EPROM Erasure Time: a suite for characterizing EPROMs
High level requirements:
- Linux
- MacOS might work but is untested and I don't have instructions
- TL866 (minipro) family of USB programmer
- Can expose chip to UV while in the EPROM programmer
Refernece setup:
- Ubuntu 20.04
- Original TL866
- Spectronics PE-140T EPROM eraser
- 3D printed adapter
- See: https://twitter.com/johndmcmaster/status/1505276794156376066
More detailed requirements:
- You'll need to run the EPROM programmer while the UV source is on
- WARNING: UV light is harmful to eyes and skin
- Take all necessary safety precautions, especially if you modify hardware
- TL866 is compact and may fit into some erasers
- Socket extension cable might work
- EPROM isn't sensitive to optical fault injection from your UV source
- If your light source is too strong it will glitch the chip
- Shouldn't be the case for commodity EPROM erasers
Install minipro software: https://gitlab.com/DavidGriffith/minipro
For basic data collection there are currently no additional python dependencies. However if you want to graph you'll also need a bunch more stuff (WIP)
numpy
matplotlib
Use minipro -l to find your device. Example:
$ minipro -l |grep 27C256
...
MX27C256@PLCC32
27C256@DIP28
27C256@PLCC32
27C256@SOIC28
...
I'm using Intel D27C256 which doens't appear in the list, so I'll use the closest match (generic "27C256@DIP28")
Start by seeing if you can talk to minipro using check.py. For example:
$ ./check.py --verbose --device "27C256@DIP28"
If all goes ok it should report whether or not your EPROM is currently blank.
Requirements:
- Slow EPROM erase vs programming time
- Can program the EPROM w/o affecting erase time too much
Example:
$ ./collect.py --device '27C256@DIP28' --passes 12 --write-init --postfix intel_d27c256
This is intended for high intensity sources.
First zero out the EPROM:
$ ./zero.py --device '27C256@DIP28'
Now turn on your UV source and immediately run:
$ ./collect.py --device '27C256@DIP28' --postfix intel_d27c256
Either filenames or a single directory can be given. Examples:
$ ./plot.py log/2022-03-19_01_intel_d27c256
$ ./plot.py log/2022-03-19_01_intel_d27c256/*.jl log/2022-03-19_02_intel_d27c256/*.jl
History:
- v1.0.0
- It does stuff
- TBD (v1.1.0)
- "erase_time" changed from 120% to 100% completion (added "run_time")
- Added metadata options
- Write initial EPROM state