Variable Current Source

Question

^{simulate this circuit – Schematic created using CircuitLab}

I need to build a variable current source that operates off a 36V supply, if possible using a single op-amp and a mosfet. I need to control current between around 10uA - 0.5mA ie very low current. This is for some iontophoresis research I'm doing.

The design I envisage is attached. The zener should give a ref voltage of 2.5V and R3 should give variable current based on I=2.5/R. That's my theory anyway. Anyone able to recommend R values, Mosfet, IC choices? My power source must be 36V.

This circuit would feed into an H-bridge, the output of which would go to an iontophoresis cell which is just two electrodes that connect to (usually) porcine skin (ears), inside a sealed chamber (like a Franz Diffusion cell). Typical resistance of pigs ears is 10k but sometimes we use skin obtained from surgical procedures or cadavers, in which case the resistance is approx 100k. My aim is to test various active molecules in terms of transdermal penetration, so it is quite important to have reasonably stable current for reproducibility. This is more important than accuracy. 36V is a convenient voltage for iontophoresis but in some cases lower voltages are used. For example, to treat hyperhidrosis (excess sweating) of the hands, it is possible to build a simple iontophoresis setup with a 12V battery. However I am dealing with quite large molecules that may require higher voltages.

Thank you!

Answer 1

There is no reason why this should not work in principle, but there are some cautions:

1) Don't use the LM393 - it's a comparator, not an opamp. Use something like the TSB611, which is a rail-to-rail opamp which can work (just) at 36V.

2) You don't show the supplies to the opamp. Assuming it is powered from the 36V supply, the negative supply can be (for example) +12V. This reduces the stress on the device considerably, so put a 12V zener in the negative supply. The amp only draws 125uA, so you need a resistor to 36V to put the zener into its correct operating region.

3) Operating with the inputs 2.5V below the positive supply is OK - it's within the common mode range of the recommended amp - but the reference voltage could be greater (in other words, further below the positive supply) if that gives any advantage.

4) Because this circuit is referenced to the positive rail, you should use a p-channel FET, such as the ZVP3306, so that the threshold voltage is well controlled and not dependent on the load impedance.

5) Surely the value of R1 is wrong?

Answer 2

^{simulate this circuit – Schematic created using CircuitLab}

As I don't know what your H-bridge looks like, I have kept to your original concept. This gives you 100:1 range of current. The LM336 is pretty stable with temperature; if you want even less drift, use LM336B. However, your 250K pot is probably going to be the biggest problem in terms of stability and repeatability, as the cell current flows through it.

A better method is to vary the reference voltage to the opamp and have a fixed sense resistance, but this is difficult in terms of common mode range when everything is tied to the positive supply. I suspect that your H-bridge is the problem in this respect - could you use a bipolar current directly to your iontophoresis cell? I have in mind a OPA454 running on +36V and -36V, with the cell current sensed by a resistor to ground. This cuts out a lot of the problematic design and, incidentally, can give a bigger voltage range across the cell.

This is the direct bipolar drive circuit. I appreciate that you may not have the power supplies for this, but it is a better way to achieve what you want. It also gets away from the H-bridge drive complexity.

^{simulate this circuit}

Power supplies for OA1 can be anything from +/-5V to +/-15V; there is virtually no load so these are not critical. (The value of R1 may need to change depending on the supply voltage.) To avoid complicating the circuit, I have not shown power supply decoupling for the opamps, but it is essential.