Unable to program 27C1024

[GiGaBiTe]

I have an IBM PS/1 that I accidentally corrupted the BIOS after dumping it (hit the program button after dumping it.)

No problem I thought, I'll just use my Xgecu T48 to reprogram it. I got out my UV eraser and erased the EPROM, verified it was empty and tried to reflash it. No such luck, the programming failed at random locations. Thinking that maybe the EPROM was bad or just old and cranky, I bought two more 27C1024 (same model AM27C1024, just 70ns instead of 120ns), and those exhibited the same behavior, random program failing at random addresses.

Doing a bit of research, some people were saying the T48 can't provide enough programming voltage to the chip during programming, and to use a powered USB hub between the PC and the programmer. I tried that, and it changed the behavior, but the programming now fails predictably at 50%. Whenever it gets to 0x8000, it will fail. Doing a bit more research, I found that some people had luck powering the Vpp pin from an external regulated voltage source, so I setup one of those to put out programming voltage of 12.75v. Same story, the programming fails at 0x8000 predictably.

Is there something I'm missing? Is the T48 a crap programmer, or is it user error? This programmer has never given me issues before this, but I've never used it to program EPROMs this large before.

 

[GiGaBiTe]

So I found a workaround that got a usable EPROM by fiddling with the programmer settings, and doing a multi-step program.

The first step was to use the USB 3.0 powered hub.

Next, in the Xgecu programming software, I set a different manufacturer for the EPROM type, specifically a NEC uPD27C1024A @ DIP40.

On 8/16 bit interface view, I set 16 bit.

For the options, "Pin Detect" and "Skip Blank" were checked, and everything else was off.

For the others, Vpp Voltage is 13.00V, VCC Verify is 6.25V, Vdd Write is 6.25V, Pulse Delay is 100us.

The programming failed in three places, 0x8000, 0xD000 and 0xF000. When it failed, I changed the "Addr. Range" in Options to the failing address for the start range and kept going. Did this for each failure, and got a full program. Verified the EPROM with the image file and it passed 100%. Put the EPROM in my PS/1 and it came back to life.

 

[shirsch]

Glad you found a workaround! But, this story makes me even more glad that I decided to spring for a used Dataman D48Pro+. Can read/write just about anything from any era.

 

[GiGaBiTe]

I ordered a GQ 4x v4 programmer with the EPROM option board in the midst of the dumpster fire. So if I'm lucky, I won't have to do all of that nonsense anymore.