No it took that many revisions to get that chip WORKING. :) This was a time before design rule checking, everything was hand checked with pen plots and small scales/rulers, an engineer might spend 60 days checking his part of the chip hunched over a table with a cup of tea and smeared ink on his sleeve. Somewhere around rev 3 it still wouldn't power up. The A10 address line had cut across A9,8 and I think 7. CLEANLY cut right across them so the design rule checker that would have complained if they had simply gotten too close stayed quiet since it was not a violation of spacing. Since they didn't want to rev the chip just to fix that and find other issues they did a neat trick: the address line short mean that they couldn't write to registers to set NTSC mode for one thing, and therefore couldn't check NTSC timings. They fund they could flip the bit by turning on the microscope light which would flood the ell with photons and flip it. Off course they had to then turn the light off since nothing worked right being bombarded with photons either. I had been there a week or two when I witnessed them doing this. I remember asking if they had just flipped the it with the light and without looking up from the scope Bruce Ahearns replied "yeah, uh-huh". I had one of those "Oh I am definitely working at the right place" moments that stuck with me. The kind of thing they fixed going from 6 to 7 dealt with things like the white noise generator. It was a 256 byte ring counter or similar and the chip guys left the cells to power up in random states and would the shift them round and round, the problem was that due to process and geometry, they weren't quite random and so the white noise had a recirculating pattern to it, on a good day it sounded like a motorboat, a bad day was a motorboat in a hurricane. So someone wrote a (Basic)program and generated a random number string and they hardcoded those as the starting bits for the white noise bucket-brigade generator. They also addressed the "raucous squawking noise" problem made famous by Terry Ryan in his passionate memo about what he considered to be still broken, I think it was something related to switching the sound sources caused it to squawk. Bil -----Original Message----- From: firstname.lastname@example.org [mailto:email@example.com] On Behalf Of Gerrit Heitsch Sent: Thursday, October 27, 2011 1:46 PM To: firstname.lastname@example.org Subject: Re: 264/TED/Plus4 Story On 10/27/2011 07:29 PM, crock wrote: > Hi Gerrit, > > The TED on the C16 board is 7360R7 week 10, 1984, ceramic. The one > inside the +4 is the same but I can't remember the week without > popping the case again. Thank you. Another little addition to my list of MOS chips where I try to collect not only the numbers but also all known revisions. R7 suggests that it took MOS a while to get that chip stable. > The one in the 232 is also ceramic but the heatsink is epoxied to the > top of the chip and I'm not prepared to chisel it off!!! I have a TED like that too. I doubt you can remove the heatsink without damaging the chip. > All the cpu's are 7501's and plastic. The earliest one is week 51, 1983. Are they all R1 or are there other revisions? I only have a single 7501 which is an R1 from 2684. All the other CPUs are 8501R1. > Two of the PLA's were ceramic, but sadly only one of them is still > functional, from week 8, 84. The ceramic PLA I have is 1084. You do know that there is someone selling a PLA replacement that can be used for the C64 and the 264 systems (selected by jumper setting)? > Thanks for explaining the lack of reset on the TED, I'm still surprised > that the cost of laying out and producing a new CPU was worth it though.... From what I read, MOS had the 6502 as a core, so the 7501 shouldn't have been too hard to create. Especially since most of the extra logic was already done due to the I/O port in the 6510. I also assume they expected better sales than they got. :) What does surprise me is that the 8501 was still made in 1990, looks like MOS kept making spares even though the 264 line was dead. Gerrit Message was sent through the cbm-hackers mailing list Message was sent through the cbm-hackers mailing listReceived on 2011-10-27 19:00:03
Archive generated by hypermail 2.2.0.