Problem Erasing a TM27C512 EPROM

[Terabyte2007]

So, getting a little Retro tonight. I have been testing some new but ancient EPROM's, and I have one that seems to have a stuck bit. Programming with random data, the EPROM verified correctly and tested good. After 20 minutes exposure in my UV box all bits were erased except 1 which I cannot seem to erase. My guess is the chip is bad. All my others checked fine. It was a chip that was never used so it may have been defective.

[SeanB]

Try 40 minutes. I have had some that had really good bits, only erasing after a few cycles of cooking. Not really going to harm a cerdip with UV in any case, and if your lamp is getting a little old it could have lowered output.

[Terabyte2007]

Try 40 minutes. I have had some that had really good bits, only erasing after a few cycles of cooking. Not really going to harm a cerdip with UV in any case, and if your lamp is getting a little old it could have lowered output.

Cool, I'll give 40 mins a shot tonight. The others did well, but this one is stubborn. Lamp is new, so that's not a problem. The spec sheet states 20 mins at something like 15000 mW/M2, don't remember off hand exact numbers right now.

[nadona]

Can you write a 0 on that address with programmer? Just asking.

[Terabyte2007]

Can you write a 0 on that address with programmer? Just asking.

No, it's not electrically erasable. UV erase only.

[amyk]

Make sure the window is clean...  a speck of dirt could be blocking UV from hitting that bit.

Interesting related article: http://www.bunniestudios.com/blog/?page_id=40

[Terabyte2007]

Make sure the window is clean...  a speck of dirt could be blocking UV from hitting that bit.

Interesting related article: http://www.bunniestudios.com/blog/?page_id=40

I believe the window is clean, but I will try to clean it before my test tonight. Thanks...

[PA0PBZ]

Can you write a 0 on that address with programmer? Just asking.

No, it's not electrically erasable. UV erase only.

Erase = make all bits 1, so an erased byte should look like FF.
Program = make bits 0, so I think nadona is asking of you are able to write to that address.

[Terabyte2007]

Can you write a 0 on that address with programmer? Just asking.

No, it's not electrically erasable. UV erase only.

Erase = make all bits 1, so an erased byte should look like FF.
Program = make bits 0, so I think nadona is asking of you are able to write to that address.

Yes I know, and I apologize to nadona, I misunderstood and thought he was talking about erasing. So the answer is still no, I tried that.

[nadona]

Sorry, I was not clear about my question.

[pickle9000]

Will your programmer allow you to burn it?

As long as the data matches the bit it will work in circuit.

[Terabyte2007]

Will your programmer allow you to burn it?

As long as the data matches the bit it will work in circuit.

Yes, I can burn the EPROM except 0000:004F, it is stuck at "B4". Unfortunately, this data will not work. I was testing the EPROM with random data, verifying and erasing in my UV box when I discovered this 1 stuck bit. If I program random data and edit the byte at 0000:0046 to "B4" which is the stuck hex value then I can get a successful burn because the programmer see's that byte as being already there and skipping over the write for that set of bits. My guess, one of the bits that make up that byte is not erasing properly under the UV. SeanB suggested I try 40 minutes which I have not had a chance to do yet but will try this weekend. All other bits were cleared in the UV box. If 40 minutes does not clear, I will probably just toss it and not spend much more time considering I don't use these much anymore.

[sync]

Maybe the EPROM is just broken. If a cell is programmed to long or twice it can be damaged and can't erased anymore.

[Terabyte2007]

Try 40 minutes. I have had some that had really good bits, only erasing after a few cycles of cooking. Not really going to harm a cerdip with UV in any case, and if your lamp is getting a little old it could have lowered output.

SeanB, 40 did the trick! Clean, tested and blank check verified!!! I guess 20-30 for most and 40 for stubborn ones. Thank man!

[nadona]

Congratulation.

[amyk]

Will your programmer allow you to burn it?

As long as the data matches the bit it will work in circuit.

Yes, I can burn the EPROM except 0000:004F, it is stuck at "B4". Unfortunately, this data will not work. I was testing the EPROM with random data, verifying and erasing in my UV box when I discovered this 1 stuck bit. If I program random data and edit the byte at 0000:0046 to "B4" which is the stuck hex value then I can get a successful burn because the programmer see's that byte as being already there and skipping over the write for that set of bits. My guess, one of the bits that make up that byte is not erasing properly under the UV. SeanB suggested I try 40 minutes which I have not had a chance to do yet but will try this weekend. All other bits were cleared in the UV box. If 40 minutes does not clear, I will probably just toss it and not spend much more time considering I don't use these much anymore.

B4 is 1011 0100, so there's actually 4 stuck bits.

Taking longer than usual to erase can mean that the cell was over-programmed (programming pulse held for too long due to some glitch), and it just takes more UV to get the electrons out.

[Terabyte2007]

Will your programmer allow you to burn it?

As long as the data matches the bit it will work in circuit.

Yes, I can burn the EPROM except 0000:004F, it is stuck at "B4". Unfortunately, this data will not work. I was testing the EPROM with random data, verifying and erasing in my UV box when I discovered this 1 stuck bit. If I program random data and edit the byte at 0000:0046 to "B4" which is the stuck hex value then I can get a successful burn because the programmer see's that byte as being already there and skipping over the write for that set of bits. My guess, one of the bits that make up that byte is not erasing properly under the UV. SeanB suggested I try 40 minutes which I have not had a chance to do yet but will try this weekend. All other bits were cleared in the UV box. If 40 minutes does not clear, I will probably just toss it and not spend much more time considering I don't use these much anymore.

B4 is 1011 0100, so there's actually 4 stuck bits.

Taking longer than usual to erase can mean that the cell was over-programmed (programming pulse held for too long due to some glitch), and it just takes more UV to get the electrons out.

Possibly, I am using the TL866CS universal programmer and the chip rating is 100uS, and that's what I was using. It could be that it's just old and tired!