Hi,
here is another 2710 which had this problem.
I bought it from a colleague who described these symptoms before repair:
- Power-on messages about NVRAM / file system errors
- Spurious characters on the display
- Useless files which could not be deleted
After exchanging battery, RAM and bus drivers on the display storage board, he used a logic analyzer to track down the error to a faulty OE- signal from PLD U352 Pin 24 to bus driver U462 Pin 19. The CPU could still write to the NVRAM, but reading was not possible.
The faulty OE- signal could be reconstructed from other existing signals by two SMD logic circuits (74HCT04 and 74HCT00).
After his repair, almost all analyzer functions worked, but many amplitude normalizations always failed.
It cost me some more hours to find out that the CPU will read the acquired signal not only from the NVRAM, but also accesses the ADC data directly. The logic signal for this is called REDAT-.
So I extended the replacement circuit, and now the amplitude normalizations also work. It seems that REDAT- is not used during normal operation, but only for calibration.
A simplified but logically equivalent circuit is attached as PDF. It uses a 74HCT08 and 74HCT00 for reconstruction.
The ICs are stacked onto U566 (right besides the bus connector) and powered by vertically soldered wires on Pins 7 and 14. All other pins have been bent to go out straight horizontally and are connected using thin isolated copper wire (as used e.g. for transformers).
Pin 19 of the bus driver U462 is lifted from the PCB, the original DROE- signal is taken from the PCB-pad and the reconstructed signal is routed to the pin.
I cannot be sure that all 2710 can be repaired this way, but since the symptoms described here are very coincident, there is a good chance to make them all work again.
Good luck,
Andreas.