We have used Ocean Optics spectrometers at our company for quite a while. We also used those same drivers (and in many cases, still do). We also got those same errors.
Best we could ever do was put all the code in the main "read spectrum" VI in a FOR loop with the following conditions:
Number of loops = 2 or 3 (configurable)
End loop condition = No error
If error was detected at end of loop and there were still loops left in the FOR loop, clear the error, re-run the Initialize VI, then return to the start of the loop
I suspect that the Ocean Optics folks have either fixed their internal software to not have these problems or have just set it to ignore errors, but never updated their LabVIEW code to do the same.
Going forwards that's the best suggestion I have if you insist on sticking with the drivers you use now. If you're willing to do a rewrite (and possibly an OS update, I don't think Win 7 will work with it), then look into their "OceanDirect" API. It's a .NET DLL that LabVIEW can interface with. I've got it working with just one spectrometer so far and have not nearly the same amount of silly errors that their LabVIEW drivers have sometimes had, but I've only used it briefly.
The downside is that you do need to contact them directly to get it as they don't have a free download on their site for it. As far as I know, as long as you can prove you bought a spectrometer from them they will send it to you for free, but if you're a 3rd party wanting to use it then there is a charge. I didn't deal with them directly though (manager did it for me) so there might be a bit more to it than that.
