How to do?
First check, if it makes sense, analyse the options, the eforts in time & cost, than kill or start the project.
That I did.
Tool offers DC voltage up to 1111V, DC current up to 111 mA (for burden voltage up to 500V).
Design is quite good from the principle, the used precision resistors, the power supply, the pre-regulator solution, the shielding. The Plug-In card design allows a partly "pimping" with acceptable efforts (you do not risk the total system).
Just the old style chopper amp and the reference prevent a much better level of precision.
A good example you may find at FM, as mentioned by Dr.Frank already.
http://www.friedrich-messtechnik.de/index.php/bauanleitungen/87-fluke335-332My Project Plan
Step 1: get it to full function / repair - Done
Step 2: measure actual short term drift - Done
~12 pmm, thus the chopper replacement is a potential option
Step 3: measure the long term drift - still running
~40 ppm/y so far according to 3458A, still inside the spec (6 ppm/month), but a potential option to replace
thinking in the LTZ1000A as I have one spare, but LM399H may also an option.
Sounds serious, isn't it?