[DSP-10] additional testing of UHF3 / UHFA / XP

[Courtney Duncan]

Bob and all,

A couple of years ago when I bought a new Mac PowerBook G4, I tried  
running the DSP-10 load and operational software under a DOS Virtual  
PC.  Although other serial-based programs (some from the 1980s!) ran  
fine in this mode (using Keyspan USB to serial hardware), the DSP-10  
would neither load from VPC nor use the VPC as a terminal.

The problems I was having then sound just like what has been discussed  
here over the last week.  Sounds like it may be worth trying again  
with UHFA 3.87.  I'll make a note to do that soon as I can get around  
to it.  Any insight anyone here might have on this in advance would be  
appreciated.  In particular, I wonder if both the handshaking at load  
time and at operate time are both "solved" now.  Some of the recent  
discussion seems like it may have touched on both, but I haven't  
followed it closely enough to be sure.

Longer term, I hope to load and run the DSP-10 native from the Mac.   
This is not because there is anything wrong with UHFA as it is now or  
any of the software, it is because I got into DSP-10 in order to be  
able to write DSP software and I would prefer to do it in the Mac  
environment which also has the benefit of breaking me away from the  
legacy so that I more or less have to write it from scratch (referring  
to legacy for hints of course) which is what I want to be doing.

I won't say anything more about that until I have something to say.   
(i.e., Don't anybody get excited about a Mac interface, apart from VPC  
or Parallels just yet.  Learning Cocoa right now.)

Courtney, n5bf/6


On Dec 19, 2008, at 12:00 PM, Bob Larkin wrote:

> Hi All,
>
> As an assist to trying Rogers's fix for running the DSP-10 under  
> post-Win98 Command Prompts, I made up a version 3.87 of UHFA that  
> inserts a delay between the PC's request for a pointer from the DSP,  
> and the time it looks for it to have arrived back at the PC.  There  
> has always been a 20 millisecond delay there, but it is now  
> programmable.
>
> The DOS file to run on the PC is at http://www.proaxis.com/~boblark/uhfa387.zip 
> .  The contents are only the one UHFA.EXE and if you loose track of  
> which is which, the R3.87 on the SCRL-F3 bottom line (the one with  
> the hex dump) tells you what it is.
>
> A new CFG variable, called serial_delay with a default value of 20,  
> has been added.  The units are milliseconds.  It is line 215 of the  
> CFG file, after version 3.87 has been run once.  Alternatively, just  
> type "serial_delay 200" anywhere in the CFG file to enter this value  
> before 3.87 is run the first time.
>
> I have tested this file under DOS.  I need to re-arrange some  
> furniture to run under XP, as the XP PC is in the wrong room.  So  
> any testing under post-Win98 PC's will need to come from you guys.   
> This file does not alter  the procedure for Roger's change.
>
> An aside is that after all this settles down, I will make ROM  
> versions available for the EZ-Kits and Flash versions for the DSPx,  
> so that the startup will be like it used to be.
>
> Testing:  I have no idea what delay would be proper.  Longer delays  
> slow down the startup.  Shorter ones *may* be the source of the  
> pointer table problem.  I suggest 20 milliseconds as a minimum for  
> serial_delay, and it results in a total delay of about a second.  If  
> a value of 200 is used, this gives a bunch more time for a pre- 
> occupied PC, but adds about 8 seconds to the startup.  And so forth.
>
> I hope this is useful!
>
> 73, Bob
>
>
> _______________________________________________
> DSP-10 mailing list
> DSP-10 at mailman.qth.net
> http://mailman.qth.net/mailman/listinfo/dsp-10

"It's science. They don't know anything; they just make a lot of  
educated guesses."  -- Viannah Duncan

Courtney Duncan, n5bf/6
cbduncan at earthlink.net
144.200, 7.040.
n5bf at amsat dot org