Re: A word about the 38.4k baud limit on QS synths.


[ Follow Ups ] [ Post Message ] [ The QuadraSynth Forum ] [ FAQ ]

Posted by John Girard on February 25, 1999 at 12:59:17:

In Reply to: A word about the 38.4k baud limit on QS synths. posted by Mike B on February 23, 1999 at 22:10:52:

I think that this may have something to do with the fact that the standard "agreed on" communications rate for MIDI is actually 31.5K (or something very close -- I can't remember the exact number). 38.4K is an improvement on that, and is only available because pc serial port baud rates typically come in multiples of 9600 (and, in fact, double from 9600 on up). So, the closest they could get to Midi was 38.4K (9600 x 4).

Perhaps bumping it all the way up to 115K raised problems with synching well with other midi devices, since there would be such a disparity between the pc-qs link and the qs-other midi link? I know this can be a problem when connecting more than a few midi devices in series, so maybe it's a problem here, too.

Just a guess . . .

-John

: Does anyone have any (other good) explanation why the Alesis QS series (the QS6.1, QS7, QS8, QSR) synths are set to communicate at a maximum baud rate of only 38.4K with a PC? I know the driver can communicate at speeds of 115.2K, which is three times faster than the current 38.4K limitation set in the QS synths. As I understand, if the QS synths could "talk" at 115.2K baud/sec, you could burn a PCMCIA flash or SRAM card in 10 minutes. This would be accomplished in 1/3rd the time that you can do it now. (How many of you working musicians using a QS synth would be interested in that capability?)

: The "weak link" seems to be the software/hardware currently used in the QS synths. I worked with engineers in R&D, and they were very conservative when it came to dealing with serial communications. My job was to install new flash in network hardware devices, and I had to establish a serial connection between a host and a network device to do this. When I would do upgrades, I would set serial communications as fast as possible, around 115.2K/sec. I did this for good reason. It would have taken forever to upgrade the network hardware if commmunications were set less than 115.2K, and testing time was always the issue. Upgrades had to be done quickly. After I upgraded the network hardware, the next day the engineers would have all the network devices set back to a much slower speed; around 96K/sec.

: I am of the opinion engineers are hesitant about establishing fast serial communications because many believe the faster speeds aren't reliable. This is mainly true when connections between the host and the client are faulty; but this is not necessarily an issue of speed. Serial communication will fail at slower speeds when there are poor connections. Why do the QS synths presently have a maximum baud rate of only 38.4K/sec? Is it because the engineers at Alesis are hesitant about fast serial communications as well? The answer would be to be more aware of the aspects that contribute to faulty serial connections in the "real world" and design cabling/hardware that addresses those problems. Are there any other ideas regarding this?




Follow Ups: