I built a super wilson current current mirror. The Iin is 4-20mA at 24Vdc. Iin to Iout accuracy was a really bad 5%. I put 100 ohm resisters on the source to ground, and that helped a little. How can i improve the accuracy between input and output? Below is my circuit

here is the simultor link for the circuit: sim link

EDIT: the 500 ohm is the load resistance of the circuit using the mirror. The 470 ohm is the max possible resistance of a PLC that reads the Iin current.

Screen shot:


  • \$\begingroup\$ I am unable to see your circuit. Can you post a screenshot? \$\endgroup\$ – hassan789 Feb 27 '13 at 18:39
  • \$\begingroup\$ link to screen shot i.imgur.com/o6ZTRnH.png \$\endgroup\$ – David Feb 27 '13 at 18:46
  • 2
    \$\begingroup\$ Do you mean you actually built it physically? Or just in the simulator? \$\endgroup\$ – Oli Glaser Feb 27 '13 at 19:04
  • \$\begingroup\$ I built it physically, accuracy was 5% as stated above. I have 2 dual Nmos chips, fds6961A. I took out one pair of Nmos and the accuracy improved to 2.5%, using the most basic mirror. \$\endgroup\$ – David Feb 27 '13 at 19:39

I would use bipolars, not FETs. The FETs have to be particularly well matched for this to work, which is difficult to guarantee with discrete parts. Use two NPNs. With enough emitter resistance, you should be able to to better than 5%. Keep in mind you can't use 5% resistors if you want accuracy better than 5%. Try something like this:

If that doesn't do what you need, a opamp circuit should do it, but will be a bit more complex.

  • \$\begingroup\$ I am getting 1% and 2% accuracy using the 390 ohm emiter resisters. 220 ohm is 5.5% accuracy. All resisters are 1% or .1%. I am using a BCM61B matched BJT chip. But 2 2n2222 BTJs gives a 1% accuracy at 390 ohms. Thank you Olin. \$\endgroup\$ – David Apr 8 '13 at 23:25

One well known way is to cascode the current mirror, this helps increase CMRR, linearizes the mirroring but also limits the range of operation.

  • \$\begingroup\$ I physically built the cascode and got a 4.7% accuracy. \$\endgroup\$ – David Feb 27 '13 at 19:41
  • \$\begingroup\$ @David so you mean you breadboarded this some how? DId you use matched pairs? unless you actually build this is Si, and then took care to have common centroidal layout, you can't expect any level of matching and therefore any level of accuracy. In fact that would be a useful exercise for you, do the calculations that show the impact of transistor mismatch. You're lucky you got 5%. \$\endgroup\$ – placeholder Feb 27 '13 at 21:55
  • \$\begingroup\$ I can see how the differences in W/L can effect the mirror. The 2 lower mosfets in the willson CM dominate. I was hoping that having 2 mosfets on the same chip would be close enough, but i guess not. \$\endgroup\$ – David Feb 27 '13 at 22:14
  • \$\begingroup\$ I also thought a 4 mosfet mirror would be better then a 2 mosfet mirror as talked about above. \$\endgroup\$ – David Feb 27 '13 at 22:15

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.