DE-5000 LCR TL-21 mod.

[mawyatt]

Nice work!!

See our earlier post (#64) where we utilized a low capacitance cable instead of RG316 to possibly reduce the variation in cable capacitance change with bending and orintation.

We developed a holder for the cables & connectors so they align with the standard LCR meter fixtures. Here's a few images.

Best

[mawyatt]

A word of caution wrt a free floating female BNC on the end of a cable like with these adapter cables/fixtures for the DE-5000. When attaching to the male counterpart sometimes the electrical pin and socket within the BNC can rotate if both connection ends can spin. Having the female end secured can prevent this and improve connection wear and tear.

Best

[coromonadalix]

i ould have used  "coax wires to bnc male"  instead of passing thru another box, at some point you add many connexions losses ... my 2 cents

[mawyatt]

The DE-5000 adapter we mentioned just above allows the use of a highly repeatable Open and Short calibration fixtures. Here's some images and also some measurements on some mica capacitors in our Direct Connect LCR Meter fixtures. These results are in excellent agreement with our lab grade bench LCR meters (TH2830 and IM3536), with speaks highly of the DE-5000 (why we have one!!!).

Best

[mawyatt]

i ould have used  "coax wires to bnc male"  instead of passing thru another box, at some point you add many connexions losses ... my 2 cents

That will certainly work and how we started out, but still have 4 clumsy coax cables that need 4 BNC barrels to be properly terminated into the fixture (the LCR meter is female, all standard LCR fixtures have male BNCs). For best results these need to be in proper order and can't be haphazardly connected.

We decided to eliminate the 4 barrels and used what direct solderable female BNCs we had available at the time and also wanted to use the low capacitance cables. Still not satisfied with the 4 clumsy cables we decided to use the box approach where each BNC is properly spaced and fixed in position. This doesn't add any additional connections and makes setup much easier and repeatable, cables moving around doesn't work as well, we know, been-there-done-that

Also this allows the use of the highly repeatable Open and Short as shown, and has shown excellent agreement with our reference mica capacitors as shown.

Anyway, we absolutely prefer this method over anything else we've tried. Of course YMMV

BTW this fixture general concept was started back in #24 by McBryce.

Best

[simcc]

Yes I think mawyat's solution above is best...all the issues he mentioned in the last few posts are relevant...that said I just wanted something easy and I don't have 3d printer...it seems to work fine if a bit fiddly.

also I checked the capacitance on the leads that come with some of these cheap aliexpress LCR bnc lead sets and it was much worse than rg316...i don't have anything flexy lying around that would be better...Mogami 2895 looks like it might be a good option though (expensive)...or rg179.

[simcc]

the TH26011CS just arrived today...I seem to be getting good results...cal went fine...and if I move the cables about at 100 kHz readings are staying stable (to +/- 0.25 pf ish)...

[simcc]

same cap direct inserted...with a recalibration first...

[mawyatt]

Yes I think mawyat's solution above is best...all the issues he mentioned in the last few posts are relevant...that said I just wanted something easy and I don't have 3d printer...it seems to work fine if a bit fiddly.

also I checked the capacitance on the leads that come with some of these cheap aliexpress LCR bnc lead sets and it was much worse than rg316...i don't have anything flexy lying around that would be better...Mogami 2895 looks like it might be a good option though (expensive)...or rg179.

Well, now you have a yearning, reason, need, excuse, requirement for a 3D printer

BTW with the fixture we mentioned, the DE-5000 has shown excellent agreement with our lab bench LCR meters over a wide range of components.

Having the same Open and Short Calibration Fixtures that work with different style LCR meters is a big benefit. Not to mention having the Reference Devices for comparisons within the same type Direct Connect fixture so the mechanical and electrical environments are identical between various meter types proves highly beneficial in reducing/eliminating variations and connection uncertainties. This shows in highly repeatable measurements and agreements with various instruments.

Best

[simcc]

small piece of 20awg direct inserted vs. using leads...

[simcc]

lol...actually I was thinking I could possibly use a 1590 box...I might try that some day if the loose cables start to annoy me...and use more appropriate cable etc.

Well, now you have a yearning, reason, need, excuse, requirement for a 3D printer

I'm pretty happy with this setup now, think it's well enough accurate for me. But thanks for the replies mawyat...cool stuff! Cheers.

[mawyatt]

Just for fun we grabbed the DE-5000 inserted the adapter and didn't even bother with an Open Short Cal since last it used the adapter fixture for this and we know this doesn't change from previous use (know this isn't advised but we are in a hurry).

No warm up, just grabbed each reference cap, inserted in DE-5000 fixture and recorded results.

Here's the results, red was just done with DE-5000 (in a hurry and took just a few minutes). Seems to be in good agreement with previous IM3536 LCR Meter readings.

Best

[simcc]

Hi mawyatt, actually slightly OT question, I noticed if the gaps between these wires change (TH26011CS) it can throw off measurements a bit...quick fix I tied them up with some string for now...should I desolder and twist them tightly, or is parallel better?

[mawyatt]

These are coaxial cables for each Force and Sense in the High and Low side LCR meter inputs. The center wire is the "active" wire while the shield becomes the BNC case, which for the DE-5000 is ground inside the TL-21.

The coaxial shields are single and not 100% shielding, so external/internal fields leak in/out, and the cable dielectric physical cross section gets changed slightly when bent, so the cable capacitance slightly changes. These contribute to the cable position, orientation & environment sensitivity, especially with low cap measurements. These are things we've known and dealt with since grad school well over 1/2 century ago when working with early Network Analyzer type measurements.

Tonghui makes some of the better LCR meters, cables and fixtures, so what you are seeing is better than some of the cheap type Kelvin clips (alto some can be pretty good, however some are not so and have really thin stingy shielding to help with flexibility, you don't know this until your measurements suffer). You can get Kelvin clips that go directly to the LCR meter with BNCs, or ones that have an extension box. This is likely because the extension box type Kelvin Clips (like you have) reduces the Kelvin cable length to the DUT before it gets combined within a much thicker and stiff cable with additional shielding going to the LCR meter. This allows finer more flexible cables to the Kelvin clips from the expansion box.

One of things we've mentioned and recommended from Hioki is to use a shielding plate below the measurements with Kelvin Clips. Connect this conducting shield plate to the Guard on the LCR meter. The works by controlling the localized fields around the DUT and Kelvin clips, improving measurement repeatability.

Another useful note is to make the Kelvin Clip Open calibration with the Kelvin Clips and cables in a similar fashion to the subsequent DUT measurement. While on LCR measurement notes, Zero Calibration with Kelvin Clips is best done with the Kelvin Clip in a Hc-Lc-Hp-Lp fashion, these notes help with High Z and Low Z DUT measurements respectively. You can find more details on both discussed here on EEVblog.

BTW are you contemplating a lab grade bench LCR meter?

Anyway, all this becomes a trade off between each of use and performance.

Best