Hey thoughtfix! Need your reply --

http://www.internettablettalk.com/fo...&postcount=208

Except I don't actually know if mine works yet! It might be a nice looking paperweight :) My little USB keyboard is in the country now according to the tracking number. Hopefully have it tomorrow sometime.

So... has anyone tried just following the USB 2.0 / USB On-The-Go 1.0 specs and connecting a self-powered USB 2.0 hub (f'rex CyberPower Systems CP-H420MP) to the Nokia 770 with an OTG host-mode cable (f'rex SerialIO.com C-OTG-6I)?

The cable supplied with the Nokia 770 is an OTG peripheral-mode cable, from what I can tell.


Per USB OTG 1.0 spec, the Nokia 770 should happily supply up to 8mA @ 5VDC nominal on VBus+ as long as the appropriate ID pin is shunted in the USB OTG 1.0 Mini-AB connector. The messages logged in `dmesg` support this hypothesis.


Although I've used the "forced" method described earlier in this thread -- plumbing the downstream VBus+ and GND lines on a self-powered USB hub (IOGear GUH-274 and some external wiring) to the upstream side by hand -- and it works (FWIW, the -data- lines should not be plumbed from downstream-to-upstream), doing it via USB OTG 1.0 just strikes me as the more elegant solution, ne?

Simon: You can always build a second one and ship it to me and I'll test it with my multimeter, keyboard, and iPod nano. ;)

JMills I must admit I had no idea what you were talking about but I found this link:
http://pd.pennnet.com/Articles/Artic...ICLE_ID=218924
that explains it I think.

Is there a way of telling which mode you're actually in via XTerminal? I vaguely understand the echo host, echo peripherial commands you use to set the mode. Can I query what it is though?

If I understand what you're saying correctly with the right USB powered hub and the right cable the 770 should automatically detect and switch itself into the right mode if it supports USB OTG?

Daniel, I'll try testing it out tonight when I am home with a USB pen drive I have. There is that fiddling I have to figure out about mounting the drive and so on. The keyboard should be a lot easier to test when it arrives.

The article meanders a bit, but yeah, that's what I was referring to. Canonical URL for On-the-Go 1.0 is:

http://www.usb.org/developers/onthego


The state data is reported in the kernel log ring-buffer (what `dmesg` sources from) so you could theoretically parse the output from `dmesg`.

I don't have the right connector on-hand yet to determine whether the SysFS node used to manually force OTG host-mode will dynamically flip from 'peripheral' to 'host' if using cable-sense.


Precisely.

How much testing would something like that entail? From what I've seen of hub packaging, there doesn't seem to be any way to determine which would be a "magic" hub. At least one person here has a hub that is supplying power to his 770 w/o any magic cable! Aside from all of us buying the same hub -- which might not even be possible -- what's the way to tell which works?

Yep, and they're doing it by shunting power with a 'Y' cable from the downstream side (USB type A sockets) to the upstream side (USB type A plug or cable-tail).

That depends on how you intend to use it. :-)

Any self-powered (i.e., internal batteries or an external power-supply) USB 2.0 hub should work fine using the shunt method, since you're going to manually switch the USB port on the 770 side to host-mode.

I'm mostly wary of this existing method from an electrical perspective... I'd rather not feed power -into- the 770 if I don't have to.

Yes, of course you're right about that. But there was someone who surprised thoughtfix without apparently needing such a cable -- that's what I meant. Post is in his Comments here:

http://www.blogger.com/comment.g?blo...67386104837654

Someone wrote...

"So it obviously depends on the hub. Mine seems to feed current. So there is no poblem. It's a cheap 4-port mini hub:"

The URL they posted requires the user's cookie to work. Iasked for the store part # instead.

*****

I'm a bit surprised that either of these (USB Type-A socket to Type-A socket "gender changer" and the hub) work... it suggests that certain USB _peripherals_ (hubs, etc...) are providing power on the upstream side, which IIRC is a violation of the USB spec.

JMills