pegram@uvm-gen.UUCP (pegram r) (03/22/90)
From article <1990Mar9.215613.17075@wam.umd.edu>, by dmb@wam.umd.edu (David M. Baggett): Description of results deleted to save B/W. > I can only assume that the opto-isolators won't handle 500K baud. It > doesn't seem like a software problem. > Dave Baggett > dmb@cscwam.umd.edu Conclusions? I hope that doesn't mean the end of your experiments. If it worked even at 125Khz, SGS net would be speeded up considerably, and 250Khz is faster than Localtalk (= Appletalk - runs at 230Khz). I haven't got the specs for the optoisolators, but a quick trip to Radio Schnook could give you a good idea of the speeds tolerated. Look for (rise) time or frequency related specs on the back of the bubble pack of isolators, or you could just proceed by trial and error (especially good if the isolators are not the only bottleneck). Hoping to hear more about it, Bob Pegram (Internet: pegram@griffin.uvm.edu, UUCP: go figure.)
hyc@math.lsa.umich.edu (Howard Chu) (03/23/90)
In article <1436@uvm-gen.UUCP> pegram@uvm-gen.UUCP (pegram r) writes: >Conclusions? I hope that doesn't mean the end of your experiments. If >it worked even at 125Khz, SGS net would be speeded up considerably, >and 250Khz is faster than Localtalk (= Appletalk - runs at 230Khz). I Well, any speed other than 500khz would mean installing a new crystal, or sticking in a divider. Unless you had some external control over the divider, both solutions would mean screwing up your MIDI and keyboard baud rates. But if that's not an issue for you, then you can go ahead and try it... It just isn't gonna be a software solution. -- -- Howard Chu @ University of Michigan
mark@ccvr1.ncsu.edu (Mark Boyd) (03/23/90)
Mark Boyd mark@ccvr1.ncsu.edu
mark@ccvr1.ncsu.edu (Mark Boyd) (03/24/90)
Ops - rn put my message in the Keywords line, highly truncated. I'll
try again.
The midi port and the keyboard port of the ST both use the same 500
KHz clock, but it is simple to make the UARTs clock inputs independent.
All it requires is cutting one trace on the PC board. I did this to
change the baud rate for the midi port. It is documented in deckeybd
which is archived in comp.{sources|binaries}.atari.st.
On the other hand, you cannout run a UART asynchronously with its baud
rate equal to its clock rate. That, with or without the optical
isolators, doomed Howard's ( I think it was .. ) experiment. You need
to use a 16X clock. I don't remember the max clock input on the ST's
UART, but I don't think it is nearly fast enough for AppleTalk speed.
Mark
mark@ccvr1.ncsu.edu dmb@wam.umd.edu (David M. Baggett) (03/26/90)
Bob Pegram, pegram@griffin.uvm.edu, writes: >Conclusions? I hope that doesn't mean the end of your experiments. If >it worked even at 125Khz, SGS net would be speeded up considerably, >and 250Khz is faster than Localtalk (= Appletalk - runs at 230Khz). I >haven't got the specs for the optoisolators, but a quick trip to Radio Schnook >could give you a good idea of the speeds tolerated. Look for (rise) >time or frequency related specs on the back of the bubble pack of >isolators, or you could just proceed by trial and error (especially >good if the isolators are not the only bottleneck). >Hoping to hear more about it, > Well, it does mean the end of my experiments. Unfortunately, you can't set the ACIA to run at anything between 32.1K baud and 500K baud. (You can either set the divide value to 64 (~7K baud), 16 (31.2K baud), or 1 (500K baud)). There's simply no way to specify any other divide values. Oh well... Dave Baggett dmb@cscwam.umd.edu