On Apr 25, 2014, at 7:35 PM, Pete Rittwage <peter@rittwage.com> wrote: > On Fri, April 25, 2014 10:25 pm, Jim Brain wrote: >> On 4/25/2014 1:22 PM, Ingo Korb wrote: >>> "smf" <smf@null.net> writes: >>> >>>> I guess commodore dropped the ball again as the 1571 connected to a >>>> zoomfloppy can run really fast. >>> As far as I can tell from occasional postings on various forums, the >>> Zoomfloppy/1571 combination seems to suffer from reliability problems >>> and the usual advice given is to use a very short, high quality >>> cable. Therefore it appears that Commodore did not "drop the ball again" >>> but instead chose reliability over speed, which most sane companies >>> would do in this situation to reduce support costs and to avoid the >>> impression that their products are unreliable. >>> >>> -ik >>> >>> Message was sent through the cbm-hackers mailing list >> We probably (if it's not an option already) should have an option for >> slowing down the SRQ transfer bit rate to create a "known good" transfer >> speed with flags to speed it up if it works. > > Arnd Menge wrote that code, but as I recall, it has to get a byte on or > off the disk in less than 32 cpu cycles (possibly ~64 in 2MHz mode), so I > believe it was tuned how it needed to be to replace the parallel cable. I > doubt it could be slowed down… As I mentioned in my presentation, you have a max of 25 microseconds per byte. At full speed, the CIA takes 16 microseconds per byte. The next slower speed would be 32 microseconds/byte, which is too slow to keep up with a 300 RPM drive. http://www.root.org/~nate/c64/xum1541/ZoomFloppy20100918.pdf You could use a slower speed for a regular accelerator, but it would not be useful for what I was targeting, which was streaming reads of an entire track at once in order to read all copy protected bits. -Nate Message was sent through the cbm-hackers mailing listReceived on 2014-05-04 23:07:13
Archive generated by hypermail 2.2.0.