# bipolar current mirror, unified control

I'm trying to obtain a high impedance, several decades current pump for charging a capacitor from 1uA to 10mA, swing +/-5V.

I arranged two current mirrors like this and is working acceptable for simplicity, cost, etc.

The problem is that I need a precise unified control (voltage) to set the output current, instead of R5/R6.

Two options I tried so far:

1) Controlled FETs, but due to P/N mismatches, control curves are completely different :) I guess worse solution.

2) Using OA's to drive voltages across R5/R6 based on the fact that when approaches +5/-5, mirrors will be "disabled" and maximum charge time will be achieved.

Still not the best deal, OA's needs to be matched and control curve will be most of the time effective at 90% of voltage (near rail) due to log effect of linear charging the cap.

So I wonder if there is a brilliant idea here :))

Thank you in advance!

• See my own question here. – jonk Feb 7 '20 at 8:50
• wow :)) thanks Jonk, but I need only +/-5V and max 10mA, so I presume there is no need for that. Moreover, my concern is for ultra low currents, I need under 1uA constant current sustained. – orfruit Feb 7 '20 at 8:54
• You wanted precision. I offered it. Make adjustments per requirements. You've added a few in your comments. You may want to include them now in your question. By the way, I only require $\pm 10\:\text{mA}$, myself. – jonk Feb 7 '20 at 8:59
• Small remark: you're confusing everyone (well, at least me) by having negative voltage at the top and positive at the bottom. I suggest that you flip your circuit up-side-down so that +5 V is at the top an -5V at the bottom and the transistor arrows all point down. That's sort of an unwritten rule in schematic drawing. – Bimpelrekkie Feb 7 '20 at 9:03
• edited, now +5 is on top :) – orfruit Feb 7 '20 at 9:05