We have a need in a new product design to be able to connect a standard
USB keyboard.  That's the only USB device we will ever want to plug in
(no, really!).  I know next to nothing about USB, having just
downloaded the spec yesterday.  I gather we want a USB Host, and that
most "USB micros" cannot function as hosts.  I also gather that Hosts
are much more difficult than hubs and functions.  What I'd like is a
chip that can take keyboard USB in one end and spit out serial or
parallel characters on the other end, taking as many shortcuts from a
universal USB design as possible given the keyboard-only requirement.
What would be the easiest approach to doing this?

Oh yeah, while I'm not completely opposed to an existing stand-alone
brick, what I'm really asking about is a chip that can be designed into
a medium-volume product.  Many thanks!

On 15 Jul 2005 09:57:48 -0700, Mike Silva <s...@yahoo.com> wrote:

> We have a need in a new product design to be able to connect a standard
> USB keyboard.
   [...]
> What I'd like is a
> chip that can take keyboard USB in one end and spit out serial or
> parallel characters on the other end, [...]

Why USB and not PS/2 keyboard ?  They are already serial.

   Vadim

Mike Silva wrote:
> We have a need in a new product design to be able to connect a standard
> USB keyboard.  That's the only USB device we will ever want to plug in
> (no, really!).  I know next to nothing about USB, having just
> downloaded the spec yesterday.  I gather we want a USB Host, and that
> most "USB micros" cannot function as hosts.  I also gather that Hosts
> are much more difficult than hubs and functions.  What I'd like is a
> chip that can take keyboard USB in one end and spit out serial or
> parallel characters on the other end, taking as many shortcuts from a
> universal USB design as possible given the keyboard-only requirement.
> What would be the easiest approach to doing this?
>

The xscale ARM pxa270 (not pxa255) can act as USB client and host, but
it would not be cheap for this only.  If you also need other functions,
it might be considered.

> Oh yeah, while I'm not completely opposed to an existing stand-alone
> brick, what I'm really asking about is a chip that can be designed into
> a medium-volume product.  Many thanks!

You can always do an FPGA for it.

On Fri, 15 Jul 2005 18:09:06 +0100, Vadim Borshchev 
<vadim.borshchev@127.0.0.1> wrote:

> On 15 Jul 2005 09:57:48 -0700, Mike Silva <s...@yahoo.com> wrote:
>
>> We have a need in a new product design to be able to connect a standard
>> USB keyboard.
>    [...]
>> What I'd like is a
>> chip that can take keyboard USB in one end and spit out serial or
>> parallel characters on the other end, [...]
>
> Why USB and not PS/2 keyboard ?  They are already serial.

On the other hand USB to PS/2 adaptor might siffice , 
<http://www.google.co.uk/search?q=keyboard+%22ps%2F2+male%22+usb>;

Mike Silva <s...@yahoo.com> wrote:

> Oh yeah, while I'm not completely opposed to an existing stand-alone
> brick, what I'm really asking about is a chip that can be designed into
> a medium-volume product.  

That begs the question: how would producing such a brick in medium
volumes be conceptually different from getting yourself a medium
volume of keyboards with a proper, non-USB interface produced?

You're taking a huge detour via USB, which will cost an enormous
amount of money, development time, space and power, for *no* useful
gain other than the mass-production advantage of standard keyboards.
Odds are high you'll overcompensate that gain in costs for the
interface brick.

-- 
Hans-Bernhard Broeker (b...@physik.rwth-aachen.de)
Even if all the snow were burnt, ashes would remain.

Vadim Borshchev wrote:
> On 15 Jul 2005 09:57:48 -0700, Mike Silva <s...@yahoo.com> wrote:
>
> > We have a need in a new product design to be able to connect a standard
> > USB keyboard.
>    [...]
> > What I'd like is a
> > chip that can take keyboard USB in one end and spit out serial or
> > parallel characters on the other end, [...]
>
> Why USB and not PS/2 keyboard ?  They are already serial.

I never question customer's request, perhaps they are doing a USB to
PS/2 converter.  Anyway, high end ARM such as PXA270 and AT91RM9200 and
others can do USB host and client.

> 
>    Vadim

Cypress sell embedded host controllers.

www.cypress.com and follow the USB links.  Around $5.

Hans-Bernhard Broeker wrote:
> Mike Silva <s...@yahoo.com> wrote:
>
> > Oh yeah, while I'm not completely opposed to an existing stand-alone
> > brick, what I'm really asking about is a chip that can be designed into
> > a medium-volume product.
>
> That begs the question: how would producing such a brick in medium
> volumes be conceptually different from getting yourself a medium
> volume of keyboards with a proper, non-USB interface produced?

That's a valid question.  I am sure they will not want to deal with
custom keyboards.  My reference to a brick was to an already-available
one, not to us designing our own.  If we design our own it would be
built in to our product, not an external brick.

> You're taking a huge detour via USB, which will cost an enormous
> amount of money, development time, space and power,

If that's the case, that making a keyboard-only USB host is so
demanding (in all the categories you state), then we will of course
rethink our ideas.  But is that the consensus?  Is a simple USB host
really so difficult?  Another poster mentioned a $5 USB host device.
Is that $5 plus a years development time, or $5 plus a weeks
development time?  Of course, I'll need to research that myself, but
I'm hoping for some guidance from this group as well.

I would think that using USB keyboards in custom devices is not an
unheard-of desire.  Is bare-bones USB hosting really such a monster?

On 15 Jul 2005 10:26:00 -0700, linnix <m...@linnix.info-for.us> wrote:

> Vadim Borshchev wrote:
>>
>> Why USB and not PS/2 keyboard ?  They are already serial.
>
> I never question customer's request, perhaps they are doing a USB to
> PS/2 converter.

Nothing in OP's post stated it was customer's request.  If it was a USB to 
PS/2 converter, bying one for less than UKP5 (USD10) would be more 
sensible economically.

   Vadim

> > Why USB and not PS/2 keyboard ?  They are already serial.

That's a possibility.  What are the chances that such keyboards will be
around for the next 5 years?

I did search for PS/2 keyboards but only found USB keyboards for
PlayStation 2s.

> On the other hand USB to PS/2 adaptor might siffice ,
> <http://www.google.co.uk/search?q=keyboard+%22ps%2F2+male%22+usb>;

Now that _is_ interesting.  And only $5-$10.  That also makes me wonder
even more about how difficult it is to host USB keyboards.  Thanks for
expanding my horizons!