>> I made a cable with a A-type usb connector on it to plug into J16 and
J15.  >>I put a scope on the lines on J16 and find no +5 volts, although
I find it >>on J15. 

>Hmm -there may be some number confusion here. Looking at the
>top_compnents.pdf  layout diag then the 3 USB connectors on the top
are:
>J15, J7, J16 (closest to J1 - serial/Jtag)

>This is different to the numbering on the clickable image map pic of
the
>board on balloonboard.org/hardware/205/index.html (which numbers them
>J20,J18,J19 (which are the equivalent duplicate sockets)

The USB-on-Molex sockets which are fitted on the top layer of the
Balloon board are (Left to right and down):
J16 - Host Port B 
J7 - SA Client port
J15 - Host Port A

>In fact I can't find J20,18 and 19 at all on the layout files - steve -
>where are they supposed to be - do they in fact exist?
>If someone (steve/dave?) can tell me which doc is correct, I'll try and

>make sure they all match up.

Yes they exist as plate through hole clusters on the bottom edge (Samtec
edge) of the board, I'm not sure but I assume the hole patterns match
standard sockets, I've never tried to use these and I'm not sure how
they would be affected by the fitting of the Samtec, I suspect care
would be needed with the pins that poke through the board, they may even
be mutually exclusive.


>Anyway, to get to the substantive issue:

>OK - yes, I am only seeing +5V on the two outer conectors - J15 and
J16,  >but not he middle (slave) one - J7. That appears to be correct
from the 
>circuit diagram, and makes sense - slave devices (can) take power from
the >host.

Yes Hosts have to supply power, but you wouldn't want the balloon to
power itself of the USB port would you :-) Well there's an idea...
(Steve note the :-))

>IS that the same as you are finding Glen?

>> When I plug a memory stick into the J15 (A type connector) I was
>> expecting to see data toggling on the +data and -data  lines, but
only 
>>see between 3.0 and 3.8 volts on the +data line (tried 2 different
memory
>> sticks).  Is there a software test I can run to see if those ports
are
>> actually working?

First some basic questions:
Since I'm not sure what sort of Balloon you have, Do you supply both 5v
and 3v3 to it, or are you using just 5v and relying on the internal 3v3
linear reg?

If you are supplying 5v then you should see 5v on the connectors.
Has the 1812 chip (next to the Host controller chip) got two extra
resistors strapped to it? (If not then you need the pullup resistors
fitted to NPSW2 & NPSW2).

Do you have one of the 1161 modules running when you do "lsmod"?

If you cat /proc/interrupts can you see an entry ISP116x or similar, and
if you cat a few seconds apart you should see a few 10,000 difference in
the count.

If these are OK then I'd expect some debug messages when the device is
plugged in. It's possible that these are turned off by default. Although
it's a while since I played with this I think there is a debug flag
option to the module, so rmmod and insmod with the debug flag set should
generate messages (check the source for details).

>I can't replicate this as I don't have an inline USB socket to make up
a
>suitable cable with. I'll get one.
>Who else has used the USB host functionality on balloon? 
>Dave your hyra hsa a USB webcam does it not?

Yes, but Nick has used more common peripherals e.g. keyboard, mouse,
hard disk etc.

You should be aware that the driver shipped in the standard distribution
does NOT support ISOCHRONOUS transfers. I have a version of the driver
that does but I believe it to be broken, so until I can be sure it works
and is compatible I'm not going to release it. (This is a version based
on Troy Kisky's driver). I have had some success getting web cams to
work BUT the Host controller is known to be "difficult" because of its
1ms interrupt service requirement.


>Wookey
-- 

David