KryoFlux - Work Started

2009-07-01

Work has started on KryoFlux (please read if you have not already done so). There is going to be quite a lot in these updates, but we want to get it all documented.

First things first though, we set up a command line environment for ARM compilation.

The main unknown at this point is the USB hanging problem with these boards. It might be better to try to find the original ARM USB sample code, assuming the USB communication is not broken in that too. Then add control commands and stream reading on top of that. Commands are also needed to return acknowledge/complete values, so the timing is reliable - there is no feedback whatsoever from the board. This will prevent the need for 250/500/1500 ms waits that were needed in Cyclone20.

There are many complaints on the official forums about the USB support on these boards when the bandwidth is maxed out, it may be something that cannot be easily fixed. This board does not have high-speed USB, it only supports full-speed, i.e. 12Mbit/s. The maximum throughput would be somewhere around or less than 8Mbit/s when communication overhead is counted in. On a normal HD disk we’d have a theoretical maximum (although unlikely) of 500K flux reversals per second. We will have to take care with encoding, as the hardware may have problems supporting fully maxed out USB bandwidth.

This is not going to be an easy job.