[Coco] Thought

[Mark McDougall]

mmarlett at isd.net wrote:

> If your device was to interface to the floppy controller of a coco then I
>  would see no speed improvemnts over the current floppy due to the timing
>  loops that are in the drivers that are in RSDOS and NitrOS-9.

I'm not sure how the CoCo drive works - I'm guessing from your comments it 
doesn't interrupt the 6809 when data is ready?!? IIRC the TRS-80 Model 3 
interrupts the Z80 when a read sector command is ready to pump data back to 
the CPU and (I'm not 100% sure of this bit) when a new byte is ready. At 
least with that setup you could (theoretically) eliminate the seek and step 
times from the equation and pump back data as fast as the Z80 could read it in.

> Making a new interface would allow you to go as fast as the coco can
> which is faster than a floppy. Then you would have to write drivers.

True. But in general are people doing things with their CoCo that are 
I/O-bound anyway? I would've thought that they're not moving a *lot* of data 
around, but rather are being restricted by the 2MHz 6809 CPU??? In any case, 
a CoCo doing PIO isn't going to be able to pump much data to/from a disk 
running at 2MHz, certainly nowhere near 480Mpbs of USB 2.0. It'd be closer 
to 800kbps if you assume ~20 clocks/IO...

> I would love to see USB devices on the CoCo.

It ain't gonna happen, IMHO. Ever looked at the linux source for the mobo 
USB controllers? I have, and you won't be doing it with a 2MHz 6809...

Regards,

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