Mjx Bugs 3

looks good when i test it.
But ultimately, i wonder how we can move forward so davdrone1 and others to be able to bind and arm?
i guess we going to need a different capture of the (different) bugs 3 arming sequence? since the binding works so far.

No luck using your model.ini on the previous build, Blade81. Interesting and useful Icon you load with it.
I'll download and flash the latest dfu now, and try again.

It seems to bind using deviation-t8sg-v5.0.0-648a226.dfu, but no reaction to any of the switches,
or pulling both sticks down and right.
I've tested with both the Blade81 model (with AIL, ELE and RUD changed from Reverse back to Normal)
and my own model for the Bugs 3.
I think the next step to encourage other Bugs 3 owners to download and test builds. I could help with
this by putting up another video about how to do this, and progress to date.

The fact that I can use any of my MJX controllers to bind and arm either of my two Bugs 3s, or my Bugs 6
makes me wonder if the test version developed only works with the Tx and RX that the captures
were done on.

I think there is a algorithm that controls which packet to send for arming based on the quads receiver.
Right now, the logic is all done and ready, just need to what is the difference between the different quads RX arm key and implement it, I think.

I remember hexfet mention he ordered a separate receiver, and hopefully once he have those he can capture those and hopefully come out with a solution.

If anyone else has a spi logic analyser and a bugs 3, please help to capture the arming and disarming sequence which would be useful for decoding

I will look around for an a1 receiver online,I have an extra controller.

Looks good. I wonder if this would be compatible with the bugs 3 mini? I ordered one and should be here soon.

Only the A2 version of the original bugs 3 has been deviated some what,there is still work to do though on the arming process because it isn't consistent or understood thoroughly.

Test build is updated (af8868a). Channel 7 should control the automatic flips. Not sure if leaving it above 0 will allow multiple flips, or it may need to be set below 0 between each flip (best done with a momentary mix on trim switch). Picture support code is removed since don't have any captures for it.

Does the stock tx always start in low-speed mode, or in the mode it was in when powered off? Need to be sure its always high-speed in the protocol code. Please make a capture of turning the LEDs on and off with a long press on the mode button. Also need to get captures for the trim switches.

The model file you posted has pitch, roll, and yaw all reversed. is this the model file you flew with? Please confirm which channels need to be reversed in the code.


The difficulty of supporting other receivers (besides Blade81's) boils down to how the radio id and/or hopping frequencies are chosen. Since davdrone1's quad responds to the bind but is not controllable it's possible the aircraft chooses the hopping frequencies. The packet sent from quad to tx at bind time would tell the tx which channels to use. The attached t8sg dfu sends a different txid to the receiver than the one from the captures. Please make a capture trying to bind to this build. The radio id and hopping frequencies chosen by the quad will tell us if it always uses the same ones, or they change based on information from the tx.

We know that Blade81's quad uses the same id and frequencies every time it binds. If they're still the same with a different txid then it's likely each aircraft is hard-coded with a different value and we'd need captures from a large number of receivers to have a chance to figure out the algorithm that chooses the hopping frequencies. If the id and/or frequencies change then it's likely the quad chooses them based on a combination of the txid and some number in the quad that's in the bind packet from the rx. In this case we can use a single receiver to make a large number of captures to try to figure out the algorithm.

My Bugs 3 seems to bind using deviation-t8sg-v5.0.0-af8868a.dfu, as the front LEDs go to steady on,
and the "Bind B3.. Press ENT to stop." is no longer displayed.
No reaction to re-setting any of the switches from -100 to +100, or to pulling both sticks down and right.
I've tested with both the Blade81 model (with AIL, ELE and RUD changed from Reverse back to Normal)
and my own model for the Bugs 3.

I dont think the stock tx can toggle lights on/off

According to the manual I found online a long press on the speed mode button will toggle the lights on / off. But it would not surprise me at all if the manual is incorrect.

The long press turns the lights on/off on my Bugs 3, as per the manual.
I'd never noticed that option before - night FPV stealth flights with a low light camera are now an option!

okay cool, i never knew this before and i just tried it and the stock tx lights on/off works with long press of the speed mode button.

I upgrade my jumper to the latest test build. Bound and armed successful. Capture in the zip file.

I use spi logic analyser on stock tx, power on with quad . Toggling speed mode and light switch in capture too.

I also did a capture of the flip button depress and stick to right (about 5 times) than to left (5 times)


For Jumper, my AIL and RUD is always reversed, it is the default template i set for all my quads. I recall deviation by default has these two channel reversed

Thanks. Makes sense for the reversed channels. I did change reverse the yaw channel in the code so you'll need to change the model before flying next.

Looks like the flip control on channel 7 and lights on channel 6 should be working. Looks like the flip channel must be reset back to less than 0 between each flip.

Please make a capture of binding to the t8sg.dfu file attached to my previous post. This is the test version that will give us information about how hopping frequencies are chosen.

wont be able to test out flip in the house and i wont get to bring the quad out till weekends.

As for the test version binding sequence, i had did a capture previously with af8868a build in my previous post, 2nd zip file?
do u need additional capture for that?

I only attached the t8sg.dfu file here in the message I posted yesterday. It's not in the test builds because I didn't want someone to download the test build that's supposed to be working and get this bind test version instead. If it's easier for you to program from a zip instead of dfu please use the one attached to this message.

Oh i see, i thought it was the test build haha.. i download the one on your message here, capture from quad the binding arm and disarm sequence

No problem. But the data in the csv file doesn't look right. Are the SPI settings correct? Please post the logicdata file.

I tried to perform the capture using another usb port and I got the error saying port is in use.

when I revert to the original usb port, the logic program clear out the previous settings and I tried to change it back based on the previous screenshot posted here but I probably missed out something.

My apologies, as I am at work now, I can only re capture tonight in another 11 hours or so.

That's okay. If you saved the logicdata file you can just upload that. Or make a new capture when you can and upload it. Thanks!