Fluke 114 DMM, accurate enough for Lith-ions?

[GarageBoy]

Sheesh, all you flashoholics worrying about 1/10000 Volt
Don't forget the resistors in your flashlight are at most 1%

 

[lrp]

Thanks guys! I really would only be using it for checking Lith-ion batteries, I just want a meter to be accurate for this task only. What other meters would do the job well?

 

[HKJ]

Thanks guys! I really would only be using it for checking Lith-ion batteries, I just want a meter to be accurate for this task only. What other meters would do the job well?

Almost any meter will do, but it is niece to have a meter with 5000 or 6000 display, because you get a extra digit in the readout.
I.e. with a 2000/3000/4000 display you will get: 4.21, but a 5000/6000 display can show 4.207.

It is not that important if the precision is 0.5% or 0.05%, as long as you are checking for a max. of 4.2 volt, the small errors in the readout will only affect the number of cycles the battery can be used, not making it dangerous in any way. And as long as you do not have the specification for the actual battery chemistry, you do not know exactly what voltages to check for anyway.

When checking the specifications for a meter, remember to check for what temperature the meter has to used within for the tolerances to be valid and how long the tolerances are guaranteed to be valid (This is usual a year). For AC measurements the tolerances also needs to include a frequency specification. If any of the above specifications are missing, the tolerance specifications are suspicious.
Also remember that serious tolerance specifications are maximum errors, that you might encounter if you are using the meter to the limit, not the typical error you get from a meter always keeps on the desktop (At least that is valid for the better meters).

One important specification (At least for me): The meter must warn about low battery, and not displays a wrong value.

 

[Justin Case]

Almost any meter will do, but it is niece to have a meter with 5000 or 6000 display, because you get a extra digit in the readout.
I.e. with a 2000/3000/4000 display you will get: 4.21, but a 5000/6000 display can show 4.207.

Almost any meter will do, but it is niece to have a meter with 5000 or 6000 display, because you get a extra digit in the readout.
I.e. with a 2000/3000/4000 display you will get: 4.21, but a 5000/6000 display can show 4.207.

It is not that important if the precision is 0.5% or 0.05%, as long as you are checking for a max. of 4.2 volt, the small errors in the readout will only affect the number of cycles the battery can be used, not making it dangerous in any way. And as long as you do not have the specification for the actual battery chemistry, you do not know exactly what voltages to check for anyway.

Again, opinions differ on this point. If you get a 4000 digit display DMM or less, then a combination of high range (gain) error and offset error (the digits) can result in IMO fairly high error. The offset error can have a big impact on these low digit DMMs, even if you have a low gain error. Yes, it will probably be safe, but IMO the confidence in the measurement accuracy is not good.

Let's say you get a DMM with an accuracy of 0.5% + 2 digits like this Fluke 114, except you settle for a 4000 count DMM.

Then a 4.20V measurement can range from 4.16V to 4.24V. For me, the potential that the voltage can be as high as 4.24V sucks.

Because of how the digits error works, it is highly doubtful that a low count DMM will read 4.21V vs. a high count DMM at 4.207V, unlike in your hypothetical example. Your example would occur only if there were zero digits of error. More likely, both DMMs will have similar accuracy specs, e.g., +/-(0.5%+2 digits). For a low count DMM (e.g., 3000), the 2 digits go into the hundredths place. For a high count DMM (e.g., 6000 counts), the 2 digits go into the thousandths place. Big difference.

Yes, there is probably a statistical distribution for the DMM accuracy, and most DMMs will not be in the error extremes. But unless you can calibrate the DMM, compare it to a known accurate DMM, or obtain a voltage reference standard, you don't know where your DMM is in that distribution.

At a minimum, to reduce the impact of the offset error, I would get a DMM with a display of at least 5000 counts. If the DMM I like (perhaps because it has other useful features) has a relatively high gain error (e.g., 0.5% or more), I would also get a 4.0960V voltage reference from Malone Electronics.

IMO, you will quickly learn how useful a DMM is and discover all sorts of uses. If you get essentially what is a fairly limited DMM (both in features and accuracy), you will soon wish you had gotten something better.

 

[HKJ]

Then a 4.20V measurement can range from 4.16V to 4.24V. For me, the potential that the voltage can be as high as 4.24V sucks.

You might not like this tolerance, but it is good enough to be safe, to get a decent number of charge cycles from the batteries and to check how much of the charge is left.
More exact measurement might be nice, but you do not really get any more useful information from it: You will not be safer, the number charge cycles depends on the exact chemistry in the battery, without knowing that, you voltage reading does not help you, and the charge left will not be more exact and even if it was, if would probably be irrelevant (Who cares if the light can run 20 or 21 minutes on the remaining charge?).

Yes, there is probably a statistical distribution for the DMM accuracy, and most DMMs will not be in the error extremes. But unless you can calibrate the DMM, compare it to a known accurate DMM, or obtain a voltage reference standard, you don't know where your DMM is in that distribution.

The error specification does not really work that way, the meter has to be much better calibrated from the factory. Then after being overloaded, being frozen, being boiled, etc. for a year, it must still be within tolerances.

That is the reason I believe that temperature tolerances and calibration time are important. It is very easy and cheap to make a meter that has 0.01% tolerance when leaving the factory (Thanks to microprocessors and EEPROMS), but can it keep this tolerance, that is the challenge.

 

[lewong]

Irp, if you want a Fluke and you're a flashaholic, you should at least get the Fluke 115. As MrGman says, you will want to be able to measure current.

 

[Justin Case]

You might not like this tolerance, but it is good enough to be safe, to get a decent number of charge cycles from the batteries and to check how much of the charge is left.
More exact measurement might be nice, but you do not really get any more useful information from it: You will not be safer, the number charge cycles depends on the exact chemistry in the battery, without knowing that, you voltage reading does not help you, and the charge left will not be more exact and even if it was, if would probably be irrelevant (Who cares if the light can run 20 or 21 minutes on the remaining charge?).



The error specification does not really work that way, the meter has to be much better calibrated from the factory. Then after being overloaded, being frozen, being boiled, etc. for a year, it must still be within tolerances.

That is the reason I believe that temperature tolerances and calibration time are important. It is very easy and cheap to make a meter that has 0.01% tolerance when leaving the factory (Thanks to microprocessors and EEPROMS), but can it keep this tolerance, that is the challenge.

I never said that there was a benefit with being safer if you had a more accurate DMM. As for cycle life, of course you will get better cycle life if you keep the float voltage lower. Even if you don't know the exact Li-ion chemistry, it doesn't really matter. It is the same battery, and thus the chemistry is held constant. If you have a accurate DMM that allows you to differentiate 4.20V from 4.24V, and you make sure that your charger goes only up to 4.20V, you will get better cycle life than if you went to 4.24V. Now you may not know exactly how many more cycles you might get because that is chemistry-dependent, but you will get more with the lower float voltage.

As for capacity, I have no idea if someone cares about whatever small amount capacity difference might result from the DMM's voltage measurement error. I agree that it is a small amount. But I don't presume that my opinion of that issue applies to anyone else either.

The issue of ruggedness is an entirely different matter than the paper spec for measurement accuracy, which is what I had been addressing.

The error spec most certainly works the way I described for the Fluke 114. With a 4.096V reference source, you are about as close to 4.20V as reasonably possible with an easily-obtained, commercial voltage reference. The Fluke will not use a different voltage range so any internal voltage divider circuitry issues don't apply here. Thus, whatever voltage error you measure for the 4.096V reference can be directly applied to your 4.20V measurement. If you are unconcerned about voltage measurement errors of 0.04V for Li-ion OCV, then any error using the above approach is far smaller than that.

 

[HKJ]

If you have a accurate DMM that allows you to differentiate 4.20V from 4.24V, and you make sure that your charger goes only up to 4.20V, you will get better cycle life than if you went to 4.24V. Now you may not know exactly how many more cycles you might get because that is chemistry-dependent, but you will get more with the lower float voltage.

You will also get a better cycle life if yo charge to 4.1 or 4.15 volt, it has nothing to do with the tolerances of the meter.


As for capacity, I have no idea if someone cares about whatever small amount capacity difference might result from the DMM's voltage measurement error. I agree that it is a small amount. But I don't presume that my opinion of that issue applies to anyone else either.

The error spec most certainly works the way I described for the Fluke 114.

Have you measured some Fluke 114? I would expect them to be much better than 0.5% from factory, as I described above.

With a 4.096V reference source, you are about as close to 4.20V as reasonably possible with an easily-obtained, commercial voltage reference. The Fluke will not use a different voltage range so any internal voltage divider circuitry issues don't apply here. Thus, whatever voltage error you measure for the 4.096V reference can be directly applied to your 4.20V measurement. If you are unconcerned about voltage measurement errors of 0.04V for Li-ion OCV, then any error using the above approach is far smaller than that.

Checking the meter against a reference is a very good idea, you just has to know which one has the lowest tolerance, at least if they disagree . But that reference is only good for one voltage range (Which is enough for checking LiIon), that is very far from being enough to verify a modern multimeter. The fact that DC voltage is correct does not in any way prove that any of the other ranges are correct.

 

[Justin Case]

You will also get a better cycle life if yo charge to 4.1 or 4.15 volt, it has nothing to do with the tolerances of the meter.

Have you measured some Fluke 114? I would expect them to be much better than 0.5% from factory, as I described above.

Checking the meter against a reference is a very good idea, you just has to know which one has the lowest tolerance, at least if they disagree . But that reference is only good for one voltage range (Which is enough for checking LiIon), that is very far from being enough to verify a modern multimeter. The fact that DC voltage is correct does not in any way prove that any of the other ranges are correct.

Of course the float voltage (and thus cycle life) has something to do with the accuracy of the meter. With an accurate meter, you know where you stand wrt your charger's terminating (float) voltage. With an inaccurate meter, you don't.

No I have not measured any statistically valid number of Fluke 114 samples. Have you? Regardless, you don't believe that manufactured products have a statistical distribution for their performance? I clearly said that most DMMs won't be at the extremes of the accuracy spec, which is what you seem to be saying. So what exactly is your complaint? Do you believe that all DMMs will have exactly the same accuracy out of the factory -- essentially a delta function distribution? That's still a statistical distribution.

I never said that comparing a DMM's measurement at a single voltage was sufficient for any of the other of the DMM's voltage ranges. The point of suggesting the 4.096V reference was to address the OP's interest in checking Li-ions. Malone Electronics checks their voltage references using a calibrated 8.5 digit HP3458A DMM. They also offer a re-cal service. If you have sources for higher voltage references so that other, higher DMM voltage ranges can be checked, speak up.

 

[HKJ]

Of course the float voltage (and thus cycle life) has something to do with the accuracy of the meter. With an accurate meter, you know where you stand wrt your charger's terminating (float) voltage. With an inaccurate meter, you don't.

No, it has not, it has to do with the chemistry in the battery, you do not know if a battery needs 4.20 or 4.15 volt to last 500 cycles, you only know that a lower voltage will give more cycles.

No I have not measured any statistically valid number of Fluke 114 samples. Have you? Regardless, you don't believe that manufactured products have a statistical distribution for their performance? I clearly said that most DMMs won't be at the extremes of the accuracy spec, which is what you seem to be saying. So what exactly is your complaint? Do you believe that all DMMs will have exactly the same accuracy out of the factory -- essentially a delta function distribution? That's still a statistical distribution.

I agree that the calibration is a statistical distribution, probably a bell curve, but it does not stretch to the limit of the tolerance specifications, it is much more narrow, because much of the tolerance is needed for aging of the components.
I suspect that many cheap meters are forgetting that in their tolerances, but I have never made a study of it.

If you have sources for higher voltage references so that other, higher DMM voltage ranges can be checked, speak up.

Fluke 9100 :tinfoil: (Sorry, that is a bad joke).

 

[Justin Case]

you do not know if a battery needs 4.20 or 4.15 volt to last 500 cycles, you only know that a lower voltage will give more cycles.

Which is what I wrote. However, I would know whether or not my charger is running to spec (e.g., 4.20V terminating voltage for LiCoO2 Li-ion) if I had a good meter.

I agree that the calibration is a statistical distribution, probably a bell curve, but it does not stretch to the limit of the tolerance specifications, it is much more narrow, because much of the tolerance is needed for aging of the components.
I suspect that many cheap meters are forgetting that in their tolerances, but I have never made a study of it.

So this is just idle speculation, despite your opinion stated as fact that the accuracy limits don't "stretch to the limit of the tolerance specifications". Got it.

 

[jayflash]

Fluke and other really good meters will protect themselves and the user when mistakes are make. The better meters will simply blow a fuse when an attempt at measuring voltage is made with the leads still in the amps position. Cheap meters can explode -- depending on the voltage.

Even if one is not an electronics professional, the cost of a good meter may be worth it for ease of use, ergonomics, accuracy, size, reliability and, of course, safety. I've have and used other decent meters, but prefer the overall benefits of my Fluke. YMMV.

 

[HKJ]

Which is what I wrote. However, I would know whether or not my charger is running to spec (e.g., 4.20V terminating voltage for LiCoO2 Li-ion) if I had a good meter.

You might like to know that (And I do, there is a reason that I have a Fluke 189), but it is not necessary to know when using LiIon.

So this is just idle speculation, despite your opinion stated as fact that the accuracy limits don't "stretch to the limit of the tolerance specifications". Got it.

It is based on my knowledge of electronic and the fact that components will change over time.
You could also check the Fluke 114 calibration manual, they require 0.125% precision on the calibrator and the meter has a microprocessor and EEPROM, i.e. it can store an exact calibration factor and is not depend on somebody turning a trimpot to "about right" value.

 

[Justin Case]

You might like to know that (And I do, there is a reason that I have a Fluke 189), but it is not necessary to know when using LiIon.

Once again, I never said that it is necessary to know this information. None of this is "necessary", including buying the DMM in the first place. But it certainly is useful information to know the terminating voltage. Simply witness all of the threads and posts regarding cheap chargers that exceed 4.20V terminating voltage and the associated concern.

It is based on my knowledge of electronic and the fact that components will change over time.

That isn't proof. That is an appeal to authority. Sorry. And in fact if the components change over time, the DMM accuracy can drift toward the extremes of the spec. You seem to be focused solely on the accuracy new and fresh out of the box.

You could also check the Fluke 114 calibration manual, they require 0.125% precision on the calibrator and the meter has a microprocessor and EEPROM, i.e. it can store an exact calibration factor and is not depend on somebody turning a trimpot to "about right" value.

This is more persuasive but hardly conclusive proof regarding the width of the distribution. And it applies only to the Fluke.

 

[Lightcrazycanuck]

Who will win the Meter Wars?????

Justin Case or HKJ.

This is better than UFC.

 

[HKJ]

Who will win the Meter Wars?????

Justin Case or HKJ.

This is better than UFC.

I am out of this discussion, it is not getting anywhere, anymore would just be .

 

[Bullzeyebill]

Guys, your arguments are not helping here, particularly for the DMM unenlightened. It looks like point for point you can contradict each other. Why don't you continue your discussions via PM, and keep it friendly.

Bill

 

[DM51]

Guys, your arguments are not helping here, particularly for the DMM unenlightened. It looks like point for point you can contradict each other. Why don't you continue your discussions via PM, and keep it friendly.

Bill

Very good post... please would you both follow Bill's advice.

 

[Lightcrazycanuck]