wireless keyboard interaction

wireless keyboard interaction

[juanslayton]

This post may be misdirected.  If so, I'd appreciate a pointer to an
appropriate forum.

My application polls 20 keyboards for an elementary school classroom,
printing the output of each keyboard to its own area of a common projected
display.  (Thanks to Linas Vepstas for earlier help in getting past the 32
event limit in the stock kernel.)  With hard-wired USB boards it works
beautifully.  However, wireless boards interact in bizarre ways. 
Sometimes a keypress on one board will inhibit reading a keypress on a
second board; sometimes the interaction will cause a burst of repeat
characters; sometimes simultaneously pressing keys on two boards will
inhibit both.  All interactions seem confined to a short interval after
the first keypress--an interval that is suspiciously close to the
autorepeat delay time.

Naturally, I am looking for the autorepeat routines.  The stumper for me
is that the problem only appears with wireless boards; as I said,
hardwired boards do not show any interaction.  I've not installed any
proprietary software with these keyboards, so it's all up to the linux
distro, but where what driver/handler/interface... would be involved?

John Slayton





-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CIDÞVDEV
_______________________________________________
Linux-hotplug-devel mailing list  http://linux-hotplug.sourceforge.net
Linux-hotplug-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel

Re: wireless keyboard interaction

[Maciej Grela]

On sobota, 6 stycznia 2007 04:06, juanslayton@dslextreme.com wrote:
> This post may be misdirected.  If so, I'd appreciate a pointer to an
> appropriate forum.
>
> My application polls 20 keyboards for an elementary school classroom,
> printing the output of each keyboard to its own area of a common
> projected display.  (Thanks to Linas Vepstas for earlier help in
> getting past the 32 event limit in the stock kernel.)  With
> hard-wired USB boards it works beautifully.  However, wireless boards
> interact in bizarre ways. Sometimes a keypress on one board will
> inhibit reading a keypress on a second board; sometimes the
> interaction will cause a burst of repeat characters; sometimes
> simultaneously pressing keys on two boards will inhibit both.  All
> interactions seem confined to a short interval after the first
> keypress--an interval that is suspiciously close to the autorepeat
> delay time.
>

It seems to me that you`re having interference between radio signals 
sent by different keyboards. Are they Bluetooth keyboards or simple 27 
MHz ? Are they all identical ? Have you tried switching channels ? When 
you put a bunch of wireless keyboards in a room working at the same 
frequency you can be sure that weirdness will occur.

-- 
Maciej Grela <jid:thermal@jabber.wroc.pl>
counter.li.org: 319794

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CIDÞVDEV
_______________________________________________
Linux-hotplug-devel mailing list  http://linux-hotplug.sourceforge.net
Linux-hotplug-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel

Re: wireless keyboard interaction

[juanslayton]

Maciej Grela wrote:
> It seems to me that you`re having interference between radio signals
> sent by different keyboards. Are they Bluetooth keyboards or simple 27
> MHz ? Are they all identical ? Have you tried switching channels ? When
> you put a bunch of wireless keyboards in a room working at the same
> frequency you can be sure that weirdness will occur.
>
Good morning and thanks for your interest.

So far as I know, all the boards I have used are 27 MHz.  That would be
Logitech, Mozart, and, I think, Interlink.  I certainly have considered
radio interference.  But with that, I would expect some scrambling of the
keycodes, and I haven't observed that; the key I press seems always to be
the key that is printed, even when I get multiple bursts.  I've tried a
number of things and I'm going from memory here, but as I recall, the
evdev value for those bursts was '2' (the normal value for autorepeat) and
I was able to suppress them by writing the application to reject an event
with value 2.  However, this still left the opposite interference, in
which keypress on one board inhibits a read on another.  It's not obvious
to me how signal interference would have either of these effects, or for
that matter how there could be signal interference at all, given that usb
devices are not peer to peer and should not be transmitting until they are
interrogated by a USB 'host' routine.  (I'm likely using the wrong term
here, but I think you will understand what I mean.)

John Slayton

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CIDÞVDEV
_______________________________________________
Linux-hotplug-devel mailing list  http://linux-hotplug.sourceforge.net
Linux-hotplug-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel

Re: wireless keyboard interaction

[Vojtech Pavlik]

On Sat, Jan 06, 2007 at 07:34:14AM -0800, juanslayton@dslextreme.com wrote:

> Good morning and thanks for your interest.
> 
> So far as I know, all the boards I have used are 27 MHz.  That would be
> Logitech, Mozart, and, I think, Interlink.  I certainly have considered
> radio interference.  But with that, I would expect some scrambling of the
> keycodes, and I haven't observed that; the key I press seems always to be
> the key that is printed, even when I get multiple bursts.  I've tried a
> number of things and I'm going from memory here, but as I recall, the
> evdev value for those bursts was '2' (the normal value for autorepeat) and
> I was able to suppress them by writing the application to reject an event
> with value 2.  However, this still left the opposite interference, in
> which keypress on one board inhibits a read on another.  It's not obvious
> to me how signal interference would have either of these effects, or for
> that matter how there could be signal interference at all, given that usb
> devices are not peer to peer and should not be transmitting until they are
> interrogated by a USB 'host' routine.  (I'm likely using the wrong term
> here, but I think you will understand what I mean.)

It's not USB that goes over the wireless link. (*) It's a custom, much
simpler protocol, where the keyboard initates the transaction on a
keypress. This is also needed because USB doesn't like transmission
errors.

This means that two keyboards will be attempting to transmit when keys
on them are depressed at the same time, resulting in interference,
failed receives and retransmits until the situation resolves. Since the
keyboards will not be implementing random backoff like ethernet does,
this can take a while.

If a key release transmit is delayed because of this, the input core
will see a long keypress, and will generate software autorepeat (value 2
events) at a default rate of 30 per second.

So, it's really most likely interference.

(*) It would not be practical to try to push a 1.5 Mbit/sec
(bidirectional!) link over a 27 MHz radio channel. WUSB (wireless USB)
uses UWB (ultrawideband) to work.

-- 
Vojtech Pavlik
Director SuSE Labs

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CIDÞVDEV
_______________________________________________
Linux-hotplug-devel mailing list  http://linux-hotplug.sourceforge.net
Linux-hotplug-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel

Re: wireless keyboard interaction

[juanslayton]

 On Sat, Jan 06, 2007 at 07:34:14AM -0800, juanslayton@dslextreme.com wrote:
>  It's not obvious  to me how signal interference would have either of
these effects, or for
>> that matter how there could be signal interference at all, given that
>> usb
>> devices are not peer to peer and should not be transmitting until they
>> are
>> interrogated by a USB 'host' routine.  (I'm likely using the wrong term
>> here, but I think you will understand what I mean.)

 Vojtech Pavlik wrote:

> It's not USB that goes over the wireless link. (*) It's a custom, much
> simpler protocol, where the keyboard initates the transaction on a
> keypress. This is also needed because USB doesn't like transmission
> errors.
>
> This means that two keyboards will be attempting to transmit when keys
> on them are depressed at the same time, resulting in interference,
> failed receives and retransmits until the situation resolves. Since the
> keyboards will not be implementing random backoff like ethernet does,
> this can take a while.
>
> If a key release transmit is delayed because of this, the input core
> will see a long keypress, and will generate software autorepeat (value 2
> events) at a default rate of 30 per second.
>
> So, it's really most likely interference.
>
> (*) It would not be practical to try to push a 1.5 Mbit/sec
> (bidirectional!) link over a 27 MHz radio channel. WUSB (wireless USB)
> uses UWB (ultrawideband) to work.

The light dawns!  I get 20 keyboards all fighting for attention. 
Obviously this is never going to work, but at least I now understand why. 
I used to say that when I retired from teaching I was going to go get
myself an education--this has certainly been educational.  But at least
the hard-wired system works.

Thanks again for your time and trouble.

John Slayton

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CIDÞVDEV
_______________________________________________
Linux-hotplug-devel mailing list  http://linux-hotplug.sourceforge.net
Linux-hotplug-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel