V202 protocol

I'm hoping i got a bad module and a new one will fix the problem. I love flying these quads with a full size radio!

Daryoon wrote: The v2x2 protocol has a lot of fail in it. For a time, various drop out or range related issues. Even with the stock TX that came with it.

However, your issue doesn't seem related. Since your models fly just fine with your stock TX.

Therefore, I would grab a new NRF24L01 module and hopefully it's just an issue of your first module being defective.

Good luck.

I hope so too. I agree...these quads with a full size radio. It's just a lovely combo.

I like cheap, hobbygrade toys. Thank god for Deviation. Or rather..thank PB and the other developers for Deviation.

Daryoon wrote: From my experience, non of the v202 protocol aircraft flies well around here. (v202, v222, v262) All has DO problems.

I had problems with one defective TX module. I use it near the home where WiFi interference is not so dramatic.

May be you're right about interference and inability of the algorithm to recover.

The packet loss ratio measured from inside of V202 running my own V202 protocol implementation is at least 10%, measured with TX at 100mW standing 20-30cm from V202. The band 2.4GHz in general is quite noisy - WiFi, microwaves etc, so it's not easy.

I suspected algorithmic defect myself sometime ago, and they did exchange of RX boards sometime ago if I remember right. May be new code in V252 and followers is better.

victzh wrote:

Daryoon wrote: From my experience, non of the v202 protocol aircraft flies well around here. (v202, v222, v262) All has DO problems.

I had problems with one defective TX module. I use it near the home where WiFi interference is not so dramatic.

May be you're right about interference and inability of the algorithm to recover.

The packet loss ratio measured from inside of V202 running my own V202 protocol implementation is at least 10%, measured with TX at 100mW standing 20-30cm from V202. The band 2.4GHz in general is quite noisy - WiFi, microwaves etc, so it's not easy.

I suspected algorithmic defect myself sometime ago, and they did exchange of RX boards sometime ago if I remember right. May be new code in V252 and followers is better.

Do you mean that loss packets management is a part of the radio protocol and not hard-coded in the RF chip ?

It's a part of a specific algorithmic implementation of the protocol. After you haven't got the packet you should've, you can decide what to do in different ways. You can wait on the same channel, or if you rely on strict timing you can speculatively wait on the next channel. The second strategy is eons better. I suspect they implemented it only in second gen - V252 and following.

Victzh, the new protocol works like a champ!!

Thanks a lot for the auto bind.

victzh wrote: It's a part of a specific algorithmic implementation of the protocol. After you haven't got the packet you should've, you can decide what to do in different ways. You can wait on the same channel, or if you rely on strict timing you can speculatively wait on the next channel. The second strategy is eons better. I suspect they implemented it only in second gen - V252 and following.

It can explain why DO have been mainly solved by changing/using the V252 TX...

Unfortunaly, we can't access to the RSSI ?

Victzh. Let's speak about your amazing project ?

github.com/victzh/bradwii

Can I flash bradwii on my JXD 385 ?

My god .... If I am not wrong, thanks to bradwii, we can flash new FC software (thanks to remaining debug pins on some V2x2, JXD quad) and by the way change also the implemented radio protocol ?

If I am not wrong, it's GREAT !!!!!!! . I believe the V262 have such pins too (SWDCLK,SWDDATA and GND). I got a "nu-link"
I am dreaming now to flash your bradwii with an optimized radio protocol (@250kbits, Improved sensibility flag, better lost packedt algorithm). Of course this bradwii protocol should be implemented in deviationTX ...
We can dream to extend range by this way ....

Didn't knew Bradwii, interesting project specially the auto tuning PID part, gotta try it. And if we can have something like that for the nano quads that's great.

Keep us posted about your progress, or even better open a thread where this can be discussed

Don't hold your breath, it flies but still needs a lot of work to compete with stock firmware. V202 protocol implemented in the way to compensate for packet loss. I planned to implement multiprotocol RX also - V202 and HiSky first, may be SLT later.

By the way, you don't need the NuLink specifically, ST-Link works great.

victzh wrote: Don't hold your breath, it flies but still needs a lot of work to compete with stock firmware. V202 protocol implemented in the way to compensate for packet loss. I planned to implement multiprotocol RX also - V202 and HiSky first, may be SLT later.

By the way, you don't need the NuLink specifically, ST-Link works great.

I guess the hardest part of this work are flying stabilisation algorithms ?

I fact, I got a chinese clone of the Nulink ....

www.aliexpress.com/item/Free-Shipping-1S...C120/1570055317.html

I got it for 12 USD at this time

I am still afraid to put my read credit card to aliexpress.

There are many problems - the radio and IMU are connected through SPI and I2C buses correspondingly. Sometimes for unclear yet reasons they fail to initialize. The result is either totally unstable helicopter (and at these sizes it can not be controlled manually - it's oscillation frequency is too high for a human to react), or totally stable but deaf to the radio one. Or both!

The radio packet loss ratio is too high to my taste, but that may be just an objective reality - 2.4GHz band is really very noisy.

The stabilization parameters in MultiWii/bradwii defaults are for much larger helicopter, for these smaller models (I managed to fly V202 and JD-385 which is a clone of V252 so far) they should be adjusted.

IMU algorithm in bradwii can be improved. All this can lead to overflowing the very small 16KB (0.5KB of which is used as EEPROM) flash memory.

I spent 3 weeks nights so far on the project, it flew first time after 2 weeks. I expect some other people will work on it as well, so may be in couple months (may be more, may be less) there will be something usable by people without strong embedded development skills.

victzh wrote: I am still afraid to put my read credit card to aliexpress.

There are many problems - the radio and IMU are connected through SPI and I2C buses correspondingly. Sometimes for unclear yet reasons they fail to initialize. The result is either totally unstable helicopter (and at these sizes it can not be controlled manually - it's oscillation frequency is too high for a human to react), or totally stable but deaf to the radio one. Or both!

The radio packet loss ratio is too high to my taste, but that may be just an objective reality - 2.4GHz band is really very noisy.

The stabilization parameters in MultiWii/bradwii defaults are for much larger helicopter, for these smaller models (I managed to fly V202 and JD-385 which is a clone of V252 so far) they should be adjusted.

IMU algorithm in bradwii can be improved. All this can lead to overflowing the very small 16KB (0.5KB of which is used as EEPROM) flash memory.

I spent 3 weeks nights so far on the project, it flew first time after 2 weeks. I expect some other people will work on it as well, so may be in couple months (may be more, may be less) there will be something usable by people without strong embedded development skills.

I already did around 20 order with Aliexpress .... Never had a problem but it's true, you have to choose carefully the vendor. They are in the best case, ommissing some product informations, in the worst case, they are liying. Take care of the processus time too

I am really impressed by your work. Is this work linked with your first education ?

By the way, I am very interested by this one ....
Are u controlling the JXD 385 "bradwiized" with deviationTX ?

Yea, that's funny when your code in transmitter controls your code in the model I tell you it's a very special feeling when your code controls the flying thing for the first time!

Of course it not only my code, admittedly the amount of other people's code is overwhelming, but anyway it's a jolt for which I in particular do this.

There are many programmers who could do this, it's more a combination of skills and interest.

victzh wrote: Yea, that's funny when your code in transmitter controls your code in the model I tell you it's a very special feeling when your code controls the flying thing for the first time!

Of course it not only my code, admittedly the amount of other people's code is overwhelming, but anyway it's a jolt for which I in particular do this.

There are many programmers who could do this, it's more a combination of skills and interest.

Stabilization is mainly performed by PID ? No more complex algorithm using Kalman filter and so on ?

We're talking with Tim - the author of the JD-385 hack - about implementing something more suitable for complex navigation. The problem is space, it's very limited and we're close ti this limit. I would like to keep the ability to use MultiWiiConfig on it, and it all takes space. Q-BOT micro with larger processor would be easier.

If we'd be able to have easily hackable ARM-based microquad, it would be amazing. The problem is everyone loves their ATmegas

victzh wrote: We're talking with Tim - the author of the JD-385 hack - about implementing something more suitable for complex navigation. The problem is space, it's very limited and we're close ti this limit. I would like to keep the ability to use MultiWiiConfig on it, and it all takes space. Q-BOT micro with larger processor would be easier.

If we'd be able to have easily hackable ARM-based microquad, it would be amazing. The problem is everyone loves their ATmegas

Actually, the original JXD 385 code is not yet dumped ?

The MCU is locked and can't be dumped.

victzh wrote: The MCU is locked and can't be dumped.

I got many of these combo mini54ZAN + BK2423 + Infineon (V212, V262, V272, V252, jxd 388, 385).... maybe for one of them is not locked ?