Woz disc II low cost approach, why did most other hone micros just shugart drive and std controller ics.

on-topic acorn-related discussions not covered by the other forums
Post Reply
B3_B3_B3
Posts: 115
Joined: Sat Apr 08, 2017 10:42 pm
Contact:

Woz disc II low cost approach, why did most other hone micros just shugart drive and std controller ics.

Post by B3_B3_B3 »

Given the expense of the standard off the shelf disc controller such as 177x and std shugart interface drive pairing in the early 80s approx, and the slow maximum speed of cassette, I am puzzled why no other manufacturers mirrored the Woz disc II low cost approach of a dumb drive and 6502 direct control, even with 2mhz equivalent or faster cpus?

Commodore did stick a 6502 in their c64 drive then use a serial link which famously ran slower than it should have (but could be speeded with special loaders)
User avatar
scruss
Posts: 390
Joined: Sun Jul 01, 2018 4:12 pm
Location: Toronto
Contact:

Re: Woz disc II low cost approach, why did most other hone micros just shugart drive and std controller ics.

Post by scruss »

Because doing it right is hard, and most people don't want their expensive CPU locked up doing disk things. Also, Apple didn't pass their savings on to the buyers. Finally, Shugart got wise to this, and buying bare drives was something you could no longer do unless buying in colossal quantities.

Commodore's parallel drives for the PET were fairly quick. Since CBM owned MOS, it was cheap for them to put a 6502 (or compatible) in everything. And CBM was all about the cheap.
B3_B3_B3
Posts: 115
Joined: Sat Apr 08, 2017 10:42 pm
Contact:

Re: Woz disc II low cost approach, why did most other hone micros just shugart drive and std controller ics.

Post by B3_B3_B3 »

scruss wrote:
Sun Sep 26, 2021 7:39 pm
Because doing it right is hard, and most people don't want their expensive CPU locked up doing disk things. Also, Apple didn't pass their savings on to the buyers. Finally, Shugart got wise to this, and buying bare drives was something you could no longer do unless buying in colossal quantities.

Commodore's parallel drives for the PET were fairly quick. Since CBM owned MOS, it was cheap for them to put a 6502 (or compatible) in everything. And CBM was all about the cheap.
Thanks, though home users using cassettes might have thought computer busy(locked up) acceptable given thats what happened with much slower cassettes.

Also, would other drive manufacturers have offered bare drives.

Perhaps if commodore hadnt messed up their drives inteface other home computers would have supported it.
User avatar
scruss
Posts: 390
Joined: Sun Jul 01, 2018 4:12 pm
Location: Toronto
Contact:

Re: Woz disc II low cost approach, why did most other hone micros just shugart drive and std controller ics.

Post by scruss »

B3_B3_B3 wrote:
Sun Sep 26, 2021 8:54 pm
Thanks, though home users using cassettes might have thought computer busy(locked up) acceptable given that's what happened with much slower cassettes.
Cassettes were always the cheapest option. You put up with them because you had to. 8-bit computers weren't very good at DMA
Also, would other drive manufacturers have offered bare drives.
There weren't many other drive manufacturers back in 1977.
Perhaps if commodore hadn't messed up their drives interface other home computers would have supported it.
Drives for the Commodore PET used the IEEE-488 / GPIB parallel standard, which in theory meant they could be used with other computers. You could certainly use other GPIB devices with the PET, if there were drivers for it. HP and Tektronix were very keen on GPIB.

The slow 1541 was a reduced price drive for a budget computer. Commodore weren't interested in interoperability. They own the fab that made all of their chips, so if they could cut costs and use a new simpler design variant, they would.

If you look at what even a simple drive controller does, it takes away a huge amount of work that is very easy to get wrong. The rationale for the Apple's weird disk format - more capacity on expensive floppies - went away as the disks came down in price. The later Apple IIgs CLV floppies were fairly slow, as the drive had to change speed and stabilize before it could read or write, and all that got you was roughly 10% increase in capacity
User avatar
sweh
Posts: 2567
Joined: Sat Mar 10, 2012 12:05 pm
Location: New York, New York
Contact:

Re: Woz disc II low cost approach, why did most other hone micros just shugart drive and std controller ics.

Post by sweh »

scruss wrote:
Tue Sep 28, 2021 2:09 am
Drives for the Commodore PET used the IEEE-488 / GPIB parallel standard... HP and Tektronix were very keen on GPIB.
At my 2nd job we had HP9000/k400 series and the printers were connected via GPIB.

I now have the absurd notion that a CBM PET floppy could be connected to the k400 as a boot device...

Hell, it wouldn't have slowed down the boot sequence at all! Those machines checked every byte of memory and would disable failing memory chips so it could boot in a degraded mode, which meant a clean boot was at least 30 minutes in duration!
Rgds
Stephen
Coeus
Posts: 2360
Joined: Mon Jul 25, 2016 12:05 pm
Contact:

Re: Woz disc II low cost approach, why did most other hone micros just shugart drive and std controller ics.

Post by Coeus »

B3_B3_B3 wrote:
Sun Sep 26, 2021 8:54 pm
Thanks, though home users using cassettes might have thought computer busy(locked up) acceptable given that's what happened with much slower cassettes.
Indeed with a single-tasking OS and reasonably fast interrupt response there is little to be gained from DMA. The reverse engineering of the 8271 disc controller revealed two embedded, specialist microprocessors in the same package, thus exceeding the transistor count of the 6502. That chip was late and expensive and possibly suffered from trying to be two general - the Intel designers were intending to make other chips in a series from the same combination of bit and byte processors but in the event there was only one other that seems to have made it to market and Intel's next FDC was licensed from someone else!

But part of the advertising was that the chip was DMA capable and could offload floppy processing from the main CPU. That is useful if you have something useful for the main CPU to do while the floppy disc transfer is taking place and would be more useful with the bigger and more complex 32 bit CPUs which had long running instructions and poor interrupt latency. This was mentioned in Why did Acorn ignore Motorola?.

But yes, for an 8bit machine with a single tasking OS, it would seem perfectly reasonable to do the operations done by the byte processor part of the 8271 on the main CPU and do the functions done by the bit processor in some dedicated logic as the main CPU is unlikely to be fast enough. Does that sound similar to what Woz actually did?
Post Reply

Return to “general”