detecting nanoamp current

Thread Starter

kender

Joined Jan 17, 2007
264
Colleagues,

I need to design a circuit that would detect a 3nA current through a slice of tissue. A crude block diagram is attached. It’s ok, if the detection takes a few seconds. The input voltage is between +2V and +3V.

Please give me some ideas, hints or reference designs.

Thanks,
Nick

P.S. This is an academic project, but not homework.
 

Attachments

Ron H

Joined Apr 14, 2005
7,014
Your drawing implies that you want a 1 or 0, depending on whether the current is greater than or less than 3mA. Why do you need to detect nA?
 

Thread Starter

kender

Joined Jan 17, 2007
264
Your drawing implies that you want a 1 or 0, depending on whether the current is greater than or less than 3mA. Why do you need to detect nA?
Apologies. That was a typo on my part. The correct text is "3nA detected
0 or 1". I've fixed the block diagram.

I need to detect 3nA to monitor the tissue (de)hydration. At the moment, I don't know the physiological details, where 3nA comes from.
 

beenthere

Joined Apr 20, 2004
15,819
Hi,

What is the nature and dimensions of the tissue sample? The excitation voltage must be truly minute to only push 3 na. What are your electrodes? Is the 3 na figure arbitrary, or chosen as a value that will not cause some effect in the tissue sample. Is your excitation going to be AC so as to avoid artifactual noise?
 

Thread Starter

kender

Joined Jan 17, 2007
264
I'm drawing up a schematic for you. I need to know your excitation voltage the current limit.
Thanks a lot!
The excitation voltage will be between +2 and +3V. It will be measured with an A/D.
The current limit will be 0.5mA (milliamps).
 

Thread Starter

kender

Joined Jan 17, 2007
264
What is the nature and dimensions of the tissue sample? The excitation voltage must be truly minute to only push 3 na. ... Is the 3 na figure arbitrary, or chosen as a value that will not cause some effect in the tissue sample.
The tissue will be rabbit. Chicken will be used for testing purposes. The dimensions are about 1cm. An important feature is that the excitation electrode will be behind a silicone membrane with micron-sized perforations. It is these perforations that will limit the current (according to my neuroscience colleagues).

What are your electrodes?
Haven't decided yet. I was thinking silver, gold of platinum plated. I'll need to experiment. Do you think that electrochemical reaction will introduce the offset?

Is your excitation going to be AC so as to avoid artifactual noise?
I'm going to try DC first. This setup is a fore-runner for an on-orbit experiment. Space and power will become a premium at some point, and I would like to avoid elaborate power supply, if I can help it.
 

Ron H

Joined Apr 14, 2005
7,014
Do you have predetermined supply voltage(s) for the detection circuit? Are you going to measure the current, or do you really only want go - no go as shown in your diagram?
 

Thread Starter

kender

Joined Jan 17, 2007
264
Do you have predetermined supply voltage(s) for the detection circuit?
At some point I will be limited to one supply voltage between +2V and +3V. But for a bench-top experiment, I can use lab power supplies. I need to make it work on the bench for the bioscience researchers.

Are you going to measure the current, or do you really only want go - no go as shown in your diagram?
I can measure the current - I have a spare A/D channel. Does measurement requires more circuitry compared to just go-nogo detection?
 

Ron H

Joined Apr 14, 2005
7,014
Thanks a lot, Ron!

If I want to measure the V(I), could I do it after R7 (possibly converting the second opamp into a gain stage) ?

I'll go take a nap and then study your circuit.
That's what I was thinking. You need the second op amp to invert and amplify the output of the 1st stage (-30mV at threshold) to get the output to fit into the range of your A/D (I'm assuming zero to 2 or 3 volts). You might need to switch to a rail-to-rail output op amp to avoid having to protect your A/D from overvoltage (depending on the A/D). Both op amps need low input offset voltage, and the first one needs very low input bias current.
This could possibly be done with a single supply. I would have to give it some thought. Maybe some of the other regulars here have ideas along this line.
 

beenthere

Joined Apr 20, 2004
15,819
Well, I'd be concerned about usinf a DC excitation and metallic electrodes because of ionic migration. There's also a chance of streaming potentials in those little pores in the membranes.

Be prepared to do some careful work to run down noise sources before you strat to collect meaningful numbers.
 

Thread Starter

kender

Joined Jan 17, 2007
264
Be prepared to do some careful work to run down noise sources [...]
I'm bracing myself for this. I know for a fact that there will be a 100kHz transmitter (up to 1W, no shielding) next to my experiment. Although, there will be a synchronization signal, and I could, potentially, do synchronous sampling. Turning of the transmitter for a short time is a possibility too.
 

hgmjr

Joined Jan 28, 2005
9,029
In view of the very small current being measured, it is important that the op-amp recommended by Ronh be used and no substitute be made unless Ronh has had a chance to review the specification..

Ronh,

Do you think it would be a good idea to put a 1 Meg resistor from the positive terminal on the first stage to ground? This would then insure that the two opamp positive and negative inputs would see very nearly the same equivalent resistance at their inputs.

hgmjr
 

Ron H

Joined Apr 14, 2005
7,014
In view of the very small current being measured, it is important that the op-amp recommended by Ronh be used and no substitute be made unless Ronh has had a chance to review the specification..

Ronh,

Do you think it would be a good idea to put a 1 Meg resistor from the positive terminal on the first stage to ground? This would then insure that the two opamp positive and negative inputs would see very nearly the same equivalent resistance at their inputs.

hgmjr
Well, I thought about that, but the 1Meg in the inverting input is just there to provide isolation from the unknown (to me, anyway) capacitance of the tissue sample, which could cause instability if it were not isolated by a resistor. The actual source resistance depends on the tissue sample, so matching it on the noninverting side is problematic. For that reason, I chose an op amp with very low input bias current. The maximum, even on the LT1464C, is only ±20pA (at ±15V supplies - not spec'ed at ±5V, but probably no worse). Even if you could match input resistances, the typical offset current is about the same as the bias current, so not much would be gained.
 

Thread Starter

kender

Joined Jan 17, 2007
264
In view of the very small current being measured, it is important that the op-amp recommended by Ronh be used and no substitute be made unless Ronh has had a chance to review the specification..
Thanks I'm researching the opamps for this circuit. I have only one desirement about the opamp: it would be nice if it could operate of a single supply rail. I've looked at the datasheet of the one that Ron have used. It's output can swing only between V(-)+1.5V and V(+)-1.5V. Ron's opamp has a JFET input stage, but the following stages and the output stage are BJT. Were BJTs chosen over FETs, because they generate less noise?

I have found a couple of other opamps:
TL082 looks very similar to Ron's LT1464
LT1013 single supply operation possible (p.8 of the datasheet)
 

Distort10n

Joined Dec 25, 2006
429
I have found a couple of other opamps:
TL082 looks very similar to Ron's LT1464
LT1013 single supply operation possible (p.8 of the datasheet)

OPA350 for single supply applications and it also has low input bias current (pA range). The LT1013's input bias current is a magnitude higher than what you are trying to measure.
 

Ron H

Joined Apr 14, 2005
7,014
I'm bracing myself for this. I know for a fact that there will be a 100kHz transmitter (up to 1W, no shielding) next to my experiment. Although, there will be a synchronization signal, and I could, potentially, do synchronous sampling. Turning of the transmitter for a short time is a possibility too.
Synchronous sampling could cause errors. Since you are in no hurry, a better solution might be a lowpass filter. The feedback network already rolls off at -6dB/octave starting at 15Hz, so at 100kHz the output of the first stage will be down at least -56dB relative to the input. This assumes that all the interference comes from the input. The circuit might need to be enclosed in a Faraday cage.
If you can afford a couple of seconds settling time, you can make the feedback cap at least 10 times larger, giving you another 20dB of attenuation at 100kHz.
 
Top