189 Calibrate or Repair

[rcross]

Hello everyone,

I am a long time lurker who finally mustered the courage to post a question.

I have recently acquired a second-hand 189 multimeter. All functions work fine, relatively clean to begin with and cleaned up easy. By the appearance nobody has been inside (except me) and minimal wear on the selector tracks. Supercap very crusty, but PCB and other components around don't have any signs of corrosion. I replaced the supercap.

Main issue with the meter is that it reads lower than any of my other meters. I do not have calibration setup or calibrated sources, but I have setup measurement of 189 compared to 289, 87V, 117 and 101. 289 is less than 2 years old, bought new; 87V second hand unknown vintage; 117 brand new and 101 bought less than a year ago.

For practical mesurements (where probe setup does not affect much) I see less than 0.2% discrepancy between 289-87V-117-101, in many cases less than 0.1%. But the 189 is consistently 1%-1.5% lower than the rest. I have checked V DC, mV DC, Ohm, mA DC and mV AC and all behave very similarly. Data table below. For current all meters connected in series with process calibrator and 24V supply. For Ohms measured decade box with each (yeah, low ohms affected by probe resistance, I did not do rel). For V abnd mV measured with each meter one-by-one. Not the most exact setup, but that is how I would be using the meter, so that is what matters to me. To be clear, I would be happy with the same level of error that I see between 289-87V-117-101.

I checked the input protection on components and visually inspected everything on the board. Cleaned the plugs. New batteries. Yet nothing changes. At this point I am starting to think that calibration is what is needed. Does this sound like there is some other defect I should be chasing before resorting to calibration?

Best regards,
rC

V DC

189	289	87V	117	101
0.1091	0.1120	0.1121	0.113	0.113
0.0560	0.5139	0.5138	0.513	0.512
0.9978	1.0136	1.0138	1.015	1.012
4.9577	5.0153	5.015	5.015	5.012
9.893	10.015	10.012	10.01	10.00
19.782	20.010	20.00	20.00	19.99

mV DC

189	289	87V	117
114.60	115.54	115.5	116.5
213.35	216.36	216.6	215.2
313.50	316.20	315.8	316.6
510.60	514.60	515.6	515.8

Ohm

189	289	87V	117	101
5.76	5.61	5.9	5.5	6.2
49.76	50.41	50.4	50.7	51.0
495.4	508.0	500.7	501.5	503
4943	5009	5000	5001	5000
49430	50120	50100	50090	50100
493100	500000	500100	499800	500000
4941000	5006000	5004000	5000000	5000000

mA DC

189	289	87V
3.880	3.999	4.01
7.849	7.996	8.01
11.814	11.994	12.01
15.775	15.992	16.01
19.737	19.991	20.00
23.697	23.989	24.00

mV AC

189	289	87V	117	101
3.691	3.651	3.92	3.8	7.6
17.72	17.844	17.94	17.9	18.8
34.879	35.165	35.28	35.2	35.3
171.72	173.03	173.12	172.9	172.3
355.81	358.82	359.5	359.1	358.6
502.8	505.50	506.3	505.7	504.6

[Conrad Hoffman]

Seems like a small error, though it's possible something could be amiss. Cal manual is here- https://www.eevblog.com/forum/repair/fluke-189-calibration-problem/msg4836668/#msg4836668
Unless you have some good standards it's going to be a PITA to cal this just based on other meters.

[rcross]

Yes, the error is small but noticeable in everyday use. Everyday use is how I discovered it in the first place. So I am sure it will drive me crazy.

I won't attempt to do calibration myself as I don't think I have the equipment to make better than it is. If I decide to calibrate then I will bring it to one of the labs in the neighborhood - Tektronix campus is just 20 minutes away. Don't know Tektronix prices, but one of the labs further out does basic calibration of 87V for $65, so not crazy.

I am just looking for opinions and suggestions about possible faults to look for before I commit to calibration.

As a side note, the 189 more pleasant to use than 289, because the measurement is faster (the settling of reading).

[bdunham7]

I'd say that your 189 has a significant error, but not a gross error that would necessarily mean something is broken.  Thus you could have a calibration problem or a something broken problem.  Normal drift should not be that great.

An important step is to determine if the error is stable over time or not.  If the error is completely consistent over time and small temperature changes, then it is possible that someone has miscalibrated the mVDC range, or that data has been corrupted.  I would think the former much more likely.  The mVDC range calibration affects all other ranges proportionally. 

I noticed that your measurements of mVDC seem to not be as consistent as I'd like to see, but if you think you can dial in your source (with good stability) to the voltages specified in the 189 calibration manual, then go ahead and just do the mVDC calibration and then take it out of calibration mode and see how all of the other ranges look.  Using your F289 as a comparison is probably good enough for now.  If you can get it reasonably close and it stays steady for 60-90 days of use, then you can send it out for calibration. 

 

[rcross]

The mVDC range calibration affects all other ranges proportionally.

Oh, interesting. I did not realize that. Re-reading clibration manual more carefully I see that note on page 22. Thank you for pointing this out.

FWIW, the calibration seal is intact on this one. Of course, there are ways around that for really industrious, but I am inclined to believe that it has not been calibrated.

[Kleinstein]

Before sending it out for calibration is a good idea to check the stability for a few months, maybe a year. There is little use in gettting it adjusted and calibrated when it is drifting and before knowing that it does not drift even after calibration it is of limited value.

Normal calibration with such meters is just a check and does not include adjustement. So one would need the adjustment as an extra point to order (may cost extra).

[bdunham7]

FWIW, the calibration seal is intact on this one. Of course, there are ways around that for really industrious, but I am inclined to believe that it has not been calibrated.

Do you mean the seal over the button?  IIRC you can just take the back off?  It has been a while since I looked at one.

If it has not been calibrated but has drifted, calibration will be a waste of money.  I have heard of one other 189 that appeared to be drifting, so it seems possible but not common.  It would probably be best to monitor it for 90 days or so.

Aditionally, you can find the AD REF43 voltage reference and measure its output directly.  If it is about 1% high, (2.525V) then that is likely your problem.

[rcross]

Aditionally, you can find the AD REF43 voltage reference and measure its output directly.  If it is about 1% high, (2.525V) then that is likely your problem.

The reference sits at 2.5010, which is beyond the accuracy of my meter, so unlikely the source of error.

[bdunham7]

Aditionally, you can find the AD REF43 voltage reference and measure its output directly.  If it is about 1% high, (2.525V) then that is likely your problem.

The reference sits at 2.5010, which is beyond the accuracy of my meter, so unlikely the source of error.

Presuming you measured that with your 289, I agree that appears to be well within the spec of the REF43.  Unfortunately, I can't think of any other faults that could be repaired practically, at least not without a parts mule.  So the only thing I can think of is to calibrate the mVDC range (if you can do it reasonably well) and then monitor it for a while to see if it is stable.   

[rcross]

Thank you for all the guidance and help. I will take the long route and see what happens.

[Kleinstein]

The resistance ranges also read low - this would not happen with the reference drifting.

[bdunham7]

The resistance ranges also read low - this would not happen with the reference drifting.

Why not?

[Kleinstein]

The driving current usually is derived from the same reference (at least this is the way it should be - otherwise one would need a good 2nd reference and still get more drift). In this case a change in the ref. voltage would have no effect on the resistance readings.

[luudee]

I see you (OP) are in the US.  I don't know how much calibration costs in
the US, but I just paid some odd $65 USD to get my Fluke 289 calibrated.
(I am in Thailand, most things are much cheaper here)

I thought $65 USD was quite worth the calibration .... I too replaced the
super cap in my Fluke.

The errors appear to be small enough to be calibration out of date, or all
calibration data has been lost somehow ....

Good Luck!

Cheers,
luudee

[rcross]

I will research the calibration topic more extensively, but a lab nearby lists handheld fluke multimeters (such as 87V) for $65. I do not know yet whether that entails full Fluke calibration procedure, or just check for OOT and PASS/FAIL answer. I will be calling Tektronix lab today to find more.

[bdunham7]

The driving current usually is derived from the same reference (at least this is the way it should be - otherwise one would need a good 2nd reference and still get more drift).

That's certainly the way most bench meters seem to work, but I'm not sure that these handhelds do the same.  If you're right, then this would be a helpful diagnostic tool for spotting reference drift that in this case would seem to favor the mVDC calibration being the issue.

[rcross]

To close the loop on lab calibration cost. I got a quote from Tektronix for 189 calibration (ANSI/ NCSL Z540.1 as described here https://www.tek.com/en/services/calibration-services/service-levels). It includes adjustment + minor repairs. The calibration is $115 + $20 handling + $60 shipping. Since the lab is near me it would cost me $115 to drop off and pick up. Cost with shipping is less palatable.

I think will do hacky mVDC calibration using 289 as a reference and if it does not drift, then will do the Tektronix calibration.