorial and gold support to access to their warranted scripts.
but I am surprised that they dont have many procedure for Keithley products, (given they owned by the same company)
Regards,
Li

[RYcal]

Hi all,
thanks for all the suggestions.
We are going to have an on side domo next week with their calibrator.
also I have requested them to add the 732C and 10K reference standard in the quote.
the quote also includes a 3 day on site met.cal tutorial and gold support to access to their warranted scripts.
but I am surprised that they dont have many procedure for Keithley products, (given they owned by the same company)
Regards,
Li

Make sure their procedures work before they leave your site...

[DavidKo]

To achieve Fluke performance specifications there is Fluke recommendation.

[NaxFM]

400+ 34470A? Wow
Mind sharing which sector requires this many high accuracy instruments?

[opa627bm]

From my experience (we are the small firm to calibrate electronic measurement devices by ourselves in contrast to mechanical measurement devices), when we sent the multimeter to external lab for calibration (hard to say what they really do) you do not see it for some period of time. In case that you have 400+ devices you must still have some devices on calibration, maybe you will not need so many of them in case that you can do calibration/adjustment in house (only the shipping will take at least 2 business days). You can have still some multimeters turned on in calibration lab (it is better for them; or just turn some on when you know that some multimeter is coming for calibration) and when the multimeter comes for calibration, you can immediately perform it on one from turned on devices and send it back with fresh calibration (another time saving, since the calibration will last longer compared to return from cal lab). Even if we speak about saving 1 week, than in the amount of devices you have, it is a lot of time. This is the way how you can get better support from top management, since they always hear on savings/speeding things up.

Our problem is, all our same type equipments are calibrated at the same time, so when the cal crew arrives on site we have to pull them off from all the running systems and slowly rotate them one by one. Thats one of the reason why we want to setup an onsite cal lab and we can slowly cal them through out the year and keep them in rotation

[DavidKo]

...

Our problem is, all our same type equipments are calibrated at the same time, so when the cal crew arrives on site we have to pull them off from all the running systems and slowly rotate them one by one. Thats one of the reason why we want to setup an onsite cal lab and we can slowly cal them through out the year and keep them in rotation

OK, I have thought that you are shipping devices out. This is different case. Slowly still means >1 piece per day .

What confused me a lot is, that the Fluke had not recommended the calibrator including necessary setup to ensure its parameters - the devices needed are clearly written in their specifications. On the other hand they had offered 8.5 digit multimeter. From my point of view I see the offer quite odd.

[alm]

also I have requested them to add the 732C and 10K reference standard in the quote.

Note that to calibrate the 5730A you also need a 1 Ohm reference standard (e.g. Fluke 742A-1) in addition to 10 kOhm and 10 V.

but I am surprised that they dont have many procedure for Keithley products, (given they owned by the same company)

Indeed. Assuming the multimeter they supply has a sufficiently high test uncertainty ratio to the 2460, the procedure doesn't look very complicated. No complicated setups to generate nanovolts, picoamps or giga/tera ohms.

To achieve Fluke performance specifications there is Fluke recommendation.

This is to calibrate the 5730A. The 90 day / 180 day / 1 year specification from calibration are still valid without them, and if you ship it out for calibration with that interval, it should maintain its specifications. Of course the set of the 732C / 742A-10k / 742A-1 is easier to ship and won't cause any downtime for your calibration, but you can still run the 5730A the same way as a non-artifact calibrated device and let Fluke worry about calibration.

What confused me a lot is, that the Fluke had not recommended the calibrator including necessary setup to ensure its parameters - the devices needed are clearly written in their specifications. On the other hand they had offered 8.5 digit multimeter. From my point of view I see the offer quite odd.

Where does it say that this equipment is necessary to ensure its parameters? It's only necessary to calibrate, but you don't have to calibrate it yourself.

[mendip_discovery]

For your purposes you dont need all the other references. Yes it's nice to have but that is only if you wish to calibrate the calibrator and ref dmm every 90 days to reduce your Uncertainties. That extra kit will set you back a pretty penny and just give a reason to management to just carry on like normal.

Keep note of it and maybe put it in the improvements part of your future plans.

[HighVoltage]

You got that instrument fast. Looks nice.
What else did you get?
More pictures please.