Strange issue with 27C512

Hey guys!

I know that parallel EPROMs are not the latest hi-tech technology but I have a problem to solve.

I have two equipments which have a 27C512 EPROM. One of them is faulty the other is OK. I found that the problem is the memory.

So, as to replace the memory, I read the dump from the good EPROM using a Genius G540 programmer. The dump was read OK (I guess).

Them, as far as I know, I used a new and blank 27C512 also from the same manufacturer (ST) to program it.

However, the programming failed. I just got a 73h on the 1st position, and the rest full of FFh. The strange is that the 1st position of the original dump is not even a 73h!!!! ^^!

Then, as I don't have (yet) a UV eraser, I tried to erase on the sun... yes I know it could take a couple days/weeks depending on the sun.

But the must strange thing is that after a few hours of sun... I got "zeros" where I had "ones". How can that even be possible? The UVC only changes "zeros" to "ones"... not the vice-versa :/

Do you think that the new EPROM is faulty? When I bought it it was supposed to be new and blank!

Thanks!

Do you think that the new EPROM is faulty?

Yes, either that or your programmer.

The strange is that the 1st position of the original dump is not even a 73h!!!! ^^!

Not strange it means all the bits did not program, write down this in binary and look at the binary of what it should be. You should see some bits were zero in the required word that did not get zeroed in the EPROM.

Grumpy_Mike:

Do you think that the new EPROM is faulty?

Yes, either that or your programmer.

Yah! I think you're right! I only hope that the problem is the EPROM and not the programmer. I never had problems with other chips... :~ PS - I never tried to program 27CXXX memories.

Grumpy_Mike:

The strange is that the 1st position of the original dump is not even a 73h!!!! ^^!

Not strange it means all the bits did not program, write down this in binary and look at the binary of what it should be. You should see some bits were zero in the required word that did not get zeroed in the EPROM.

Yah... it could be that

But, and about the sun? How could the sun set a couple "FF" (11111111) into "7F" (01111111). For me it's really really strange. The sun can only change "zeros" into "ones", not "ones" into "zeros".

This is a demon EPROM XD

How could the sun set a couple "FF" (11111111) into "7F" (01111111)

Well that sounds like a faulty chip to me.

If a cell is borderline it could read as 0 or 1 more or less at random depending on supply voltage, temperature etc...

You need the right kind of UV lamp, its hard UV.

MarkT: If a cell is borderline it could read as 0 or 1 more or less at random depending on supply voltage, temperature etc...

You need the right kind of UV lamp, its hard UV.

Oh! I see.

Thanks ;)