New EX330, bad calibration?

[jimdeane]

I am putting together my essential electronics workbench, and have several bits of new equipment now. Korad supply, two Extech multimeters, and other things.

My Korad power supply came in, so I checked it's output with my EX330. It showed a lower voltage than the PS display, so I was suspecting the display wasn't very accurate.

I also purchased an Extech MN35 multimeter, so I checked the output with that one, and it was in agreement with the Korad, even with different leads on the multimeters. I need to dig out my older multimeters to check against them.

So, with two out of three agreeing, I'm looking at the EX330 as maybe out of calibration. Before I exchange it or try calibrating it, I would like to know if there is a good boot-strap way to check the meter calibration (without a really known good meter or vintage source). Maybe Vdrop across a diode?

[rs20]

Can you provide some hard numbers? Rather than just saying "agree" vs "not agree".

...even with different leads on the multimeters....

Lead length is unlikely to be relevant unless the lead is severely damaged; multimeters typically have a 10 megohm input impedance.

Maybe Vdrop across a diode?

Unless your meter is off by 5%, a diode isn't going to tell you much; way too much uncertainty and variation between diodes.

[crispy_tofu]

I hear that the quality control on the Extech EX series is quite bad...

[jimdeane]

I'll get some specific numbers next chance I have at the bench (could be a few days).

[jimdeane]

Okay, so some numbers, and perhaps (?) even a bad power supply.

I decided to do a comparison between the EX330, the MN35, and for a couple of the measurements a cheap multimeter that came in an experiment kit from a workshop.


Battery 1: EX330 9.29v, MN35 9.35v
Battery 2: EX330 9.26v, MN35 9.32v
Three batteries in series: EX330 27.82v, MN35 28.1v, cheap DMM 27.9v
Korad power supply running at 5 amps into a bulb, measured at bulb terminals: EX330 13.19v, MN35 13.28v, Korad display 13.30v, cheap DMM 13.25v

So at 9v the difference between EX330 and MN35 is 0.06v, at 13v the difference is 0.09v, and at 27v the difference is 0.3v. If that tells you anything. (Not quite linear with voltage.)

I also measured the AC component of the power supply output with a scope while the power supply was driving a bulb at 5 amps. I noticed that if I dropped the voltage below 7v (under load), with the next relay click I saw a huge jump in running noise on the scope, going from less than 10mv peak to peak up to 125mv peak to peak. I can take some screenshots and/or measure it if that is of interest (or maybe I start a different thread on the power supply). I checked this to see if maybe an AC component on the PS output would be causing a difference between the cheaper and the more expensive Extech meters.

[retiredcaps]

Nothing in the readings above suggest either meter is out of their DCV 1.0% + 2 counts (EX330) and 0.5% + 2 counts (MN35), respectively, range wrt to each other.

For example,
EX330 - 27.82 * 1.01 (1.0%) = 28.0982, round up  => 28.10 => add 2 counts => 28.12
MN35 = 28.1
cheap DMM 27.9

The real question is how the meters, over time, will drift, if any.

[XFDDesign]

I'm developing some classes and labs for a first term electronics course, and I've opted to direct the students to pickup the EX330 for their first "decent" intro meter.

I have found, that while they're all in the ballpark of their specified accuracy, the meter-to-meter agreement is lacking. For example, I have a transfer standard resistor of 10.00016kOhm. My main meters are all Fluke bench meters. This same standard reads as:
Fluke 45 Unit 1: 9.999kOhm
Fluke 45 Unit 2: 10.000kOhm
Fluke 45 Unit 3: 10.000kOhm
Fluke 8846A: 10.00016kOhm
(All at 20C)

They all generally agree with one another, with one meter being in disagreement by 1mV (0.01%). How do my two EX330s stack up?

EX330 Unit 1: 09.95kOhm
EX330 Unit 2: 09.98kOhm

Both are "within spec," but their agreement with eachother is only to about 30mV or 0.3%. More over, neither actually reads 10k, but rather "low." It's tolerable for an entry level meter, but as far as "good" test and measurement gear goes, you will want something higher caliber.

I actually suspect it has to do with the trimpots on the EX330's pcb. During factory test, they adjust to a point then call it a wash.

[lowimpedance]

@ XFD,  mV or ohms ?

[jimdeane]

Thanks for checking against the published specs for accuracy, I should have done the calculations myself (oops).

I am considering a couple of options. One is to have the 330 NIST-reference calibrated, either by Extech or a local service. Then calibrate the other meters to the 330 so they come close to agreement.

Alternately, I've seen reference standards from sources like Voltagestandard.com that I could use to keep all of my meters in some common calibration.

[ModemHead]

Keep in mind that checking DMMs with a resistance standard will probably not reveal anything about their DCV calibration.  The dual-slope ADC converters in most handheld DMMs measure resistance by passing a test current through a reference resistor and the unknown resistance in series.  The unknown resistance is determined by the ratio of the voltages across each resistance, not an absolute voltage measurement.  Since the voltage divider resistors are often dual-purposed as reference resistors, a resistance check may reveal the tolerance and precision of the voltage divider components, but will be unrelated to the DCV calibration.

[XFDDesign]

Thanks for checking against the published specs for accuracy, I should have done the calculations myself (oops).

I am considering a couple of options. One is to have the 330 NIST-reference calibrated, either by Extech or a local service. Then calibrate the other meters to the 330 so they come close to agreement.

Alternately, I've seen reference standards from sources like Voltagestandard.com that I could use to keep all of my meters in some common calibration.

You understand that NIST cal is not going to make the meter more accurate, right? This was the point I made in my post above.  Moreover "Calibration" is not adjustment. A calibration is simply "yup, this is within spec" or "it is no longer in spec." It is not "we'll adjust it until it gives us a particular number."

You need to watch this: https://www.youtube.com/playlist?list=PLvOlSehNtuHstfCI8cQYweKVjgVw0mZBy before you continue down this rabbit hole.

[IanB]

I am considering a couple of options. One is to have the 330 NIST-reference calibrated, either by Extech or a local service. Then calibrate the other meters to the 330 so they come close to agreement.

Alternately, I've seen reference standards from sources like Voltagestandard.com that I could use to keep all of my meters in some common calibration.

It would be as waste of money to have an expensive calibration done on a cheap meter.

I have various inexpensive meters with adjustment trim pots and I have been very satisfied with aligning them to match a voltage reference module. I get all me meters reading consistently with each other and making best use of their available accuracy.

[jimdeane]

I would certainly have assumed that if a calibration service found the meter to be out of spec, they would adjust it to be in spec. I've only ever had a torque wrench re-calibrated, and my understanding was that it was adjusted to bring it back into spec. Adjusting if out of spec is also part of the procedure for calibration that Extech sent to me.

I think for my purposes, purchasing a DMMcheck+ http://www.voltagestandard.com/DMMCheck_Plus.html and keeping track of meter agreement with the various standards would be more than adequate to keep me happy.

[rs20]

I recommend the EEVblog videos where Dave visits Agilent's calibration facilities -- the distinction between calibration and adjustment is discussed at some length.

[Tim F]

I would like to know if there is a good boot-strap way to check the meter calibration (without a really known good meter or vintage source). Maybe Vdrop across a diode?

Just use a cheap shunt voltage reference like an LM4040A (0.1% initial tolerance). Choose a resistor and supply voltage to be in the ballpark of 0.3-1.0mA and you'll have a voltage reference which is more than good enough to check the meters you have.

[retiredcaps]

I think for my purposes, purchasing a DMMcheck+ http://www.voltagestandard.com/DMMCheck_Plus.html and keeping track of meter agreement with the various standards would be more than adequate to keep me happy.

I don't have the above, but I noticed it is $84 USD.  That is likely more than the cost of all your 3 meters combined.  Does that make sense given the 1% DCV specification of your most expensive meter (EX330)?

A better option might be this < $20 USD option as discussed below

https://www.eevblog.com/forum/beginners/t17704/

It doesn't have resistance, but you can easily get 0.1% resistors for cheap (<$1 USD).

BTW, if you do have to adjust the EX330 periodically, it is an unreliable meter because it drifts.

PS. I know Doug at voltagestandard makes a quality product, but given your current meter's specifications, it is overkill.

[XFDDesign]

I would suggest that if the point is certainty better than the +/-(1%+2counts), the EX330 is the wrong route. It's taking a Toyota Camry, and trying to polish it into a Ferrari.

If knowing, to the millivolt, with certainty is the goal, it's time to pick up a 4.5 digit dmm. Then when you find that you're bothered about 100uV, you go to the 5.5 digit dmms. 10uV not enough? 6.5digit dmms.

Then, of course, we've gone from the $40 meter, to the ~$200 meter, to the $800 meter to the $1200 meter.

And that only assumes one stops at 6.5 digits. Still two more to pursue.

The question that has neither been asked, nor answered is, "to what certainty is required of power supplies?"