kick-back on the diode voltage current mirror

Hi all. How do you overcome a click back on the diode voltage of a current mirror when you enable it?

"Enable it" means turning on the current source into the current mirror or what? Where do you see a "kick back"?

I see it on the diode voltage that bias the mirror. The switch is on the mirror side and it is enabled by a signal that controls it.

You would prefer to "steer" current softly, rather than
hard-switch it. Switches have charge injection and a
scheme that connects, disconnects abruptly can make
charge "find its own way out".

For example, use a diff pair as your current source
that feeds the mirror, and you can steer the current
with maybe 100mV of emitter / source voltage
instead of 5V CMOS gate voltage. Take a look at
older bipolar DACs to see how they get low glitch
energy bit-current switching. Or CMOS IDAC if
you can find one.

If the mirror is current-fed, then you might find
some benefit in adding series resistance inline;
not so much that it bothers low-Vdd fidelity, but
it will "mellow out" switching edges against the
mirror input capacitance.

Some current mirror styles are easy to pump and
have no bleed besides base current (or MOS drain
feedback). Ones using a booster can over-drive
the reference rail and take time to return to setpoint.
You might lash up every kind of mirror you can, and
bang them all the same, to see which has the best
close-in settling behavior.