I think this would be improved if you replaced R3 and R7 by a current mirror, and added a trimpot in the emitters of Q6 and Q7; and add the 99 ohms in series with Q3.
It may also benefit from an oscillation suppression cap base-collector of Q8.
The trimpot can be used to maximise the current range linearity.
It may also benefit from an oscillation suppression cap base-collector of Q8.
The trimpot can be used to maximise the current range linearity.
I have finally found a satisfactory solution: simple and high performance.
The current ratio of 1:100 holds for a 1:1,000 range, ten times better than what I required.
Q3 is the gain stage/error amplifier, and Q8 provides buffering and Vbe compensation.
R9 needs to be tweaked wrt. R8, because the transistors aren't perfect compliments, and a schotkky has to inserted in Q8's collector, to accommodate for the opposite polarity, leaving it only 0.4V Vce to operate at the minimum current, but it seems to work well enough in sim
The current ratio of 1:100 holds for a 1:1,000 range, ten times better than what I required.
Q3 is the gain stage/error amplifier, and Q8 provides buffering and Vbe compensation.
R9 needs to be tweaked wrt. R8, because the transistors aren't perfect compliments, and a schotkky has to inserted in Q8's collector, to accommodate for the opposite polarity, leaving it only 0.4V Vce to operate at the minimum current, but it seems to work well enough in sim
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.