🚀 go-pugleaf

In article <R3sTqB1w164w@sys6626.bison.mb.ca>, flux@sys6626.bison.mb.ca (John
Kamchen) says:
>
>I am having major problems trying to program a GAL chip.  I'm using a

[stuff deleted...]

>I have blown three GAL chips so far.. the last one came a a shock.. I
>placed it in the circuit, issued the command to program it, and for the
>first row, the EDIT voltage was a steady 16.5v .. then came the /STR

Edit voltage requirement is on the order of 1 - 2 mA.  I generate mine
with a capacitor charge pump with 74HC04.  BTW I notice the chip went
into latch-up during programming.  Luckily I have current limiting in
the power supply as a side effect.  That might explain about your
2.76V at Edit.

>I'm using part number  PALCE16V8H-25  from AMD.  I could find NOTHING on
>these chips at my local store.. nothing (I read that AMD sends the
>programming info ONLY to registered manufactures of PLD programmers)
>ANY speck of help would be a ray of light in my bleak tunnel.

Yeap.  I can confirm as I spent $$$ trying to talk to the marketing
Engineering.  I was told they decided (I think they were the ones
started all this) not to release programming spec after customers
complains about bad chips.  I think they should have full programming
specs plus sample circuits for programming instead of the other way
around.

If you look back a couple of weeks, you can see me requesting PALCE
programming info.  I would bet $10 you bought the PALCE chips assuming
that they are compatible with GALs from Active Component (aka Future).
The PALCE chips are NOT GALs in spite of what they told you.  I learnt
that lesson after spending $50 on a major order for 20 chips.  I was
told that it was my fault and can only get about 50% restocking with
purchase of equal amount of goods.  (ie I spent 150% if only they
have Lattice in stock.)

Normally I am pleased with them, but in this case they are seriously
at fault.  BTW GAL is a tradmark, that's why only true GALs from
Lattice, NS (& possible SGS Thomas) would have GAL as prefix.
Accept no substitute (ie. PEEL, PALCE etc) no matter what they tell you.
You'll thank me for that.

>(Quick info: my programmer shuts off Vcc and EDIT before a chip is placed
>in the socket. I cannot see anything wrong with my hardware or software)

I was going to do that, but when I read about the Latch-up protection
paper in my Lattice GAL databook, I was too lazy to put that in.
They claim their chip won't go into latch-up under live insertion and
a bunch of conditions that blew up the output drivers.  That's why I
suspected PALCE has a different programming interface than GAL.

>;--- (John Kamchen) a user of sys6626, running waffle 1.64

K. C. Lee
Elec. Eng. Grad. Student

P.S.  Before anyone ask me where I got my GAL programmer.  I designed
mine from spec under sort of non-disclosure from National Semiconductor.
I guess these guys deserve a mentioning whenever it is appropriate.
Afterall I have the freedom of not picking manufacturers that I don't
like for various reasons (eg. too secret about things that user should
know.)