Suggest a Pico-ampere measuring circuit

[itp]

Dear all,

Can anybody suggest  Pico ampere measuring circuit or IC for a measurement range of 1pA to 50nA.

Thanks & Regards.
Ittoop.

[Ahmad_k]

Directly you can't.

First, what is the required sampling rate ?

Supposing it is low, you have to use an amplifier + Low pass filter (x1000) (Multiple stage is better) then feed output signal to an analog to digital converter (Sigma Delta is great).
you can use MCP3421 (I2C 18 bits Sigma Delta converter with internal Programmable gain amplifier 1 - 2 - 4 - 8 )

[itp]

I am planning to use an analog front-end, followed by an A/D converter.
The analog front-end and its pcb designs are challenging to me.
Can anybody share any document/circuit of this analog section.

Thanks and regards
Itp

[PICker]

If you want very high precision in the lower pA range I suggest the LMP7721
http://www.ti.com/product/LMP7721
this is useful for making electrometers. But the choiche of the feedback resistor and the lowpass capacitor is crucial.
You can also use a switching integrator instead a classical transimpedance amplifier (I/V converter); plese have a look to this appnote:
http://www.ti.com/lit/an/sboa034/sboa034.pdf

[mexpcb]

Hi, for those Ranges it may be better if you use TIA (Transimpedance amplifier)

You can take a look at TI OPA859..

not sure if the full details but after that you may need to amplify it like an VCA824 also from TI

Regards
mexpcb

[BharatSujanani]

You can use any low noise operational amplifier in Trans impedance mode. And use very high resistance like 500Megaohm  to measure your current range. If you signal is noisy than use low pass filter after this opamp output.

[bigtoy]

As a general comment: Try to avoid super-high resistors (> 10 Mohm) in your circuit. Because leakages on the PCB become significant sources of error. Putting guard traces on your PCB (google it) can help reduce errors due to leakages.

[OscarH]

I also found these super-high resistor are sensitive to humidity and moisture.
It does mean, on top of Bigtoy comment, you can see value changing as a function % humidity in the air.
I don't remember the specification of the resistors I used, but I remember I didn't had too much choice in models, so took what I found...
OH

[PICker]

I have experience with transimpedance I/V converters with a feedback resistors in the range of 50-100 MOhm and I confirm that sometimes is the resistor the key component of the circuit. I suggest quality carbon or precision (0.1%) metal oxide thick film resistors  (i.e. https://www.ohmite.com/) with very low inductive effect.
I confirm also the sensitivity to temperature changes (sometimes counteracted by using a diode that responds in the opposite direction).
Another interesting approach is to use diodes  (LEDs) as feedback components for log I/V converters:
https://pdfs.semanticscholar.org/2ecd/fdcbedf2ea6d171cbc0b3f6336758457a58f.pdf

[bellona]

I built several pico-ampere amplifiers about a decade ago, and I suspect nothing significant have changed in this field since then. To really get to pico-ampere or femto-ampere range, you have to deal with three things beside what are mentioned in the informative replies above:

First, you have to cut down interference, and the most significant contributor is the 50 Hz or 60 Hz interference depending on where your amplifier operates. Shielding can help a lot in this regard, but you really have to understand some basic ideas like the ground-loops to do it right. You might want to refer to the classic book "Grounding and Shielding Techniques in Instrumentation" by Ralph Morrison. Pico-ampere amplifier and Electrometers typically use Low-noise Triaxial cables. When all else fails, a handy trick to cut down the 50/60 Hz interference is to sample your signal at 100/120 Hz and add two-consecutive results together, this results in a sharp notch filter around 50/60 Hz.

Second, you have to pay attention to your PCB layout and your test fixture. Cut down possible leakage into your current input path, this involves slotting your PCB around the sensitive current-input trace (This cut down the leakage due to humidity) and/or surround the input trace with an additional guard trace which is kept to the same potential to your input. Low input-bias-current amplifiers like the ADA4530-1 has dedicated pin for driving the guard. The input-bias current of the amplifier also drifts with temperature, I've seen people using dummy amplifiers (unpowered) connected to the input trace to compensate the main amplifier.

Finally, choose the right amplifier for your circuit. You would like to have an ultra-low input bias current amplifier like ADA4530-1 mentioned above. The mean square value of noise current due to amplifier input bias equals 2*q*I*delta_f where q is the electronic charge and delta_f is the measurement bandwidth. With these ultra-low input bias amplifiers, you'd like to use high resistor in your TIA circuit since low resistor generate high current noise. I've used resistors up to 10 G ohm, and you really should try to keep your resistor clean. I used to immerse them in 100% alcohol and ultrasonic clean them.

If you are careful, you can get your sensitivity in the 10-100fA range if you keep your measurement bandwidth low (1Hz or so).

[PICker]

In my experience, battery-powered circuits reduce the noise and the ripple related to 50-60Hz mains.

[bellona]

Yes, that definitely helps as it cut down potential ground loops in your system. But the interference is still there, and it can easily get into your system especially if you have long signal cables.

[kreutz]

try https://www.eevblog.com/projects/ucurrent/

[bellona]

About the triaxial cable:
It turns out the friction between the metal conductor and the insulator could generate noisy current when the cable is bended or twisted, just like the electrical charge accumulated on you when walking on carpets when humidity is low. High-quality low-noise cables add graphite between the insulator and the shield to mitigate this problem.
If I remember it correctly, the current generated due to the friction in low-cost coaxial cables could reach pA levels. So, I switched to those costly low-noise triaxial cables. Try googling "low-noise cables" and you'll find some useful information and suppliers for this.

Posted on: February 09, 2020, 09:50:37 21:50 - Automerged
An informative and approachable book on low-noise measurement is available from keithley (Now Tek):
https://download.tek.com/document/LowLevelHandbook_7Ed.pdf

[PICker]

The wire-related noise is often called triboelectric and it is a problem in medical and precision devices (mainly for low amplitude signals)
https://experience.molex.com/triboelectric-noise-in-medical-cables-and-wires/

[sadman]

About the triaxial cable:
It turns out the friction between the metal conductor and the insulator could generate noisy current when the cable is bended or twisted, just like the electrical charge accumulated on you when walking on carpets when humidity is low. High-quality low-noise cables add graphite between the insulator and the shield to mitigate this problem.
If I remember it correctly, the current generated due to the friction in low-cost coaxial cables could reach pA levels. So, I switched to those costly low-noise triaxial cables. Try googling "low-noise cables" and you'll find some useful information and suppliers for this.

Posted on: February 09, 2020, 09:50:37 21:50 - Automerged
An informative and approachable book on low-noise measurement is available from keithley (Now Tek):
https://download.tek.com/document/LowLevelHandbook_7Ed.pdf

your share link is dead but manage to find 6th edition here is link

https://web.mit.edu/8.13/8.13d/manuals/LowLevMsHandbk.pdf

[bellona]

your share link is dead but manage to find 6th edition here is link

https://web.mit.edu/8.13/8.13d/manuals/LowLevMsHandbk.pdf

The link is working fine, maybe you should try it again?
https://download.tek.com/document/LowLevelHandbook_7Ed.pdf

[Vineyards]

The usual tricks are starting with aan opamp having an Ib of a few femto amperes as well as ultra low input capacitance and, implementing shielding techniques as well as either a stand off input implementation or disconnecting the input pin of the opamp from the PCB and soldering it as it stands in the air (I usually trim it a bit and clean the entire area with pure IPA.) The reason for that is standard PCB leakage levels which are considered excellent for everything else (like 10 Tera Ohms) are extremely degrading for the femto ampere range and will render the measuring circuit completely useless. We are trying to stay close to infinity in this case.

You could also use a small rectangle you would cut from a copper PCB and solder one side to ground for shielding the input area of the opamp.  Always clean thoroughly whatever you may add. Should you use any small value capacitors in the hi-z zone choose low leakage types if you can't avoid them altogether. Hi-Z opamps are usually CMOS types like LMC6001 and they are a bit more difficult to work with. You need to read the datasheet thoroughly. Use a large value resistor to protect the input, if your design involves a voltage follower  do not connect the inverting pin and the output directly as one would normally do, use a very large value resistor. instead. Cleanliness and maintaining cleanliness in the long term is of essential importance.

Hi-Z circuits are especially prone to noise and ground loops. For that reason, you will need to seperate the ground of the hi-Z side of the circuit by using isolators and DC to DC converters with isolation. There are digital and analog types, if you opt for the analog type the ADC chip will stay on the processor side if not it will have to be placed on the measuring side. The digital approach is more efficient but voltage ranges and protection considerations may drift you into the analog domain as well.

[bellona]

Should you use any small value capacitors in the hi-z zone choose low leakage types if you can't avoid them altogether.

That's true
Sometimes a tiny feedback capacitor is needed for stability concerns and/or damping the response. With feedback resistance in the range of Gohms, the opamp is almost working as an open-loop amplifier. I've seen capacitors built with a plated thru-hole in their PCB: the plated side serves as one plate of the cap, teflon is used as the dielectric material and they insert a free-standing conductor inside the teflon plug as the second plate.