Y Combinator is restoring one of Alan Kay's Xerox Alto machines

Josh Dersch derschjo at gmail.com
Mon Sep 5 11:30:45 CDT 2016


On 9/5/16 8:55 AM, Al Kossow wrote:

> This is why Alto restoration with one set of boards is extremely difficult.
> The machine wasn't designed to be serviced. Almost everything has to work
> before you can do anything, and there is a very high probability that in the
> process of bringing it up you will corrupt the disk pack. Hopefully the ethernet
> bridge boards will come through soon so you can try to etherboot a diagnostic
> once you stop getting parity errors. The ECC memory subsystem is fairly complicated
> so it wouldn't surprise me there is a problem with it. You might try reducing the
> number of installed memory boards.
At the LCM, I used an Apple II to test out the Alto's memory -- the Alto 
II XM uses 4116 RAM chips.  I swapped in a row at a time and wrote a 
little BASIC program to test for obvious errors.  This was 
time-consuming, but eliminated the obviously bad chips, which helped 
immensely.

> If you can get it to the point where it hits the
> disk it should sequentially read a bunch of tracks, seek to zero, then start to boot
> which changes the pitch of the 5v switching supply. You can tell a lot about what's
> going on just by the sound of the power supply. If the power supply isn't chirping
> the machine isn't doing much.
The (faint) noise patterns on the screen will also vary depending on 
what it's doing... (at least on our machines... maybe others have better 
shielding :)).

- Josh

>
> On 9/3/16 11:08 PM, curiousmarc3 at gmail.com wrote:
>> Episode 5, still does not boot, but we are starting to follow long why:
>> https://youtu.be/Wr7vDZpniNIr
>>
>



More information about the cctalk mailing list