Mould King 33043 Super F Quad - HS6200 RF chip

More
26 Feb 2016 15:57 #43681 by mdon
SPI sniffed: AE A6 60 A8 98 56 AA 40 CB
received by nrf: 2E 53 5B A5 F5 7C 53 FC 9A

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

More
26 Feb 2016 16:02 - 26 Feb 2016 16:03 #43683 by goebish
Hey, looks like it matches the xor table I posted on the previous page :)

2E 53 5B A5 F5 7C 53 FC 9A xor 80 f5 3b 0d 6d 2a f9 bc 51 = AE A6 60 A8 98 56 AA 40 CB
Last edit: 26 Feb 2016 16:03 by goebish.

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

More
26 Feb 2016 16:06 - 26 Feb 2016 16:53 #43684 by goebish
next thing to do is to check if CRC is xor'ed (scrambled) as well, then find from what the CRC is generated, are the 2 bytes guards included ? (PCF is, as the CRC changes when packet ID bit changes), and is the CRC generator (CRC16 CCITT) fed with scrambled or unscrambled payload, in which order, with bits order reversed or not etc...

Need a hand with that ? :)

Also, we've to hope this scrambling table is static, as with the xn297. We'll also have to find a HS6200 donator so we can send longer payloads with an actual chip to guess the rest of the scrambling table.
Last edit: 26 Feb 2016 16:53 by goebish.

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

More
26 Feb 2016 17:00 - 26 Feb 2016 17:16 #43689 by goebish
From the datasheet:

"The CRC is the error detection mechanism in the packet. It may either be 1 or 2 bytes and is calculated over the address, Packet Control Field and Payload."

initial value: 0xffff
polynomial: 0x1021
xorout: if we need one (other than 0x0000 or 0xffff) that probably means crc is scrambled too.

looks like the 2 byte guard is not fed to the crc generator (or maybe it's an omission in the datasheet ...)
at least that's one piece of the puzzle ;)

and we need a crc generator that we can feed with an arbitary number of bits, not only bytes because of the PCF (again... I hate this thing :p) as this one:
github.com/RFStorm/mousejack/blob/master/src/radio.c#L107

hmmm I've to check how I can feed a payload to crc reveng that's not a multiple of 8 bit.
Last edit: 26 Feb 2016 17:16 by goebish.

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

More
26 Feb 2016 18:28 - 26 Feb 2016 18:29 #43699 by goebish
Can you post a couple of raw packets, as received by the nrf24 please, with guard, pcf, payload, crc and remaining 7 bit noise ? And also tell me which address is used.
Last edit: 26 Feb 2016 18:29 by goebish.

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

More
26 Feb 2016 18:46 #43701 by mdon
for now I have 3 raw packets:

Control power up:
address: 55 42 9C 8F C9
Payload sent (9 bytes): AE A6 60 A8 98 56 AA 40 CB
32 bytes received: 55 55 27 97 29 AD D2 FA BE 29 FE 4D 5A DE 3B FB 7E BB DD DD AA 39 CA 71 CD 66 AF 9E EE 72 CD 28

Landing:
address: AE A6 60 A8 98
Payload sent (9 bytes): 00 80 40 80 80 40 40 00 BF
32 bytes received: AE AE 25 C0 3A BD C6 F6 B5 5C DE 77 59 50 EF 7E FA EF EB DA 1F 71 AB 22 5A 51 CD 34 BD D2 A8 DD

Full throttle
address: AE A6 60 A8 98
Payload sent (9 bytes): FF 80 40 80 80 40 40 00 BF
32 bytes received: AE AE 25 BF BA BD C6 F6 B5 5C DE 48 D8 F2 2F 7F FA 77 FF FD FF FF 77 FE F7 F5 2A 33 6E C8 55 52

conclusions for now:
guard bytes = first address byte
flight address is sent as power up payload
flight first payload byte refers to throttle

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

More
26 Feb 2016 18:48 #43702 by goebish
Thanks, that should help me to check / sort out CRC.
I'll have a look at that tonight.

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

More
26 Feb 2016 18:50 #43704 by goebish
can you confirm that xor'ing your decoded payload with
80 f5 3b 0d 6d 2a f9 bc 51
yields good results ?

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

More
26 Feb 2016 18:54 #43705 by mdon
Yes, it seems perfect.

const byte xor_table[9] = {0x80,0xf5,0x3b,0x0d,0x6d,0x2a,0xf9,0xbc,0x51};

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

More
26 Feb 2016 19:02 - 26 Feb 2016 20:51 #43707 by goebish
Cool, so it's probably static, I mean it's always the same table whatever the address is ... (well, as victzh suggested for the xn297, there's probably no hardcoded table inside the chip, my guess is those xor are probably the result of the 8 bit CRC generator fed with 0s or something like that, this has to be checked, the fact that the 1st value is 0x80 is interesting ...)

So now that's only a matter of generating the CRC :)
Last edit: 26 Feb 2016 20:51 by goebish.

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

More
26 Feb 2016 23:03 - 27 Feb 2016 14:35 #43722 by goebish
I think I got it, it matches with your 3 raw packets:



So, the CRC generator (lfsr) is fed with:
Address (in reverse order) + pcf + payload (scrambled)
(guard bytes are not used in CRC computation)

.. and the CRC itself is not scrambled (xorout = 0).
Last edit: 27 Feb 2016 14:35 by goebish.

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

More
27 Feb 2016 14:45 #43771 by goebish
I'm trying to code something to generate the packets to be sent by the nrf24... all this shifting is giving me headaches, but I should succeed, eventually :P

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

More
27 Feb 2016 15:01 #43772 by planger
Hi Goebish, If you look at my LT8900 emulation, the bit shifting of the buffer is implemented. You can give it the number of bits you want which is just 1 in your case if I'm correct (LT8900 is anything between 4 to 18bits).

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

More
27 Feb 2016 16:43 - 27 Feb 2016 18:25 #43774 by goebish
Need a bit of clean up, but I got it:

File Attachment:

(only the msb is relevant in the last byte of the generated raw packet)

crappy POC code:
gist.github.com/goebish/a7b5607dc36af06b0cdd

don't forget to alternate the value of pid between packets, especially if 2 successive packets are identical, or the 2nd and following ones might be ignored by the rx.
Last edit: 27 Feb 2016 18:25 by goebish.

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

More
29 Feb 2016 14:41 #43881 by goebish
mdon, do you need more help ?
You didn't tell which machine you were working on :P

For the MK33043, I'll either have to wait that you sacrifice the HS6200 in your stock TX (well, no really, this can be done in a non destructive way) so you can send and sniff longer packets to fill the scrambling table, or I'll have to order one myself ... or 'just' crack the scrambling table generation algo (there's probably one), but I'm lazy ;)

Pascal, yes, when we'll add these new chips emulation in DeviationTX we'll try to use common code as much as possible for all the chips (CRC algo, bit shifting ...)

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

More
10 Mar 2016 09:59 #44367 by goebish
No reply ?
I'll wait 1 or 2 days then order a mk33043 then.

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

More
10 Mar 2016 13:16 #44373 by mdon
Hi.

I was traveling. ;)

Now I'm working on another project, but I'm warm to resume Nrf->HS6200 hobby asap. sorry

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

More
12 Apr 2016 19:05 - 12 Apr 2016 19:05 #46376 by goebish
Looks like someone has ordered a MK 33043 for me (you know who you are, thanks ;)), so I should be able to continue the work on that in a few weeks :)
Last edit: 12 Apr 2016 19:05 by goebish.

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

More
12 Apr 2016 22:20 #46388 by dc59

goebish wrote: Looks like someone has ordered a MK 33043 for me (you know who you are, thanks ;)), so I should be able to continue the work on that in a few weeks :)


:woohoo: Good news for me!

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

More
14 May 2016 22:01 #48459 by goebish
Hmmm I've a bad news, looks like not all the MouldKing Super-F are the same, mine has only 1 chip in the TX and is compatible with the CX-10 protocol, please read from here for more info :(
www.deviationtx.com/forum/protocol-devel...m9916?start=80#48435

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

Time to create page: 0.065 seconds
Powered by Kunena Forum