Direct LCR Meter Connect Test Subjects

[mawyatt]

Per request we've created this, hopefully it's not repeating something we've shown elsewhere.

This is more about LCR meter confidence and certainly not metrology grade, rather than resorting to expensive calibration which is necessary for actual client use, and will be applied when necessary.


It's about creating a direct BNC connection for reference components (mostly capacitors at this time) to a quality bench LCR meter. The general idea is to remove most of the uncertainty in LCR measurements attributed to the clips, cables, connectors and fixtures. Our goal was/is to show repeatability over time that will allow more confidence in LCR instrument behavior for our home lab instruments (Hioki IM3536 and Tonghui TH2830).

The idea starts with using 4 direct BNC connectors to the LCR meter Lc, Lp, Hp, and Hc terminals. A 3D printed container box with separate front and back cover and a rectangular "tube" to house the DUT within. The covers are attached with 3mm screws and all the "insides" of the covers and tubes covered with copper foil to create a somewhat Faraday cage for the DUT to reside within. This provides an effective E-Field shield which has shown to eliminate the effects of local distributes such as hands moving around the DUT when setting the LCR meter parameters, it's not an H-Field shield tho!!

To facilitate good Open and Short Calibrations, we created a dedicated Open and Short fixture as well. The Short uses a thick continuous copper shunt across the 4 BNC terminals.

Anyway, here's the actual fixtures showing the Open, Short, and various reference capacitors. All the caps are select mica except the 0.1uF and 1uF which are select PP film types. The special purple 10nF has a 5kΩ thermistor attached to the Soviet metal case for remote temp readings.

Best

[mawyatt]

The setup and Short Fixture.

Best

[KungFuJosh]

Very nice! Thanks for sharing.

[coromonadalix]

have you tried copper paint in spray to see if it gives good shielding properties ?

[mawyatt]

If we were producing these then the spray paint would be in order, as it's much quicker than putting the tape down, however paint is also a little messy (the good stuff is expensive) and the tape seemed OK for our needs. The copper tape (or paint) provides a good E-Field shield (doesn't look pretty tho) which is the intent, however it's not a good H-Field shield.

You can move your hands around the reference cap fixture (even the 2pF cap) and this has no effect on the LCR readings, without the fixture this can influence the readings.

Best

[jogri]

I'm working a lot with impedance analysers and had to built quite a number of test fixtures/cal fixtures for them, my personal go-to are junction boxes. Get one that fits the BNC spacing of your particular impedance analyser, drill four holes into the side, screw in the BNC plugs, done.

(Learned that trick from Novocontrol, their calibration sets use the exact same boxes)

I used a lot of those in the past, but they might be a bit short for the typical BNC spacing so you might need a slightly longer box to fit all 4 BNCs:
https://no.rs-online.com/web/p/junction-boxes/2539525?gb=s

Usually for labelling i get my 24mm label printer and just stick the label on the lid.

[graybeard]

Nice work!

[TERRA Operative]

I did a similar thing to the Hammond box, but I used Takachi brand die-cast aluminium enclosures instead.

These are resistance standards I made (and I made a BNC-banana plug adapter for my multimeter for direct connection), but I haven't made capacitor standards yet, as I have no way to measure them except on my uncalibrated LCR meters. I'm in a chicken and egg scenareo. I need a calibrated meter to check my standards that I need to calibrate the meter....






Does anyone knoe how to use not an LCR meter to precisely measure a capacitor?
I have all the usual test gear at my disposal...

[eliocor]

VNA?

[TERRA Operative]

I have a Siglent VNA, but I've never tried to use it for measuring capacitance. I'll have to do some research..

[mawyatt]

Depending on what size capacitance your reference (small) a VNA may be a good choice. However, don't think you can achieve the level of precision with a VNA that a lab grade LCR meter can produce at lower frequencies, especially so for larger values of capacitance.

We have two lab grade LCR meters (TH2830 and IM3536) purchased new for this reason, and we get very good agreement between the two on various DIY home brewed capacitance references we've created.

Aged Mica are for the lower values, from 2pf up to ~10nF, and aged PP for higher values up to 1uF which is the limit of our DIY references.

If you want we can measure your reference caps with both our LCR meters for comparisons.

Best 

[jogri]

If you have known resistors you can measure it via the time constant in a RC circuit with a function generator and some way to measure phase angle (scope or whatnot). Cheap trick to measure capacitances of liquids since you're usually dealing with pF, just slap a decent sized resistor in parallel with the capacitor.

Also if you're doing anything with a VNA do keep in mind that capacitance is a function of frequency (dielectric losses), so in my opinion if you're going for steady-state near DC values (kHz and the like) everything above maybe 50-100 MHz may already give you drastically different values.

[mawyatt]

Continuing from over here:

https://www.eevblog.com/forum/testgear/genrad-decade-inductance-capacitance-and-resistance-boxes/msg6280460/#msg6280460

Some good discussions over there but deviated from the OP topic, so bringing it back here. Please visit there, then return for a better overview.

We are considering offering these direct BNC capacitive references with select Mica capacitors as kits if anyone is interested. We've done enough test/evaluations to know they are quite stable and repeatable, so comfortable to let others experience the benefits of such. With the shipping/tarrifs uncertainty and such the expected cost for most cap references is hard to predict, but should fall between 30~$40.

One advantage of this Direct BNC Connect fixture not obvious is the Short and Open Calibration achievable, these represent a close environment (mechanical and electrical) the DUT "sees" from within the fixture, and produce better Short Open references than can be achieved with Kelvin Clips.

We can provide more details and images if folks are interested as the results we've achieved have been impressive

Best