WLToys V912/V915/? protocols

More
22 Mar 2015 14:20 #30046 by goebish
Replied by goebish on topic WLToys V912/V915/? protocols
This is starting to get mind blowing, can you try this firmware, the only difference with the last one is that it has no hardcoded id.
www.dropbox.com/s/m1u5xtzvqm4n8lt/deviat...0.1-d5e268a.zip?dl=0
( gist.github.com/goebish/fd6212b6e99360c9bc63 )

Please Log in or Create an account to join the conversation.

More
22 Mar 2015 15:58 #30047 by Deal57
Replied by Deal57 on topic WLToys V912/V915/? protocols
I hear ya. I really appreciate your patience and persistence with this!

The firmware from post #30013 worked fine with no FixedID. I walked around the house and didn't see any dropouts, all tested at 30mw setting.

This latest firmware from post #30046 works great without a fixed ID. All collective movements are perfect and function buttons work exactly as expected. I took a basic walkaround with no dropouts noticed.

I think I have a lead on what is happening with the fixed ID. I have been using Model 10, so I used "1010" as the fixed ID. If I use that fixed ID with this version, the V912 has the stutters and hesitations. If I use other fixed IDs, like "10101", "1111", any of the 256-271, and the blank fixed ID, there are NO stutters. I captured both streams from the RX. In "V912 RX 1010 fast elevator" I have stutters, but in "V912 RX 10101 fast elevator" the collective is smooth, no stutters. I scaled the capture so it shows about 2 seconds on my screen, and on the "1010" there seems to be breaks in the data (beginning at about the 6 second mark), but in "10101" there aren't any visible at that scale.

dl.dropboxusercontent.com/u/31578062/V91...20elevator.logicdata
dl.dropboxusercontent.com/u/31578062/V91...20elevator.logicdata

I guess "1010" isn't good to use as a fixed ID. :dry:

Deviation Devo7e 3way switch mod, A7105, NRF24L01
Devo6s 2x2 switch mod, trim mod, haptic, multimodule, A7105, NRF24L01, CC2500
Devo12e 4-in-1 with voice mod -- it speaks!!

Please Log in or Create an account to join the conversation.

More
22 Mar 2015 16:16 - 22 Mar 2015 16:18 #30049 by goebish
Replied by goebish on topic WLToys V912/V915/? protocols
It's clearly visible that most packets are missed in capture #1, that explains the stuttering and lose of bind, and shows the RX listens on channels # that we're not expecting for this TX id.

So, I was right, it works with some IDs, but not other... will have to stare at the tables again :p

We're making some progress ...
Last edit: 22 Mar 2015 16:18 by goebish.

Please Log in or Create an account to join the conversation.

More
22 Mar 2015 16:36 #30050 by PhracturedBlue
Replied by PhracturedBlue on topic WLToys V912/V915/? protocols
maybe something funny with the upper txid bits? we tried the lowest 4 bits and it seem to bind reliably for all those. we tried several of the upper 4 bits and those looked right too. So maybe it is doing something with the upper 8 bits (which normally are not used for channel selection)

Please Log in or Create an account to join the conversation.

More
22 Mar 2015 16:38 - 22 Mar 2015 16:45 #30051 by goebish
Replied by goebish on topic WLToys V912/V915/? protocols
There's still something weird in the "working" capture (id 10101 decimal, 0x2775), the Rx listens on:

8E, 3E, 7A, 2A, 84, 34, 66, 16, 5C, 0C, 98, 48, 52, 02, 70, 20 (row5, offset -8 )
while with this ID we transmit on:
8F, 3F, 7B, 2B, 85, 35, 67, 17, 5D, 0D, 99, 49, 53, 03, 71, 21 (row5, offset -7)
(eg good row, but bad chanoffset by 1)

So I think the RX receives the data because the channel is close enough, but we don't emit on the right one.
Last edit: 22 Mar 2015 16:45 by goebish.

Please Log in or Create an account to join the conversation.

More
22 Mar 2015 16:59 #30052 by PhracturedBlue
Replied by PhracturedBlue on topic WLToys V912/V915/? protocols
goebish. This is normal for the A7105. It is typical to listen one channel down from the transmit channel. the off by one is not part of the issue.

the 1010 data is using row 2 with an offset of 9, but the computed offset should be 15. this is why it isn't working.

I vaguely remember some discussion over at rcgroups about my algorithm being wrong for some txids, but I don't think nayone ever told me what was wrong. I'll go search and see if I can find the discussion.

Please Log in or Create an account to join the conversation.

More
22 Mar 2015 17:02 #30053 by goebish
Replied by goebish on topic WLToys V912/V915/? protocols
Ok thanks, I didn't know it is typical to listen on rfchan-1.

That's why I love this community, I learn tons of things each time I come here :)

Please Log in or Create an account to join the conversation.

More
22 Mar 2015 17:05 #30054 by PhracturedBlue
Replied by PhracturedBlue on topic WLToys V912/V915/? protocols
www.rcgroups.com/forums/showthread.php?t=1626362
some users are seeing a maximum offset of 9

Deal57,
can you capture with the following txids?
272
288
304
320
336
352
368
384
400
416
432
448
464
480
496

Please Log in or Create an account to join the conversation.

More
22 Mar 2015 17:25 #30057 by Deal57
Replied by Deal57 on topic WLToys V912/V915/? protocols
Its gonna take me about an hour to get these done... my batteries are all charging. I'll post as soon as I get a few done.

Deviation Devo7e 3way switch mod, A7105, NRF24L01
Devo6s 2x2 switch mod, trim mod, haptic, multimodule, A7105, NRF24L01, CC2500
Devo12e 4-in-1 with voice mod -- it speaks!!

Please Log in or Create an account to join the conversation.

More
22 Mar 2015 18:13 #30060 by Deal57

Please Log in or Create an account to join the conversation.

More
22 Mar 2015 19:53 - 22 Mar 2015 19:55 #30068 by PhracturedBlue
Replied by PhracturedBlue on topic WLToys V912/V915/? protocols
Indeed it looks like the maximum offset is 9 for your rx.
I think the best thing to do is to truncate the offset at 10 for all models. I don't know if any support values > 10, but all will work between 0 and 9.
There are 2 options:
1) just use: 'if offset > 9 {offset = 9}
2) use something like id = (Model.txid & 0xff00) | (Model.txid % 160) which would never get an offset > 9. I wouldn't actually use modulo, since it requires needless division

Either way it is backwards incompatible, but I am ok with that. There is no reason to have an option to enable this configuration. I'll let goebish decide what he wants to do and have you test it.
Last edit: 22 Mar 2015 19:55 by PhracturedBlue.

Please Log in or Create an account to join the conversation.

More
22 Mar 2015 20:32 #30070 by goebish
Replied by goebish on topic WLToys V912/V915/? protocols
Here a firmware with option 2)

If we use 1), there's a risk a Rx can allow offset >9 and doesn't work ?

www.dropbox.com/s/3rph570j3kou8xn/deviat...8a_txidhack.zip?dl=0
( gist.github.com/goebish/f1879228b8cf93333009 )

Please Log in or Create an account to join the conversation.

More
22 Mar 2015 20:44 - 22 Mar 2015 20:53 #30071 by goebish
Replied by goebish on topic WLToys V912/V915/? protocols

PhracturedBlue wrote: 1) just use: 'if offset > 9 {offset = 9}


or maybe you mean only modify high nibble from id low byte if it's >9 ... I didn't understand at first, I thought you wanted me to change chanoffset without modifying txid.
Last edit: 22 Mar 2015 20:53 by goebish.

Please Log in or Create an account to join the conversation.

More
22 Mar 2015 21:10 #30072 by Deal57
Replied by Deal57 on topic WLToys V912/V915/? protocols
Ok, the firmware from #30070 works with no problems noted. I tried the Fixed ID of 416 and 1010 as well as no fixed ID, and everything worked exactly as expected.

Nice work! What next?

Deviation Devo7e 3way switch mod, A7105, NRF24L01
Devo6s 2x2 switch mod, trim mod, haptic, multimodule, A7105, NRF24L01, CC2500
Devo12e 4-in-1 with voice mod -- it speaks!!

Please Log in or Create an account to join the conversation.

More
22 Mar 2015 21:18 - 22 Mar 2015 21:18 #30074 by goebish
Replied by goebish on topic WLToys V912/V915/? protocols
Tidy up the code for V6x6 and V91X (not sure this is a good name) then create a pull request so it's reviewed and included in the nightly builds :)
Last edit: 22 Mar 2015 21:18 by goebish.

Please Log in or Create an account to join the conversation.

More
22 Mar 2015 21:23 #30078 by PhracturedBlue
Replied by PhracturedBlue on topic WLToys V912/V915/? protocols
Thanks both of you for pushing through with this. I know testing can be tedious, and here was a lot of back-and-forth before the problem was identified.

Please Log in or Create an account to join the conversation.

More
22 Mar 2015 21:28 #30079 by goebish
Replied by goebish on topic WLToys V912/V915/? protocols
My pleasure :)

The only problem I see with the latest patch is as you said it is backwards incompatible, so certain FlySky Rx may need rebind.

Please Log in or Create an account to join the conversation.

More
22 Mar 2015 21:56 - 22 Mar 2015 21:57 #30083 by Deal57
Replied by Deal57 on topic WLToys V912/V915/? protocols
We definitely need a new name for this option. The new option doesn't appear to be compatible with the V911.

I tried this new firmware with my V911 and it flies great as long as WLToys ext. is Off. I DID have to rebind the first time I used it. When I tried to use the (current) V91x option, the heli spins. I set channels to 4 and 8 and it just spins. So we wouldn't want someone selecting this for the V911. We also need to test it on a V913 and V915 but I don't expect a problem there.

On the other hand, the V912 flies nicely. And the camera functions worked! :woohoo:

Deviation Devo7e 3way switch mod, A7105, NRF24L01
Devo6s 2x2 switch mod, trim mod, haptic, multimodule, A7105, NRF24L01, CC2500
Devo12e 4-in-1 with voice mod -- it speaks!!
Last edit: 22 Mar 2015 21:57 by Deal57. Reason: Clarity

Please Log in or Create an account to join the conversation.

More
22 Mar 2015 22:02 - 22 Mar 2015 22:03 #30084 by goebish
Replied by goebish on topic WLToys V912/V915/? protocols
That's what I thought, glad that you had a v911 to test...

Other than the code, it would be cool if someone with a better English than me updates the manual regarding new WLToys options & channels features for v6x6 and now v912.
(I can update the French manual if it exists)

For the 912, I'll use channels 5 & 6 to simulate top and bottom buttons, 6 & 7 were a mistake.
Last edit: 22 Mar 2015 22:03 by goebish.

Please Log in or Create an account to join the conversation.

More
22 Mar 2015 22:50 #30085 by goebish
Replied by goebish on topic WLToys V912/V915/? protocols

Deal57 wrote: On the other hand, the V912 flies nicely. And the camera functions worked! :woohoo:


How do accessories work ? Are they controlled only with top and bottom buttons ? What if you've more than 1 accessory on the copter ?

Please Log in or Create an account to join the conversation.

Time to create page: 0.081 seconds
Powered by Kunena Forum