Re: PAL Checksums

[Fred Townsend]

Now I better, although still not fully, appreciate your problem.  I am not familiar with your programmer. On my programmer http://www.xeltek.com/pages.php?pageid=5  if I need another interface I pick the phone and order one. Then I run down the street and pick it up. No so with the Irish I take it. I also take it you are not able to extract a set of known good PALs. Are they soldered in? No second machine? Are the security links blown? None of my customers use fuse link PALs now but when they did most found it was more trouble than it was worth to blow security links. Something about needing two test programs, one for blown and one for unblown security links. I still think you are likely to have propagation problems since the old AMD pals are no longer made and the subs are much faster than the old. Do you have schematics? Fred Townsend Declan Moriarty wrote: Recently, Somebody Somewhere wrote these words I suggest you quit messing around with check sums and get to work. ;-) Unfortunately, that's a no can do. I need the checksum of that PAL to ever get a replacement from these guys in the future. This is to fix a 1999 version of a machine sold from 1992 - 1999. Those guys are 2 model revisions on. Who was it that offered the defination of insanity as repeating the same experiement and expecting different results? Your boolean equations, and therefore your check sums, must be right or couldn't get it to boot with one of the old PALs. Ordinarly PAL programmers automaticly checksum so if there was a problem I think the programmer would say so. Maybe my explanation lacked. One of my PALs has a pin blown (Pin 23) The manufacturers won't replace the one, they feel I should replace the four. They can't supply me correct PALs. Here's where the checksums of mine become relevant. They haven't a clue really, and I have the only working part for one of them. Your symtoms suggest you are running ripple logic with faster PALs. If so, you are trying to fix a bad design rather than a simple compoent failure. In short a hard road to hoe. I Agree totally - fixing a bad design is what I am trying to do. They do have 74BCT (Texas bus logic) all over the place. Here is where I'd start. 1. Check your power supply. Make sure it is absolutly clean. Faster devices are more sucesptable to noise. Add bypass caps if you don't see them on the PALS. Check out the power supply electrolytics. The power supply isn't perfect, but it's reasonably good. I'll add bypass. But this fault developed with heat (Cold everything worked - warm it failed) and then became permanent. I am not thinking of power supply issues, even thou that has had a ressurection. 2. See if you can find low power equivalents for your PALs. Quarter power versions are ideal. That does not mean switch to CMOS. The idea of using low power bipolar versions is they are much slower devices. It's not to save power. Remember speed kills. I can't choose the PALs they program! They won't give me the code. My guess is they are programming in the wrong files. I have reason to believe this (They hunted for weeks and then came up with these). So I'm about proving they have the wrong parts and pointing them at the right ones.