how to avoid the switching transients genertated by ADG1206 multiplexer

There are two effects, charge injection as Harry outlined and capacitive load on the mux.

The mux will connect input and output through a low impedance switch which will quickly equalize the voltages on both sides of the switch. Any capacitance will generate a significant current that will be reflected as a voltage spike based on the impedances involved. A voltage buffer might actually make things worse as you are upsetting the feedback node, by suddenly pulling current out of the OPAMP testing the stability of the feedback.

In any case you can separate the two issues by switching between two nodes of equal potential (two shorted inputs).

Then all you will be left with is charge injection.

Klaus

Dear harryh

I agree it is quite common to see spikes during switching.

As iam giving sufficient time 100 microsecond after switching
the switching spike is not captured by ADC.

But the switching noise which is seen at the input of the multiplexer which is connected to the AD588 reference output.This Reference output is used in the analog output genertation.

So the spikes are seen at the analog output.

How to avoid the same

i did an experiment to feed the same voltage as the reference output in the previous channel of the multiplexer.

Which means two consecutive channels are at the same potential. but the switching noise spike is still seen.

this switching noise spike is reflected in the reference output (which is connected in one input channel of the multiplexer) which is used during the analog output voltage generation. hence in all my analog outputs the noise spike is seen. Actually i want to avod thte nosie spike at all the analog outputs whose source is form the multiplexer swiching

It would be helpful if you could post your schematics to clarify what you are having trouble with and maybe some pictures of where and how you are doing your measurements.

The ADG1206 has less than 1pC of charge injection, so at this point I'm a bit suspicious of your measurement setup...

Klaus