Date   : Mon, 18 Feb 2008 10:01:55 +1100
From   : msmcdoug@... (Mark McDougall)
Subject: BASIC error message - "Mis..."?

Jonathan Graham Harston wrote:

> Have you got a real, hardware BBC to compare your results with and
> see what actually should be happening?

No, I have no hardware (never even seen one in-the-flesh!). A quick check on 
eBay shows there's none for sale here in Australia BTW... not that I'm 
looking for one atm...

So far my resources have included the Advanced User Guide, the Service 
Manual (including schematics), the MOS ROM disassembly and the MESS source code.

I haven't had a chance to look further into this problem just yet - I just 
thought I'd put it out there and see if there was anything "obvious" that 
happens on a 'Mistake' that I was unaware of.. like checking some I/O 
location or something like that...

I've been able to simulate a few hundred ms of BBC operation in my 
simulation tools, but it's a tedious process. For instance, I spent several 
*hours* tracking down the keyboard problem a few days ago by single-stepping 
through MESS whilst looking at waveforms from the simulation... not much fun 
at all. The other problem with simulation is that it's difficult to provide 
any keyboard stimulus, for example.

Anyway, it's early days yet and I've by no means exhausted all avenues. I've 
still got some enhancements to do on other areas of the design before I need 
to worry about this particular problem in any case. Right now the design 
uses dual-port RAM for the video, and the CPU runs at a fixed 1MHz, so 
there's no interleaving or speed control atm. Also the video ULA only works 
in 1 bit/pixel mode, so only modes 4,6 are working. And with only 16KB RAM 
and no SAA5050, I can only run modes 4-6 anyway...

My next task is to get the video ULA working for 1,2&4 bit modes, start to 
add the palette logic for colour output, and then switch to external SRAM 
and interleave CPU & video - then I'll have 32KB RAM and can finish the 
video for modes 0-6 at least. I have a few options with the CPU speed - 
there isn't any real reason why I'd need to switch down to 1MHz in this 
case... but of course it'll then run faster than the real thing...

Meanwhile I'll ask my colleague to do the sound chip for me...

Regards,

-- 
|              Mark McDougall                | "Electrical Engineers do it
|  <http://members.iinet.net.au/~msmcdoug>   |   with less resistance!"