If you used to service the things, I have to take your word for it, but I've seen several with 2114s and I don't think I've seen any with 6550s. Oh well. [...] OK, so you are sure that it is the chip that's dead. I still think it's a fault that affects the whole chip... I'm afraid I have a minor quibble here. To write a byte to every block would be pointless as a memory diagnostic, although it would indeed find the top of RAM. However, the PET actually writes two test patterns to EVERY BYTE. PETs I've seen with dodgy RAM come up with really strange numbers of bytes free, not just multiples of 256 minus 1. Early PETS like yours (and my first one) leave 36 in each location when they've finished. Goodness knows why. Later PETS seem to be using 01010101 and 10101010 as test patterns, since they leave 170 (decimal) in each location. (As an aside: vintage PET users, on old machines at least, will have come across the corrupted links in the program listing, which instead of pointing to the next line point somewhere silly. The most common crash then is a program listing that just prints $ signs and doesn't respond to the stop key... Can't remember what keyword 170 gives you on the newer machines, though. When I was involved in developing an adventure game for the BBC micro, we typed a lot of the text on an 8032 (can't remember why). 170 ended up as the token for "HA" when we compressed it. Result - a corrupted address made the machine laugh at us. No, this was not intentional!) ********** Expansion box. The old PETs don't AFAIK have the internal expansion connector - they have an edge connector sticking out at one side. Very useful for toggling RESET with a pair of tweezers! That aside, on this edge connector are brought out most of the block select lines from the 74154 I mentioned in my previous post. In particular, brought out are lines 1 to 7, 9, A and B. It takes only three four-input AND gates to re-encode any eight of these you choose into three upper address lines for a 62256. On my PET I used 2 to 7, 9 and A, and got 32K main memory (including the 8K it already had) plus 8K of RAM in ROM space, above the screen memory. My suggestion is: disable the select to the upper 4K of memory within your PET, and encode lines 1 to 7 and 9. This will give you 32K of main memory, several spare 6550s, and 4K of RAM above the screen. Essentially, you will have a 4K PET fully expanded. Power, as always, comes from the second cassette port... Note block select 0 is NOT present on the expansion connector, so you can't readily disable all the internal RAM. Philip. ********************************************************************** This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you have received this email in error please notify the system manager. This footnote also confirms that this email message has been swept for the presence of computer viruses. Power Technology Centre, Ratcliffe-on-Soar, Nottingham, NG11 0EE, UK Tel: +44 (0)115 936 2000 http://www.powertech.co.uk **********************************************************************