Single-Board Universal Module

the java-based upload tool won't work in its current form. I'll fix it one of these days.
I think the generic linux-based dfu-util will work for this though. it should, since this is just a stock dfu upload without any of the walkera specific encryption.

mikemacwillie wrote: After populating the back side of the board, I now have power to ground shorts on both. Looking at the eagle files, it looks to be in the regulator pads. The tab is connected to VSS, when it should be connected to +3V3. It's a common mistake if you're used to working with 78xx regulators, and I've made it myself more than once. I'm going to lift the tabs on the boards and see if that removes the short. I'd bet that's the problem with the board you built a little while ago.


Edit: Yep, that did it. dfu-util recognizes it now.

Well crap. I'm glad I didn't build the last board before you discovered that. I'd be pulling my hair out. I'm going to try reinstalling the stm32 on my built board, maybe I can salvage it. I'm very curious whether you have any luck powering up the transceivers. With the xtals on the bottom of the board, it may be necessary to reduce the caps for them to get stable oscillation.

I reattached the stm32 and fixed the 5v regulator. For some reason even without the boot0 jumper installed, I'm getting 3.3V on the boot0 pin, and the mcu won't exit the bootloader. If I pull the boot0 pin low, the board boots up and I see on the PPM pin:Which is the same thing SadSack was seeing. Now I'll start investigating whether the transceiver xtals are toggling or not, and we'll see where it goes from there.

I should have added probe pads for the various csn lines.

I was able to verify that the A7105 and CC2500 xtals were oscillating. I couldn't see any activity on the other 2.

I then scraped the resist off the cc2500 csn pin and attached a probe. I could see spi working as expected but no signal on MISO. I was thinking that maybe one of the other transceivers was holding the MISO line high because it wasn't staring properly. I was considering cutting the MISO lines to the other oscillators. As I was writing up the report in this post, I started to explain why the A7105 would still work. That is because it uses 3-wire SPI instead of 4-wire SPI. And then I looked at the schematic...The A7105 has MISO tied to GIO1. I did that so that so that an external board could use 4-wire SPI which is a lot easier than switching between 3 and 4-wire. But I'd forgotten about it, and had never configured the software that way. which meant the A7105 was holding MISO high and preventing proper communication.
A 2 line fix, and check this out:

So, we're in business
I have to start writing code to control the board now, but I don't have time tonight. Anyhow, This is a major milestone.

Mike, here is the new dfu file to ensure your board also works:
dl.dropboxusercontent.com/u/52854517/uni...altx_DOA_working.dfu

The changes are:
in universaltx.c, move A7105_Reset() to the top.
add this to A7105_Reset() (in a7105.c..It must be after the SetTxRxMode call):

So the next step is to measure the RF performance of the chips. I'll do that by hooking them up to an RF meter and sending a constant data stream and sweep the channel. I'll then do the same on the pre-built boards I have and compare the output. Since the meter hooks directly to the antenna port, there should be no worry about getting any interference from external 2.4GHz sources.

I haven't used the meter hooked up to a PC yet (for recording), so I have a bit of learning to do there.

I have an RFExplorer. I think I mentioned it earlier in this thread. Not exactly a serious spectrum analyzer, but I think its about as good as you can get anywhere near that price-point.

did you try adding a dc blocking cap between the mux and the PA? I haven't made that change (I'm doubtful in my ability to do so without botching it).
I'm just crossing my fingers it won't bite me.