Fluke 8840A - 4 wire vs 2 wire accuracy

[int2str]

I just picked up a nice, used Fluke 8840A meter on eBay. Looks like I got lucky as the display is in really good shape and nice and bright. Really nice unit - like it so far. Help me understand one thing please:

Overall, the unit seems to be very well calibrated (last calibration sticker from 2012). Compared to my DMMCheck and any other reference I have, any reading is well in spec. Specifically the resistance.

For example, in 2 wire mode with the leads zeroed out, the meter reads 0.99981kOhm for the 1k reference, which according to the calibration sheet of the DMMCheck is spot on to the exact digit - nice.
However, in 4 wire mode with Kelvin clips, the 100k reference reads 100.066k vs the 100.096k of the actual reference. In 2 wire mode that same resistor reads 100.095k.

So, could it be that the meter was only calibrated for 2 wire measurements?
Is there a separate adjustment for 4 wire? Or am I missing anything else?

 

[Performa01]

2- and 4-wire ohms are separate measurements functions which require separate calibration.

In general, using 4-wire ohms measurement is not recommended for resistance values above some 10kohms, so it might well be that only the lower ranges have been calibrated.

Other than that, congratulations on getting the meter! It doesn't have any bells and whistles, but has excellent long term stability and is very convenient to operate and has a very nice display. Plus, it maintains high input impedance up to 20V, with a nice low input current <50pA, which is hard to find nowadays.

I'm sure you'll enjoy it in practical use.

[nukie]

The 8840A is one of the most stable meter I have. Maybe because of its age or perhaps it's low count digit. Unfortunately VFD is getting dim as it ages, hopefully I will be able to mod it into LED one day.

In general, using 4-wire ohms measurement is not recommended for resistance values above some 10kohms,

Can you please explain why?

[mzacharias]

Not an engineer so pardon me, but it would seem to me that at higher resistance values, such small differences are probably irrelevant even in most laboratory environments. 4-wire is mainly used for situations where even a small variance in resistance becomes important; shunts, switch and relay contacts, etc. I have an 8840A (one of my prized possessions with a NASA cal sticker) but I generally just use an 87V in high res mode and rel'd out lead resistance. It's just easier.

[Performa01]

Can you please explain why?

That’s a good question.

From a users point of view, it shouldn’t make a difference, yet some higher end meters won’t allow you to use 4-wire ohms on high resistances. For example, the 7.5 digits Keithley 2001 lets you chose the max. autorange for 4-wire from 20 ohms up to 200 kiloohms and will generally refuse to measure resistance >200k even at manual range when in 4-wire mode.

It doesn’t make sense either, as contact and lead resistances normally don’t exceed some 100 milliohms, which would cause an error of just about 1ppm at 100 kiloohms.

The actual reason for this might be in the calibration process. Some more advanced resistance standards use a small selection of precision resistors, e.g. from 100 ohms up to 10 megaohms to synthesize any desired resistance value from <1 milliohm up to 10 gigaohms. Such devices have buffered 4-wire sense outputs, thus swamping the error caused by the input current of the meter that is to be calibrated. That means that even if perfectly calibrated, the meter would still show an error proportional to the measured resistance value.

[xwarp]

1vdc/.001 ohms = 1000 amps,
1vdc/.01 ohms = 100 amps,
1vdc/100.001 ohms = .010 amps
1vdc/100.010 ohms = .00999 amps

A while back, I picked up few 8840's and an 8842a. Repaired several of the 8840's and sold them for a decent profit. Kept the 8842a for me after replacing the VFD with a really bright VFD out of one of the parts units.

I use that meter occasionally, but I have it on for at least 5-6 hours before I make any 4 wire measurements with it.