EEVblog Electronics Community Forum
Products => Test Equipment => Topic started by: RoboTechEd on April 20, 2015, 06:05:55 pm
-
Hi All,
I got a DMMCheck Plus in the mail today and was wondering what calibration practices the forum recommends (how-to, frequency of calibration, labeling, etc)? I listed the items that I wish to calibrate and their status below. :-/O
EXtech 430 DMM - Cant find manual, assume there is an internal pot to tweak the screen's reading?
Fluke 8012A DMM - I believe there are internal pots for this
Xantex 20-3 PSU - Internal precision resistor needed to replace to update screen's reading. Whats the best way to do this? Replace with a 10 turn pot for precision? Get a small precision pot? Replace with a simple film resistor that is measured at the right value, order a bunch of new precision resistors?
HP 54602B DSO - May not be easy to calibrate using the DMM check. Just want to verify measurements etc. Is this something I should calibrate?
-
My recommendation is to leave them alone.
The DMM check is not a calibration standard, it is used to verify that the meter is working properly.
Randy
-
The DMM check is not a calibration standard, it is used to verify that the meter is working properly.
That's fair, and it will be useful for verifying some of the DMMs I have lying around. It may not be a standard, but I figured it would be appropriate for a low level hobbyist. I'm not going to send stuff out to get calibrated, but would like to make sure everything stays within spec. What would be the best way to accomplish that? I figured comparing it to a well regulated and verified source would be best?
-
I'm afraid that the DMM check is just a gateway drug that will lead you on a long and sometimes painful journey.
Next thing you know, you will be monitoring the Voltnuts group and bidding on lots of standards and calibrators on eBay.
Proceed with caution my friend.
Randy
-
Calibration is not an old man with a long grey beard, who strokes it while twiddling pots.
https://www.youtube.com/watch?v=bN1y1my4gEE&index=2&list=PLvOlSehNtuHstfCI8cQYweKVjgVw0mZBy (https://www.youtube.com/watch?v=bN1y1my4gEE&index=2&list=PLvOlSehNtuHstfCI8cQYweKVjgVw0mZBy)
-
Not sure how I missed Dave's video when searching :palm: . But I haven't tweaked yet. I did verify all of my devices, and my results can be found at the link below. I took all of the equipment and verified their specs and measurements. Now to decide if I want to tweak anything at all. Thanks for the link!
https://www.eevblog.com/forum/testgear/lab-calibration/msg656771/#msg656771 (https://www.eevblog.com/forum/testgear/lab-calibration/msg656771/#msg656771)
-
Those results are pretty sweet, I wouldn't bother with adjusting anything already within spec, especially if it involves pot tweaking.
Keep in mind, you have to calibrate each range separately (200mV, 2V, 20V etc), so while you have a good 5V DC reference, you can't be sure the other DC voltage ranges are within spec. Also, remember that you generally want your reference voltage to be close to the full scale range for best accuracy (check manuals for details), 5V is okay when testing a 6.6V range, but it's a little low for the 20V range on your Fluke 8012, as it wants 19.00V ideally for calibration.
So like K5HJ said, the DMMcheck isn't reaaaally suitable for proper calibration of your gear, but for hobbyist use......ehhhh she'll be right. ;D
For my home lab, I send out the best multimeter I have out for calibration and use that as a reference for all my other multimeters. I usually wont adjust anything to be right on the money because just knowing that something is within spec is good enough for me. I can just fire up the reference multimeter whenever I need the greater accuracy.
I don't think I've ever needed anything better than a bog standard 6600 count 0.5% accuracy handheld meter, but it sure is fun seeing a 10.0000001V reading on a DMM even if only for a brief moment.....you've been warned. ;)
-
Those results are pretty sweet, I wouldn't bother with adjusting anything already within spec, especially if it involves pot tweaking.
Keep in mind, you have to calibrate each range separately (200mV, 2V, 20V etc), so while you have a good 5V DC reference, you can't be sure the other DC voltage ranges are within spec. Also, remember that you generally want your reference voltage to be close to the full scale range for best accuracy (check manuals for details), 5V is okay when testing a 6.6V range, but it's a little low for the 20V range on your Fluke 8012, as it wants 19.00V ideally for calibration.
So like K5HJ said, the DMMcheck isn't reaaaally suitable for proper calibration of your gear, but for hobbyist use......ehhhh she'll be right. ;D
For my home lab, I send out the best multimeter I have out for calibration and use that as a reference for all my other multimeters. I usually wont adjust anything to be right on the money because just knowing that something is within spec is good enough for me. I can just fire up the reference multimeter whenever I need the greater accuracy.
I don't think I've ever needed anything better than a bog standard 6600 count 0.5% accuracy handheld meter, but it sure is fun seeing a 10.0000001V reading on a DMM even if only for a brief moment.....you've been warned. ;)
Yeah I have been doing more and more research. Dave's video above was very informative. I think I will build up over time a drift record for each device before trying to actually tweak anything (if I decide to). I just thought it would be a cool hobby to learn calibration practices and to get really precise instruments :)
And good information on the different ranges. That makes sense. I would have figured that would be the case for each with manual adjustments, but is it for auto adjusting meters too? I would think so because they still have ranges, it just automatically chooses for you.
I dont think I have a DMM yet thats worth sending out, but maybe if I upgrade to an 87V or something I will go through with the process. How much does that typically cost? At work it seems fairly expensive, but maybe thats just due to the nature of the work.
-
Volt nutting is pretty fun! ^-^
Yep, you still have to test each range for autoranging meters, you're spot on.
It's hard to say how much it'll cost, it varies with the multimeter, what lab you go to, and to what standard you need it calibrated to. Something like a Fluke 87V(4.5 digit) will cost <$100 to calibrate as it's quite low resolution.
I recommend getting a cheapish used bench meter like the HP 3478A (5.5 digit), Dave has a couple videos on it. Calibration on a 5.5 digit meter is still very affordable, I think it should only cost a hundred bucks at most.
Calibrating a 6.5 digit multimeter starts to get a bit pricey, I've been quoted $300 for a HP 34401A at Hewlett-Agisight, but then again, everything is a bit expensive here in Australia.
Going further than that, A HP 3458A (8.5 digit) will cost ~$700 for a standard calibration, and about ~$2000 for calibration against a Josephson Junction (The Holy Grail DC voltage reference =P). Yeah...it get's expensive real fast.
For hobbyists, I don't think it's necessary to send out a meter every year for calibration. Once you have a good calibrated bench meter, you can just build a DIY voltage reference to track it's drift, and there are heaps of volt nuts out there who would be happy to measure something you can use as a transfer standard with their calibrated gear. You'd be surprised how little quality gear actually drifts, especially when only used intermittently like a hobbyist would do.
I have a calibrated Advantest 8.5 digit meter, good to about 5ppm per year on 10VDC, I'm happy to measure something for you but I am quiiite faaar away down under. Offer is open to anyone local, shipping is just too crazy to/from Australia. ;)
-
Volt nutting is pretty fun! ^-^
Yep, you still have to test each range for autoranging meters, you're spot on.
It's hard to say how much it'll cost, it varies with the multimeter, what lab you go to, and to what standard you need it calibrated to. Something like a Fluke 87V(4.5 digit) will cost <$100 to calibrate as it's quite low resolution.
I recommend getting a cheapish used bench meter like the HP 3478A (5.5 digit), Dave has a couple videos on it. Calibration on a 5.5 digit meter is still very affordable, I think it should only cost a hundred bucks at most.
Calibrating a 6.5 digit multimeter starts to get a bit pricey, I've been quoted $300 for a HP 34401A at Hewlett-Agisight, but then again, everything is a bit expensive here in Australia.
Going further than that, A HP 3458A (8.5 digit) will cost ~$700 for a standard calibration, and about ~$2000 for calibration against a Josephson Junction (The Holy Grail DC voltage reference =P). Yeah...it get's expensive real fast.
For hobbyists, I don't think it's necessary to send out a meter every year for calibration. Once you have a good calibrated bench meter, you can just build a DIY voltage reference to track it's drift, and there are heaps of volt nuts out there who would be happy to measure something you can use as a transfer standard with their calibrated gear. You'd be surprised how little quality gear actually drifts, especially when only used intermittently like a hobbyist would do.
I have a calibrated Advantest 8.5 digit meter, good to about 5ppm per year on 10VDC, I'm happy to measure something for you but I am quiiite faaar away down under. Offer is open to anyone local, shipping is just too crazy to/from Australia. ;)
Yeah it looks like something I would really like to learn more about! A good rough cost is fine. When I get something a bit more precise I will get it professionally calibrated. No sense in doing that for an extech430 or fluke8012A yet though. The HP Bench meter you recommended is next on my list!
I was curious though, in the video dave pushed above it states that calibration just (normally) verifies accuracy and doesn't actually tweak anything to be spot on. Wouldn't the rate of error transfer over if I was to make a voltage reference based off of this? Wouldn't this rate of error compound the more times it was transferred as well? Maybe once I get a nicer DMM I will start to get more serious and make a voltage standard.
Do you have any sources that would be useful to learn more about this? I would love to learn how these high-end meters work, and what it takes to create accurate voltage references on the cheap at a hobby-level
-
Just to lay down a reference point, here is the pricing I pay for my calibration:
Fluke 45 (4.5 digit dmm) $48
Tek TDS5032B $85
LeCroy WaveSurfer 3054 $85
Fluke 8846A $110
If I cal my DS815 (ahhahahaha) $245
You're basically paying for the lab tech's time. $100/hr is a pretty typical number (though some fancier shops can cost way more) and a 1hr min, so something basic like a Fluke 115 might take 15 minutes. If you have 4 of them, it's $25/meter to cal. On a meter that costs $100, you're paying 25% (to keep math pretty) of the cost for lab-grade certainty on something that really won't bring back any value.
-
I was curious though, in the video dave pushed above it states that calibration just (normally) verifies accuracy and doesn't actually tweak anything to be spot on. Wouldn't the rate of error transfer over if I was to make a voltage reference based off of this? Wouldn't this rate of error compound the more times it was transferred as well? Maybe once I get a nicer DMM I will start to get more serious and make a voltage standard.
Do you have any sources that would be useful to learn more about this? I would love to learn how these high-end meters work, and what it takes to create accurate voltage references on the cheap at a hobby-level
Yup, while you may not be able to measure the absolute value of your DIY voltage standard to better accuracy than your best multimeter, it's actually very possible to make a voltage standard that is far more stable over time, and that's still useful for checking drift.
In modern DMMs, you don't need to tweak any pots to perform adjustment, it's all done in software. All you have to do is input and enter a known exact value for each range, that's it!
In general, 'calibration' verifies accuracy by measuring against a known standard, and 'adjustment' is the procedure performed to tweak the device to read bang on against a reference. Calibration doesn't imply adjustment, which is pretty confusing for a newbie as it feels like the terms are used interchangeably on the forums and such. The cost of calibration doesn't always include adjustment, so be careful.
If you want to learn more about errors and such, look up measurement uncertainty, Agikeysight had a few videos about it IIRC.
For voltage references, there's tons of info on the forums and datasheets/application notes for voltage references from Linear Tech, good stuff. I think the free released chapter from The Art of Electronics 3rd edition book had a few beginner friendly sections about voltage references, check it out, you'll find it somewhere on the internet.
-
you can just build a DIY voltage reference to track it's drift
That is okay, but what about the drift in that DIY voltage reference, who's going to keep track of that? :phew:
-
you can just build a DIY voltage reference to track it's drift
That is okay, but what about the drift in that DIY voltage reference, who's going to keep track of that? :phew:
That's why I got the DMM Check! I can get it re-cal'd yearly pretty cheap though
Yup, while you may not be able to measure the absolute value of your DIY voltage standard to better accuracy than your best multimeter, it's actually very possible to make a voltage standard that is far more stable over time, and that's still useful for checking drift.
In modern DMMs, you don't need to tweak any pots to perform adjustment, it's all done in software. All you have to do is input and enter a known exact value for each range, that's it!
In general, 'calibration' verifies accuracy by measuring against a known standard, and 'adjustment' is the procedure performed to tweak the device to read bang on against a reference. Calibration doesn't imply adjustment, which is pretty confusing for a newbie as it feels like the terms are used interchangeably on the forums and such. The cost of calibration doesn't always include adjustment, so be careful.
If you want to learn more about errors and such, look up measurement uncertainty, Agikeysight had a few videos about it IIRC.
For voltage references, there's tons of info on the forums and datasheets/application notes for voltage references from Linear Tech, good stuff. I think the free released chapter from The Art of Electronics 3rd edition book had a few beginner friendly sections about voltage references, check it out, you'll find it somewhere on the internet.
the guy who makes them. Its not 100% kosher calibration, but I think for hobby purposes it meets the needs rather well.
Good to know! I know that the DMMCheck was on constantly for ~300 HRS before I bought it to 'burn in the settings' and keep make sure it doesn't drift initially much.
Is this software used to tweak DMMs given for free? I believe I still have pots in my Extech430, but if I get a new benchtop DMM should I be looking for one thats easy to configure? Is it worth trying to learn proper adjustment myself? I would love my devices to be as spot-on as possible (although I get that its not necessary at all).
And thanks for the starting points! I'll have to start my Google-fu after work tonight and get studying.
Just to lay down a reference point, here is the pricing I pay for my calibration:
---
You're basically paying for the lab tech's time. $100/hr is a pretty typical number (though some fancier shops can cost way more) and a 1hr min, so something basic like a Fluke 115 might take 15 minutes. If you have 4 of them, it's $25/meter to cal. On a meter that costs $100, you're paying 25% (to keep math pretty) of the cost for lab-grade certainty on something that really won't bring back any value.
Thanks for the info! This put it in perspective a bunch for me. Thats a great explanation. I definitely wont bother getting any of my current equipment calibrated, but if I start moving up I will be sure to consider it. If you want adjustments done to them to make them more aligned to the references, I assume that would be the same price/hr and just add some more time (which would depend on the item)?
-
I have in the past wondered exactly this topic, or more specifically the NEED for absolute measurement for general hobbyist and repair work.
IMO if the need arises the DMM Check is perfect for ones 'piece of mind'.
Volt nuts excuded. ;)
A customer when sourcing some Flukes through me also got a DMM Check as a periodic check.
Interesting after 1 year all 15 Flukes remained well within factory spec but he :-/O 2 units to better match the others readings.
https://www.eevblog.com/forum/testgear/my-first-'expensive'-dmm-thoughts-fluke-17b/msg613061/#msg613061 (https://www.eevblog.com/forum/testgear/my-first-'expensive'-dmm-thoughts-fluke-17b/msg613061/#msg613061)
Really what is the absolute accuracy level required for TE for other than traceable/professional use?
Personally to ~5 mV does all I require.
-
I have in the past wondered exactly this topic, or more specifically the NEED for absolute measurement for general hobbyist and repair work.
IMO if the need arises the DMM Check is perfect for ones 'piece of mind'.
Volt nuts excuded. ;)
A customer when sourcing some Flukes through me also got a DMM Check as a periodic check.
Interesting after 1 year all 15 Flukes remained well within factory spec but he :-/O 2 units to better match the others readings.
https://www.eevblog.com/forum/testgear/my-first-'expensive'-dmm-thoughts-fluke-17b/msg613061/#msg613061 (https://www.eevblog.com/forum/testgear/my-first-'expensive'-dmm-thoughts-fluke-17b/msg613061/#msg613061)
Really what is the absolute accuracy level required for TE for other than traceable/professional use?
Personally to ~5 mV does all I require.
Maybe one day I'll be able to be a Volt Nut, but right now a DMMCheck has to be my option. And yeah, 5mV does all I require too. I mainly do digital work so 100mV would probably be fine even. I dont know though...There's just something special about a piece of test equipment that is capable of measuring spot on...Even though I dont need it
-
I mainly do digital work so 100mV would probably be fine even. I dont know though...There's just something special about a piece of test equipment that is capable of measuring spot on...Even though I dont need it
Carefull, these are the first symptoms of the Volt nut bug.
Get innoculated against it as soon as you can, a 2 kV shock should do the trick. :-DD
-
Yup, I think the 'bug' is starting to bite him. Can be expensively bad!
-
Yup, I think the 'bug' is starting to bite him. Can be expensively bad!
Luckily it is recent (within a week, since I researched the DMMCheck), and as a recent-grad I don't have the funds to get too into it as a hobby....yet. Hopefully the excitement will fade as my reasoning takes over and I eventually convince myself that it isn't necessary/practical
Carefull, these are the first symptoms of the Volt nut bug.
Get innoculated against it as soon as you can, a 2 kV shock should do the trick. :-DD
And you mean 2.0000kV right? I wont settle for 2.01kV.
-
Yup, I think the 'bug' is starting to bite him. Can be expensively bad!
I had a small infection of this bug recently and got a 5 1/2 digit bench DMM a few months back. :-DMM
Sure, my buy price was pretty good and hopefully I've been cured for the time being.
https://www.eevblog.com/forum/testgear/siglent-new-bench-dmm-sdm3055/ (https://www.eevblog.com/forum/testgear/siglent-new-bench-dmm-sdm3055/)
Hopefully this small infection has gained me resistance to this disease for the rest of my days....
And you mean 2.0000kV right? I wont settle for 2.01kV.
Any signs of relapse we can up it to CRO PDA levels of ~10-15kV. :o
-
I'm afraid that the DMM check is just a gateway drug that will lead you on a long and sometimes painful journey.
Next thing you know, you will be monitoring the Voltnuts group and bidding on lots of standards and calibrators on eBay.
Proceed with caution my friend.
Randy
How Very True
-
Its a sickness. Just got a 7 1/2 Keithley. AC measurements are off by about 12% but DC and resistance, etc are all spot on.
You know its bad when you start researching calibration gear for your home lab.
-
Its a sickness. Just got a 7 1/2 Keithley. AC measurements are off by about 12% but DC and resistance, etc are all spot on.
You know its bad when you start researching calibration gear for your home lab.
Yes, I have the same fever as well. Keysight and I currently maintain a seasoned & pristine 3458A along with a couple of other standards that cost me as much to keep in annual cal as my house insurance every year. I try to recap on some of my expenses but it can be quite the challenge.
-
Thanks for the info! This put it in perspective a bunch for me. Thats a great explanation. I definitely wont bother getting any of my current equipment calibrated, but if I start moving up I will be sure to consider it. If you want adjustments done to them to make them more aligned to the references, I assume that would be the same price/hr and just add some more time (which would depend on the item)?
You're paying for someone's time, but none the less the point of calibration is not to "adjust" it back into spec.
So let's lay out the point: You're a small business who makes a particular kind of widget. One pass/fail check measures the reference voltage that feeds an ADC on your widget. You've produced 10,000 units in the last two years. In the last 3 months, you've started to have field failures. They're not consistent in serial number, rather it's pretty spotty. You re-run them through your manufacturing test line. Sure enough, they still pass. How do you know your equipment isn't lying to you about what it measures?
The lack of confidence is called "Certainty."
Since you haven't spent the money for calibration in the last two years, you decide to submit the whole line for cal. Of three test stations you have for the line, one of the meters fails its calibration. This does not mean that an old man with a long grey beard has "fixed" it for you. Rather, they have reported that your 0.0025% ± 2 counts meter has drifted well beyond spec. Some time in the last two years, this meter has gone south. Now, the question is "when did it go south?" - if it went south shortly after the last cal, then you're likely to have a HUGE number of field returns. If it went south recently, then you might be seeing a different number of returns. You have two years of no certainty, for this meter. However, the other two meters are all within cal. This does not mean that they're the same ±0.0025% ± 2 counts as the day they came out of the box. Now that they're 5 years old, they have drifted to ±0.0025% ±2 counts ± 10µV/year.
This is now where the cal history becomes useful: you have a record of how the units have aged. You have more certainty about the measurements taken with this unit, than a brand new unit. Every unit which passed under these two meters, will have a high level of certainty that they did not ship with a defect. The other meter, well that's the current problem of the day. This bad meter, now that it has been put into calibration, is deemed to be out of spec. Now, you may send it off to be adjusted back into spec if it isn't broken, but once adjusted, you would send it back into calibration to start a history on it.
-
If this is for home/hobby use then employing a professional calibration service for many types of test gear is just a waste of time and money.
I'd argue that it adds little value. However, the downside risks are quite significant. eg your gear could get damaged or lost and you run the risk of having a spotty technician 'learning his trade' inside your test gear if it needs internal adjustment or repair.
The chances are, they will take away your gear in a van, make sure it complies within a certain tolerance and then send it back to you unadjusted after it's been bouncing halfway around the country in the back of a van for a few days.
There's no guarantee your gear is still 'calibrated' the moment it is unpacked at your house. It's mostly an illusion.
-
If this is for home/hobby use then employing a professional calibration service for many types of test gear is just a waste of time and money.
I'd argue that it adds little value. However, the downside risks are quite significant. eg your gear could get damaged or lost and you run the risk of having a spotty technician 'learning his trade' inside your test gear if it needs internal adjustment or repair.
The chances are, they will take away your gear in a van, make sure it complies within a certain tolerance and then send it back to you unadjusted after it's been bouncing halfway around the country in the back of a van for a few days.
There's no guarantee your gear is still 'calibrated' the moment it is unpacked at your house. It's mostly an illusion.
Do any commercial calibration labs (e.g. Keysight) offer a service where they actually adjust the instrument (vs. just check and adjust only if necessary)?
I assume the 3458A $700 calibration actually includes adjustment?
-
If this is for home/hobby use then employing a professional calibration service for many types of test gear is just a waste of time and money.
I'd argue that it adds little value. However, the downside risks are quite significant. eg your gear could get damaged or lost and you run the risk of having a spotty technician 'learning his trade' inside your test gear if it needs internal adjustment or repair.
The chances are, they will take away your gear in a van, make sure it complies within a certain tolerance and then send it back to you unadjusted after it's been bouncing halfway around the country in the back of a van for a few days.
There's no guarantee your gear is still 'calibrated' the moment it is unpacked at your house. It's mostly an illusion.
Do any commercial calibration labs (e.g. Keysight) offer a service where they actually adjust the instrument (vs. just check and adjust only if necessary)?
I assume the 3458A $700 calibration actually includes adjustment?
And IMO it is just as you say "assumption".
Why would one even consider disturbing any adjustment that was "within spec".
Where a wiper sits on a track is best left alone unless adjustment IS needed.
The old skeletal presets were buggers if disturbed, put an old one on a bench and observe how noisey they can be.
-
Yup, I think the 'bug' is starting to bite him. Can be expensively bad!
I had a small infection of this bug recently and got a 5 1/2 digit bench DMM a few months back. :-DMM
Sure, my buy price was pretty good and hopefully I've been cured for the time being.
That's what I thought. First I acquired a 5.5 digit bench meter, a Keithely 199, and all was good. Then I got another on ebay for a steal. The problem is that while the two agreed to within their quite good specifications (0.007%), they didn't agree exactly. Indeed, there was a good 20 counts difference. Ugh. A man with one precision meter knows the voltage; a man with two does not!
So I toyed with the idea of getting a Fluke 5440B calibrator which I could have had for under $1k ... but instead got a Keithley 2001 7.5 digit meter. I used this as a reference against which to calibrate my other 5.5 and 4.5 digit meters. Then... I stumbled across a deal and acquired three more 2001's! You might think that the viscous cycle would continue, with endless small disagreements between meters that would bug me, but actually no. Those four 2001's of different vintage and origin agree within +/-2 ppm on all DC voltage ranges! Most of the disagreement is drift/noise (one is higher for a while, then the other one...). I am very happy to know that my meters are accurate with a high enough statistical probability that I don't care any longer. It wasn't cheap getting here, but I spent much less than a single used 3458A would have cost.
p.s. If anyone in the Ottawa area wants to verify their meter(s) against mine, I'd be happy to oblige. I don't have any traceable calibration certificates, just a few closely agreeing meters.
-
And suddenly I'm glad I have one 4.5 digit meter, and another "6.5 digit" meter. (I use quotes, because only in a rare occasion does it give me more than 5.5 digits.) :-DD
-
Its a sickness. Just got a 7 1/2 Keithley. AC measurements are off by about 12% but DC and resistance, etc are all spot on.
You know its bad when you start researching calibration gear for your home lab.
-Is it possible that it is because it is not an RMS meter? I would expect something similar with imperfect waveforms (square etc)
p.s. If anyone in the Ottawa area wants to verify their meter(s) against mine, I'd be happy to oblige. I don't have any traceable calibration certificates, just a few closely agreeing meters.
We should get a local calibration post going ongoing. Would help noobies get a cheaper alternative with at least some level of certainty! I dont like the idea of trusting the specs out of the box
And IMO it is just as you say "assumption".
Why would one even consider disturbing any adjustment that was "within spec".
Where a wiper sits on a track is best left alone unless adjustment IS needed.
The old skeletal presets were buggers if disturbed, put an old one on a bench and observe how noisy they can be.
I understand the concept of 'if it's not broke don't fix it' but is there really a disadvantage of further tweaking budget multimeters? Would tweaking them to learn the concepts cause other issues? And what if you want a different spec per-se? I know this would be a bad decision, but for the sake of an argument, what if you have a .1% meter and you want it to be .05% Is there a harm with tweaking it to be within that spec? Although we all know that in most cases we wont need anything much more accurate practically, it would be nice to have a super-accurate meter.
And also, @XFDDesign, thank you for the comprehensive write-up!
-
I understand the concept of 'if it's not broke don't fix it' but is there really a disadvantage of further tweaking budget multimeters? Would tweaking them to learn the concepts cause other issues? And what if you want a different spec per-se? I know this would be a bad decision, but for the sake of an argument, what if you have a .1% meter and you want it to be .05% Is there a harm with tweaking it to be within that spec? Although we all know that in most cases we wont need anything much more accurate practically, it would be nice to have a super-accurate meter.
Robo,
Just remember you might have a 'very accurate' device and you will still have errors or uncertainties that can be created by a multitude of circumstances .
- there is no such thing as a perfect or 'accurate' measurement.
with this:
Accuracy and Precision are two terms that can be misunderstood and widely misapplied.
What you strive for is Precision which is simply the repeatability in your measurement data. Another way to examine "Precision" is to view it as a measurement of variation.
-
Robo,
Just remember you might have a 'very accurate' device and you will still have errors or uncertainties that can be created by a multitude of circumstances .
- there is no such thing as a perfect or 'accurate' measurement.
with this:
Accuracy and Precision are two terms that can be misunderstood and widely misapplied.
What you strive for is Precision which is simply the repeatability in your measurement data. Another way to examine "Precision" is to view it as a measurement of variation.
In terms of multimeters, isn't precision the number of possible values that can be displayed (ie, 6.5 count display), while accuracy is the percentage off the display can be (ie .1% +- 2units). Wouldn't I want the accuracy to increase (so that it is .05% for example), rather than have it more precise? I must have gotten something mixed up somewhere if I have these terms backwards. :-// .
And yes, I want the repeatability of the measurement to be consistent, so I guess I do want precision. I would assume there is no way to tweak that though?
In an Extech 430 teardown posted here today, you can see ~7 trimpots meant for adjustments to the meter. Just for clarity, what would those adjust, precision, or accuracy?
-
...
We should get a local calibration post going ongoing. Would help noobies get a cheaper alternative with at least some level of certainty! I dont like the idea of trusting the specs out of the box
...
Parrrrrrrrtayyyyyyyy!
https://www.eevblog.com/forum/chat/eevblog-%27calibration-party%27/ (https://www.eevblog.com/forum/chat/eevblog-%27calibration-party%27/)
-
In terms of multimeters, isn't precision the number of possible values that can be displayed (ie, 6.5 count display), while accuracy is the percentage off the display can be (ie .1% +- 2units). Wouldn't I want the accuracy to increase (so that it is .05% for example), rather than have it more precise? I must have gotten something mixed up somewhere if I have these terms backwards. :-// .
https://www.youtube.com/watch?v=U4JFeU-o2kc&list=PL3C5D963B695411B6&index=8 (https://www.youtube.com/watch?v=U4JFeU-o2kc&list=PL3C5D963B695411B6&index=8)
-
:-+ +1
A good site to help on those terms. Enjoy!
http://www.interactagram.com/physics/PrecisionAndAccuracy/ (http://www.interactagram.com/physics/PrecisionAndAccuracy/)
-
:palm: Doh! I was assuming accuracy to be constant! I figured a measurement that is +6% off will always remain 6% off and not fluctuate. If the spec is +-10%, it could be anywhere between -10 and 10 at any given time. Precision is what measures this consistency. I guess what I need to do is build up a log to get a good idea of the precision. If I find it to be relatively constant, I THEN can tweak it to have a more strict accuracy. Until then though, I could be just making it worse. I think everything clicks now.
I guess my question should be, what should the time increment between measurements be? I know stuff is often tested yearly, but how quickly do these things change? Can it vary much per week, day, hour? Is there any advantage of me testing it daily for a month and building up some data trends for the precision?
-
:palm: Doh! I was assuming accuracy to be constant! I figured a measurement that is +6% off will always remain 6% off and not fluctuate. If the spec is +-10%, it could be anywhere between -10 and 10 at any given time. Precision is what measures this consistency. I guess what I need to do is build up a log to get a good idea of the precision. If I find it to be relatively constant, I THEN can tweak it to have a more strict accuracy. Until then though, I could be just making it worse. I think everything clicks now.
I guess my question should be, what should the time increment between measurements be? I know stuff is often tested yearly, but how quickly do these things change? Can it vary much per week, day, hour? Is there any advantage of me testing it daily for a month and building up some data trends for the precision?
:palm:
Oh dear.....
Shock therapy required.
Latest addition to the Volt Nut Club: RoboTechEd
-
That could be quite a loaded question, there are so many variables to consider. It all depends on the required conformity or adherence to a dictated set if guide lines. Someone correct me if I am using the incorrect terms, but I believe they refer to these as "Standards" Trust me, as a hobbyist you won't ever see them or need them. Good to follow if you can though to become better prepared down the road.
For example, my HP 3458A is calibrated (not meaning they tweak anything inside - only if needed) based on a gold-transfer guard banded standard in Loveland Colorado. Some might ask, why such a stringent calibration and why every year? If you have a track record of your instrument you can follow the drift curves on your calibration and reduce the levels of uncertainty or increase your chances of making more accurate and precise measurements. Adjusting for errors in your test setup, your environment and all the associated parts involved.
I hope this helps
Best of luck,
Jon
-
Oh dear.....
Shock therapy required.
Latest addition to the Volt Nut Club: RoboTechEd
I better get in that shock line then. :-\
I've spent the last two days, off and on, trying to get this HP 5315 counter to the last 0.5 Hz after warm up. I actually had to get a magnifying glass to get it right, because it required so little of the turn on the trimmer that I couldn't tell if it was moving or not by feel. :palm:
-
:palm: Doh! I was assuming accuracy to be constant! I figured a measurement that is +6% off will always remain 6% off and not fluctuate. If the spec is +-10%, it could be anywhere between -10 and 10 at any given time. Precision is what measures this consistency. I guess what I need to do is build up a log to get a good idea of the precision. If I find it to be relatively constant, I THEN can tweak it to have a more strict accuracy. Until then though, I could be just making it worse. I think everything clicks now.
I guess my question should be, what should the time increment between measurements be? I know stuff is often tested yearly, but how quickly do these things change? Can it vary much per week, day, hour? Is there any advantage of me testing it daily for a month and building up some data trends for the precision?
:palm:
Oh dear.....
Shock therapy required.
Latest addition to the Volt Nut Club: RoboTechEd
LOL tautech, I guess Robo better break out that check book.. :-DD he has been bitten.. grins
-
Oh dear.....
Shock therapy required.
Latest addition to the Volt Nut Club: RoboTechEd
I better get in that shock line then. :-\
I've spent the last two days, off and on, trying to get this HP 5315 counter to the last 0.5 Hz after warm up. I actually had to get a magnifying glass to get it right, because it required so little of the turn on the trimmer that I couldn't tell if it was moving or not by feel. :palm:
....and a puff of air rushed by.. oh crap.. it just moved 1.5 Hz.. :-DD sucks when that happens
and the "Time nuttery begins"
-
....and a puff of air rushed by.. oh crap.. it just moved 1.5 Hz.. :-DD sucks when that happens
and the "Time nuttery begins"
LOL - yep. When I took the top cover off to make a tweak it took about 15 m to see if it made a difference. :o
-
:o :-DMM :-/O :-BROKE :-/O :-+ :phew: ....... :wtf: :-/O :-BROKE :palm: x 10
:bullshit: :bullshit: :bullshit:
-
Oh dear.....
Shock therapy required.
Latest addition to the Volt Nut Club: RoboTechEd
Told you so.
Hook, line, and sinker.
Randy
-
I guess my question should be, what should the time increment between measurements be? I know stuff is often tested yearly, but how quickly do these things change? Can it vary much per week, day, hour? Is there any advantage of me testing it daily for a month and building up some data trends for the precision?
Ahh, now welcome to the world of precision, where the answer is a confusing, "Yes."
There are two main components: short-term drift, and long-term drift.
Short-term drift is all the stuff that is sensitive to your environment. Give this precision amp a look and jump to page 3.
http://www.analog.com/media/en/technical-documentation/data-sheets/ADA4805-1_4805-2.pdf (http://www.analog.com/media/en/technical-documentation/data-sheets/ADA4805-1_4805-2.pdf)
"Offset voltage drift: 0.2uV/ºC relative to 25ºC" -- so at 21C, you have an uncertainty of 0.2uV/C * 4C or 0.8uV of uncertainty. There are questions about whether humidity and other things also affect the short term too. Look at the old classic LM741 http://www.ti.com/lit/ds/symlink/lm741.pdf (http://www.ti.com/lit/ds/symlink/lm741.pdf) - they don't even specify a maximum drift, but only a typical of 15uV/C. For that same 4C it's walked 60uV. All of the short-term stuff will affect your measurement with it's haunting of altered variables.
Then there is long-term drift. This is the quantity that represents how the part ages. Suppose today you measured the offset voltage of your amplifier to be 0.1uV. What will it be after 5 years of active use? LTD is an emerging area of study within a lot of the major semiconductor manufacturers. It's why the LTZ1000 is so popular as they characterize how it ages over time. http://cds.linear.com/docs/en/datasheet/1000afd.pdf (http://cds.linear.com/docs/en/datasheet/1000afd.pdf) According to the front page, it's promised not to age more than 2uV over the first 1000 hours. This doesn't mean it stays flat at 2uV after 1kHr, but rather it gives a certainty of aging.
-
Did anyone introduce him to "thermoelectric effects" yet?
-
NO!!! We were waiting on you to set the hook. :-DD
Wait until he gets his hands on a quality pair of Kelvin clips with a quality instrument. Oh boy!
-
I guess I may have to break out that check-book. Although it would be a shame to have voltage standards worth more than my measuring equipment...With that said, I can still make this a science experiment with the lower-tier equipment! I also wonder if there are any hacks I can do to the internals to make them a bit more accurate :-/O...
And I know hobbyist don't have to follow standards, but why shouldn't we!? As a Systems Eng at work, I am familiar with strict standards. Maybe thats why I get excited at this idea
There are two main components: short-term drift, and long-term drift.
So its similar to a waveform in that sense. There is short-term drift which is similar to the noise on a line. It is relatively unpredictable, but there are ways to reduce it. In terms of equipment it would be constant temp, humidity, etc. Whereas long term drift is some type of line. Each device would have a different slope for that line (but it can vary depending on use, environmental conditions etc). The point of lab calibration is to track the long term drift to give you an amount of certainty that the equipment is within spec.
I would assume it is relatively useless to track short-term drift then? Unless you make a science out of it? You would need to get yourself a controlled environment and edit different parameters to predict the short term drift.
Did anyone introduce him to "thermoelectric effects" yet?
I'm listening....
-
Meanwhile... back on planet reality, I have several cheapish and basic DMMs here and I don't think I've adjusted any of them (more than once?) in 15-30 years depending on how long I've had them. I would expect them all to pass a basic DC voltage test that proves they are as useful today as they were when I first bought them. There would be virtually ZERO added value in having them checked professionally. I don't think I've adjusted any of them this century!
I bought a couple of decent DMMs fairly recently on ebay and I doubt that I will ever adjust either of them and I expect that they too will be just as 'useful' in xx years' time as they are today despite not having them formally calibrated.
-
Meanwhile... back on planet reality, I have several cheapish and basic DMMs here and I don't think I've adjusted any of them (more than once?) in 15-30 years depending on how long I've had them. I would expect them all to pass a basic DC voltage test that proves they are as useful today as they were when I first bought them. There would be virtually ZERO added value in having them checked professionally. I don't think I've adjusted any of them this century!
I bought a couple of decent DMMs fairly recently on ebay and I doubt that I will ever adjust either of them and I expect that they too will be just as 'useful' in xx years' time as they are today despite not having them formally calibrated.
You're 100% right. I really don't see any practical use for this. Anything I'm doing is digital logic or robotics, so my tolerances are very loose. With that said, I think learning this stuff is very valuable. The concepts/knowledge can be applied in many places, and it seems like a fun science. It definitely is not useful though.
The way I see it though, I can spend hours playing video games/surfing online to relax and zone out (which I do), or I could spend some of that time learning about calibration etc. Seems like its only disadvantage is it being a time-sink, which at the moment is better than my other time sinks.
-
The best thing about all this discussion is the learning experience you get from it. There are people on here that are very well versed in various fields of study and research. I say do not ever feel embarrassed to ask a question, you will be amazed at what you can pickup from the forum. Look through Dave's blogs and enjoy, some of them are definitely the "Bees Knees".
-
or I could spend some of that time learning about calibration etc. Seems like its only disadvantage is it being a time-sink, which at the moment is better than my other time sinks.
Yes, that's fair enough :)
If you find calibration (and measurement uncertainty etc) interesting then read a few papers on the subject...
The best advice I can give you is to never forget that the most important thing in your workroom is YOU and what you learn and experience from playing with electronics. Calibration has its place but I would suggest that it shouldn't become an all consuming hobby in its own right ;)
I get the impression that the voltnut hobby usually involves buying products based on someone else's (design) success in the form of very accurate commercial DMMs and calibration standards. The goal is to 'achieve' unrealistic levels of accuracy and this whole aspect of the hobby smacks of pointless chequebook based elitism in my opinion. But it is just my opinion :)
-
If you are looking for reading material, try this book http://www.amazon.com/Calibration-Philosophy-Practice-Fluke-Corporation/dp/0963865005 (http://www.amazon.com/Calibration-Philosophy-Practice-Fluke-Corporation/dp/0963865005) from Fluke.
-
Seems like its only disadvantage is it being a time-sink, which at the moment is better than my other time sinks.
+1 for this line of reasoning :-+
-
I guess I may have to break out that check-book. Although it would be a shame to have voltage standards worth more than my measuring equipment...
...
Look at it this way, instead of have to spend $100 for a DMM you can trust, DMM Check allows you to save money by giving you that confidence with your $1 DMM.
I spend more on DMM check+ than my UT61E. Then I got a couple of $1 DT830B ($4 shipping from China). DMM check+ gave me the confident that my $1 DMM is working nicely. In fact, those DT830B is rather a nice bugger.
Rick
-
I guess I may have to break out that check-book. Although it would be a shame to have voltage standards worth more than my measuring equipment...
...
Look at it this way, instead of have to spend $100 for a DMM you can trust, DMM Check allows you to save money by giving you that confidence with your $1 DMM.
I spend more on DMM check+ than my UT61E. Then I got a couple of $1 DT830B ($4 shipping from China). DMM check+ gave me the confident that my $1 DMM is working nicely. In fact, those DT830B is rather a nice bugger.
Rick
I guess you make a point....I have a DMMCheck+, which is more valuable than all of my other gear. Didnt think of it that way
-
I guess I may have to break out that check-book. Although it would be a shame to have voltage standards worth more than my measuring equipment...
...
Look at it this way, instead of have to spend $100 for a DMM you can trust, DMM Check allows you to save money by giving you that confidence with your $1 DMM.
I spend more on DMM check+ than my UT61E. Then I got a couple of $1 DT830B ($4 shipping from China). DMM check+ gave me the confident that my $1 DMM is working nicely. In fact, those DT830B is rather a nice bugger.
Rick
I guess you make a point....I have a DMMCheck+, which is more valuable than all of my other gear. Didnt think of it that way
So it's not a gateway drug, it's a vaccine?
-
For example, my HP 3458A is calibrated (not meaning they tweak anything inside - only if needed) based on a gold-transfer guard banded standard in Loveland Colorado.
Mate, do you know off hand what was the uncertainty of your 3458A after calibration? I know they use the golden 3458A for full verification all ranges, but I wonder what they use for the 10V and 10kOhms artefact calibration.
After artefact calibration, the uncertainty of the 10V/10k range on my R6581 was 2.4ppm/1.2ppm (95% confidence), which isn't very good, but I'll live with it because I don't have the big bucks (nor the need =P) to send it to a JJ array.
I still lust after a 3458A, and I just wanted to know how good the normal calibration was with Keysight if when I pick one up....
-
Deadly,
Interesting you ask that specific question. I am not sure at this point, initially I sent my unit in for repair/rebuild so looking at my Certs the Loveland lab stated they had no previous data acquired in their report. This will be the 2nd year to cal with my ownership, so I presuming I don't have enough samples as of yet to give a valid answer.
-
@jlmoon
Thanks for checking for me. ^-^
Looking at the Keysight website (http://www.keysight.com/main/editorial.jspx?cc=AU&lc=eng&ckey=2364633&id=2364633), I think one would actually have to pay extra for the uncertainty data to be supplied. That's a bit slack considering its already like $700+ for calibration of the 3458A...
My report didn't have previous data either, but it mentioned the uncertainty of the standards used and the calculated uncertainty after calibration.
-
If these devices used to be calibrated is it possible to get the old calibration data and track it? Like if it WAS calibrated at Tektronix by ABC company, and then sold to me, can I call tektronix and get a report of their data (maybe for a small fee)? If the point of cal is for a traceable history to a standard, then shouldn't I be able to trace it myself?
Also, on that note. I dont have any cal data but I'm confident that three of my devices were calibrated before. Is it possible to request the old data to see the long-term precision?
-
Robo,
I don't know about other equipment manufacturers. Keysight/Agilent maintains a history on all their equipment provided you use their repair or calibration labs. As far as I know the only way to obtain old data is if the equipment in question has been registered with Keysight, regarding calling Tek for a piece of equipment cal'd by a 3rd party they have no way of maintaining these records. You're pretty much at the mercy of the 3rd party, their traceable certifications and record keeping. If you have old calibration stickers affixed to your equipment you might be able to trace back through the "Cert Id" Numbers posted on that cal sticker. I have some equipment that is maintained by Tektronix Service Solutions and they provide a Traceable Calibration listing :
Model Number , Serial Number, Date of Cal, Due date of Next Cal, Lab where cal was performed, the Calibration Technician and the Cert Id. Keysight provides about the same information.
Hope this was a good answer.
Jon