I try to develop a special kind of transimpedance amplifier which is able to servo out a huge DC part of a AC signal. The signal of interest is modulated with 200 Hz up to 20 kHz with an amplitude of about 100 pA to 1 mA. The DC part varies between a few mA and about 300 mA. The whole electronic circuit should produce as low noise as possible. Temperature stability is, in turn, not a big issue. I´ve spend already some time to find appropriate solutions. One might be to use a integrator to servo out the DC component. However, I tried some simulations, unfortunately without success. One reason are the huge difference between AC and DC amplitude. Another way could be the use of 10 or more relatively percise operational amplifier which amplify the AC as well as the DC component conected to the source in parallel. I guess the biggest deal is now the layout and the exact current distribution. Moreover, I´m not really sure if it is possible to build such a system in reality due to all the stability issues.
Due to the measurement accuracy in the later application, I won´t use a simple shunt-instrumentation amplifier combination to measure the current.
Is there another (maybe more elegant) way to overcome my DC cancellation problem?