[SOLVED] Current Amplification (nA => uA)

[flote21]

Hello. i am new in this forum and first of all I woul like to introduce myself. I am Enrique and I am working in the electronic department of a small company which develops camera systems.

My problem is that I dont know if there is any IC which can amplifies low current with a very low noise. I need to amplify the low current generated by an ISASSB photovoltaic detector (around nA) to something bigger (around uA). And i can't use a transductor amplifier....

I was thinking about a mirror courrent source with the WIDLAR strucuture. But I don't know which trt't use in order to achive a properly working with nA current...Can you give me some advise?

Thank you very much for your help!

B.R.

Enrique P.

 

[crutschow]

Why can't you use a transconductance amp? That's the typical way to amplify a small current.

What is the frequency of the signal?

 

[flote21]

Hello thank you for your quick answer.

I can't use a transducer because I am already using one. But This transducer works also like a reset circuit for the photo diode. And it provides 100uA. Then the nA produced by the photo diode have an offset of 100uA....

Then I need to connect current amplificator between the photo diode and my special transducer to amplify the low current of the photo diode...

Any idea?

Thank you!

 

[crutschow]

What is this "transducer"?

I still see no reason you can't use a transconductance amplifier.

Post a schematic of you setup.

 

[flote21]

Sorry I meant Transconductance amp. I was writting through my smartphone and it made a mistake.

Yes, please find attached the schematic of my circuit. As you can see the photodiode and the IVC102 (Switched integrator + reset circuit) must be together for a proper working of the photodiode. The IVC102 already works like a Transconductance amp. but it is also provided with a reset circuit to reset the photo diode. And this reset circuit needs current to work fine.So this is the main reson becuase I can't use directly a simple Transconductance amp. I need one but with a reset circuit....I was reserching for a long time and this IC was the only IC valid for me propose.

But the problem of this IC is the low sensitvity. I need to amplify nA from the photo diode and this IC does not detect nA variations. So I need to put some current amp between the InAsSb Photovoltaic detector and the IVC102....Or maybe there is an IC very similar like IVC102 wich a higher sensitivity.

Thank you so much.

 

[crutschow]

There is no attachment.

 

[flote21]

Re: Current Amplification (nA => uA)

Now see the previos post..

Thanks

- - - Updated - - -

Ahh we want to take mesurements of ms. So the max freq would be around 300KHz.....

 

[godfreyl]

If you want to amplify the nA signal and ignore the 100uA offset, then you need to use AC coupling.

 

[crutschow]

ms and 300kHz are not compatible values.
Do you mean µs?
How long are you integrating the signal?

 

[FvM]

Your description of the existing circuit is rather unclear and can be hardly related to IVC102 specification. The term "reset the photo diode" needs clarification.

If the photo diode generates 100 µA DC current, the respective shot noise will be probably the dominant noise source. At worse case, the current sink will produce a similar noise amount. Sinking the current to a "noiseless" integration capacitor is basically a good idea, but must not necessarily work under the operation conditions of your system.

AC coupling might be a solution, but a low noise DC current sink will be required though. In terms of IVC102 input noise current, I don't expect that an additional amplifier would improve anything.

 

[flote21]

Re: Current Amplification (nA => uA)

Sorry you were right. I was hurry and the time was uS to have a maximum freq of 300kHz. The maximum integration time of the signal is arund 3.3uS...

- - - Updated - - -

the photdiode generates a dark current around 100uA (and it depends of the temperature). And the maximum input current for the reset circuit of the IVC102 is 100uA. So I need to remove the dark current of 100uA before driving it to the IVC102. AC coupling would be also a nice solution. But I would like to ask if it would be also valid compensate this dark current with a high quality digital potenciometer (x9118tv14z-2.7) which drives -100uA...Take a look to the picture attached.

 

[FvM]

The digital potentiometer solution refers to a resistive current sink. That's probably acceptable in noise current terms. An inductor would be a "noiseless" alternative, but I don't know the intended lower cut-off frequency and if it can be implemented with an inductor load. In case of a resistive load, there could be also an active feedback circuit that avoids the need for adjusting a potentiometer.

 

[flote21]

Yeah I need a low noise solution for this current compensator. And for this reason i dont know if the AC Coupling would be good to achieve my propose with low noise. If you could advise me some AC coupling IC/circuit would be very nice....

thank you for your help.

 

[Analog Ground]

You say there is 100uA of "dark current". Is it dark current from a reverse bias on the detector? If this is the case, you do not have an amplifier problem. You have a signal-to-noise ratio problem at the detector. What is the frequency range of your signal or the frequency characteristics of the signal? The signal-to-noise must be improved by filtering out the noise in some way. Simply offsetting the dark current does not improve the signal-to-noise ratio since the dark current noise and the offset signal noise are uncorrelated. In fact, the offset makes it worse. The place to start is with a discussion of the frequency characteristics of the light signal. Again, my comment applies if there is a true dark current from the detector.

If the "dark current" is due to background light, it is a different problem. Here optical filtering and some means of subtracting the background are done. So, is it true detector dark current or background light (or both)?

 

[flote21]

No iti is a DC offset current provided by the photo diode. As you can see in the picture attached the dark current is lower as lower is the temp. We are able to keep the temp of this photodiode to -60°C in a constant way. So the dark current mesured at -60°C is -100 uA always. So Iout is going to be like I show you in the picture. Iout = Data (nA) + I_DC offset (100uA). And I need to remove the 100uA offset current before driving Iout to the input of the reset Circuit. Because the max input current for the reset circuit is 100uA.

 

[Analog Ground]

No iti is a DC offset current provided by the photo diode. As you can see in the picture attached the dark current is lower as lower is the temp. We are able to keep the temp of this photodiode to -60°C in a constant way. So the dark current mesured at -60°C is -100 uA always. So Iout is going to be like I show you in the picture. Iout = Data (nA) + I_DC offset (100uA). And I need to remove the 100uA offset current before driving Iout to the input of the reset Circuit. Because the max input current for the reset circuit is 100uA.

I see you are using the Hamamatsu device which has a very, very low shunt resistance. Let me approach the problem at little differently. From the Hamamatsu data sheet it looks like the response of the detector is not dependent on a small bias. Then, let's model the bias problem by putting a shunt resistance of about 25 ohms in parallel with the photodiode. The principal source of bias would be the input offset voltage of the amplifier. If the input offset voltage was very, very small we have an easier problem.

Now, why have you chosen a charge amplifier like the TI part which is really an integrator and needs periodic reset? Is there some special sampling you are doing with the light source? Why not use the typical current->voltage amplifier for the photo-current? If one of these is used, the problem is then finding an op amp with a low enough input offset voltage or implementing an "auto-zero" circuit which can constantly adjust for changes in the detector over time and temperature. Is this going in the right direction?

 

[flote21]

Exactly!!! You are going in the right direction! We want to use this device for a special application. And for this application I need to reset the hammatsu device after each integration time (See datasheet of IVC102. For this reason I choosed the TI device becuase this device provides to me exacly what I need...

As you figure out, the hammamatsu device produces a dark current which depends on the temp. And the perfect solution for this problem would be to implement a zero-adjust circuit. But I don¨t know if there is something in the market which can do this function or where i can find some scheme to help me to create one...

On the other hand, you spoke about adding a shunt resistor of 25 ohm in parallel with photodiode and then put a typical current->voltage amplifier. This is also a smart solution, but then, how can i reset the photodiode??

Thanks!

 

[crutschow]

What is the lower frequency limit of the signal? Do you need to go to DC?

 

[Analog Ground]

Exactly!!! You are going in the right direction! We want to use this device for a special application. And for this application I need to reset the hammatsu device after each integration time (See datasheet of IVC102. For this reason I choosed the TI device becuase this device provides to me exacly what I need...

As you figure out, the hammamatsu device produces a dark current which depends on the temp. And the perfect solution for this problem would be to implement a zero-adjust circuit. But I don¨t know if there is something in the market which can do this function or where i can find some scheme to help me to create one...

On the other hand, you spoke about adding a shunt resistor of 25 ohm in parallel with photodiode and then put a typical current->voltage amplifier. This is also a smart solution, but then, how can i reset the photodiode??

Thanks!

The 25 ohms is not an addition to the circuit. I include it in the discussion as part of a model of the detector. In a very simplistic way, it models for the low shunt resistance of the detector. In reality, I think the dark current is due to lots of magical device physics like electron tunneling.

I don't like the TI part because it has high input offset voltage and input offset voltage drift with temperature. Not a good thing for this detector. So, here are my suggestions:

1. Use a precision op amp with low initial offset voltage and low offset voltage temperature coefficient. It should also have low bias current. I would start with Analog Devices and T.I. which make likely candidates.
2. Put in a precision offset adjustment to zero out the input offset voltage.
3. Build a charge amp around the op amp with a discrete capacitor and an analog switch with low charge injection.
4. Do at least a rough noise analysis to get an idea of the signal-to-noise with the 5 nA signal.

 

[Analog Ground]

FYI. There is a commercial camera (which shall be nameless but easily found) using the same detector and a low frequency cutoff of 30 Hz. Pulsed or chopped light is required. In our case above, DC operation is difficult but I think it might be done with sufficiently low input offset voltage. Perhaps Hamamatsu can comment or provide more data to extend their dark current graph down to microvolts? Good luck!