Testing for shorted bits

greenspun.com : LUSENET : ece342 : One Thread

I'd like to know when testing our DRAM chips, if two bits are shorted would they necessarily reside beside one another in a byte.. or could they be the same bit number in two different byte addresses?

-> Gino Filicetti -- ginof@interlog.com

-- Gino Filicetti (ginof@interlog.com), February 09, 1999

Answers

I know Gino has already finished the lab, but I'll answer for the sake of general interest.

When we speak of shorted bits we are talking about physical shorts either of wires or the breadboard/headers. (If you leave a lot of exposed metal when inserting a wire into the breadboard and two wires are next to each other they can touch.) We aren't talking about shorts inside the DRAM or Altera chips. These are possible but not very likely. If a chip gets blown up it is usually catastrophic and the chip will get very hot when power is applied.

Remember that these shorts can happen anywhere along the chain of connections. Thus although the high and low nibble of the byte in the DRAM lab are in two separate chips (which might lead you to think that there could not be a short between any of the high 4 bits and any of the low 4 bits) you probably put the two nibbles next to each other on the 60-pin header, so the 4th and 5th bits could be shorted.

The best way to find shorts is to remove drivers and see if the wire voltage level floats. So let's say you have two wires being driven by pins on the 60-pin header. If you pull up one of the wires at the header side (so it is not being driven by anything) and the logic probe says it still has some valid logic value (i.e. one of the lights on the probe is on) then you know they are shorted at the other end.

robin

-- Robin Grindley (grindley@eecg.toronto.edu), February 11, 1999.


Moderation questions? read the FAQ