Current mirror with stabilizing emitter resistors

[selotelefon]

When i used 3.3k resistor for Rl i observed major change in the current. What accounts for this change?

[magic]

Saturation. Q4 has too low collector-emitter voltage to work normally.

[mawyatt]

Since the current mirror is derived from the same voltages with same emitter resistors and same device multiplicity, then RL can't be any larger than Rref, otherwise Q4 will saturate. With RL equal to Rref, then Q4 has a similar Base to Collector voltage of ~0 volts as forced by Q3's Base to Collector connection. As RL increases beyond Rref then Q4 enters Quasi-Saturation, then full Saturation.

If one requires RL to be higher in value (with lower current), then changing the ratio of RE1 and RE2 allows scaling, ultimately becoming a Widlar Current Source as RE1 approaches 0 ohms, or changing the active device ratios allows the output current range to be higher or lower than the current supplied by Rref.

Many options with resistor and/or device ratios.

Best,

[ArdWar]

Depending on the power limitation of your chosen components - you could (theoretically) increase the source voltage to 15+ volts to sustain the Iq current level.

I guess you mean lowering Vee to -15V? Not really since you'll also change the ref current that way. And the load side transistor will stay saturated. As a rule of thumb assuming transistor and supplies are identical you can't put larger load resistor than the set resistor (minus some base current so actually slightly larger).

However you may supply the load with different voltage. (i.e. "lifting" the R_L ground)