Depending on your specific systems that may not even be possible. The general principle is that you want the lock-in to be able to accurately determine the phase shift between applied [voltage] and measured [cuurent], but it needs to be the potentiostat that is controlling that potential. If there's some pass thru mode where you can take an external signal in (you probably need a signal generator too) and hand it over to the potentiostat to apply to the cell, then you'd get the current value back from the potentiostat and use the lock-in to get the phase shift--the amplitude would come from the potentiostat.

There's really a good reason why people don't do this anymore: it isn't easy to setup, you either need to run one frequency at a time, compile data, then adjust to a different frequency or do some programming to talk to all the various systems to automate it, and you also need to account for system artifacts/contributions yourself, which likely means manually adjusting filtering and range settings, and doing all this generally results in data that isn't very accurate relative to modern commercial setups that are designed to do EIS.

Thanks Jacob for your valuable answer

In fact there is a modern EIS system in our laboratory which works based on FRA. I need to assemble the lock-in amp based EIS for specific purposes and my own experience. I need to know how can I use internal voltage controlled oscillator of the lock-in amp to make excitation signal and how can I feed the resulting current to obtain in-phase and out of phase signals?

Your best bet would be to contact the manufacutrer(s) of the lock in (and potentiostat).