I know that SI issues (noise, non-monotonic edges) can readily
cause DRAM errors, but none of the usual candidates seem to be
the source of the problem I'm trying to fix..
My precise question concerns the effect of a slow-rising
data bus. But first, a brief rundown on the background to
the question:
We have a board which has two identical DRAM memory systems:
- same type of address buffers
- same type of ASIC for data buffering
- same type of controller (control signals)
- same number/type of memory devices
- same routing (my opinion)
One memory system (A) has many more errors (x100) than the other (B).
The errors are "soft" and only a single bit is in error each time.
On average, a board will have a soft error every 6 months.
Obviously, alpha particles cannot explain all these soft errors.
All the timings are OK. The refresh (CBR) is OK.
That leaves SI, which is my job....
I've checked and double checked, but all the signals
are noise-free, with nice clean edges.
One of the few leads - and its a long shot - is as follows:
A difference between the two memory systems is that the data bus
of memory system A is left in a high impedance state for
much longer intervals (10's of microseconds) than system B.
So the signal level is more likely to float into the input
threshold zone of the memory and the ASIC.
So my question is:
Can a floating bus create enough noise inside a memory device
to give rise to a soft error?
Thanks in advance for any comments/advice,
John
-- John Fitzpatrick <[email protected]> Alcatel Telecom, 4 rue de Broglie, 22304 Lannion, France Tel: +33(0)2.96.04.79.33 Fax: +33(0)2.96.04.85.09