Building a nA-Current Source for a Project.

[Aeon]

Hi,

I just started a Project and one of the things I have to build for it, is a nA-Current Source. The Source has to be in the range from 1nA to 1000nA.

In the last couple of Days I have read about current sources and how to build them. But I found nearly no Information on how to build one in the Nano-Ampere-Range.

I found out that I need an OpAmp that has its input currents in the pA-Range (that makes sense )

My Goal is to control the Current via a Controller by a Computer. So my Vin (see Attachment) comes from a DAC.


My first approach with the Current source is the Circuit you can see in the Attachment. Before I start building the circuit I thought I get a second Opinion on it. perhaps some of you Guys has a better Idea.

My Current I should be in the Range 1nA-1000nA. I choose a Resistor with a Value of 1M --> A Voltage at Vin of 1mV is 1nA and 1V is 1000nA.

I hope I could explain my Idea clear enough. English is not my first language if I wrote something that is not understandable please ask

[Dr. Frank]

Hi,
if you connect + input to GND, and reference the input voltage to GND also, your circuit will work in principle.
The load will be virtual earthed at - input, and the OpAmp is capable of driving 1µA easily .

Practically, your circuit probably will be disturbed greatly, if you - for example - connect the DAC to a µP with external supply.
That means, your output current will have a lot on noise, meaning, it will fluctuate a lot.
At nA level, the fluctuations will exceed the nominal output value.

For such low currents, let's say below 10µA, you must provide very good shielding of the OpAmp, the reference resistor, the power supply - perhaps better realized with batteries , and the output jacks for the load. You'll need perhaps an active shield for the components, and also for the load itself. The reference voltage also has to be free of noise, especially via GND.

Perhaps you download manuals for Keithley nA / pA sources (e.g. search for Keithley 220), or have a look at Daves tear down of Keithley low current source and DMM. (#406)

You did not explain, for what kind of load you are building your circuit.
Perhaps the output needs an earth / GND referenced or a floating output.. in those cases, you need another circuit.

Frank

[robrenz]

Along the same train of thought you may need triaxial connections/cables with a driven guard to minimize cable leakage also like a SMU would have.

[Aeon]

Hi,
if you connect + input to GND, and reference the input voltage to GND also, your circuit will work in principle.
The load will be virtual earthed at - input, and the OpAmp is capable of driving 1µA easily .

Practically, your circuit probably will be disturbed greatly, if you - for example - connect the DAC to a µP with external supply.
That means, your output current will have a lot on noise, meaning, it will fluctuate a lot.
At nA level, the fluctuations will exceed the nominal output value.

For such low currents, let's say below 10µA, you must provide very good shielding of the OpAmp, the reference resistor, the power supply - perhaps better realized with batteries , and the output jacks for the load. You'll need perhaps an active shield for the components, and also for the load itself. The reference voltage also has to be free of noise, especially via GND.

Perhaps you download manuals for Keithley nA / pA sources (e.g. search for Keithley 220), or have a look at Daves tear down of Keithley low current source and DMM. (#406)

You did not explain, for what kind of load you are building your circuit.
Perhaps the output needs an earth / GND referenced or a floating output.. in those cases, you need another circuit.

Frank

I am building a Test System for an IC that is in development. My System has to Emulate a Sensor the IC has to Control. Because the Sensor ist currently also in Development  and we want to test our Control-IC when we have the first prototypes, and they will be ready before we get the gen2-Prototypes of the Sensor, so I am building a Test-System that can emulate the sensor.

As for the load: the load will be a also an Op Amp, it will be connected to an Transimpedance amplifier to be precise.

I hope that makes sense what I wrote


I also looked more around what I can find about nA-Current Sources.
I found this Circuit: http://www.linear.com/solutions/1185  I am trying to understand this circuit (not that easy with nearly no explanations on how it works) perhaps it helps.


What you also should know: I have build a couple of digital Circuits, but its the first time I am building an analog circuit with Op Amps so I am sure I will have many questions.

[tszaboo]

Hi,
if you connect + input to GND, and reference the input voltage to GND also, your circuit will work in principle.
The load will be virtual earthed at - input, and the OpAmp is capable of driving 1µA easily .

Practically, your circuit probably will be disturbed greatly, if you - for example - connect the DAC to a µP with external supply.
That means, your output current will have a lot on noise, meaning, it will fluctuate a lot.
At nA level, the fluctuations will exceed the nominal output value.

For such low currents, let's say below 10µA, you must provide very good shielding of the OpAmp, the reference resistor, the power supply - perhaps better realized with batteries , and the output jacks for the load. You'll need perhaps an active shield for the components, and also for the load itself. The reference voltage also has to be free of noise, especially via GND.

Perhaps you download manuals for Keithley nA / pA sources (e.g. search for Keithley 220), or have a look at Daves tear down of Keithley low current source and DMM. (#406)

You did not explain, for what kind of load you are building your circuit.
Perhaps the output needs an earth / GND referenced or a floating output.. in those cases, you need another circuit.

Frank

I am building a Test System for an IC that is in development. My System has to Emulate a Sensor the IC has to Control. Because the Sensor ist currently also in Development  and we want to test our Control-IC when we have the first prototypes, and they will be ready before we get the gen2-Prototypes of the Sensor, so I am building a Test-System that can emulate the sensor.

As for the load: the load will be a also an Op Amp, it will be connected to an Transimpedance amplifier to be precise.

I hope that makes sense what I wrote


I also looked more around what I can find about nA-Current Sources.
I found this Circuit: http://www.linear.com/solutions/1185  I am trying to understand this circuit (not that easy with nearly no explanations on how it works) perhaps it helps.


What you also should know: I have build a couple of digital Circuits, but its the first time I am building an analog circuit with Op Amps so I am sure I will have many questions.

The bottom two opamp is the classic two opampt instrumentation amplifier to amplify the curent going through the shunt, the top opamp is an integrator, or PI controller, depending how you look at it. I would increase the shunts value if possible.

[Aeon]

Hi,

sorry for not answering, but I was reading and gathering informations for my Project. I found a Book about Current Sources and VOltage references. It helped me understanding more of this stuff.

With the help of the Book I came up with the circuit I attached to this post.

as Resistor Values I choose:

R1=R2=1M
R3=R6=R5=R4=100k
RL is my Load.

Accourding to the Book my Current ILoad is IL = - (Vin * R3) / (R1 *R5)
Also the following equations apply: R1 = R2,    R4+R5 = R3+R6

With all that in Place an Input Voltage of -1mV translates to an Ouput Current of 1nA and an Input Voltage of  1V becomes 1uA Output Current.

I hope the use of smaller Resistors on the Output side of the Op Amp makes the circuit less sensitive to disturbances, but I think I still have to put a shield around it.

Next step is building a Prototype circuit on a breadboard to see if it works how I think it schould.

As I wrote before I want to use a DAC as input-voltage-source. I found the DAC8562  it has a Resolution of 1mV per LSB and a maximum output of about 4 Volts, I also have to put an Inverter circuit between DAC and input (Vin) of the current source circuit, for that I use a inverting Amplifier-circuit with an amplfication of 1.

I am still searching for a fitting Op Amp.
I allready found the OPA129 ( http://pdf.datasheetcatalog.com/datasheet/texasinstruments/opa129.pdf ) I think I can use that, but perhaps someone knows a better fitting one.

And if I made an error in the circuit or possible problems please point them out. And also thanks for all the help I appreciate it very much. 

[dannyf]

That means very high resistance, possibly into gohm range. Very difficult to do with semiconductors.

[Aeon]

That means very high resistance, possibly into gohm range. Very difficult to do with semiconductors.

What do you exacly mean by that? Can you explain statement a little further?

[Marco]

I am still searching for a fitting Op Amp.
I allready found the OPA129 ( http://pdf.datasheetcatalog.com/datasheet/texasinstruments/opa129.pdf ) I think I can use that, but perhaps someone knows a better fitting one.

The sages of usenet like the LM662 for low leakage current (now LMC662). Lower leakage and cheaper.

PS. why not just build the circuit from Linear? (maybe with a slightly larger shunt).

[Aeon]

The sages of usenet like the LM662 for low leakage current (now LMC662). Lower leakage and cheaper.

Cool thank you for the Tip. I will look at that bad boy. lower leakage and cheaper ist very good

PS. why not just build the circuit from Linear? (maybe with a slightly larger shunt).

I thought about building the  circuit from linear, but I still havent fully understand that circuit. Linear has only one Sentence explanation which is very little for an Analog-Noob like me. This project is my first Analog-Project and its difficult for me to decipher Analog Circuits without more than one sentence of explanations.

[Aeon]

Hi I startet my first tests with the circuit I attached to this Post. I wanted to share my first findings with you guys.

I used the following Resistors. (all 1% if not otherwise noted):
R1 = R2 = 10M
R3 = R4 = 470k
R5 = R6 = 47k

RL = 1k

for more details on this circuit see this post: https://www.eevblog.com/forum/projects/building-a-na-current-source-for-a-project/msg386261/#msg386261

I build the Circuit for testing on a Breadboard (see the second Attachment to this post), I know its not precise, but it should be enought to test the concept and if I can get Currents down to 1nA.

As Op Amp I am Using a LMC662.

I am powering it with a plus and minus 5 Volts through a Agilant E3631A DC Power Supply.

For measuring the Output Current I am measuring the Voltage drop at RL with an Agilant 34401A Multimeter and Calculate the Current.

My first tests are very promising. when I am putting 1V on the input of the circuit I got about 1,1 µA and with 1mV on the input I got about 5 nA. I didnt use any shielding and only 1% resistors, so I know the result are not precise but I am in the right ball park.

wuhuuuu  it works (that had to be said  )

Next is testing the circuit from Linear. if they perform the same then I am using this one because the Op Amp is cheaper and its easier to build.

[Rufus]

As for the load: the load will be a also an Op Amp, it will be connected to an Transimpedance amplifier to be precise.

If your load is the virtual earth of a transimpedance amplifier then why would you need more than a voltage source and a high value resistor?

What is the expected input offset of the amplifier? That compared with the voltage source is the only error an active current source would avoid.