WLToys V912/V915/? protocols

Ok, no joy with the first DFU, Testing the second one now. I'm just beginning to get what's going on (I think!), so it's going slowly.

With the 7e I don't have MIX, so I have been using SW A and B (3-position switches). I tried the channel 8 mixes above (on SWB/fMODE), and it just blinks the bind light slowly. I can bind it back no problem. Oh, yeah, motors are disconnected!

As I understand the second DFU, I need to use channel 10, so I'll assign it to SWA0/1/2 as normal, 40 and 80. Channel 6 remains unused.

I'll let you know what I find.

Loaded the second DFU and used the same file system.

Found my problem with the bind lite blinking on Channel 8; I had used 40 and 80 instead of 10 and 20. Set them correctly, and now the 10 value doesn't get a response, and SOMETIMES the 20 value makes the bind blink. Since the V912 doesn't show lights or turbo mode, I can't see anything else happening.

I also tried 40 and 80 on Channel 10, but nothing happened. The protocol seems to restrict the number of channels it shows in the Mixer menu to 8, so while I can see channel 10 in the Channel Output monitor, I can't edit it in the Mixers pages

Here is the Channel 8 and 10 sections:
[channel8]
template=complex
[mixer]
dest=Ch8
switch=SW B0
scalar=0
muxtype=multiply
curvetype=fixed
[mixer]
src=AIL
dest=Ch8
switch=SW B1
scalar=10
usetrim=0
curvetype=fixed
[mixer]
src=AIL
dest=Ch8
switch=SW B2
scalar=20
usetrim=0
muxtype=add
curvetype=fixed

[channel10]
template=complex
[mixer]
src=AIL
dest=Ch10
scalar=0
usetrim=0
curvetype=fixed
[mixer]
src=SW A1
dest=Ch10
switch=SW A1
scalar=40
usetrim=0
curvetype=fixed
[mixer]
src=AIL
dest=Ch10
switch=SW A2
scalar=80
usetrim=0
curvetype=fixed


So now that I have a better grasp on the process, I'll go back and repeat my steps from the beginning. I'll let you know if there are any changes.

The mixer page shows the output channels configured for the model, which defaults to 8. You need to go to the Model Setup page and change the number of channels from 8 to 10 to get 10 to show up on the mixer page

Model setup won't let me go past 8.

I set the number to 10 in the Ini file, but you still can't get to it in Mixers. You can see it in the Channel output display

And BTW, thanks for your efforts fixing the output display to show all the virtual channels. I'm not using it here, but it's helpful on other models.

While we're here, I looked at the outputs from the V912 (new) board connectors with a voltmeter (and the camera was hooked up). The ports respond to the V912 TX MF Up and MF Down buttons in different ways:

MF Up:
Camera: Single Photo
Sprayer: +5v (can hold - allows you to keep spraying!); this also operates the bubble maker according to the docs.
Hoist: -5v (can hold so it drops/lifts the basket to your needs)
Missile: (from document - don't have a way to test) Fire 1 missile
MF Down:
Camera: Video on/off (toggle)
Sprayer: No measurable voltage
Hoist: +5v (can hold)
Missile: (from document - don't have a way to test) Fire missile barrage

Left Shoulder button - this triggers the camera single photo like the MF Up button. It seems to let you do time lapse, taking a photo every 10 seconds or so, but that was inconsistent on my setup.

I've got a full set of the original V959 accessories, and have done some poking at them while working on controlling them from an arduino. They all work just fine on the v959 with the v9x9 extension.

FWIW, on the V959 they use 3.3 volt logic, but the 4-line devices (Camera, Missile Launcher and Hoist) all have a 5 volt line as well.

The Camera and Missile Launcher have two behaviors, and holding the button down doesn't change their behavior. The missile launcher either fires one missile or all of them.

The Hoist can run in two directions, and holding the buttons down means it runs continuously.

The bubble blower & water gun both only have one action, and holding the button down makes them run continuously (which is why it's on a switch on my Devo 10).

I don't believe any of this matters for the firmware, though when you configure a model using them, you might want to have both a switch and a toggle button for both possible actions.

I repeated my setup for the V912 using my Devo 7e, and I got the same results as I got before. With the "deviation-devo7e-v4.0.1-59ef305" version (using your modified Devo7e.dfu), with the Flysky "Bits" option, I get no camera activity when I set channel 10 to 40 or 80. I also tried the "9x9" option with the same results.

As I mentioned, the Model Setup # Channels on the TX lets me set up to 8 channels, but will not go further. In the Model10.ini I set the num_channels=10, and I can see the Channel 10 output in the channel monitor, but it can't be edited on the TX. I presume it's actually sending out the codes... but that's a guess because nothing seems to be happening.

I haven't connected the Logic Analyzer yet, so let me know if you think that's needed.

Thanks for your thoughts on the accessories... there doesn't seem to be a lot of info on them, and they seem kinda fun.

I don't believe the receiver gets channel 10, so whether it sends it or not is immaterial. So with the 9x9 option, setting channel 10 won't do anything. With the bits option, setting channel 10 to 40 or 80 should change the output of channel 6 to use values that won't show up in the channel monitor. if they don't do anything on the v912, then possibly it's the wrong bits.

The next step would be to try setting channels 5, 6, 8 9 and 10 each of 10, 20, 40 and 80 and see if those do anything. Though you can skip the two that control lights & boost mode on the v915.

I was working through setting up the channel testing when I saw some odd behavior. The V912 motor would just stop then restart immediately, or the sticks would briefly stop responding then pick back up. It took me some time to notice because I wasn't actually flying during the first few days of testing (safety and all that!), but after a while I needed my "fix". I tried to hover and she responded poorly. I had similar results with my V911, so I thought maybe it was the TX hardware, but I have no problem with Hubsan4 protocol, and the CYRF and NRF24L01 exhibit no problems.

I haven't seen any reports of problems with FlySky in the nightlies, so it's probably something I'm doing with it. Sorry if I'm rambling with these thoughts...

1. I downloaded the nightly (3/9) to verify the behavior (DFUse, Format, then download just the delivered file system), and I'm still seeing it. I tried setting up a new Model.ini from the default template, and I continued to see the stops.

2. The birds fly fine with the stock V912 TX. So I can probably rule out a problem with the PCBs.

3. It's possible my power settings are set too low... I do my hovers in the garage (about 8mx12m) so I'll tinker with those settings next. I'm using the A7105-PA-D03. Or maybe the antenna position is bad?

4. It is possible I've tried to get too fancy with the switch combinations, so I'll use as basic model.ini files as possible.

5. I'll go back to the August nightly version 92d1705 that I was able to use before all the testing.

It will probably take a couple of days to report the results. Thanks for your patience and support.

The thread on the v686 protocol may be relevant.

Will do. I should be able to get that done this afternoon.

Byte 19 seems to be the high byte with 18 the low byte of packet word 10, if I'm reading the latest Flysky.txt. So bytes 16 & 17 are packet 9. Packet 8 is where all the fun is, in bytes 14 and 15. When I looked at them in that order the pattern was consistent (at least in the couple captures I looked at).

Packet word 9 (bytes 16-17) 0x02F4 to 0x0321
Packet word 10 10 packet sequence, around 1.1 seconds
sequence begins with one of 0x0213, 0x0214, 0x0215, 0x0219 repeating about every ~1.1 seconds Each instance is followed by 9 packets starting with 0x0031, 0x0040, 0x0049, 0x0049 ... to the 9th iteration, then the 0x02.. value shows up.
Packet word 11 0x0000

I'll get going on those captures.

Yes, sorry, I've struggled to get the coding terminology... I'm about four or five years behind! But I think I almost understand you

I am thinking that the V912 TX uses just the highest two bits for the two buttons at the top and bottom of a little "multifunction" circle on the TX lower panel. The top button sends a 0x40 in that byte. The bottom button sends a 0x80. They appear to transmit as long as you keep them pressed.

There aren't any other function buttons that appear to be sent out the TX, but I'll check everything I can try.

I am curious how you are getting the nice format on these files. I have an excel template that does most of the work for me, but my PC struggles to run it and it takes a lot of time. I do a lot of VBA and VB, but I'm not a programmer, more systems-oriented. I'm curious how you do that, and it should let me have the same points of reference you use.

@goebish what did you use for decoding A7105 SPI traces?