EEVblog Electronics Community Forum
Products => Test Equipment => Topic started by: bitman on April 19, 2018, 01:23:46 am
-
So long story short, since the Eevblog meter doesn't seem to arrive anytime soon, if at all, I couldn't wait any longer and invested (spend a gift-card on) in a Fluke 87v to become my main meter, replacing several very cheap ones in my possession (<$40). Precision hasn't been too important to me in the past, but lately I've found 6000 count meters just isn't cutting it when I try to figure out why things work they way they do (or why they don't).
However, I am not sure the calibration is right. Comparing capacitor testing to the other meters, there seems to be quite a difference. The Fluke seems to constantly measure high - a 3300uF cap measures around 3500 for instance. A 470uF measures almost 500uF - of course it could be the capacitor itself, but the old meters are much closer to the written value. I've tried about 10 different caps, and only one seemed within a reasonable range of the printed value for the cap.
So my dumb question is pretty simple: Do I need to do something to dial things in? What is a good way to find out if the meter or other things are producing wrong results (and if they do show there's a variation, is that something I can set a compensation for on the multimeter itself?). I frankly think the old meters were way OFF - but still the difference on the fluke seems high?
-
I've tried about 10 different caps, and only one seemed within a reasonable range of the printed value for the cap.
So my dumb question is pretty simple: Do I need to do something to dial things in? What is a good way to find out if the meter or other things are producing wrong results (and if they do show there's a variation, is that something I can set a compensation for on the multimeter itself?). I frankly think the old meters were way OFF - but still the difference on the fluke seems high?
You might be interested in this thread:
https://www.eevblog.com/forum/testgear/multimeter-bench-meter-delimma-(-fluke-87)/ (https://www.eevblog.com/forum/testgear/multimeter-bench-meter-delimma-(-fluke-87)/)
-
I've tried about 10 different caps, and only one seemed within a reasonable range of the printed value for the cap.
So my dumb question is pretty simple: Do I need to do something to dial things in? What is a good way to find out if the meter or other things are producing wrong results (and if they do show there's a variation, is that something I can set a compensation for on the multimeter itself?). I frankly think the old meters were way OFF - but still the difference on the fluke seems high?
You might be interested in this thread:
https://www.eevblog.com/forum/testgear/multimeter-bench-meter-delimma-(-fluke-87)/ (https://www.eevblog.com/forum/testgear/multimeter-bench-meter-delimma-(-fluke-87)/)
Not sure how a thread about someone wanting to buy the meter I have helps? I scanned it, and it seemed to just be a lot of people who like the meters they know and everyone recommends "just use what I have". Doesn't help me much, except if there is a limit on the capacity I can test reliably. I'll definitely look into that.
-
Well you want to be sure, send your meter for calibration that way you will have an definitive answer, or find someone near you with an calibrated one and compare your values of dc ac ohms current etc ...
When i buy new meters, i put them against my calibrated Gossen 29s freshly done, to have an certitude
-
Not sure how a thread about someone wanting to buy the meter I have helps?
It has a lot more than that. Read the parts about buying an LCR meter instead of a Fluke 87V if you're really interested in capacitors.
Also the parts discussing what multimeters can and can't tell you about capacitors.
-
What Fungus and others are trying to tell you is that dmm measurement of capacitance is not that godd, even from high end dmms like the Fluke87V !
Most dmms measure capacitance by measuring RC time, ie measure how long it takes to charge capacitor to known V through a known resistor. While this is a good indication of capacitanve value, does not tell you much else.
My point is that even the best dmms capacitance measurement is not that good and cannot be relied upon. None of my meters agree either measuring C (one of them us a Fluke 187).
Do not worry about the C measurement, just enjoy your quality meter ;)
-
I was going to say that capacitance is about the least suited measurement for checking the calibration of a DMM. By stating this, I have a Mylar capacitor with <2.5% tolerance in mind. But of course matters can get even worse when using electrolytic capacitors that might have +50/-20% tolerance from the outset plus a huge dependency on frequency, voltage and temperature.
Capacitance measurement in a DMM is not particularly useful and most likely only there because so many users demand it.
Some will argue that it's still useful for identifying defective capacitors that have lost their capacitance, but I would not rely on even that.
In contrast, a proper LCR meter that measures capacitance together with quality/loss-factor/phase angle and/or ESR at some frequency representative for the conditions in the actual application provides a true picture about the condition of any capacitor - electrolytics (and tantals) in particular.
BTW, I would not doubt the calibration of any Fluke meter. In my experience they stay well within spec. even after decades.
-
@ performa
I can understand that some meters may keep an good cal over time, but what certainty do you have when you have only one or two meters at home and they haven't been calibrated in the past years, if you dont have acces to an dc ac volts standard
I would not personally be sure of everything at 100%, for home use it can be okay, but if you take your meter in an industrial / comercial environment it may get problematic. All my meter i bring at my job will get calibrated for ISO purposes and reliability. As i said on other posts, i replace my 2x fluke 83 III with fluke 189 with extended battery packs, kinda sick to change the 9v cells to often.
Even digging precision parts <1% resistors and some precise capacitors may help a little, but ac-dc needs to be checked against something precise, like the reference voltages kit available, but they dont get higher than 10 volts ... that leave the current measurments to check ...
-
3300uF cap measures around 3500 for instance. A 470uF measures almost 500uF
Do you compare it directly to the instrument with better specs on the same caps?
References for performing some sort of home made performance test without the above mentioned comparison should be at least the same tolerance as meter specs or better.
-
FWIW I forget the 87V has the capacitor test function till the mode appears when I'm toggling about in the Ohms mode
It's not my 'go to' cap tester, but if you discharge the capacitor/s fully, the meter gives a reasonable go/no go/suspect cap reading
All my meters with the 'bonus' half baked cap tester function thrown in disagree on cap accuracy,
but they all come close enough, which is good enough considering cap values can be out by +/-20% of the written value from the factory, and be perfectly fine
Peak Atlas ESR70 rules for me for cap testing, almost a no brainer to use :-+ and slots in the case easily next to the the 87V and troubleshooters artificial rabbits foot ;D
-
@ performa
I can understand that some meters may keep an good cal over time, but what certainty do you have when you have only one or two meters at home and they haven't been calibrated in the past years, if you dont have acces to an dc ac volts standard
Of course you can never be sure about the calibration of any (ab)used meter. But as I've understood it, the OP has a new instrument, so my mention of long term stability is just meant to give some additional confidence, assuming that the happy owner will be aware that heavy abuse might void my claim.
Personally I have bought several Fluke 8840A/8842A bench meters, all coming from industrial environments, one 8840A was covered with dirt and last calibration way more than a decade ago, but still spot on within ~20ppm (which would have been good even for a 8842A). Needless to say that the other, less abused meters were even better.
I personally am not a big fan of Fluke handhelds (or handhelds in general), but their low accuracy makes it all a lot easier anyway and I firmly believe that they would remain within spec for ages just as the Fluke lab gear does. My handheld Metrix MX54C certainly does it, even after 20+ years without calibration. That's one of the major differences when compared to some el-cheapo handheld meter (yes, I do have a few of them as well).
-
So long story short, since the Eevblog meter doesn't seem to arrive anytime soon, if at all, I couldn't wait any longer and invested (spend a gift-card on) in a Fluke 87v to become my main meter, replacing several very cheap ones in my possession (<$40). Precision hasn't been too important to me in the past, but lately I've found 6000 count meters just isn't cutting it when I try to figure out why things work they way they do (or why they don't).
However, I am not sure the calibration is right. Comparing capacitor testing to the other meters, there seems to be quite a difference. The Fluke seems to constantly measure high - a 3300uF cap measures around 3500 for instance. A 470uF measures almost 500uF - of course it could be the capacitor itself, but the old meters are much closer to the written value. I've tried about 10 different caps, and only one seemed within a reasonable range of the printed value for the cap.
So my dumb question is pretty simple: Do I need to do something to dial things in? What is a good way to find out if the meter or other things are producing wrong results (and if they do show there's a variation, is that something I can set a compensation for on the multimeter itself?). I frankly think the old meters were way OFF - but still the difference on the fluke seems high?
A brand new meter from Fluke out of cal, seems a bit odd.
If you follow this link, message #1687 shows eight different values of capacitors tested on some of my meters including my two BK RLC meters. I bought a brand new 87V this year. The second link is to a video I made of this meter checking the same capacitors.
https://www.eevblog.com/forum/testgear/hear-kitty-kitty-kitty-nope-not-that-kind-of-cat/msg1245564/#msg1245564 (https://www.eevblog.com/forum/testgear/hear-kitty-kitty-kitty-nope-not-that-kind-of-cat/msg1245564/#msg1245564)
https://youtu.be/DA77DYGE6IM?t=1559 (https://youtu.be/DA77DYGE6IM?t=1559)
-
... Fluke ... 20+ years without calibration.
That's one of the major differences when compared to some el-cheapo handheld meter (yes, I do have a few of them as well).
Nobody's doubting the expensive stuff.
The question is: Do you have a cheapo handheld that's gone out of calibration? :popcorn:
(or anybody else here...)
-
This is a rev 10 PCB (very old 87V) that I had damaged the front end on and repaired, then ran more tests on. This link shows it measuring the three lower values from the table.
https://youtu.be/GhslaLKOpKE?t=1858
-
3300uF cap measures around 3500 for instance. A 470uF measures almost 500uF
Do you compare it directly to the instrument with better specs on the same caps?
References for performing some sort of home made performance test without the above mentioned comparison should be at least the same tolerance as meter specs or better.
Nothing that advanced. I realize that I don't have a known value to test against - I just have multiple instruments, mostly digital meters and the "component identifier" you can build yourself or get cheap on Ebay - and I've never really cared about precision before so the fact that they all disagree on the measured value of the same component has never really been an issue.
Btw. I think my old MS88 (which was my "main multimeter" before I got this one) decided to get jealous and stop working - I've never seen better timing - exactly on the day I got the Fluke :) So I cannot compare to the source I used to take for granted was "close enough" - but instead of I have a few other radio-shak like devices that I'm comparing against. I expected to see something a bit more "reliable" - ie closer to the printed value on the cap. In particular my high-voltage caps - but none of it gets close. I found a different thread here on eevblog that stated that the meters weren't meant to measure more than a few uF - I looked for that and didn't see that in the specs for the Fluke87v - it should do fine to 9999uF according to the specs https://www.alliedelec.com/m/d/8cf21be95af06ae474858dd790972ba3.pdf (https://www.alliedelec.com/m/d/8cf21be95af06ae474858dd790972ba3.pdf). Even if we say half or a 1/3 is reasonable, none of the caps I've tested have gotten close to the limit yet.
I appreciate the feedback here - I have a very cheap ESR meter that I'll use if need be for more confirmation on the value. What's most important to me is precision on resistors - color blind doesn't make reading those small suckers' color codes - even with a microscope :) So I have to measure pretty much every resistor (thru hole at least) I use *sigh* - I'm slow but at least it's fun.
-
The question is: Do you have a cheapo handheld that's gone out of calibration? :popcorn:
Of course this should not happen, as it would take a remarkable amount of incompetence to design a ±0.5% base accuracy meter in a way that it can drift out of spec – even if it’s cheap – and like anything else, “cheap” is relative anyway.
Yet certain manufacturers have managed to come close, simply by using crappy trimmers together with a much too wide adjustment range.
The much more widespread problem is meters not properly adjusted at the factory, hence not within their specified tolerances right out of the box. While not strictly related to the price, it is related to manufacturers and certainly never happens to A-brand meters.
A good example is my experience with an EXTECH EX730 clamp meter, bought around 2002: ridiculously high tolerance (±1.5%) - I’m talking about the direct (not clamp) measurement ranges – yet 4000 counts, barely within spec right out of the box, crappy potentiometer with ridiculously wide adjustment range, so it’s very hard to adjust and happily drifting all over the place within a couple days. I performed another careful adjustment after a couple days and never touched it ever since.
A re-check today reveals a rather unstable zero reading, fluctuating up to 1mV in the 400mV range. Fed with 3.9V it displays 3.893V; this is an error of -0.18%. Oh well, quite obviously I got lucky and managed to find a reasonable stable setting for the crappy potentiometer back then…
An EX520 (also 2002) was borderline when it was new back in the late 90’s, calibrated once immediately after purchase (which was fiddly, but not nearly as bad as the EX730) and found it to be still within its ±0.09% spec at -0.053% today.
Another Example is the Uni-T UT61E purchased in 2/2012, slightly out of tolerance right out of the box, then calibrated. Now, 6 years later, I did a quick check: 1.9V is displayed as 1.8976V. This is -0.1263%, hence exceeding its already high (twice as usual) tolerance specification of ±0.1% again.
Metex M4650CR, a defective meter bought at a -50% discount around 1990 out of the Conrad Electronic bargain bin at the ham radio exhibition in Friedrichshafen, Germany. Repaired and calibrated back then and never touched ever since. Quick check now, more than 25 years later: reading 1.9004V for 1.9V, this is an error of +0.021%, well within its specified ±0.05% tolerance.
Finally the A-brand Metrix MX54 from around 1992. Checked at 4.9V, reading 4.8988 hence an error of -0.0245%, well within its specified ±0.05% tolerance even after 25+ years without a single calibration in its entire lifetime.
Conclusion: For low accuracy meters it is certainly very unlikely that they will drift out of spec, no matter how cheap and what brand. Yet the Uni-T UT61E managed to do exactly that and it did not meet its specs right out of the box on top of that. In general, for non A-brand meters it is mandatory to check their calibration right after the purchase.
Back in the last century, I’ve found that cheap digital meters (“cheap” still means some $40-50 in pre 1990’s money) even have been totally unadjusted, hence off by up to 10%. I haven’t encountered such severe and obvious issues in this century yet, but wouldn’t blindingly trust any B/C-brand meters even today.
The big problem is, that folks on a budget who buy such a cheapy as (one of) their main meter(s) might not have the means to easily check the accuracy and correct the adjustment. I have been in a similar situation when I started using digital meters in the early 80’s last century. The first one was a cheapy (cheap back then = still at least $100,- in today’s money) and it was quite obviously way off. I had to use my analog DVM to adjust it and asked myself what’s the point to buy a ±0.8% digital meter if I need my 2.5% DVM (which was in fact more accurate than that) in order to calibrate it. Soon after that experience, I got my first reasonable accurate (±0.05%) DMM, a Metex M4650 (even older than the one tested above, but I still have it, just not readily accessible since it is in my old lab) which was some $150,- back then, equivalent to more than $300,- in today’s money. The only reason why I got it was the desire to have a reference for checking newly purchased cheaper meters, because back then I didn’t have any applications that required more than 2.5% accuracy. Yet once we get an instrument that claims a certain accuracy, we want to be able to rely on that, don’t we?