Author Topic: Unexpected LCR meter behavior change - dependent on measurement lead positioning  (Read 53312 times)

0 Members and 1 Guest are viewing this topic.

Offline maelhTopic starter

  • Regular Contributor
  • *
  • Posts: 97
  • Country: de
Hello,

I have a Peaktech 2170 LCR meter, which comes with Kelvin clips/leads.

I am measuring an inductor in the 100-200nH range (measured at 100kHz), and I noticed wildly changing values, between 40nH and 600nH. It seemed erratic until I found a way to reproduce it.

When I hold the clips in a steady position (so that the inductor coil doesn't get squished/stretched, and the connection from the clip to the DUT remains stable, too), but I move around only the leads going to the LCR meter, the measured inductance changes a lot.

When I keep the leads in a relatively stable position and recalibrate in this position, then the measurement is in the right ballpark. But if I reposition the leads the measurement is off again.

Is that normal behavior or is that a sign there is an issue with the leads?

If it is an issue with the leads, is there a test I can perform on the leads themselves to find out if maybe some strands are broken or something else is wrong (for example using a signal generator and a scope)?
« Last Edit: January 11, 2024, 02:53:05 am by maelh »
 

Online KungFuJosh

  • Super Contributor
  • ***
  • Posts: 3166
  • Country: us
  • TEAS is real.
Sounds like the leads to me. Can you unscrew the covers over the plugs and inspect the wire connections?
"Right now I’m having amnesia and déjà vu at the same time. I think I’ve forgotten this before." - Steven Wright
 

Offline floobydust

  • Super Contributor
  • ***
  • Posts: 7678
  • Country: ca
Try shielding the inductor, it might be acting like an antenna for something. Keep your cell phone, cordless phone, WiFi, iPad etc. away from the test setup.
 

Offline TERRA Operative

  • Super Contributor
  • ***
  • Posts: 3168
  • Country: jp
  • Voider of warranties
    • Near Far Media Youtube
Are the leads shielded?

If not, it might be worth replacing the test clip cables with thin coax and terminate the meter end of the shield to a banana jack to plug into the guard terminals.
Where does all this test equipment keep coming from?!?

https://www.youtube.com/NearFarMedia/
 

Online Kleinstein

  • Super Contributor
  • ***
  • Posts: 15154
  • Country: de
The inductance is tiny and the shield is only effective against the electric field, not much against a magnetic field. Even with Kelvin probing there is magnetic coupling between the wires.
So the behaviour is probably normal. To measure such small inductance it needs a more ridgit setup, more like the SMD part fixtures and than compare to a short.


Probably not a factir here, but a possible problem with higher value inductors:  ferrite material, especially that with high permebility reacts to mechanical stress. This is especially a factor for inductors with a closed core and thus a large dependence on the µr.
 

Offline johansen

  • Super Contributor
  • ***
  • Posts: 1181
Are the kelvin clips gold plated steel?
 

Offline maelhTopic starter

  • Regular Contributor
  • *
  • Posts: 97
  • Country: de
To answer some of the questions: I tried the LCR meter in various rooms, so I don't think it's radio waves / noise that is being picked up, but that is of course hard to fully rule out.

The leads do not seem to be shielded, they look like banana plug cables you would have in your benchtop powersupply.





The test clip jaws are isolated from each other, in the sense that if you open the jaws of a Kelvin clip, the two parts of the jaw are electrically isolated and each go to an individual lead (I tested the jaw parts are correctly isolated when spread open with a multimeter). I checked the manual again fully, and it does not mention if the leads are shielded, their visual aspect makes me assume they are not.

I thought about using other banana plug cables to compare measurements to the Kelvin clip leads, but I wonder what the purpose of the guard plug is, and how I would replicate its effect with individual leads.
 

Offline maelhTopic starter

  • Regular Contributor
  • *
  • Posts: 97
  • Country: de
Sounds like the leads to me. Can you unscrew the covers over the plugs and inspect the wire connections?
It seems to be soldered in place, disassembling would likely damage the leads/clips. But a resistance test returns values in the 0.01 to "0.00" Ohm domain (tested with a cheap multimeter, my main one is currently not available).

Are the kelvin clips gold plated steel?
Hard to say, it's not specified in the manual, but I attached an image in the post above.

Are the leads shielded?

If not, it might be worth replacing the test clip cables with thin coax and terminate the meter end of the shield to a banana jack to plug into the guard terminals.
I am unsure if they are shielded, see my post above (manual doesn't specify). I wanted to make a test setup with banana plug cables, but not sure how to attach them correctly and where the guard cable would come into play in this setup.

I assume your approach would mean I need to make my own cable out of thin coax and then solder banana plugs onto one end, and aligator clips on the other end? It would become a 2-wire measurement setup only though, so I am not sure it would be compatible with my 4-wire (plus guard cable) LCR meter.
« Last Edit: January 11, 2024, 10:25:53 am by maelh »
 

Online Kleinstein

  • Super Contributor
  • ***
  • Posts: 15154
  • Country: de
The leads look like not shielded.  Anyway the shield would only be effective for capacitors (especially small ones) and maybe a little at the very large inductor end.

For such small inductors one should build a kind of jig with shorter cables, or maybe a small PCB with directly mounted banana plugs. The position of the cable relative to each other can make a difference. Beside the normal inductance from the part there is a coupling (similar to a transformer) effect from the cables. These small inductance values don't simply add up like resistors - there is inductance depending on the loop area and the return path thus also matters. So it is not even easy to define the inductance exactly in the this low range.
 

Online iMo

  • Super Contributor
  • ***
  • Posts: 5570
  • Country: va
The inductors in the 100nH-200nH range are measured with special meters, at higher frequencies, or with VNA (like nanovna) with a special jig.
FYI - 100mm of a straight wire is aprox 100nH..
Readers discretion is advised..
 

Offline TERRA Operative

  • Super Contributor
  • ***
  • Posts: 3168
  • Country: jp
  • Voider of warranties
    • Near Far Media Youtube
Are the leads shielded?

If not, it might be worth replacing the test clip cables with thin coax and terminate the meter end of the shield to a banana jack to plug into the guard terminals.
I am unsure if they are shielded, see my post above (manual doesn't specify). I wanted to make a test setup with banana plug cables, but not sure how to attach them correctly and where the guard cable would come into play in this setup.

I assume your approach would mean I need to make my own cable out of thin coax and then solder banana plugs onto one end, and aligator clips on the other end? It would become a 2-wire measurement setup only though, so I am not sure it would be compatible with my 4-wire (plus guard cable) LCR meter.

Nah, use four lengths of coax, one coax to each side of each kelvin clip, four banana plugs terminated to each center conductor at the meter end of the cables and then tie the shields together per kelvin clip, using two banana plugs for the shields, one per kelvin clip and plug both in to their respective guard terminals.

The shields stay unterminated at the kelvin clip end.

Using dual banana plugs for the signal wires would make a neat arrangement less prone to tangling too.
« Last Edit: January 11, 2024, 11:24:21 am by TERRA Operative »
Where does all this test equipment keep coming from?!?

https://www.youtube.com/NearFarMedia/
 

Online chilternview

  • Regular Contributor
  • *
  • Posts: 198
  • Country: gb

Your meter spec says it has an accurary of +/- 1% on its 20uH range at 100kHz. That's 200nH.
Although the meter resolution may be 1nH on that range, the last 3 digits are likely to be noise and/or test lead effects.

Consider the impedance you're measuring - a 200nH inductor has  Z = jwL = 2*pi*1e5*2e-9 = 0.001 ohms. As iMO says, you need to be measuring it at a frequency at least 2 orders of magnitude greater, using something like a VNA.
 

Offline maelhTopic starter

  • Regular Contributor
  • *
  • Posts: 97
  • Country: de
Thanks all for your insightful replies.

I actually do have a NanoVNA and was trying to validate the measurements of the LCR against the VNA.

Another issue is that when using Kelvin leads to measure, you can put a coil into plated throughholes, then attach the ends of a coil to a meter, and fine tune it the coil (by squishing/expanding it). With the NanoVNA you need to keep it in the test fixture, and cannot measure it in the final board.

 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf