DIY cable for picoammeter

[serg-el]

Advantest triax cables have a coating on the inner insulation.  Conducts current, or not - did not check.

[Kirill V.]

Decent quality solid polyethylene is fine. I use the H&S low noise coax (G_01130_HT) with a polyethylene isolation and a PVC conductive layer and it is good for femtoamp level leakages even at considerable length (a few meters), for low voltages. I've measured the leakage current below 10fA at 100V DC for 2m of that cable at room temperature and about 40% RH. There is some DA so you have to wait some minutes for the leakage to settle.

Thanks Alex. I found this cable in Mouser and the price suits me perfectly. It is clear that this is ideal for measuring picoamps. Now I'm wondering how cheap polyethylene cables behave.
In any case, when I start building my picoammeter I will be forced to buy parts abroad, because such things are not sold in Russia(teflon standoffs and TO-99 socket for AD549). At the same time buy 1 meter HUBER + SUHNER cable.
But first I'll make a simple prototype with a cheaper and less sophisticated op amp to test my ideas. And nothing but TV cables made in China in our stores do not sell.

[serg-el]

My google-fu tells me - low noise triax cable!

[Kirill V.]

Thanks, but unfortunately this cable is out of stock in Mouser

[David Hess]

Specifically, directly clamp the inverting input with low leakage diodes and place a current limiting series resistor between the inverting input and the summing node.  For stability, feedback capacitance will need to be added to cancel the RC time constant between the series resistance and capacitance from the clamp diodes.

I'm not going to use external protection diodes. I will limit the input current to 2 milliamps in case of an input overload with a protection resistor.
So, I can make the second important conclusion: I can quite use simple or" low-noise " cables with polyethylene which do not cost mad money. Am I right?

Yes, that will work if you know that you can rely on the input protection diodes.  Some low input bias current operational amplifiers completely lack input protection diodes and would have to rely on external ones for protection.

There are more complicated circuits which bootstrap external protection diodes to reduce their leakage but I doubt you will require that.

[Gyro]

Using the internal protection diodes is a valid thing to do with certain low input current opamps which already contain bootstrapped protection diodes such as the LMC6001 or, more usefully the low cost LMC662.

The maximum rating for the LMC6001 input diodes is 10mA and the LMC662, 5mA. As long as you resistively limit the maximum overload input current to well below these levels (and ensure that the series resistor has sufficient voltage rating and minimal parasitic capacitance) then you are fine.

Opamps with internal bootstrapping of the protection diodes makes things a whole lot easier these days.

[Kirill V.]

AD549 hasn't internal protective diodes. However, the same rules apply here as when using amplifiers with a JFET input, the input current must not exceed 5 mA. I have already received advice from ADI experts about this.
Of course, it is possible to add external protection, I thought of JFET transistors as diodes
And you can use bipolar transistors, but I do not understand by what criteria to select transistors for such protection.

[Alex Nikitin]

AD549 hasn't internal protective diodes. However, the same rules apply here as when using amplifiers with a JFET input, the input current must not exceed 5 mA. I have already received advice from ADI experts about this.
Of course, it is possible to add external protection, I thought of JFET transistors as diodes
And you can use bipolar transistors, but I do not understand by what criteria to select transistors for such protection.

1) The AD549 is a poor (and expensive) choice unless you really need a metal case or +/-15V capability. The ADA4530 will be much better or the LMC622 on the cheap will be better in most respects (input current, drift).

2) The BAV199 diode from a decent manufacturer (TI, NXP) will give you a much better performance than either BJTs or JFETs as diodes.

Cheers

Alex

[serg-el]

Took measurements on his cable.
Advantest A01009-200.

Autopsy results in photos.

[Kirill V.]

Great-great thanks Alex for diodes information and serg-el for cable disassembly!
I know the latest op-amps are much better, but it's not critical for me because I'm not aiming for the femtoampere range yet.

[David Hess]

Of course, it is possible to add external protection, I thought of JFET transistors as diodes

JFETs only protect themselves when the junction is forward biased so an additional diode is required.

2) The BAV199 diode from a decent manufacturer (TI, NXP) will give you a much better performance than either BJTs or JFETs as diodes.

As with 2N3904s used as low leakage diodes, the BAV199 is only suitable if you grade them yourself; they are only tested down to like 2 nanoamps or something.

If you want a guarantied low leakage diode, 2N4117 or 2N4117A JFETs are the least expensive option.  But personally I just grade 2N3904s which only takes a couple minutes and can be done with a power supply and any 10 megohm input voltmeter.  Use the 0.2 or 2.0 volt scale as a picoammeter unless you have an electrometer handy.

[MiDi]

Took measurements on his cable.
Advantest A010009-200.

Autopsy results in photos.

Interesting, Keysight N1415A seems to use same triax connectors

[Kirill V.]

BAV199 specifications says about 3 pA minimum at 75 Volts reverse, in my case they will work at 1 mV maximum.
In older electrometers, a pair of bipolar transistors can be found as diodes, and often only one pn junction is used.

[Echo88]

https://www.google.com/url?q=https%3A%2F%2Fwww.dropbox.com%2Fs%2F11wp4nlumil1koq%2F4x.3_TIA_final-DRAFT.pdf%3Fdl%3D1&sa=D&sntz=1&usg=AFQjCNHlXTnDRvTaCN_XTzJ73ML6ytRQNg

On page 9 they mention a trick on how to short the integrating cap of an integrator by manipulating the supply rails of the OP-Amp to get the ESD-protection-diodes to conduct.
Can anyone share a suitable schematic (couldnt find any info on this by Mr. Gottschalk) for this and is it a better solution than just using a relay to short the cap for example?
I assume the ESD-diode-trick doesnt produce charge injection on the cap, unlike the shorting-switch?

[Kirill V.]

I want to ask forum members to measure the actual leakage current in RG type coaxial cables. Then I can make prototypes without having to buy cables on the other side of the world.
1 meter cable with decent pure copper screen.

[David Hess]

BAV199 specifications says about 3 pA minimum at 75 Volts reverse, in my case they will work at 1 mV maximum.
In older electrometers, a pair of bipolar transistors can be found as diodes, and often only one pn junction is used.

The maximum specification is what the part is actually tested for.  The typical or minimum leakage is not guaranteed.  From the datasheets:

BAV199
   Typical   3pA
   Maximum   5nA
2N4117
   Maximum   10pA
2N4117A
   Maximum   1pA

Now like most 2N3904s, most BAV199s will leak at around 3pA but the only way to know is to grade them yourself because unlike with 2N4117 JFETs, the manufacturer is not testing to that level.

In older electrometers, a pair of bipolar transistors can be found as diodes, and often only one pn junction is used.

Newer instruments may still do that.  The manufacturer graded the transistors for low leakage.  Or if they got tired of doing that, bought 2N4117s or 2N4117As which they did not have to test individually.

True low leakage diodes can be had but it comes down to paying the manufacturer to do the testing for you and without the economy of scale of the 2N4117, they end up costing an awful lot.  Contrast the price between the LMC6081 and LMC6001 operational amplifiers; they are the same part tested to different leakage specifications.

[Alex Nikitin]

https://www.eevblog.com/forum/metrology/measuring-nanoamps-and-below-like-a-ninja/msg1119731/#msg1119731

The spread from good makes is surprisingly small. You need to keep things very clean though.

Cheers

Alex

[iMo]

Metal 2N4117A costs 14Euro, TO92 epoxy PN4117A 1.60E..
Are those comparable in the max leakage??

[Alex Nikitin]

I want to ask forum members to measure the actual leakage current in RG type coaxial cables. Then I can make prototypes without having to buy cables on the other side of the world.
1 meter cable with decent pure copper screen.

I've measured what I had available - a ~180m coil (18nF total capacitance) of the RG58C (LSZH outer isolation, tinned copper screen and solid PE internal isolation). At 100V DC after settling the leakage of the coil is about 0.1pA  (100fA) .  This is a high quality mil-spec coax. I've also measured ~15m of RG-62 but it was terminated with a nylon isolated BNC, which leaks badly, so the result was the connector leakage (~0.5pA at 100V). In my experience at a room temperature and reasonable humidity (<50% RH) quality standard coaxial cables with solid PE isolation are rather good for low leakage.

Cheers

Alex

P.S. my measuring setup (K617 + 1m triaxial cable to an external box with PTFE isolated BNC connectors and 1G test resistor) leakage with nothing connected to it is below 5fA at 100V.

[Kirill V.]

Great thanks for measurings
Me available for purchase RG-58C/U. I will proceed from the fact that my picoammeter must have a resolution of 1 pA so that the leakage in the cable is an order of magnitude smaller and does not affect the measurement results.

[Alex Nikitin]

Great thanks for measurings
Me available for purchase RG-58C/U. I will proceed from the fact that my picoammeter must have a resolution of 1 pA so that the leakage in the cable is an order of magnitude smaller and does not affect the measurement results.

About 3 orders of magnitude, as for 2m cable length the leakage at 100V DC would be below 1fA and obviously less for lower voltages. 

Cheers

Alex

[David Hess]

Metal 2N4117A costs 14Euro, TO92 epoxy PN4117A 1.60E..
Are those comparable in the max leakage??

Yes, I have never noticed any difference.  Plastic packages have been outstanding for decades now.

Metal can packages in theory could be better however Bob Pease recounted an incident where light leaking through the bottom of a TO-72 can caused high leakage.  I have had the same problem with glass packaged low leakage diodes.

[Kirill V.]

No one else here has mentioned InterFet products and their low leakage JFET pairs. Have you tried this?

[David Hess]

No one else here has mentioned InterFet products and their low leakage JFET pairs. Have you tried this?

I have not but Linear Systems also produces low leakage JFETs and JFET pairs as well as low leakage diodes.

[Kirill V.]

I have not but Linear Systems also produces low leakage JFETs and JFET pairs as well as low leakage diodes.

These diodes are actually JFET connected as diodes to two external pins, right? Unfortunately, this is a rare and not cheap product. At least for prototypes or the first attempt, as in my case:)