Protection diodes can be noisier than the signal they protect

The probe input is exposed to a spacecraft charging environment, so it needs transient protection, and D1 and D2 are there to provide it. The trouble is where they have to be: directly on the electrometer input. Every protection device leaks, and the currents this front end is meant to resolve are small enough that clamp leakage can be comparable to or larger than the plasma current itself.

I wrote the protection requirement as a two-sided constraint rather than a single one — the clamp network must protect U3 and the converter *without* producing excessive leakage — and set the acceptance criterion so that it cannot be satisfied by a protection device alone: the front end has to still meet its original noise and leakage limits after transient testing, not before.

The numbers that bound it are targets, not measurements: input leakage below 10 pA under controlled laboratory conditions, and input-referred current noise below 100 pA RMS over the primary science bandwidth. Verifying clamp-diode leakage against those is a Rev B item precisely because it has not been done.

Nothing here is experimentally resolved, and it cannot be until there is hardware. What the review produced is the shape of the trade and a test that would expose it, rather than a part choice that assumes it away.

Verify the clamps against the picoamp signal they protect, with transient pulses applied and precision current measurements repeated afterwards.