In the question.
The question isn't where, it's when. And the answer as far as I know is "not yet"
JMO/YMMV: My guess is that with Model B still selling out, the RasPi Foundation and distributors don't see any reason to complicate their lives by managing inventory for two products. It's risky having two different builds, because if the more popular one sells out you're stuck with unsold inventory of the other one. OTOH, if someone wants to buy and resell thousands of Model As, the picture could change.
Out of curiosity, any particular reason you want a Model A?
Hello Fergus,
If you are building a power constrained robot controller why on earth would you use any kind of RPi ?
The RPi has very limited IO and is, by design, pretty much obliged to run Linux (because of the secret hardware). You could make a much more power efficient robot controller based on (for example) the STM32F4 Discovery board, (< £10 and an ARM M4 with 168Mips that belong all to you not the OS !, plenty of IO etc).
Michael Kellett
In this RasPi Forum thread, jamesh says:
Don't worry - the A's are still on the way, certainly by end of year, and hopefully much sooner than that.
Now they're saying "by Christmas": http://www.raspberrypi.org/phpBB3/viewtopic.php?f=63&t=19480
In an engineering sense, the Pi is a poor choice for a mobile robot. A moving application does not have any need for an HDMI display with high powered media abilities, which is the only area in which Pi shines. A moving application also benefits from low power consumption, whereas the Pi has high consumption compared to many other ARM boards because of its high-power media and graphic functions. There is also limited I/O on the board, whereas a mobile robot can clearly put ample I/O to very good use.
I would have chosen a different Linux board for a mobile robotic application, perhaps something like Olimex's iMX233-OLinuXino-MICRO which at 23.95 euro is certainly price-competitive with Pi. Its 64 meg of RAM is ample for a robotic application since the memory-hungry X11 is not required, and it has much more I/O capability and will run longer in a battery-powered rover.
John,
In that thread, Liz wrote:
>It requires a bit of a bump in manufacturing capacity - the same PCBs are used for both versions, so to produce any Model As we have to sacrifice some Bs at the moment, and we aren't prepared to do that while there's still a giant backlog with one distributor.
I'm trying to figure out how this squares with the notion that they have licensed the RPi
design to the distributors to manufacture and sell. Does one distributor have to wait until
the other is ready?
It seems that the lifting of the "one per customer" limit was a similar situation, where Farnell
was ready, but RS apparently got dragged along, and still hasn't delivered all its orders from
before that transition on July 16.