ultra low current measurement

[krivan]

low current measurement

Hi,

I would like to sense current in the range of ~10pA - ~100nA. I am looking for some clever techniques which enables to do it on PCB.

I have a chip, I put it into a socket and I have to measure the current of the working electrode[flowing through the socket, of course]. I thought that I could do it with a current sensing resistor+current sensing amplifier+one resistor[setting the amplification] but it wouldn't be a smart idea in a noisy environment. Regarding such low currents, everything is noisy though.

So, please suggest a smart way to do it, I appreciate every comments, links, schematics, etc.

Thx,
Krivan

 

[agunther]

Try a CMOS current conveyor in series instead of a resistor. Maybe you can even find a current conveyor with plenty of current gain.

If the noise is random, take a Agilent meter (3340 i believe) and crank up the averaging. You should be able to eliminate the noise with enough averaging cycles, assuming your current is constant.

 

[Siniša]

ultra ultra low current amplifier of insulation

First you said "sense". Do you require to detect if current is present, detect if current passes some threshold or measure it? If measurement or comparison is required, what kind of resolution you hope to achieve?

It is obvious if you require 10pA range and if we asume 1pA resolution, that this is going to be very difficult to achieve. If you have noisy enviroment, you have very litle chance doing this succesfully. If you can't get enviroment that is low noise for this measurement, your whole aparatus will have to be enclosed in heavy gauge steel enclosure with RF shielding even maybe using special magnetic shielding materials if noise is at low frequencies. This you will have to find out experimentally.

Measurement aparatus will have to be very well isolated from surroundings electrically to avoid creation of current loops. Power supply should be either batteries or low noise, medical grade isolation power supply.

Source of signal and complete trace to first IC that processes signal has to be protect by driven shield and on top of that should be electromagnetic shield. Signal should be transported as balanced to reduce possibility of interference.

Whole aparatus will have to be protected from any vibrations that might produce triboelectric or microphonic effect in cables or piezoelectric effect in components. Setup should have minimum thermal differences in critical areas to reduce Tc effects.

Measurement could be done in number of ways, most common being integrator. Small capacitor is used as "bucket" to collect drips of current over time. You can measure time it takes to fill the capacitor to predeteremined level or measure level in the capacitor after predetermined time. You discharge capacitor at the end or charge it in the oposite way.

To grasp what you are measuring, 6x10^6 electrons every second is not that much. If you had 1V voltage across insulation that leaks less then 1pA, that insulation has to be over 1TΩ.

There are integrator chip available from Texas Instruments for such purpose. ACF2101 and IVC102 . They include integrating capacitors and discharge switches.

There are a lot of techniques involved in building electrometers in pA range, so first search on internet for electrometers to see ideas how it's done. Printed circuit is generally not good medium when you are in pA range. Air is very good isolator so a lot of high impedance ciruits bend pins of IC's into air and make connection o ceramic or other suitable materials posts of sufficient legth.

There are few articles arround...
Bob Pease: **broken link removed**
Keithley:**broken link removed** (Hard Copy) and **broken link removed**

 

[FvM]

low current mesurement

The term working electrode seems to indicate an EC sensor, but the question is rather vague in many respects. I don't think, that it's necessary to assume accuracy and bandwidth on my own.

 

[krivan]

very low current measurement

Thanks for the answers guys. Special thanks for Sinisa for the detailed discussion.
I think I will try the method you suggested. The circuits you wrote should fit well for the application.

Regarding the problem, yes I have to measure it but I still haven't thought about the resolution. According to the specification I would have to measure a current from 10pA to 100nA but the pA seems to be a dream on a PCB...on-chip it would be of course different...

 

[FvM]

low current measurement circuit

According to my experiences, e. g. an OPA129 with 1 GOhm feedback resistor is basically able to achieve pA accuracy with usual guarding techniques. Depending on the enviromental conditions, the PCB may need an appropriate coating.

 

[sureshamathew]

Hey!

I would like to measurement current in the range 10uA to 100mA. I was hoping to use LTC Low Current sense amplifier LTC6100 (http://cds.linear.com/docs/Datasheet/6100fb.pdf). However, I am not sure if this is the right one to go with. From what I understand from the specifications, the V_SENSE voltage (voltage across the sense resistor) must be at least 300mV (VCC = 5V) for the op-amp to measure correctly. But, later in the specification, there is a mention that VSENSE can be 50mV. I don't quite understand how this possible! I am guessing my understanding of this is incorrect. I was hoping someone could explain what the minimum sense voltage (or sense current) for this op-amp is? Please assume a gain of either 20/50 and precision of at least 10uA.

If 300mV is the minimum sense voltage, then to measure 1uA of current, I would need a very large resistor(300K appx). If it's a large resistor, then, when I measure 100mA of current, the voltage drop will be too high for the remaining part of the circuit to work. My intuition was that sense resistors were generally very small valued resistor so that the voltage drop was negligible.

If you believe LTC6100 is not the right op-amp for this purpose, can you suggest another?

Kindly help!

- Suresh

 

[snafflekid]

Your chip should be installed in a low leakage socket. these are made of teflon and other highly resistive materials. Everything should be shielded, including the probe wires leading to your ammeter. The easiest way of doing this measurement is to use an HP 4145 parameter analyzer with the low leakage sockets inside of the shield box. But that is expensive

 

[FvM]

When you say 100 mA with 10 uA precision, then the 100 mA full scale is defining the shunt dimensioning. This isn't actually
an ultra low or a low current range, so your post seems to be rather off-topic in the present thread. Also the previous comments
regarding low leakage don't apply in my opinion.

Dynamic range is however an issue. LTC6100 offset voltage will limit the (unadjusted) precision to about 100 uA. I don't
know, if you need at all a high-side sense amplifier like LTC6100, otherwise standard instrumentation amplifiers will provide
higher accuracy.

 

[sureshamathew]

Thank you!
Could you explain what you mean by "shunt dimensioning"? I didn't quite understand what you meant by

When you say 100 mA with 10 uA precision, then the 100 mA full scale is defining the shunt dimensioning

? Does that mean 10uA resolution is too low?

What is classified as ultra low current measurement? If I was to use LTC6100, what is the "ultra low current" and what is the resolution that it's capable off? What is the maximum current that it can measure?

On a general note, how do I find out the resolution of a current-sensing op-amp?

If I should post these comments on a separate thread, do let me know...

- Suresh

 

[FvM]

Shunt dimensioning is another word for what you considered regarding a 300k resistor.

For the other questions, you should simply calculate the effect of amplifier error terms, particularly offset voltage