EPROMs - blank but won't program

Solder

Member
Joined
Apr 29, 2007
Messages
862
Reaction score
7
Location
UK, United Kingdom
I've got about a dozen old AM27C512-255DC EPROMs that I've successfully blanked but which I can't program. They verify as blank but I can't program them at all (have tried two totally different programmers). Some of the same type (taken from the same old bootleg board) programmed okay, but this dozen stubbornly refuse to program..

Is the only option to bin them, or is there anything that can be done to 'revive' them?

Thanks
 
By "won't program", what do you mean? Just fail the verify check, or something else? Is any data at all getting written to them? - i.e. do they still check as blank after attempting to program?

Couple things to check. Clean the legs to ensure they're making a good connection to the programmer socket. An empty programmer socket will check as blank too...

If it is programming some data, but not all, and the legs are clean, program the entire device full of zeroes (verify off), then erase it, and try again.

-Ian
 
I've got about a dozen old AM27C512-255DC EPROMs that I've successfully blanked but which I can't program. They verify as blank but I can't program them at all (have tried two totally different programmers). Some of the same type (taken from the same old bootleg board) programmed okay, but this dozen stubbornly refuse to program..

Is the only option to bin them, or is there anything that can be done to 'revive' them?

Thanks

I have some EPROMs here that do the same thing, I haven't been able to fix them. You can try to program them at a little higher voltage and see if they'll take, but I've only had a few of my chips (none are 27C512's) program at a higher voltage. If you're doing them at 12V, try 15V, if you're doing them at 15V, try 21V. That's if they're not programming... and going for a higher voltage CAN fry the EPROM, so... try at your own risk.
 
By "won't program", what do you mean? Just fail the verify check, or something else? Is any data at all getting written to them? - i.e. do they still check as blank after attempting to program?

Couple things to check. Clean the legs to ensure they're making a good connection to the programmer socket. An empty programmer socket will check as blank too...

If it is programming some data, but not all, and the legs are clean, program the entire device full of zeroes (verify off), then erase it, and try again.

-Ian

Sorry, I mean that no data is written to them at all, the programmer software (for each programmer) gives a "program fail" message and the EPROMs remain blank if then blank checked.

Legs have been well cleaned (and look clean), etc.
 
I have some EPROMs here that do the same thing, I haven't been able to fix them. You can try to program them at a little higher voltage and see if they'll take, but I've only had a few of my chips (none are 27C512's) program at a higher voltage. If you're doing them at 12V, try 15V, if you're doing them at 15V, try 21V. That's if they're not programming... and going for a higher voltage CAN fry the EPROM, so... try at your own risk.

Thanks. Just checked and don't seem able to change the programming voltage ........ but as you say, I'd probably fry them anyway. :)
 
Thanks. Just checked and don't seem able to change the programming voltage ........ but as you say, I'd probably fry them anyway. :)

You never know... I have some 27C32s that are "supposed" to program at 12V, none of them will take any data until I pump up the voltage to 21V... That should be fry'ng them, but they program and blank and program again just fine at 21V. My 27C64s won't program at anything under 15V. It's all about how many you have and how lucky you feel.

Plus, look on the bright side... if you can't adjust the voltage for the chips, then you're not working with a Willem. Be thankful.
 
You never know... I have some 27C32s that are "supposed" to program at 12V, none of them will take any data until I pump up the voltage to 21V... That should be fry'ng them, but they program and blank and program again just fine at 21V. My 27C64s won't program at anything under 15V. It's all about how many you have and how lucky you feel.

Plus, look on the bright side... if you can't adjust the voltage for the chips, then you're not working with a Willem. Be thankful.

Every cloud has a silver lining. ;)
 
Couple of things come to mind... they could have become damaged somehow in the past (fried, whatever) that prevents them from being programmed again. If so, then they're bad, and can't be used.

If you have a chip that only programs at the wrong voltage... now that's weird. The only thing I can think of there is that the chip is mislabelled - perhaps a "bootleg". AMD is a well known maker of quality parts, so perhaps some unscrupulous company simply doctored their crummy parts up to look like AMD's. Try programming them with the generic 27c512 algorithm, and see if that works. I can't say if I ever remember seeing an NMOS '512, but I guess it's possible - see if there is a setting for that.

EPROMs are pretty resilient critters, but I've run across a handfull of dead ones. You may just have a board that got good and fried and nothing can bring those chips back.

-Ian
 
Do you have a known good 512 to test with? If it passes on known good 512, I would just pitch the bad ones. Roms are inexpensive (compared to other things IC related). How much time and agervation you waste, be better to pitch and get some new ones.

Kevin
 
Do you have a known good 512 to test with? If it passes on known good 512, I would just pitch the bad ones. Roms are inexpensive (compared to other things IC related). How much time and agervation you waste, be better to pitch and get some new ones.

Kevin


+1. You can waste alot of time trying to save a few cents. Stick 'em in a bag, mark it as "bad?" and move on.

ken
 
Oh yes, no problem with other 512's, I just didn't want to 'waste' a dozen 512's if there was any viable way to resurrect them. :)
 
Do you have a known good 512 to test with? If it passes on known good 512, I would just pitch the bad ones. Roms are inexpensive (compared to other things IC related). How much time and agervation you waste, be better to pitch and get some new ones.

Kevin

+2 No matter how badly I need an EPROM for something, I keep a couple of each type that are known good for testing. 1 for verification testing, another for programming testing. Since my programmer is shit, I have to make sure I remember how I had to tweak the settings to make it work... that's the programming test EPROMs job (poor thing), and since sometimes the programmer will program bad information, then read it incorrectly (same way that it programmed it bad) and it will pass a verification... that's where the other chip comes in. If the programmer is reading incorrectly, then it'll read the known good chip as bad and it won't pass WinRomIdent (usually a reboot and a cool-off for the programmer will snap it back to where it should be).

But, it'll never pass a blank test on a ROM if there's information on it... if it's not programming anything, then it's just not programming anything.
 
I've had the same scenario where a chip from particular manufacturers that would not program without bumping the VPP voltage up regardless of what the datasheet says. Granted I only do that when all else fails. Figure if I can't program it now, then there's nothing to lose if it cooks it. So far I haven't damaged one yet. Recommend starting with what the datasheet says and then go from there.

And I am working with a Willem.....PCB3 ..... and I can adjust the voltage.
All depends on which one you have.
 
I've had the same scenario where a chip from particular manufacturers that would not program without bumping the VPP voltage up regardless of what the datasheet says. Granted I only do that when all else fails. Figure if I can't program it now, then there's nothing to lose if it cooks it. So far I haven't damaged one yet. Recommend starting with what the datasheet says and then go from there.

And I am working with a Willem.....PCB3 ..... and I can adjust the voltage.
All depends on which one you have.

Yeah, same... But I still have issues with 2732 and 2532's regardless... 2532's are just a nightmare, but 2732's will program at the proper voltage IF I adjust the "timing" on them a little... and even then I have for program them 4-5 times in a row to get the code to "stick". But, whatever works, works.

An easy way to verify the chips are good from the start is a .bin of all 00's. If it'll erase all the way to all FF's, and program all the way to 00's... then erase to all FF's again, it's a good chip.
 
+1. You can waste alot of time trying to save a few cents. Stick 'em in a bag, mark it as "bad?" and move on.

ken

+2

Best advice of the day. If these are used chips they may have been abused by someone (like plugging them in backwards or UV over-exposure) or if they are old they did their job (most were not designed to last as long as we are using them - a testimony to how well they were designed or how little was known about long term reliability / useful life of ICs). If they are all identical (from same manufacturing lot) and one or two don't work, they all have the same issue. You've already spent too much time with them. I've had similar situations come up and after a quick verification that my programmer and eraser are ok, and one more erase/programming try (in case of operator error), I pitch 'em. I've tried to resurrect eproms before and spent way too much time only to come to the same conclusion - file them in the round file or save one or two for their legs to repair another ailing IC.

Bill
 
Back
Top Bottom