Dominik
Member
Hi all,
I´m wondering whether MCU+FPGA is still something that people are looking for or there´s no need for this kind of designs/boards?
I´ve heard about the Maple Oak long time ago but still haven´t seen any boards. Maybe when compared to Arduino etc. Maple abandoned this idea? Or maybe if there were some tools which would make it easier to use (as it was with Arduino) people would give it a try?
Thanks,
Dominik
I think that Oak is still on the cards, however they're concentrating on releasing the new Maple II/Maple Native which uses a more powerful microcontroller. I'll see if I can get a leaflabs staff member to come here and let us know what is happening.
There is a blog post from February that mentions they've got a fuctional prototype of an Oak but are planning on redesigning it to use the STM32F4.
http://leaflabs.com/2012/02/state-of-the-maple/
There is still a need for pairing a microcontroller and a FPGA. You could potentially create a co-processor on the FPGA if you need assistance in processing data or for it to handle IO needs while the microcontroller handles the processing. Or you could implement a DSP on the FPGA for signal processing while the microcontroller does other important tasks. These are just a couple of examples what can be done with a microcontroller and an FPGA.
mbolivar
LeafLabs
yes, we still very much would like to make oak. it's embarrassing how long we've wanted to do it, and there's still no ETA, and i can't exactly say why it hasn't come to pass.
we are indeed focusing on F4. the current development branch of libmaple (the wip-family-support branch in the leaflabs tree on GitHub: github.com/leaflabs/libmaple/tree/wip-family-support) is basically F2-ready at this point (only i2c.h, dma.h, spi.h not ported to F2 in libmaple proper, and only HardwareSPI from wirish -- though, thanks to hardware changes, we're going to launch without USB support on F4). check out the code for yourself if you're interested ;). F4 support will then hopefully be a trivial matter of adjusting the clock configuration, though there may be some problems getting an arm-none-eabi- toolchain working on all three platforms (we're pretty allergic to windows, if that hasn't been made clear by this point).
as soon as those loose ends are tied up (which will hopefully be this week), there'll be a blog post with the gory details.
as for oak, let's just say we've got some FPGA projects we're working on behind the scenes. as usual, no timeline for how that translates into something you can buy.
juzza_s
Member
Funnily enough, I've been looking at the Oak's vague details for a while wondering if the idea would be growing. :)
Personally, I'd like something like the maple, with all its gpio pins routed through the FPGA as default. Then I can program the FPGA to route the IO pins to where I want them, or to a softcore, or to an FPGA-implemented logic circuit, or to what ever is floating my boat on that particular day.
For instance, for the project I'm working on, it would be nice to have either the UART TX or a PWM to output at in 38KHz pulses (I'll be using the UART RX, identically, in both modes) coming out of the same pin, software selectable. This would require just one output pin rather than potentially 3 (PWM output, with UART TX, and the third pin attached to a switching circuit of some kind).
I'm a software engineer by trade (thus, constantly changing my mind), so the less track cutting and soldering needed, on veroboards, the better! LOL!
I've heard of one person on the freenode channel #leafblowers using a CPLD (they're similiar in a way to a FPGA but don't have as many resources available generally and don't need external components to hold the settings, although there might be FPGAs with this feature built in) to route all of the digital signals to the outputs. However, if you have analogue sensors/circuitry, then you'll need to use something else to do that, like arrays of analogue multiplexers.
You must log in to post.
