[vcf-midatlantic] Good source for eproms?

Herb Johnson hjohnson at retrotechnology.info
Fri May 27 15:23:21 EDT 2016

I'm persisting, because this "x-rays to erase EPROMS" keeps getting 
talked about, decade by decade. But it is just - plain - impractical, 
and never WAS practical, but most certainly was not approved by the 
manufacturers in products. Please note my qualifiers.

But as generally asserted - "you could do it" or "it was done", I 
believe is improper history in particular, and therefore appropriate to 
discuss here.

Dave McGuire posted a response, but he overlooked my point. Dave's 
comment can be summarized in this excerpt:

>  We've also read the documentation you
> referenced below and know [x-ray erasure] had been done, and was originally the
> intention for the product, but the approach had problems, so they moved
> to an alternative method of erasure.

Dave, my point is to debunk the idea that X-rays were EVER intended to 
be a way to erase EPROMS, except as some early lab result or research 
idea. "Intention for the product" says to me "we are developing a 
product that will include use of X-rays to restore the product to an 
unprogrammed state", with the implication that it would be within its 
*production design* and *approved customer use* to be erased that way.

But it wasn't. So X-rays were never recommended or used *in practice*. 
Practice, Dave, not that someone didn't TRY it at some point.

The "erasable" 1702 was designed for production with a quartz window, so 
the customer or factory could erase the EPROM, by specified light, time, 
and energy. That's "production", that's "accepted practice". It's on the 
data sheets! Dave, do you see my point?

Conversely, even the references I barely read today - one a second-hand 
account from the Intel inventor! - suggests the inventor (or one of 
them) determined early on, that X-rays sufficient to erase the device 
inside then-conventional packaging (plastic or ceramic I suppose), so 
disrupted the crystalline structure, that "annealing" was needed (high 
temperatures which would melt any plastics) to restore the device to 
some determined condition-for-sale and use. Therefore, the 
then-unconventional use of a quartz window, to admit UV photons of 
sufficient frequency to "knock" electrons out of the virtual well of 
charge (no physics lecture today).


reports  a 1971 article by Frohman (Intel's designer of the EPROM) is 
quoted to describe how a "package is sealed, information can still be 
erased...by X-ray...with commercial X-ray generators".

BUT - the above Web page carries an interview - by the Web page author 
so it's second-hand - with the author in 1993. And Frohman said in 1993 
as paraphrased: (quoting the reportage) "they ended up deciding to ship 
the part with a window because erasing the EPROM with X-rays created 
surface states in the silicon that required annealing." Annealing is 
described as a process above 450 degrees Celsius.

Need I explain the impact of 450 C upon a plastic package?  Or guess the 
impact upon a sealed ceramic package? the melting temperature of solder, 
expoxies, etc? Dave, I can probably erase an EPROM with *fire*, but I 
don't expect it to work afterwards!

That's MY "read of the documentation", David.

Frohman's patent, filed in 1971:


discusses EPROM structure and programming - not erasure, so it's not 
helpful. And it could be, Dave, that "erasure" was an afterthought. Most 
"programmable memories" of the period were not "erasable".

While the patent cites many published research papers, I'm not gonna 
find those papers today, they may be paper-only and not Web accessible; 
and I have other things to do.

But my assertion is: Intel NEVER released the part to be erased by 
X-ray. I assert nobody else did, either. *That is my point.* Never 
released. For use. To be erased. By customers. With X-rays.

It's possible I'm wrong, but I'll take that bet. I think the burden is 
on you, to give a positive example. Not on me, to "prove a negative".

There's no point in saying "X-rays can in principle can erase EPROMS", 
if the result of erasure is to so damage the device that it's not 
reliable, won't meet specification of use it not *lifetime*. Even the 
patent I referenced gave a "ten year" lifetime for programmed content. 
Again - programming it once, in the field, may have been considered 
sufficient for use.


I"ve posted the same statements twice, so I'm done with the (sub)topic. 
I've stated and clarified my position. Dave McGuire is welcome to 
respond, perhaps he misunderstood the limits of my counter-claim. Or has 
positive evidence.

Those interested can do their own research. If anyone finds positive 
evidence for use of X-rays to erase EPROMS as a matter of approved 
product use of production EPROMS, please please direct me to that 
evidence. A weaker assertion, you can do it anyway and not goof up the 
EPROM - with proof of lack of damage (tough to establish) - would be of 
interest, too. Thanks for your patience, if you read this far.

Herb Johnson

Herbert R. Johnson,  New Jersey USA
http://www.retrotechnology.com OR .net
preservation of 1970's computing
email: hjohnson AAT retrotechnology DOTT com
alternate: herbjohnson ATT retrotechnology DOTT info

More information about the vcf-midatlantic mailing list