I have been trying to debug a development board but I can not get the circuit to fully work.
The issue appears to be my replacement for the Samsung K6X4008CIF-BF55 which is fitted to the RCM2250. My device is an Alliance AS6C4008-55SIN.
Most of the diagnostics program work fine, but when I clone our own software, the RTL device is continually cycled on it’s reset line from the Rabbit but the rest of the program does not work correctly. If I use the DynamicC to compile the software into the board, the software works correctly but the ethernet is held permanently in reset instead of cycling. The software works fine using the original SRAM.
I have checked the data sheets and the devices appear identical apart from some minimum read/write times which are only off by 5nS. Note these are the MINIMUM write times not the typical times.
Can anyone shed any light on the problem?
One thing I noticed in looking at the data sheets is that the minimum input high voltage is different if you’re running them @ Vcc=5V. For the Samsung, it’s specified as 2.2V, for the Alliance, it’s 0.7*Vcc (=3.5V). At Vcc=3.3V, they’re virtually identical.
Other than that, the timings that you mentioned could be causing problems. Have you tried running with a slower clock to see if that makes the problem go away (or otherwise verifying the timing)?