Keithley DMM6500 question 2w resistance behavior

[core23]

Info for anyone searching for Keithley DMM6500:
Problem identified, it was not caused by the DMM6500 
Reason was -> User / cable failure
If you are still interested, read on... 
Hello forum,


I am following this forum for a few years but never had any content to share, so never registered.

Background:
Since about 5 years i wanted to get hold of a decent bench DMM, but never bought one. In summer this year i discovered the Keithley DMM 6500 and was collecting every information i could get.
Finally in october, I decided to buy.

So, here I am, proud owner of a factory new Keithley DMM6500!

To give you guys a short info about my level of experience: I'm more like a type of hobbyist just having fun with equipment, most time not producing or creating anything. .

I've been playing around a bit and I'm really pleased with the DMM, except my confusion with the following:

The attached screenshots show a measurement with 2w resistance function, DMM was allowed to warm up for > 1 hour.
The measured resistor is a Dale, 1ohm, 1%, 25w  (4wr measurement result is ~1.0003 ohm).
The measurement was not zeroed / referenced to the test lead resistance.

I was expecting noise because of 2 wire (and the uncertainty of 100mOhm as of specs), but why the downwards drift and those spikes? (Equipment was not touched during measurement, no other devices nearby turned on or off).

I have nothing comparable in terms of resolution or accuracy (only 2 pcs Uni-T handhelds), so I don't know if this is normal.

Is this supposed to show such behavior for 2wr?

[Vgkid]

I have never done low resistance readings with  a 2 wire setup. But the high temperature coeficcient of copper makes your results believable.
What does the graph look like as a 4 wire setup?

[core23]

Could you explain a little bit more regarding the spikes (I want to understand this behavior)?

The intention for my post was not to discredit the DMM, but to understand the behavior / graph of the 2wr measurment.
Regarding temperature: The measurement duration was ~30 minutes where the temp did not really change (+- 0,3 °C).
I did not (and am unable to) measure the temperature of the junction / connectors, but they are sitting in the same room for the last week 
I also made measurements with 5 minutes duration (find screenshots attached), the peaks were also present.

As i am no where near average, nor expert, my question is: what effects cause this?

For clarification: I don't ever need 2wr measurment, but because I am a "noob", i want to be shure that the error is on my behalf, and not on the gear 

Attachments show a ~ 5 minute measurment 2wr and the requested 4wr measurment of the same resistor.

Correction:
I attached the wrong 4wr measurments, they show a 0.1 ohm Dale 1% resistor.
Additional info regarding attachments:
1st image show 2wr reading ~5 minutes of 1 ohm dale resistor
2nd & 3rd image show 4wr measurment of 0.1 ohm dale resistor, for requested 1 ohm 4wr measurment see next post.

[core23]

Find the 1ohm 4wr measurment attached (not same duration, but should show requested data)

[lukier]

First of all, as others said don't measure low ohm resistors with anything other than 4 wire and OCOMP.

With 2 wire you've made a thermometer with all this copper (and probably an antenna as well). Note the difference in standard deviation in your measurements.

The spikes seem quite regular, so either your AC or heating is kicking in or maybe it is because your FILTER settings.

Here I attach short measurements of VHP4T 1R0000 0.1%, both 2W and 4W, NPLC 5, AZ ON, OCOMP ON (in 4W), LINESYNC ON, no filters!

[core23]

First: Thank you all for your input

I have no intetion to do low ohm 2 wire measurment anyways, just want to understand the results.

Comparing your 2w measurment with mine, your graph show no spikes, that is my main confusion.

For your input regarding AC or heating:
Here: No AC present, heating is by underfloor heating (ceramic tiles). If the heating would "kick in", it would take about 4 hours to raise the ambient temp (passive mass, as i am no native english speaker im not sure about the correct wording).
These spikes show up only in 2w measurment. I made voltage and / or 4wr measurments > 8 hours, showing no spikes (but as expected noise in uV/mOhm range).

The downwards or upwards drift would be ok, but what about the hard spikes?
I am nowherer neat understanding this, therefore I posted here to request help  and / or answers to this result.

I'd have no problem if it is my fault, i would appreciate it. But that also means im am dependent on you guys to tell me if I am wrong or right

[core23]

Here another measurment:
As stated in Keithley docs: "...for 2 wire measurment take relative offset and add 100mOhm..."

I tried to measure relative offset with the leads, but as the measurment was drifting, I waited for the reading to settle. See attached SC (leads shorted for ~24 minutes):

I understand the upwards and downwards drift, but not the spikes (could other DMM6500 owners reproduce this?).
Addition:
My setup description was not clear: For this measurment i shorted both test leads.

[iMo]

Those pulses are pretty clean relaxation oscillations, with almost no sign of measuring something real (except maybe the first 2.5mins in your last picture).

It does not seem it comes from outside, unless there is a pretty strong EMI source with ~1.5minute pulse activity period, which makes your DMM totally crazy.

Try to measure a 100-1000ohm 2w (or similar resistor from your junk-box), what will you get.

[core23]

I'll be back home tomorrow and make the suggested measurments in the > 100 ohm range.

Attached you find the same resistor, measured with the rear terminals (same day, same environmental conditions).
It is different to the front terminals and is showing no peaks (this is the graph i was expecting initially).

Could you guys point me in the right direction to understand this effect? Is this some sort of capacitance build up?

[core23]

Ok, so now I'm feeling really stupid:
I am unable to reproduce this   

Same setup, same settings, was trying for hours but unable to catch those spikes
Most probably one of those "error in front of device" situations, I would really like to understand what I did or what was causing this 

If i can reproduce it I will report back, presumably with some "stupid me" kind of thing... (or one of my neighbours was playing with high power EM gear )

Thank you guys for sharing your time!

[Kleinstein]

Those spikes could be from a poor contact in some from. This could be a bad cable or maybe oxidized switch / relay contact inside the instrument. Just using the switch / relay a few times could have cleaned the contacts.

The Keithley meters also use quite some numerical compensation of the thermal effects. So in times when the internal temperature changes a lot there can be some odd jumps in the measured curve, like in the first few minutes after turn on. The compensation usually works quite well to reduce the needed warm up time, but it causes some odd looking jumps in the very beginning, when there is no good reading anyway.

[core23]

@Kleinstein: Thank you, I think that is/was the problem!

Follow-up to my mystery measurements:
After testing around to reproduce the peaks in the infamous 2w low ohm measurements I probably found the reason:

• me - not double and triple check everything
• you may have guessed it: the cabling

For the measurements I used one of 4 pairs of coax to clip / grabber type (don't know how these are called exactly) cables, used them since few years for my DSO, never had problems (no DIY, but no brand visible so I am unable to tell you guys the brand/origin).
Edit: Just re-read this and think this could be misunderstood: I am not referring to DSO probes, just the common cables COAX to Grabber/Hook type.

My failure was to not lable them, like "1", "2", ...
Today one clip/grabber of one of the cables came loose (solder joint failure), so I assume when I did my inital measurements few days ago I used that cable - and after "freaking out" because of the results I grabbed a similar, but not the same, cable the next day and was not able to reproduce the spikes.

That is no where near a proof for the peaks, but the best reason i can find.
Perhaps one of you guys could explain the effects of a "beginning to fail solder joint" related to the peaks in my 2WR measurements?
Don't hate me for this: but could it be some form of capacitance build up between the solder joint and the "grabber" copper plate (partial air gap because of failing joint)?
Would this explain the fast rising edge and slow fall afterwards?