6! Who knew I’d get this far and still not have a flying quad? At least I’m getting closer and learning a lot along the way…

Satellite Receiver

I’ve updated the spektrum_serial.py code to have code that now converts the split channels into appropriate values and having watched it what it produces seems sane. Using this is a basis I changed the code in paparazzi to do the same conversions and rebuilt. All was well and so the next step was to flash the new firmware to the board – as usual.

Denied!

The flashing process is sometimes a little fussy and from time to time needs a few attempts, but after 30 or so attempts it was still not working. This is very unusual and checking the board it seemed as though the firmware I had flashed a while back has been removed – I have an empty board ready and waiting for code! The bootloader is still in place and when connected the USB device appears, so it’s just the upload that’s causing the problem. Given I’ve always uploaded using this laptop the sudden change is a little strange.
The KroozSD board uses the luftboot bootloader (or a modified version thereof) and a custom uploader which implements the DFU protocol. Running with debugging switched on didn’t show much, so I dug a little deeper. After adding some code to see the state being returned by the bootloader the problem became apparent (with my additional debug messages).

Using device : ID 0483:df11 S.Krukowski - KroozSD CRC - Bootloader Programming memory from 0x08004000... [0%=====================50%=====================100%] State = DFU Download Busy [04] State = DFU Download Idle [05] State = DFU Download Busy [04] State = DFU Download Idle [05] State = DFU Download Busy [04] State = DFU Download Idle [05] State = DFU Download Busy [04] State = DFU Download Idle [05] State = DFU Download Busy [04] State = DFU Download Idle [05] State = DFU Download Busy [04] State = DFU Download Idle [05] State = DFU Download Busy [04] State = DFU Download Idle [05] State = DFU Download Busy [04] State = DFU Download Idle [05] State = DFU Download Busy [04] State = DFU Manifest [07]

The DFU process should only enter the final “Manifestation” state when the code is complete and a zero length download is sent. That’s what the DFU specification says and what the luftboot code looks for, so quite why it’s entering it so early is unclear. Each write until then appears to be sent OK and there is no sign of any extra zero byte requests being sent.

What next?

My options appear to be switching to another computer for uploading the firmware, getting an external USB interface that doesn’t cause this problem or figuring out what’s going on and fixing it. Of those I suspect the last may be a little beyond me 🙁 I guess when I have time next week I’ll try again with a different computer.