6485 overranging for signal at 1% full-scale

I am using a 6485 to measure DC photodiode currents around 90pA-100pA. When the photodiode is in photovoltaic mode (no reverse bias), the 6485 overranges when put into the 1nA and 10nA ranges, but doesn't overrange when in the 100nA range or higher. When the photodiode is put into photoconduction mode (reverse bias applied) with a modest voltage (500mV to 1V), the 6485 no longer overranges in the 10nA and 1nA ranges. The 6485 also doesn't overrange when 100pA is applied using a voltage source and 10M resistor. Why does the 6485 overrange when the photovoltaic mode? Is it perhaps due to excessively high source impedance?

Is likely to be due to the different capacitance of your photodiode if at 0V vs some reverse bias.
The DUT C will be larger at 0V.

In 6485 Instruction Manual, look at section noise and source impedance in Appendix I.
It gives a nice treatment of the topic in how the R and C of the DUT interacts with picoammeter gain.
If it is too much C, adding some additional series R can mitigate in some cases.  Just need to account for the I*R drop on the added R.

What sort of cable do you use?  How long?  It too is contributing some C.

This solved it! Adding resistance inline fixes the issue. I tried 10k to 10M and they all worked. I used the BNC cable that came with the 6485 (0.5m I think?)

Where would I find the 6485 instruction manual? I couldn't find it in the 6400 family downloads section. I did manage to find a copy hosted by Purdue University.

As for the manual section I referenced, I see it is gone from what is presently posted on tek.com for the 6485.
I have some older revisions on my computer.

I'll upload the relevant pages for the topic of how the source R and C of the DUT interacts with the picoamp meter (or electrometer) current measure functionality.

Good to know that I'm not going blind. Thanks for the help.