I think the startup NVM test is documented in the service manual. I think it tries to load from one page, if that fails, tries to load from the other, but I might not be remembering correctly. In any case, once you get comms working, you can examine the memory. If it's readable at all, hopefully you can get the zener bias value, it's the most critical. (I wrote mine down on a label and put it on the back of my meter) Also, if the meter comes up enough that you have front-panel control, you can try a refresh from there. That copies the cal data from the secondary page to the primary. If you've lost the Zener cal constant, you've got a problem. You'd really need a temp-controlled enclosure that you can vary over a signficant range. The goal of setting the proper cal constant is to bias the Zener into its best zero-delta tempco point, so in theory if you can vary the temp, you should be able to fiddle the cal constant until you minimize the tempco. -- Bill Ezell ---------- They said 'Windows or better' so I used Linux.

Actually, DON"T do a refresh; that copies the current live values back to the NVM. If you can get comms and you can still read the zener cal, you can set it in working memory and then force a write-back to NVM. That might be enough to solve your problem. On 04/25/2012 11:53 AM, Bill Ezell wrote: > I think the startup NVM test is documented in the service manual. I > think it tries to load from one page, if that fails, tries to load > from the other, but I might not be remembering correctly. > > In any case, once you get comms working, you can examine the memory. > If it's readable at all, hopefully you can get the zener bias value, > it's the most critical. (I wrote mine down on a label and put it on > the back of my meter) > > Also, if the meter comes up enough that you have front-panel control, > you can try a refresh from there. That copies the cal data from the > secondary page to the primary. > > If you've lost the Zener cal constant, you've got a problem. You'd > really need a temp-controlled enclosure that you can vary over a > signficant range. The goal of setting the proper cal constant is to > bias the Zener into its best zero-delta tempco point, so in theory if > you can vary the temp, you should be able to fiddle the cal constant > until you minimize the tempco. > -- Bill Ezell ---------- They said 'Windows or better' so I used Linux.