I ran some tests to see how the speed compared to all of this theory.
I created a basic program with 10 bload test,A$4000,L$4000, so 10 x
16K of data. This took 10.6 - 10.9 seconds. so my calculations make
this about 15KBytes/s. Not that fast, but its a start. This time was
from hitting the enter on the run, until the prompt came back. It
took 28.3s for writing, so this was considerably slower. I'll have to
have a look why this is so much slower. The internal floppy took 29.5
for reading, and 148 for writing.
Its interesting that smartport devices where only ever connected to
cards with the IWM on them. Looking at the original disk ii interface,
I don't see why this would not have supported smartport devices, as
they only use the normal timing as per a disk II drive. I suppose they
just used the iwm with the liron, as it needed to be a new card
because of the firmware rom to run the smartport,
Also, i wonder why they have not used the 2x mode available on the
iwm? Although, looking at the code in the iic, it does a good job of
decoding the packet as it comes in. If it where to run the interface
at double speed, then it would probably have to concentrate on reading
the data, and then decode it later. ( if it can read it fast enough)