The best CP/M machine ever? was Re: Kaypro II

I was one of the very early adopters. I got the January 1975 PE [they actually arrive in late November to the first week in Decemeber for subscribers.] and since I'd worked with the 8008 and was in process for my on 8008 I took all of maybe 4 days to decide what I'd order and wrote the order letter. I had SN00200 in my hands and operational by the third week in January. I have the copy of PE that I'd scribbled all my notes on. The initial order was CPU, 4K ram, PIO, 88acr, front panel, case and power supply. It took me one sleepless weekend to assemble Case, panel, 4k ram and test. The next weekend was the 88ACR and the PIO. Then another weekend to debug the design errors (clock and front panel circuits had oneshots!) and make it run small programs from the front panel. I would interface a SWTP CT1024 to it during Feburary and early March. The next order was for the another 4K and 4K basic. Things I'd learned by summer. The power supply was crap. With the 5 boards I had it was barely able to power them (8V line was actually 7.8V with excess ripple). the fix for that was NOT free from MITS So I attacked it and rewound the transformer and added more caps to take out the ripple plus rectifiers heavy enough to stand the load. The front panel oneshots were a PITA and would remain so even after replacing all with the then believed better TI parts. the CPU clock was replaced with the 8224 and 18mhz crystal, I was seeing stability for the first time. The backplane rang like a struck bell, it was replaced with a third party 16 slot that was far cleaner two sided construction. The 88ACR was one of the worst tape interfaces ever used. I would design and build my own saturation cassette tape system using a Redactron drive. 4K basic load time went from 2-5 tries at 10 minutes each to first time and about a minute and most of that was toggling in the loader. The saturation system ran at 15kb/s compared to 300! At 16K of ram the power supply transformer for the 16V lines was an issue. In all I was spending a lot of time correcting original design errors and hardware flubs little to no programming was done. While dealing with that I would get the first copy of MITS Programming Package II, it was terrible, most of the issues were documentation. By fall of 1975 there were the new synchronous 4K memory boards (I bought 4 "upgrades" and they worked much better). Soon after there was a new 88CPU-B that featured the 8224 and a better layout with less timing and noise issues. I bought a NS* Z80 board instead Right about then Billy Gates put out his Hobbyists are thieves letter. Despite that I bought and paid for the BUGGY MITS BASIC V3.2 and then later bug reduced versions. Then I got a copy of PT 5K basic, way better. Also the PT assembly programming system. By the summer of 1977 I was using a NS* floppy disk system and in that late fall bought the NS* horizon chassis. That by far was the greatest improvement over the ALTAIR. However all the memory bought for the ALTAIR had to be replaced as they only work with the 8080 in an ALTAIR backplane, 4 SEALS 2102 based 8k static board were the fix for that. I would add another 2 Processor tech 8K 2102 memories to make 48K ram. That would lead to CP/M 1.3 and later 1.4 and decent software tools. Spring of 1978 the Altair was retired and considered part of my first great waste of money in computing hardware. I sold some of the memory over time and recovered much of my initial investment thanks to ALTAIR mainia, I still have the CPU, ACR, BOX, PIO, and some memory. I learned a great deal about building reliable micros but not near as much as I'd have liked about programming to that point. Over the years I'd worked with other ALTAIR owners and users plus a lot of 8080 based systems including the Hazeltine H500 series terminals and Never saw a bad CPI or any intel part (other than user destroyed). But the early MITS CPU timing was so bad that the poor 8080 never had a chance. The 8800A and later 8800B systems were far better. I have a 8800BT I rescued that is a solid machine still despite being one of the first deliveries. By time MITS had matured the design, many people had soured to the product and the market had proclaimed the Z80 was the cheaper and better way to go. Allison On 08/24/2012 03:18 AM, jimpdavis wrote: