From: "Pete Turnbull"
<pete(a)dunnington.u-net.com>
On Aug 13 2004, 11:27, Dwight K. Elvey wrote:
You should try to create a more exhaustive RAM
test.
Although, time consuming, GALPAT is just about the
most intensive.
I know Dave doesn't need it now, but a GALPAT is a very slow test for
any sizeable amount of RAM. It's also very old, and not 100% effective
against certain errors, though much better than some of the common
march tests.
Hi
There are many kinds of test. The simplest effective
test for stuck-at and address problems is the March C.
It lacks many possible problems that are not simple stuck
at. Worst case column test require various types of walks
( GALPAT is one of these ). Still, most simple RAM test
Used by POSTs will do about as well as a March C. Now
we get into some of the other problems of address line
switching, retention and pattern problems.
One need not include the entire memory in there GALPAT
test. One can work in blocks and then work on a block to block
sampling. It is a long test but then it does a lot of
things that take a long time.
I only suggested it because it seemed that he had an
issue that a simple RAM test wasn't showing.
I am involved with creating BIST engines for use on
arrays in uP's I would never put a GALPAT in a production
test but if I can get two address generators, it is
easy to implement as a programmable element. It, along
with other test, can often be used for diagnostic operations
to help understand a fault until a simpler/quicker test
can be used to detect that particular fault.
Dwight