# Designing current mirror

I have a 105V/3A output current source, and I am trying to get a current mirror circuit to pass 10-100 uA DC across 5 lanes that have very high impedance (around 1 MOhm). What is the best/quickest way to design this and what transistors should I be getting?

Thanks in advance. I've never designed such a circuit but my project is on hold until I figure it out.

• It sounds like you want a "current loop" with high voltage. 100uA x 1M Ohm = 100V. It can be just a scaled current loop. Find 4-20mA current loop examples, understand it, and then scale it up.
– jay
Nov 22, 2021 at 16:25
• @jay could you please elaborate a little bit more on why this is a useful solution? I'm familiar with current loops from my old college classes, but I don't see how I can create 5 channels form a single current source using current loops.
– Bee
Nov 22, 2021 at 16:57
• I agree with you, and did not say to drive 5 ch from a single channel. I didn't focus on "current mirror" part, sorry. You would create that current mirror at low voltage, than driver stage takes it to high voltage. 1MOhm is rather high impedance for a distance signalling, will suffer from noise, and costly medium.
– jay
Nov 22, 2021 at 17:25
• Is this a school assignment? Can you tell us more about "my project"? Nov 22, 2021 at 18:03
• It's not a school assignment, it's a project I'm doing for fun where I'm passing current through a saline solution and studying the electrochemical reactions. I want to test the setup with 5 lanes at once -- one passing 10 uA, the next passing 20 uA, the next passing 30 uA, then 40 uA, and finally 50 uA. But I only have one current source to do this.
– Bee
Nov 22, 2021 at 19:03

Okay, worst-case you have 100uA at 105V = 10mW even into a short. So any transistor that can handle the voltage (and that has relatively low leakage compared to 10uA) will do. Your 300W+ power supply is a bit scary, so maybe we want some series resistances in there in case something goes wrong with the transistor, say 20K 0.5W on each channel.

simulate this circuit – Schematic created using CircuitLab

Maybe something like the above. The zener forms a power supply for the op-amps (rail-to-rail and very low current). The 20K resistors can be 1W and rated for 200V or more.

I see what you are doing now, "saline experiment". I thought, you were working on a "signaling". I just become interested for seeing puzzles. This is not an opinion, but an answer, if you are looking for a solution; the other way around, if you are looking for yes/no.

I have a 105V/3A output current source, and I am trying to get a current mirror circuit to pass 10-100 uA DC across 5 lanes that have very high impedance (around 1 MOhm).

It is possible. You may get 100uA through 1M ohm, assuming the power and the medium stay stable. But, in practice, electrical side will suffer from the operating margin, if the accuracy was desired,or the conductivity/voltage-source cause noise/variation. It is simple to see it as an EE, but will take your intuition.

-. The voltage developed on the DUT/subject = 100uA x 1M ohm = 100V.
-. Operating margin, given to the circuitry, without considering other variables:
105V - 100V = 5V
-. The freedom of control, of the 100uA on 'varying' variables:
5V/100V => 5%

What is the best/quickest way to design this and what transistors should I be getting?

Transistors will only reduce the operating margin. All you need is a few resistors, assuming many parameters are fixed stable.
-. For an example: 105V / (1M + R) = 100uA --> R = 50K ohm

I'm passing current through a saline solution and studying the electrochemical reactions. I want to test the setup with 5 lanes at once -- one passing 10 uA, the next passing 20 uA, the next passing 30 uA, then 40 uA, and finally 50 uA. But I only have one current source to do this.

That implies many.. things.., and, I will try to cover many, dealing with a Chemist :).
-. What you described is more of voltage source, not a current source. Thus, no current mirror is needed.
-. That is only true, if the medium gets separated for each current path, like separate tube for each current source.
-. A constant current as a source works, if the condition of the medium stays in operating range while giving enough margin to the control circuitry.
-. If the 3 probes go in the same 'bath', then there is no way to control the current path away from the source.
-. However, you may be able to force the source and sink point, multiple of them in a bath, by isolating (electrically) each probe sets.

My point is: If you show us more of what you are doing and how it works, then EEs can provide better/right solutions.Besides, you did not explain how and what to detect from the "reactions".