LCR tryin to dig up pF's

[travLTTL]

First post so bear with me on this one.
   Ive been tasked to measure a contact capacity. We are talkin small! .05pF small. . . Ive been given a Keysight U1733 here is the product datasheet and a Agilent U1733C (screenprint is the only difference). Both are in calibration and
    I simply don't trust the data I'm pulling with this meter, as its all over the shop from unit to unit and never get the same measurement twice on the same unit. I can get a bang on reading from a 1pF and a .1pF cap every time.. I have a controlled environment lab... we even tried turning off the lights and putting the thing in a lead box . .
    Any kind of input would be great as I'm being told that all previous testing was fine and that I must be doing something incorrectly.

      Thanks in advance,
               travLTTL

[nctnico]

Are you using a 4 wire measurement method? What kind of accuracy do you need?

[alm]

What is your fixture like? Any variation in position of the wiring could account for a 0.05 pF change in stray capacitance. Does switching between a 10 kHz and 100 kHz test frequency make a difference? Keep in mind that a zero reading on the 20 pF range is allowed 0.01 pF or 0.02 pF offset, so reading a value of 0.05 pF as 0.03 pF should be considered in specification. You are definitely operating at the absolute limit of what the instrument is capable of.

[HighVoltage]

For such small capacitance to be measured correctly, you need to have an instrument that can set the compensation for the test leads to zero, while they are not connected to the DUT. But keep them very closed to the DUT. Only when the instrument is showing 0 connect the DUT by only moving the test leads clips very slightly. This might give you a stable reading.

I use this method with a Fluke PM6306 for anything below 1 pF.
But then, I never had to measure anything accurately below 0.1 pF

[travLTTL]

Im using a pair of these that came with the meter. Im not sure if these are four wire or not, I need a little more knowledge in what exactly that entails (off to find a rabbit hole). We also purchased the cal cert with both meters.

[travLTTL]

Im not using a fixture, but the terminals are floating in open air; not touching anything and is in cased in a can or housing that is grounded/shielded from all internals.

[alm]

What kind of readings do you get when you move the tweezers close, but not on the contacts? What if you move your hand along the cable or near the tweezers?

[travLTTL]

There is certainly a "Null" feature included with the meter and I am zero'ing the instrument. Also the meter seems to float around slowly rising or falling in its reading when the DUT is being looked at and when the leads are 'floating' in air... Shorting the leads together immediately O.L.'s the thing and returns immediately to its Null'd state and can be used again. Ive also done the same between the terminals with a piece of solder to try and clear any charge that may be residual. (Let me know if Im way off base on this).

   Also thanks to everyone who has and is looking at this post. Love the Blog and this is my first outing in to the forum.

[travLTTL]

when the tweezers close I get the O.L. which should be expected. Nothing in regards to the cable affects the readings.

[alm]

I was talking about moving the tweezers close to the terminals, not short them (which is pointless for a capacitance measurement).

I imagine you are zeroing it, but how stable is this zero value as you move around as you may during a measurement?

Not sure what the meter drifting represents. It uses AC, so charge should not be a factor. Capacitance is mostly a function of geometry (think the classical two parallel plates), which is why I was asking about the variation in open circuit reading. I imagine moisture in the air will affect the dielectric constant, so breathing on it may also very well change the value.

Can you reproduce the drift in open circuit reading if you just lay the tweezers on the bench without any movement or people nearby? Note that the meter is allowed to drift up to 0.02 pF for a zero reading, depending on the test frequency.

[TheSteve]

Are you performing the open/short cal with the tweezers connected and what frequency are you using? I have used my u1733c to measure 1.5 pF caps, it reads high but it is pretty consistent.

[floobydust]

Oh the SMD tweezers use a guard terminal.

Perhaps your bench/test fixture is too noisy. Cell phones, mains power cords have to be away. The bench surface, if metal needs to be grounded.
The guard terminal might have to connect to your DUT, I don't know how big it is or stray capacitance with it.

Try close the contacts on a thin sheet of paper or plastic, to increase their capacitance and see if the meter settles down.

If it's a relay, the coil excitation power supply might be adding too much noise.

[WastelandTek]

I call this "chasing Heisenberg out of the shack".

I do not have anything productive to add as I have no experience trying to measure capacitance at this level.  I am just posting to sub as I think this is going to be interesting.