WLToys Q303

Unless you've a Q303, there's no reason to replace those parts, it will work the same as it is.

Ok, replacing the 2 capacitors should be enough then, ajtank, can you take a picture to show where they're located on the board ?

I just received your FC, thank you very much for already having soldered leads to the RF board, I only have to connect it to my analyzer

Now I'll try to understand the tx id / rf channels algo, or at least generate enough to avoid interferences if many people fly Q303 with Devo transmitters at the same time (unlikely, but still )

Also, we've now agreed to add protocols to Deviation even if only 1 tx id / channel hops pair is known, so I'll add CX-10D / CX-10WD / CX-35 / CX-20 and H20H soon.

ajtank wrote: Goebish can you tell where you purchased the chips for your module ?[/attachment]

banggood or ebay when bg is out of stock, also got a set from taobao once, those modules are quite expensive (there are other modules with pa+lna+antenna that cost $3 ...) but I never had an issue with them ...

I do not have raw chips, but those module boards:
www.banggood.com/DEVO-Transmitter-Tuner-...ys-V922-p-89812.html
That's what I use to make nrf24 modules for Taranis and other transmitters.

Work in progress for the Q303 ... that's a real pleasure to crack a protocol when we can tap onto the RX to retrieve expected frequencies with crafted transmitter IDs
I still have to check a few stuffs but the algo seems really simple.

So, it's extremely simple ...
TX id is 32 bit long (arbitrary), but only 2 bits are used to generate the frequencies sequence (4 hops), yes, there are only 4 possible sequences ...

and to answer your previous question, the lower frequency it can use is 2470 MHz and the higher one is 2479 MHz.

I suppose the CX35 is using a similar pattern but with a different base frequency (0x46 here), we'll have to check that

edit: if this theory is verified, the cx35 is using 0x14 as base channel and channels are spaced 3 MHz apart instead of 2MHz on the Q303, ie:That matches with what you captured from the CX35 transmitter:
ID: 24 37 46 89 (offset = 0x24 & 3 = 0)
RF: 14 17 1A 1D

Test builds for Devo 7e & 10 available here:
www.dropbox.com/sh/zfnlskw1p6ivjth/AAC-5...ikgRQKn6QXN1Eka?dl=0
source: github.com/goebish/deviation/blob/protoc...ocol/q303_nrf24l01.c
If you confirm everything is working fine it will be added to the nightly, please try with a few different transmitter IDs ...

Next is the CX35, if my theory is right, that's 2 birds with 1 stone

ajtank wrote: PS: it seems ID cannot begin with '0'.

That's normal, trailing zeros are removed, 012345 is the same than 12345.

The CX35 writes only 1 byte to register 0x19 (enable whitening), so yes, that's a XN297L, register 0x19 is a multi-byte register on non L version (DEMOD_CAL), also this protocol runs at 1 Mbps and whitening is enabled, so it should not have the same issues than the Q303 ...

The JXD-506 and H26WH are not using a xn297 but a Beken RF chip (bk2421 or bk2423 or something like that ...).