Eachine or JJRC H8 mini ?

dc59 wrote: Yes, you are right ! long press headless button is "RETURN HOME" function , I captured this function at new file as below,please check it out!
www.mediafire.com/download/zaq4rzet6l9zb...8mini-ReturnHome.zip

I installed test build , but it could not bind to quad!
Normal procedure:
1, power on quad, quad LED slow blink.
2, power on stock TX one beep , quad LED fast blink.
3, TX second beep, quad LED on solid.

This build, quad LED always slow blink, I tried to set different ID '12345678' / '111111' / '11' and no ID,
it's all the same.

Thanks a lot!

The 98 capture only shows the one bit, same as 99. The 06 capture has the second (suspected) "long press" bit. Anyway I added a control on channel 10 for (the suspected) RTH in the new test build. My understanding is that "RTH" on these toy quads really means "reverse" and is sort-of return-to-home if headless mode is also on. So for it to work as expected you'll probably want to have headless and RTH both on. Which is maybe why the "long press" on the stock tx turns both on.

I found one bug that may have been the problem. But I also put in code so that if you set the fixed ID to None the bind packet will be the same as sent by your stock tx. If that works, please try a couple different fixed IDs, which will change only the txid. I'm hoping the quad just takes the rf address and channels from the bind packet data, but if there's some special relationship between those values and the txid there'll be more detective work to do.

The test build is updated. Thanks for testing!

Richard96816 wrote:

Durete wrote: I vote for "BayangToys" because is the older model using this protocol

Will this protocol work with other quads?

Mine doesn't say Bayang anywhere. But they're all "H8 mini". That's the prevailing name used in the forum. Suggest that's a better name to use. Unless the Eachine H8 mini is a different bird.

Durete listed some other models it's expected to support. I went with Bayang because it's fun to say With more than one brand already the full list will have to be in the supported models spreadsheet.

hexfet wrote:

dc59 wrote: Yes, you are right ! long press headless button is "RETURN HOME" function , I captured this function at new file as below,please check it out!
www.mediafire.com/download/zaq4rzet6l9zb...8mini-ReturnHome.zip

I installed test build , but it could not bind to quad!
Normal procedure:
1, power on quad, quad LED slow blink.
2, power on stock TX one beep , quad LED fast blink.
3, TX second beep, quad LED on solid.

This build, quad LED always slow blink, I tried to set different ID '12345678' / '111111' / '11' and no ID,
it's all the same.

Thanks a lot!

The 98 capture only shows the one bit, same as 99. The 06 capture has the second (suspected) "long press" bit. Anyway I added a control on channel 10 for (the suspected) RTH in the new test build. My understanding is that "RTH" on these toy quads really means "reverse" and is sort-of return-to-home if headless mode is also on. So for it to work as expected you'll probably want to have headless and RTH both on. Which is maybe why the "long press" on the stock tx turns both on.

I found one bug that may have been the problem. But I also put in code so that if you set the fixed ID to None the bind packet will be the same as sent by your stock tx. If that works, please try a couple different fixed IDs, which will change only the txid. I'm hoping the quad just takes the rf address and channels from the bind packet data, but if there's some special relationship between those values and the txid there'll be more detective work to do.

The test build is updated. Thanks for testing!

Hi Hexfet,
Thanks for great job, I will try new build later today and report result soon!

Tryed this build against X7, X9 and H8mini. Very quick and dirty test as I am really sleeping now. Results:

1. All the three have absolutely the same results. Briefly - they bind and fly.
2. Seems to be "always flip mode". I did not find the channel to disable flipping. But after some channel changing sequences flipping may stop - but not sure.
3. Ch9 - RTH.

Did not find any other dependencies between channels and behavior. I mean channels 5-8 and 10-11 - do not know what they are doing.

vasiliy_gr wrote: Tryed this build against X7, X9 and H8mini. Very quick and dirty test as I am really sleeping now. Results:

1. All the three have absolutely the same results. Briefly - they bind and fly.
2. Seems to be "always flip mode". I did not find the channel to disable flipping. But after some channel changing sequences flipping may stop - but not sure.
3. Ch9 - RTH.

Did not find any other dependencies between channels and behavior. I mean channels 5-8 and 10-11 - do not know what they are doing.

I think channel mapping perhaps is same as V2x2 protocol,
Ch5 - Light (unused for H8 mini)
Ch6 - Flip
Ch7 - Still camera (unused for H8 mini)
Ch8 - Video camera (unused for H8 mini)
Ch9 - Headless
Ch10 - RTH
Ch11 - X-Axis calibration (unused for H8 mini)
Ch12 - Y-Axis calibration (unused for H8 mini)

Yes, you are right.
The channel order at Hexfet's code is:

Durete wrote: Yes, you are right.
The channel order at Hexfet's code is:

// For code readability
enum {
    CHANNEL1 = 0, // Aileron
    CHANNEL2,     // Elevator
    CHANNEL3,     // Throttle
    CHANNEL4,     // Rudder
    CHANNEL5,     // Leds
    CHANNEL6,     // Flip
    CHANNEL7,     // Still camera
    CHANNEL8,     // Video camera
    CHANNEL9,     // Headless
    CHANNEL10,    // Return To Home
    CHANNEL11,    // Calibrate

But I can't see where is the flip channel at code.

@dc59. Did you capture the flip button?

Durete wrote:

Durete wrote: Yes, you are right.
The channel order at Hexfet's code is:

// For code readability
enum {
    CHANNEL1 = 0, // Aileron
    CHANNEL2,     // Elevator
    CHANNEL3,     // Throttle
    CHANNEL4,     // Rudder
    CHANNEL5,     // Leds
    CHANNEL6,     // Flip
    CHANNEL7,     // Still camera
    CHANNEL8,     // Video camera
    CHANNEL9,     // Headless
    CHANNEL10,    // Return To Home
    CHANNEL11,    // Calibrate

But I can't see where is the flip channel at code.

@dc59. Did you capture the flip button?

Yes, I captured the file button!
Where can I find hexfet's code?

Here is the Bayang branch:
bitbucket.org/hexfet/deviation/branch/protocol_bayang

vasiliy_gr wrote: Tryed this build against X7, X9 and H8mini. Very quick and dirty test as I am really sleeping now. Results:

1. All the three have absolutely the same results. Briefly - they bind and fly.
2. Seems to be "always flip mode". I did not find the channel to disable flipping. But after some channel changing sequences flipping may stop - but not sure.
3. Ch9 - RTH.

Did not find any other dependencies between channels and behavior. I mean channels 5-8 and 10-11 - do not know what they are doing.

Good news! Thanks for testing! Next chance you have please try some different fixed ID values

Don't see a separate flip mode in the captures (the 05 capture shows the flag changing from normal to expert mode). They're probably enabled whenever expert mode is on, which is always enabled. Did I misunderstand the meaning of expert mode - does that button increase the max angle (max speed) of the quad? or does it just enable/disable flips?

The channel assignments in the code comments match the proposed convention for the toy quads. This makes using a common model easier. Since not all quads implement all functions some protocols will have unused channels (which seems a bit awkward).

There's also no proposed channel for "expert mode". If that switch controls both max angle and flip enable, then please try reducing the scale in the deviation mixer to see if there's a point where it no longer flips.

Hi hexfet,
Report test result, the second build works great!
I tried to set 'Fixed ID' in those different value "None" / "3" / "77" / "777" / "3333" / "123456"
all of them work fine, no problem!

CH9 Headless : sometime works as Headless , sometime is RTH, I can't figure out what difference with it at this moment,keep trying.
CH10 : Filp ............. it should be my fault , I am very sorry.

Now my H8 mini flown excellent, thank you very much.

Expert mode is separate from flip mode. Toggled by the yaw-trim button. It increases angle/speed and yaw rate (from super-fast to crazy-fast.)

Flip is enabled by pressing down on the right stick. Disabled by tilting that stick and causing a flip.

Would you please check if channel 10 is controlling the flip function (chan 10 > 0 to enable flips)? I think that's what dc59 is saying.

And then does that mean that headless mode is toggled by channel 9 being greater than zero for a short time, then back to zero? And RTH by setting channel 9 on for a longer time?

How is headless mode turned off? Same for RTH?

If expert mode only affects max angle/yaw rate I'd prefer to leave it on all the time, and the quad response adjusted by mixers.

In addition to my previous test.

Ch9 - RTH always. Not once managed to get pure headless on.

Ch10 - flips. Really working. But there is one feature in it. If you turn it it on with switch (i.e. FMOD1) - it works while switch is on plus one more time after it is off. It is not a big problem but I have not seen this variant of behavior in any other quad I have.

Fixed ID - works.

Rate mode corresponds the high mode of Bayangtoys TX (so it should be also the high mode of H8mini's TX - but I hide it from myself somewhere). Setting rudder limits to 40/75/125 on SWA makes it close to yaw rate on three bayangtoys rates.

hexfet wrote: And then does that mean that headless mode is toggled by channel 9 being greater than zero for a short time, then back to zero? And RTH by setting channel 9 on for a longer time?

No. Tried several times - only RTH or nothing.

How is headless mode turned off? Same for RTH?

On X7/X9 TX left vertical trim up - headless on/off, down - RTH.

If expert mode only affects max angle/yaw rate I'd prefer to leave it on all the time, and the quad response adjusted by mixers.

Tried this approach with rudder channel - seems to be ok.

So we've got the flip and expert mode controls sorted. Still a bit of confusion on the headless/RTH control. It's possible the H8 and Xx models work differently here. Need more captures to figure this out. One that captures both turning on and off headless mode, and another with the same for RTH.

Is there a calibration procedure that should be captured?

I've uploaded a new test build that moves the flip control to channel 6. It also goes back to using random txid, rf address, and rf channels. Please give it a try and let us know.

If that works then we can try some dynamic trims

Anyway H8's headless mode is working fine with X9's TX. As for the calibration - it is simply sticks to lower left both for headless calibration. And it is already working.

why flips are not in CH6 ?

hexfet wrote: I've uploaded a new test build that moves the flip control to channel 6. It also goes back to using random txid, rf address, and rf channels. Please give it a try and let us know.

Now flips are in place. No more visible differences.

1. Flip 'mode' is always cancelled after one flip on each of my H8minis.
2. Headless mode is toggled with the left stick. On and off. And it works quite well. Saved my quad from loss over the weekend. Always calibrate headless mode before outdoor flights now. Easy to switch it on when you lose orientation. No need for RTH. Headless mode is the solution. Not much fun for flying, but great for recovery.
3. Staying in expert mode sounds good on the surface. But the yaw rate is crazy fast. Too fast. I hope we don't lose resolution with expert-only.
4. There are two different calibrations. (This bird has everything!) Both sticks to bottom left calibrates headless mode. To bottom right is standard calibration.
5. At least one of my birds is a little touchy on binding. More reliable with the transmitter held close. Range on all birds is otherwise quite good. Out of sight long before control is lost.

Edit: Knew I forgot something ...
6. Flip is accessible in both rates, low and high (expert.)