measuring small capacities around 1 pF

[TimFox]

The 4332 is newer than the Tektronix 130, but it also measures capacitance to ground with a guard terminal.
The main thing I like about the 130 is the analog nulling of cable capacitance before the 3 pF full-scale measurement.  Does the 4332 offer this feature?

[jfphp]

Cable capacitance is eliminated by the special test leads : p. 3.1 manual
https://nscainc.com/wp-content/uploads/pdf/A_4332A.pdf

[Physikfan]

After the measurement with the HP4277 did not work out,
I would like to measure the internal capacities of these electrometers with this LCR meter:

[TimFox]

Does that 4-terminal LCR meter allow one side of the capacitor to be grounded?

[bob91343]

The nanoVNA does the job.  You can select the frequency, and should make it high to avoid error due to computational resolution.  I made a little cable with RG-58/U about two inches long with a connector to the vna and some clips.  I take a reading with the clips open then put the unknown on.  It measures to around 3 digits with resolution very good.

If the frequency is too high, the unknown's lead inductance must be factored in.  So take the open circuit measurement with the clips spaced for the unknown's size without leads.

I looked at the kit from Roman's work and it's almost as much money as a brand new nanoVNA.

[Henrik_V]

Seems like the actual application if more in the nearly static.
So I question the way to measure it with HF.
Assuming the capacity of the e-static voltmeters is constant and the (temp) and humidity is constant ..
I would  apply a voltage ramp and measure the current.
(or, since the thread starter want's to measure HV-high value resistors, apply high voltages and measure currents without the e meter )

Or add a known 10nF(+x) HV cap in parallel and ignore the 1 pf 

[jonpaul]

Bonjour  PhysicFan, just saw this.

We designed special low capacitance transformers and high voltage ignitors for arc lamps, and avainoc CRT Power supplies.
 CRTs. The electrometers are great as AC, DC or HF are all treated the same.

We have a collection of  electrometer VMs including the "milking stool" 100 kV from Sensitive Research.

A few notes please:

1/ Guarding and shielding need attention

2/ use Higher bridge F, below 10 kHz the Zo is too high. We have bridges to 10 MHz. Wayne Kerr 3245 goes to 300 kHz

3/ Network analyzers are an alternate: Our HP 4195A has a special impedance test set,  100k -  500 MHz.

4/ Surprisingly the small HP  RLC meter, 4332A, has lowest C range 3 pF FS!

Easy zero adj and guard but freq is not adjustable. In R&D and prod test we had good and repeatable test down to 1-3 pF on 4332A, 

http://prc68.com/I/HP4332.shtml

5/ We made a test fixture for the SMD partfor  HP4195A, Setup:

GEN>>HP 4195A TX>>cable>>50 Ohm term>>High of cap -[]- low of cap >> direct to HP  4195A 50 Ohm RX.

We swept 100 k - 500 MHz, and plot the resulting output on db scale.

Its of course a line of slope proportional to the C, due to the cap displacement current. From the slope one can infer the capacitance.

We calibrated with a precision low C, perhaps a 3 or 10 pF mica capacitor.

Using this technique we had repeatable measurements of 0.3 - 1.0 - 3 pF.

BUT it had BOTH electrodes floating, unsure how this can work on an electrometer with one side at chassis ground.

Hope this note solves your problem!

Just the ramblings of an old retired EE!

Bon Chance

Jon

[jonpaul]

http://mavin.com/pictures/bw/sensitive_esh_1.pdf

[Physikfan]

Ad TimFox
"Does that 4-terminal LCR meter allow one side of the capacitor to be grounded?"

I think that my two small electrostatic voltmeters can be used without grounding one side.

Ad JonPaul
"We have a collection of  electrometer VMs including the "milking stool" We have a collection of  electrometer VMs including the "milking stool" 100 kV from Sensitive Research."

Did you ever measure the insulation resistance of one of your electrometers, especially the electrometers from sensitive research?
And please, can you give some comments of your experience with and the performance of your collection of electrometers?

[TimFox]

The basic circuit of the 4263 and 4277 is the same as in the DER DE-5000, but the latter is a battery-operated plastic-cased smaller unit.  I have used it to measure grounded impedance by floating the DE-5000 on a polyethylene box.  You could try carefully floating your electrostatic voltmeter in the same way, to minimize capacitance from the “low” end to the chassis of the LCR meter.  When using a QuadTech 1920 LCR meter in my previous company’s lab (very similar to the two HP units), we had to carefully float the DUT, which was difficult for large circuit boards.

[jonpaul]

Bonjour,  Tim Fox and PhysicFan: A few notes please:

0/ Tim, the 4 term RLC meters generally are used with both sides floating but the low side terminals are near ground. For HP 4332A,  both terminals must float.

1/ The smaller EL meters indeed have two floating terminals. The large units 20-100KV and up must have a ground due to a metal case.

2/ We used the Sensitive Research Electrostatic VM in applications of measuring 12 kV Avionic CRT power supply, with 100 uA current capability. The leakage was never an issue.

3/ The large meters were for testing arc lamp ignitors up to 55 kV of RF bursts 100k - 10 MHz. Again leakage was not relevant, only capacitance.

4/ Ultra high resistance needs a special bridges and guarded electrodes. all of these apply a high test voltage, the GenRad to 1000V

We have several methods,

GenRad Megohm Bridges and meters series 1862, 1863, 1844A  https://www.ietlabs.com/pdf/Manuals/1644_im.pdf

Danbridge (Danish)  JP30A 30 kV insulation tester, up to 30 kV, 1 uA FS indications can read ~ 2-10 nA least)

Lab HVPS, with Keithley elecrometer set on current range  in series at return end of DUT.

5/ The Sensitive Research devices are excellent BUT dust, dirt film and exposure to humidity naturally degrade the leakage.

We cleaned with  99% iospropyl alcohol  every insulator before every measurement, had room humidity below 40% , used Teflon or similar insulated wires and jumpers.

6/ All electrometer voltmeters are VERY delicate, move with care, store upright and NEVER drop even a few inches! CAREFULLY LEVEL AND ZERO ADJUST BEFORE EVER MEASUREMENT!

I hope this info and experience can be interesting and benefit your work.

Bon Chance,

Jon

PS: my background.. since 1957  built tesla coils.  in TCBA since 1980.
Fascinated by vacuum systems, photomultipliers, CRTs, arc lamps.
At Lawrence Berkeley Labs, 1967.

PPS: For the highest and craziest HV on a very professional scale, See my old friend,  Greg Leyh, lod.org


 

[jonpaul]

Genrad 1862, 1863 1864 are Megohm METERS

There is a megohm bridge, 1644

[jonpaul]

Confusing that you have this thread and " HP 4277A" for the same topic, testing of electrometers.

The moderator should combine these two topics  into one eg " HV Electrometer capacitance and leakage"

Jon

[TimFox]

The "low current" terminal on four-terminal meters is usually a virtual ground (op amp inverting input), so it is near ground.  I don't believe it is wise to short it to ground.
I worry about the capacitance between the case of the electrometer (I have no experience with these electrostatic units) and the case of the large LCR meters.
As discussed above, there are LCR meters designed to measure capacitance to ground.

[Physikfan]

Ad jonpaul

Thank you very much for your very interesting comments on your collection of sensitive research electrometers.

For the two small electrostatic voltmeters I could
measure now finally the self-capacitances at 100 kHz,
taking into account the capacities,
due to the geometry of the measuring cables:

Trüb and Täuber, Zurich, 0 - 500 V
HP 4277A: 7.1 pF
HP 4263B: 6.2 pF

GB instrument, 0-300 V.
HP 4277A: 10.5 pF
HP 4263B: 10.0 pF

Within the experimental error there is a good agreement of the values of  these two different LCR instruments.
The way is now free for the measurement of the
insulation resistances due to observing the decrease of a short time applied voltage as a function of time.

[jonpaul]

Hallo  Physikfan

I must unearth my Sensitive Research meters and test the Cap and resistance.

They are in a 3rd sub basement!

1/ The Sensitive Research large meters are multirange, the stator can slide on a detent. different capacity per voltage range.

2/ yes 100 kHz bridge measurenets should be fine

3/ what is your exact application and voltage you shall measure that needs such high resistance?

Jon

[Physikfan]

Hallo Jonpaul

"I must unearth my Sensitive Research meters and test the Cap and resistance."
I am very interested to compare your values of capacities and insulation resistance of your electrometers with
the corresponding data of my electrometers.

"The Sensitive Research large meters are multirange, the stator can slide on a detent. different capacity per voltage range."
Your 100 kV electrometer is also a multirange instrument?
There should be also a 150kV electrometer available.

"what is your exact application and voltage you shall measure that needs such high resistance?"
My main goal is to measure the resistance, around 10**15 ohm, of teflon and polyethylene samples at 20 kV, rods and disks, in the lab and finally in the free air.
 

[jonpaul]

Dear Physikfan!

The meters are in a subbasement, and hard to access up 3 flights. Next week!

From memory: One single range in square wood box, as you have pictured probably 3 , 5 or 10 kV

One multirange in the small black metal case with side window,   5-10-15 or 5-1--20 kV

One huge,  beehive or milking stool, 4-5 ranges 50 cm dimensions with    20 cm insulator,   50 or 100 kV. I have never seen the 150 kV.

" measure the resistance,   10**15 ohm,   teflon and polyethylene  20 kV, rods and disks

1/ The volume resistivity of these materials is well documented.

https://omnexus.specialchem.com/polymer-properties/properties/volume-resistivity

Notice that volume resistivity in ohm.cm is a materials property and does not change with shape or size as resistance will.

The resistance is greatly affected by surface contamination and preparation, and how the sample is prepared and fixtures.

Please note that HP, Yokogawa, and other fine instrument firms make special volume resistivity fixtures for this exact purpose with the required shielding and guard rings!

Normally a disc is used not a rod! The discs must be machined to exactly fit the fixture.

HP 16339A, also see keysight  16008B resistivity-cell

For a DIY measurement, suggest to use the electrostatic meters just to monitor test voltage and to use this simple setup:

HV supply....current limit R....(fixture and DUT)....return of fixture.....Keithley electrometer set for lowest current.


Then you can test at 20 kV and see the leakage.

Hope this proves of use to you!

Jon




 

[jonpaul]

Physikfan, I found my original 1980s Sensitice Research catalog and app notes, in a 1" black binder.

Full specs on ever unit, with capacitance etc.

Mentions R ~ x 10 exp 15 ohms.

My memory was bad, I have: Multirange 30 kV and multirange 50 kV, NO 100 kV.

Each has 4-5 ranges EACH RANGE HAS DIFFERENT CAPACIANCE.

So testing will take time.

SR does NOT spec any exact resistance.

You should be able to find the exact SR specs on web, you need the model numbers you have.

I could make a scan if necessary.

Best Regards,

Jon

[Physikfan]

Hello Jon

It would be absolutely great, if I could get such a scan from you.
I will start today loading my small electrometers with about
300 or 500 V respectively and afterwards I will observe the decrease of the
voltages.

Best regards
Physikfan

[jonpaul]

Hello again Physikfan:   

1/ Sensitive Research has NO exact spec on insulation R, just a magnitude ~ order of 10 exp 15 ohms.

2/ Big project to move all my meters to the lab and test on 30 kV power supply.

3/ Will try to scan the Sensitive Research cat, use PM or email so i can send you a link.

Kind Regards,

Jon

[jonpaul]

http://mavin.com/pictures/bw/sensitive_esh_1.pdf

[jonpaul]

PhisikFan:

Scanned Sensitive Research catalog and apps 1980

http://scientificonversion.com/SensitiveResearch1980.pdf

 Impossible to send to ancient Yahoo emal addr you should gert a modern alt like gmails.

Jon