LCR meter test parameters Cp-Rp help.

[Jeremy M]

I have an old HP 4263A LCR meter that does not support Cp-Rp it does support Cp-D.
My question is what is the difference?  And do I need a new LCR meter?

We have a customer complaining about a part we make not passing their test.
They are using a B&K Precision 894 that meter is $3800 new. 

[TimFox]

With the parallel Rp Cp model, Q = Rp / Xp, where Xp is the reactance of Cp at the measurement frequency.  D = 1 / Q.

[bob91343]

The math works but it's nice to have a meter that reads directly, such as my GR1658.  If the frequency is high, the nanoVNA works but isn't especially accurate.

[Kosmic]

I have an old HP 4263A LCR meter that does not support Cp-Rp it does support Cp-D.
My question is what is the difference?  And do I need a new LCR meter?

Rp = parallel resistance
D = dissipation factor: "Loss tangent. The ratio of energy lost to energy stored; the reciprocal of Q"

A mode like "Cp-Rp" measure 2 parameters (parallel capacitance and parallel resistance) at the same time. But you can probably individually measure Cp and Rp with your HP 4263A. Putting the instrument in R-X mode should give you Rp I guess.

Some documentation:
http://www.techni-tool.com/site/ARTICLE_LIBRARY/BK%20Precision%20-%20How%20to%20Use%20an%20LCR%20Meter.pdf
https://www.ietlabs.com/pdf/application_notes/030122%20IET%20LCR%20PRIMER%201st%20Edition.pdf
https://www.keysight.com/main/redirector.jspx?action=ref&cname=EDITORIAL&ckey=722686&lc=fre&cc=CA&nfr=-536900197.536897554.00

[Kosmic]

Funny thing is that the 4263B can do Cp-Rp. Hmm maybe it's not possible directly with the 4263A but you can always calculate the parameter yourself.

[Jeremy M]

I appreciate all the help of everyone here.  I have never used an LCR meter before and really don't understand all the math just yet.
Yes, a meter that would be capable of reading Cp-Rp simultaneously would help.  I have the step by step instructions using a B&K Precision 894.

I may need a new LCR meter.

[Jeremy M]

The math works but it's nice to have a meter that reads directly, such as my GR1658.  If the frequency is high, the nanoVNA works but isn't especially accurate.

What brand is your meter?

[Kosmic]

What are you trying to measure exactly ?

Depending of the DUT (device under test) R might be really close to Rp (meaning even if you measure Rs and not Rp the difference between the 2 might be negligible).

[TimFox]

What are you trying to measure exactly ?

Depending of the DUT (device under test) R might be really close to Rp (meaning even if you measure Rs and not Rp the difference between the 2 might be negligible).

If Q is high, say > 3, Rs is much less than Rp, but Cs is approximately the same as Cp.  Q is the same for both series and parallel models at the single measurement frequency.

[Tomorokoshi]

It looks like the 4263A has Cp-G mode. G = 1/R, so you can calculate Rp as 1/G.

If you have some example capacitors, measure them and list the results by manufacturer, value, type, etc. Those results can be compared to similar ones to partially validate your measurements.

Make sure the measurement frequency and DC bias match what your customer is doing. You may also want to do the short and open corrections.

Otherwise that unit should have more than enough capability.

[Jeremy M]

It's a screen printed flexible circuit board.  We are one of 3 vendors supplying them and the only one having issues.
Visually our parts look good and the Ohms reading on the traces are lower than our competitors, but within specs (this may be an issue).
We are all supposed to be using the same inks and materials yet only our parts are failing according to the customer.
   

[Jeremy M]

It looks like the 4263A has Cp-G mode. G = 1/R, so you can calculate Rp as 1/G.

If you have some example capacitors, measure them and list the results by manufacturer, value, type, etc. Those results can be compared to similar ones to partially validate your measurements.

Make sure the measurement frequency and DC bias match what your customer is doing. You may also want to do the short and open corrections.

Otherwise that unit should have more than enough capability.

If this works I can just recalculate Rp to G and just use the meters reading.
I will look into this thank you.

[Tomorokoshi]

Customer acceptance tests with multiple vendors. Not fun.

1. Do they tell you which parameter is failing and in which direction?

2. Do you have examples of the alternate parts?

[Jeremy M]

Customer acceptance tests with multiple vendors. Not fun.

1. Do they tell you which parameter is failing and in which direction?

2. Do you have examples of the alternate parts?

Yes, on both counts.

As you can see below I thought were doing fine.  Until they connect it to their circuit and an LCR meter.
the Rp falls below their spec for the completed circuit and the alternate parts from the other vendors pass and have a higher ohms reading.
But we are buying all the same materials from the same vendors not sure why we are falling short.

Spec                 Actual
<13 ohms            4     -   4.5    ohms
<37 ohms          11.15 - 12.64   ohms
<10 ohms            1.68 -   1.82   ohms

I'm thinking of just getting a better LCR meter with a Cp-Rp simultaneous read out so we can do the exact same test as the customer.

[bob91343]

My GR1658 is made by, surprisingly, GR, or General Radio.  At one time they were the premier test equipment company, after Weston and before HP.  The brand still exists.  Do a google on it.

They made many models.  I also have a GR1650A which is a manual bridge.  And a 1644 which is a megohm bridge, very useful for electrolytic capacitor testing.

The 1658 is digital and displays either series or parallel readings to several places, at either of two frequencies.  I use it frequently.

[TimFox]

Customer acceptance tests with multiple vendors. Not fun.

1. Do they tell you which parameter is failing and in which direction?

2. Do you have examples of the alternate parts?

Yes, on both counts.

As you can see below I thought were doing fine.  Until they connect it to their circuit and an LCR meter.
the Rp falls below their spec for the completed circuit and the alternate parts from the other vendors pass and have a higher ohms reading.
But we are buying all the same materials from the same vendors not sure why we are falling short.

Spec                 Actual
<13 ohms            4     -   4.5    ohms
<37 ohms          11.15 - 12.64   ohms
<10 ohms            1.68 -   1.82   ohms

I'm thinking of just getting a better LCR meter with a Cp-Rp simultaneous read out so we can do the exact same test as the customer.

These numbers seem absurdly low for Rp.  Are you sure this isn’t Rs?  Also, as worded here, you pass the spec since your “actual” values are, in fact, less then the quoted specs.  Normally, Rp would have a minimum spec and Rs would have a maximum spec, e.g. < 10 ohms.

[Jeremy M]

Customer acceptance tests with multiple vendors. Not fun.

1. Do they tell you which parameter is failing and in which direction?

2. Do you have examples of the alternate parts?

Yes, on both counts.

As you can see below I thought were doing fine.  Until they connect it to their circuit and an LCR meter.
the Rp falls below their spec for the completed circuit and the alternate parts from the other vendors pass and have a higher ohms reading.
But we are buying all the same materials from the same vendors not sure why we are falling short.

Spec                 Actual
<13 ohms            4     -   4.5    ohms
<37 ohms          11.15 - 12.64   ohms
<10 ohms            1.68 -   1.82   ohms

I'm thinking of just getting a better LCR meter with a Cp-Rp simultaneous read out so we can do the exact same test as the customer.

These numbers seem absurdly low for Rp.  Are you sure this isn’t Rs?  Also, as worded here, you pass the spec since your “actual” values are, in fact, less then the quoted specs.  Normally, Rp would have a minimum spec and Rs would have a maximum spec, e.g. < 10 ohms.

I would agree our parts meets specs.  However when our printed circuit is added to the customer device it fails Rp.

The spec has 5 test points (point 1) Rp >25 ohms = fail  (points 2-5) Rp > 1000 ohms = fail.

Yet, two other vendors are not having the same issue.
I have rented a B&K Precision 894 the exact same meter the customer is using and I have a golden part from the customer.
When the meter arrives I want to see the failure the customer is complaining about duplicated using their own test.

Truthfully I suspect something is not right on their end.

Update:

It turns out they interface the printed circuit with a hydro-gel acting as a capacitor.  The trouble seems to be with way hydro-gel contacts the circuit.