Fellow Newtonians,
There is a lot of talk about "Pretec" cards being the root problem here, but
I can say for the record that my dead 32MB linear flash card was NOT Pretec!
It was in fact an AMD D-series 5v linear 32MB flash card that I had used
without problem in my Newton for two years before it died. And yes, I did
probably 20 or so hard resets during those two years without problem.
*** Now here's the story on how my AMD flash card died...
-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/
While working in Clarity, I got a lockup. This is nothing new as Clarity has
many bugs that, sadly, will never be fixed. But it suits my purposes so
well though that I've learned to live with the resets. Anyway, sliding the
power switch did nothing, so I took the stylus as I've done many times and
pressed the Reset button on back. The Newton reset but after everything
loaded the wait icon at the top of the screen stayed on forever. I was
forced to reset again via the reset button on the back of my 2100. Still
the same wait icon when all the packages loaded.
I then resorted to killing power by removing my battery pack, waiting a few
minutes, and then powering up the 2100. (Yes, I've done that many times
before even with my cards still locked in their slots.) But this time I got
a very frightful message that said:
"There is a problem with this card. (Newton cannot recognize this type of
card.)"
-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/
*** PCBman kindly offered to run full diagnostics on my dead AMD card, so I
shipped it to him and he then offered me this report:
-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/
Here are my findings:
The card is physically in good shape except for the back cover. I
noticed it had been pushed in to the extent that the shape of the
chips underneath was embossed in it. Not good. Although there is a
plastic film between the case and the electronics, there is the
potential for a short. However, I don't think that is your problem.
I was able to read and write successfully to common memory (i.e. 32MB)
but all attempts at operations on the attribute memory failed.
A brief explanation:
In addition to the 32MB of flash memory there is a small amount of
attribute memory. This is usually a small eeprom.
It holds the CIS or card information structure. It tells the host
device about the card, how big it is, speed, formatting etc.
When you insert a card in a device, the device reads the CIS and can
then make some informed decisions about what to do with it.
The problem with your card is that the CIS is corrupted and cannot be
written to. So despite the fact that the actual memory is fine, the
Newton cannot read the CIS and so won't go any further because it
doesn't know what you have just plugged in!
It's not software corruption it's hardware corruption. The eeprom is
toast.
You could probably still use it with your PC if you get the right
drivers. Something like the True Flash File System would work. The PC
ignores the CIS in this case. Or you could run it under Linux.
Bottom line, I'm sending it back to you. Sorry I couldn't fix it. If
it was just a matter of rewriting the CIS I could have done that but I
can't get the numbers in there.
If you were really adventurous, you could peel the lid off and replace
the eeprom...
-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/
*** PCBman later offered me even more details...
-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/
The attribute memory is what holds the CIS. Without that, the Newton
doesn't know what you have inserted and will not power up the card.
There are some some situations where attribute memory is not required
but the Newton is not one of them. It will not be recognized when you
insert it and so you will not be
able to use it.
A linear flash card with attribute memory looks like this:
Common memory (0-32MB)
Attribute memory (256 Bytes in your particular card)
The attribute memory occupies memory locations 0-256.
The REG pin (pin61 of the card) determines if common or attribute
memory is accessed. high for common, low for attribute.
When the card is inserted into the Newton the REG pin is pulled low an
so memory accesses are from attribute memory. Once the info has been
read and parsed, the REG pin is pulled high and memory accesses now
apply to the common memory space.
In your case, the faulty eeprom could not be programmed and held
corrupt information so the Newton could not go any further.
Just for interest, I am including some scans of the inside of your card.
The AM29F017B chips are the common memory. There are 16 x 2MB chips =
32MB
The AT43101 chips are the PCMCIA interface and contain 256 bytes of
eeprom.
-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/
*** It was after this (one year ago exactly) that Woody Smith kindly offered
to sell me another 32MB Linear card (which is what I am currently using),
and this new card is a Pretec brand:
Pretec 32MB Flash
P/N FAD032-P
S/N 098101243
So far so good, but I have been careful to eject it before pressing the
button on the back of my 2100.
I hope this information serves to educate the Newton community on 32MB
Linear card death.
Sincerely,
James Wages
====================================================================
The NewtonTalk Mailing List - http://www.newtontalk.net/
The Official Newton FAQ - http://www.splorp.com/newton/faq/
The Newton Glossary - http://www.splorp.com/newton/glossary/
WikiWikiNewt - http://tools.unna.org/wikiwikinewt/
====================================================================
Received on Tue Aug 18 20:23:05 2009
This archive was generated by hypermail 2.1.8 : Sat Aug 22 2009 - 11:30:01 EDT