Re: Clearing EPROMs: How much time to lay in the sun?

From: Ethan Dicks (
Date: 2006-07-17 06:09:26

On 7/17/06, Rainer Buchty <> wrote:
> > - A mid-summer Northern emisphere location at Sea level, 60% humidity,
> >   sunny exposure (75% of total sunlight):  ~ 15 days
> I doubt that the EPROM is *really* erased after that period.
> Have a look at this:
> Apart from the erase patterns (see animated GIFs)...


> ...the really interesting part is that even when using an EPROM eraser until the EPROM
> is erased (3 minutes worst) data comes back when you lower the supply
> voltage from 5V to 3V.

I wonder if it would be reasonable to rig up a special socket for an
EPROM burner with a resistor on the Vcc line... you test-erase and
EPROM, check it as blank at +5V, then stick the EPROM in this socket
and re-read the EPROM at close to +3V and see if it's still blank.

I suggest this because I'm at the South Pole, have some non-blank
EPROMs (mostly recycled BIOS chips), and no proper erase light.  The
sun doesn't come up for over two months, and even then, it'll be weeks
until the sun is more than 10 degees above the horizon, and on top of
that, it's normally -60F around sunrise, meaning you wouldn't want to
leave the chips outside for risk of thermal damage (next time, I'll
bring an eraser...)

The closest I've come to a UV-A erase lamp is a 365mn (UV-C?)
inspection lamp (boxy, with a handle on the back so you can wave it
around at walls/floors to get blood and other bodily fluids to
flouresce for safety and cleanup purposes.  It has a 6"-long bulb
surrounded by a reflector, so pretty much all the light comes out the
bottom. I dropped a couple of 27C256s under it for a few hours, and at
5V, they look blank, but as the light is of the wrong wavelength, I
can't be sure they are _really_ erased.  I'd like to calibrate the
process somewhat so that chips get erased without cooking them.  50%
too long is no big deal; I'd rather not guess too conservatively and
roast them for 300% or more.  I've read that can adversely affect
access time.

       Message was sent through the cbm-hackers mailing list

Archive generated by hypermail pre-2.1.8.