Anyone have any experience with using an onboard ITX motherboard in place of a basic stamp/atom? I know the hardware is sort of overkill, but it provides a ton of room for expansion/potential
Onboard wifi, remote administration/control over your home network, unlimited program space (could just use SSC sequencer to interface to servos and/or a basic stamp/atom for sensors if needed), access to PC based voice recognition software and/or voice emulation, webcam setup… all with easy hookup. The motherboard I linked only draws 15w as well, so powering wouldnt really be too hard with a 12v battery.
Probably WEIGH a TON too! And then there’s the increased battery weight. 'Hope you’ve got the big servos!
A few of us are working to put different uP boards on our 'Bots. One of my early 'Bots (wheeled) had a bay for a full XT motherboard. That was a while ago.
Well the Pico-ITX board im looking at measures 10 cm x 7.2 cm, so I cant imagine it weighs much. It can be powered from a 12v source, so I’d imagine a NiMh solution would work. Overall I think it would only add a little over the weight of an additional 6 cell battery.
For my particular implementation, Im looking at adding it to the Johnny 5 kit, I’m pretty sure those treads could handle the extra weight.
You can even power the ITX boards off of a 6v battery. Get the picoPSU. The Nano/Pico-ITX will both require a battery that is about $200-300 and weighs like 1-3lb. If you want it to run a few hours. If you want 5-10 minutes you can get a smaller battery. Lithium based batteries are the best bet I think because of their low weight and high capacity. But they come at a cost $$.
Excuse me if I’m not understanding this correctly…
But lets say that the VIA EPIA PX10000 draws 15w (its like 13-17 from what I’ve read) and I power it with a 12v battery @ 1600mAh via the picoPSU-120 (e-itx.com/picopsu-120-mini-d … upply.html)
Wouldn’t the 12v battery @ 1600mAh = roughly 19.2w? Shouldnt that in theory provide me with about an hour worth of runtime on the pico-itx board?
If its a 12v source and 15w then the amperage is 15/12 = 1.25A total. If you have a 1600mAh battery, thats 1.6A/1.25A = 1.28 hours, or 1 hour 16.8 minutes.
Since it isn’t exactly 15w all the time, then expect more/less runtime depending on the load.
Don’t forget the power consumption of the motors and servos!
Unless you’re using a different battery for them, they will reduce your run time substantialy.
If you have linux knowledge then these babies are smaller, lighter, lower voltage, lower power (cheaper?) and give you most of the functions (tcp/ip, usb, bluetooth, mass storage, etc) that an x86 system will.
Although I don’t see them as replacing microcontrollers, rather leave the low level pwm, sensors, servos, etc on the micro and have the embedded system run network servers, mapping, vision processing and other high level software.
Both Gumstix and virtualcogs have microcontroller daughterboards available.
Parallax’s Propeller is another interesting possibility.
I actually just started using the VirtualCogs stuff at work for some of our MR research and have a bit of experience (though I still have lots to learn).
The main Virtual Cogs board has no problem running without Linux (much like a simplier microcontroller) and this is how Ive used it so far. My background was mainly in AVR and it has been fairly straight forward to transistion to something like this (even though its much more powerful). I had image capture from the camera running in an afternoon.
Im not sure if you saw, but one of their daughterboards also has a simple microcontroller than can connect to the more powerful ARM9 main SBC (with motor drives, servo outputs, etc), though I haven’t bought/tried one of these yet (on my list of things to get to eventually!)… (the BB won’t let me post a link to it)