GAL programming help

[Iggy_Drougge]

Does anyone here have experience programming GALs? I bought an NS GAL22V10-15 as a replacement for a Lattice GAL22V10-25. When I put it into my ALL-07 programmer, the first "blank check" works fine, but programming it with a JEDEC file fails miserably, and after that, the software just reports that manufacturer number doesn't match. Checking the GAL in another programmer gives no better results; after programing with the same JEDEC file, the verify fails and the GAL is just filled with 11111111111.

[Castellen]

I've programed the Lattice devices on my universal programmer at work.  Including replacement GAL devices for the A3640, etc.  I only had to make a suitable PLCC to DIL adaptor, which was easy enough.

Haven't tried the NS versions and I'm not familiar with the ALL-07 programmer.

The Lattice devices are of couse flash based, so they're re-programmable.  Surely the NS devices would be the same?

Perhaps your programmer has blown the security fuses, which basically means you can no longer read/write to the device.  Might explain why you can no longer read the manufacturer ID.  The default for some programmers is to blow the security fuses, but I would have expected this data to be read from the JEDEC file?

Before programming a device, can you view the buffer contents?  With CPLDs you usually get a large table of O and X indicating cell states.  Doesn't mean much, but you can get a good idea if the JEDEC file appears to have been read or not.

Are you using the correct adaptor?  If you get the Lattice datasheet, you can quickly work out the pin conversion between the PLCC and DIL devices.  It's industry standard, not a Lattice specific thing.


Just for the record, the GAL22V10B-10 devices (as used in a lot of C= stuff) is no longer available, but the readily available GAL22V10D-10LJ devices work fine as replacements and are of course compatible with 22V10 JEDEC files.
The Farnell part number for these is 413-9793.

[Iggy_Drougge]

Yes, surely the NS devices would be the same? I'm sure I read that they had an EEPROM mask and could handle a hundred programming cycles.

I certainly haven't blown the fuses willingly (there is a command to do so, but I haven't heard that a JEDEC file can enforce it).

I have checked the buffer contents, they look good, and the checksum given by the software matches up.

The NS GAL22V10-15LNC are DIL packages, so I just put them with the notch in the direction indicated on the ZIF socket on both programmers.