DSM2/X and "Chinese Roulette"

er9x is the best known firmware. It has a superb range of features and is well supported by the community. Well worth trying out.
Camboui
Posts: 79
Joined: Wed Feb 15, 2012 3:01 pm
Country: -

DSM2/X and "Chinese Roulette"

Post by Camboui »

I posted on DSM2/X mod forum, in the "The end of DSM mod..." thread.
It's about conflicts between transmitters that could control the same DSM receiver simultaneously when using the Orange compatible DSM module sold by hobbyking.

I'm cross posting here because you, er9x developpers that implemented DSM2/X inside the firmware, may know better what is this about.
The "chinese roulette" is an article in a german RC magazine "Modell" which suggests that the Orange DSM module uses identifers as per DSM protocol that could not be unique (randomly assigned by Orange). So pilots on field using legit Spektrum tx and others using Orange modules could interfer and crash.
Is that possible ?

User avatar
ShowMaster
Posts: 4327
Joined: Thu Dec 29, 2011 3:44 am
Country: -
Location: Los Angeles, CA USA

Re: DSM2/X and "Chinese Roulette"

Post by ShowMaster »

Good question?
I posted my concern here a while ago and in fact dsm2 concerns in general.
So far it's not generated any comments or discussions but maybe with your post it will.
It's been shown that dsm2 can suffer lockouts and Spektrum quietly phased in DSMX to replace dsm2 but never posted why!
Now with the orange module from HK and the ability to change the ID at random, I personally have abandoned using my Spektrum systems except for very small aircraft that can't do much damage or costs much.
SM



Sent from my iPod touch using Tapatalk.
Helle
Posts: 577
Joined: Sat Jul 21, 2012 7:08 am
Country: -

Re: DSM2/X and "Chinese Roulette"

Post by Helle »

Hy,

I am from Germany and it was a discussion and warning at RCnetwork.de 1 or 2 month ago
the original artikel at the forum is from DD8ED and he wrote it at "Modell" 3/2013 again

http://www.rc-network.de/forum/showthre ... e-Warnung-!

But:
No one has realy measured and delifered facts that there are Problems with the orange DSM2/DSMX Modul only!
(only with the orange DSM2/DSMX Moduls , not with original SpectrumTX )

They "think", they "beleave" that "could be" because of the 5 or 8 GUID you can select at each Modul

After my ask in details none of them could give a measerment or a protokoll or demonstration or proofs

DD8ED wrote some "secret people" told it him that HK has only made a list of 5 or 8 same GUID
because they were not able to hack the hopping code generated with the GUID at the cypress chip


My simple question:
Who can demonstrate it?
Where are the measurments of the GUID
Where are the proofs?

Helle
User avatar
tilmanb
Posts: 347
Joined: Thu Oct 11, 2012 9:36 pm
Country: Germany
Location: Karlsruhe, Germany

Re: DSM2/X and "Chinese Roulette"

Post by tilmanb »

OMFG, that would make those modules actually dangerous for other Orange DSM2 TX users.
Regardless of when the model got binded even, because as far as I understand is binding done on a per GUID basis. With such a small key-space conflicts are almost inevitable. And impossible to avoid with practices similar to frequency pegs.

I don't like those modules personally (because they are PPM only) but I still hope this turns out to be FUD only.
User avatar
Kilrah
Posts: 11108
Joined: Sat Feb 18, 2012 6:56 pm
Country: Switzerland

Re: DSM2/X and "Chinese Roulette"

Post by Kilrah »

But that's the point. There are suppositions that the modules have a limited pool of IDs, but nobody actually managed to get a conflict even when trying to.
Which is the whole problem here. Some are just being cautious, some are using that uncertainty as an occasion to blame HK as a whole, some are saying there's no problem until a conflict has actually happened... i.e. it's a mess.

While saying there's no risk is probably wrong, going out on every site and in the press and screaming for scandal without a single trace of evidence is also unjustified.

The problem is that HK never said a word on the subject, so suppositions are prevalent. And it would be so easy for them to say so and explain why if there was no problem that people kind of assumed that there is one.

Helle
Posts: 577
Joined: Sat Jul 21, 2012 7:08 am
Country: -

Re: DSM2/X and "Chinese Roulette"

Post by Helle »

Hy,

does HK realy know about this discussion in Germany and at "Modell" about the orange DSM2/DSMX Moduls
because there are only diskussions in German Language by now.

And Yesterday I wrote to DD8ED personal and to the German DMFV, Mr. Perkhuhn, to give me a detail proof

Lets seee whats happen

If its true, its would it be a criminal act form HK, because they now that and
the orange DSM/DSMX are selling to the hole world

if not, it would be a criminal act doing from DD8ED and "Modell"

So easy it is!

And here is the forum for hackung DSMX with GUID start at side 1 orange is at side 8
http://www.rcgroups.com/forums/showthre ... 502&page=8

Helle
User avatar
rperkins
Posts: 1422
Joined: Sun Jan 08, 2012 12:51 pm
Country: -

Re: DSM2/X and "Chinese Roulette"

Post by rperkins »

I'd be the last to defend dsm,dsmx,HK,orange et al.
I know there are much more robust,desirable, feature laden solutions.
however
By my limited understanding - Seems like a fair amount of FUD going on here.

I'm sure I'm making some wrong assumptions and am trying to gain knowledge as to which assumptions are invalid -- thanks
1. dsm had been deciphered long before the orange TX. Maybe HK/orange were too stubborn to use the publicly available data, probably not. No one is saying that dsm2 is suspectable, are they ?
2. The issue resolves around dsmx.
3. The issue is that HK/orange did not decipher the entire code sequence,PN code, hence the button on the back.
4. This in no way affects 'genuine' dsmx Tx. They have the entire code sequence .

So the issue seems to be if you have more than one orange module running in dsmx mode at the same time and they happen to be on the same PN code ( which the orange module utilizes ?? 5-10 of them ), and the timing is such that the 2 signals are synchronized and stomping on one another, and the interference is so great that the receiver cant deal with it as all ISM band devices account for other devices using the same frequency, and the interference lasts for a long enough time for the rx to lose signal, then the plane falls out of the sky.

I'm not saying it's something to ignore, and yes at a large field it may be an issue.
So if you go to a busy field how many 9x's do you see that have an orange module sticking out the back ?
That's something I know nothing about
JustinTime
Posts: 140
Joined: Tue Jan 29, 2013 4:12 am
Country: -
Location: SoCal

Re: DSM2/X and "Chinese Roulette"

Post by JustinTime »

Is this problem only with the module that you buy and plug in the back? If I make the Dx4 mod I'm fine?
User avatar
tilmanb
Posts: 347
Joined: Thu Oct 11, 2012 9:36 pm
Country: Germany
Location: Karlsruhe, Germany

Re: DSM2/X and "Chinese Roulette"

Post by tilmanb »

JustinTime wrote:Is this problem only with the module that you buy and plug in the back? If I make the Dx4 mod I'm fine?
yes
User avatar
ShowMaster
Posts: 4327
Joined: Thu Dec 29, 2011 3:44 am
Country: -
Location: Los Angeles, CA USA

Re: DSM2/X and "Chinese Roulette"

Post by ShowMaster »

I mentioned this at my local field, not a private one BTW, to the turbine flyers. Mostly I got the answer it can't happen to them. To be fair, we don't know that it will, but I'm not betting $7-15k it won't.
At least it was nice enough of HK to warn us it could happen, right???
SM


Sent from my iPod touch using Tapatalk.
User avatar
jhsa
Posts: 19480
Joined: Tue Dec 27, 2011 5:13 pm
Country: Germany

Re: DSM2/X and "Chinese Roulette"

Post by jhsa »

I think that if they put that button there and warn about the problem, it is because it already happened to them..
My er9x/Ersky9x/eepskye Video Tutorials
https://www.youtube.com/playlist?list=PL5uJhoD7sAKidZmkhMpYpp_qcuIqJXhb9

Donate to Er9x/Ersky9x:
https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=YHX43JR3J7XGW
User avatar
rperkins
Posts: 1422
Joined: Sun Jan 08, 2012 12:51 pm
Country: -

Re: DSM2/X and "Chinese Roulette"

Post by rperkins »

jhsa wrote:I think that if they put that button there and warn about the problem, it is because it already happened to them..
SM wrote: At least it was nice enough of HK to warn us it could happen, right???
2 good points. Thanks
User avatar
Rob Thomson
Site Admin
Posts: 4543
Joined: Tue Dec 27, 2011 11:34 am
Country: United Kingdom
Location: Albury, Guildford
Contact:

Re: DSM2/X and "Chinese Roulette"

Post by Rob Thomson »

I believe the reason for the id change button is simple.

With dsm, spectrum to date have controlled the release of the uid values.

Now.. We have two manufacturers releasing uid's. But no chance of them cooperating with each other.

So... Assuming dsm uid keys follow a standard format, orange would have no choice but to keep to the uid format the protocol requires.

So.... How do you prevent a possible clash with a spectrum tx, if you don't have cooperation with horizon?

You put a button on the module that allows you to change your id, to cater for the slim chance of a collision at a flying site!



Sent from my GT-I9300 using Tapatalk 2
Slope Soaring, FPV, and pretty much anything 'high tech'
...........if you think it should be in the wiki.. ask me for wiki access, then go add it!
User avatar
jhsa
Posts: 19480
Joined: Tue Dec 27, 2011 5:13 pm
Country: Germany

Re: DSM2/X and

Post by jhsa »

Rob Thomson wrote:
So.... How do you prevent a possible clash with a spectrum tx, if you don't have cooperation with horizon?
Easy answer.. don't sell without being sure.. It's dangerous
You put a button on the module that allows you to change your id, to cater for the slim chance of a collision at a flying site!
As if that would work.. well, unless you ask the guys that fly DSM not to turn their equipment on at the field without confirming with the people that own orange RF modules that there is no interference..
Also make sure that no one is flying a bnf plane somewhere on a near field..

Actually that would be worse than going back to the crystals system.. at least you could see the channel you're in and tell the others, or get a peg from the frequency board and put it on your antenna.. if the peg was not on the board you knew that someone had the same frequency as you
you... didn't have to turn your radios on to check for interference..
My er9x/Ersky9x/eepskye Video Tutorials
https://www.youtube.com/playlist?list=PL5uJhoD7sAKidZmkhMpYpp_qcuIqJXhb9

Donate to Er9x/Ersky9x:
https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=YHX43JR3J7XGW
User avatar
ShowMaster
Posts: 4327
Joined: Thu Dec 29, 2011 3:44 am
Country: -
Location: Los Angeles, CA USA

Re: DSM2/X and "Chinese Roulette"

Post by ShowMaster »

I know, we put a colored flag in our little orange module antennas warning given we DSM flyers need to take turns. That should really cut down on the amount of Spektrum fliers being in the air at one time to (1) at my local crowded field!
Humm, so now to think about the negative impact that creates... Thinking ... Thinking ....
Ok, I know this is serious stuff but the fact that so many orange modules have been sold that fit any JR module pinout, not just the 9x and 9xr series, and my box has a checkmark for Futaba also, it is a joke, serious, but a joke!
What about the orange $68 DSM tx that HK's been selling before the orange module was released? I'm sure all those have the same change I'd button on those as well?
Anyone own one that can comment?
SM
SM


Sent from my iPod touch using Tapatalk.
User avatar
rperkins
Posts: 1422
Joined: Sun Jan 08, 2012 12:51 pm
Country: -

Re: DSM2/X and "Chinese Roulette"

Post by rperkins »

dont forget about microwave ovens, wifi,bluetooth device, and anything else in the ISM band. Inferring any protocols use on the ISM as 'chinese roulette' is just over the top IMHO

By the very nature of ISM band "communication devices using the ISM bands must tolerate any interference from ISM equipment, unlicensed operations are typically permitted to use these bands,"

Over and above that all versions of 'dss' inject a known table of noise into their transmission which is removed at the receiver to improve the s/n ratio that is inevitable in the ISM.

Comparing rc usage in the ISM to the crystal days is not an applicable comparison, IMHO.

Earlier i outlined a narrow scenario in which an issue could occur based on assumptions that have not been challenged. Since then I took the time to gather some more information such as 'There are over 4 billion unique ID's assigned to OrangeRx modules.',albeit only 5 per module. Additionally it appears the model match further uniquely identifies each packet sent out.

I've never been to a large rc field but am getting the impression that dsm planes are just falling sky left and right and there is evidence pointing to orangerRX equipment being at fault. And if I show up there with spektrum gear I am shuned, orangerx gear and I am ran out of town :)

I know spektrum/dsm is not popular/robust/modern/etc. Protocol and i cant carry a political conversation as to its merits. I would however like to carry a technical conversation as to the affect orangerx products have on other dsm based products over and above the normal interference all ISM devices are designed to overcome.

I do have an orangeRX module coming with my 9xr. So this does have a practical aspect besides my typical disdain for a FUD.
pmullen503
Posts: 275
Joined: Sun May 06, 2012 8:37 pm
Country: -
Location: Fond du Lac, WI

Re: DSM2/X and "Chinese Roulette"

Post by pmullen503 »

If there really are 4 billion OrangeRx UIDs, then the chance of conflict is negligible. The addition of the change ID button and Orange/HK's silence on the true risks (or lack thereof) have allowed rampant speculation on various fora. I've yet to see a report of an actual OrangeRx UID conflict.

But with the cost of a "real" Spektrum module from a donor radio so cheap (I got mine from a HP6DSM for $35), why use the OrangeRx module at all?
User avatar
jhsa
Posts: 19480
Joined: Tue Dec 27, 2011 5:13 pm
Country: Germany

Re: DSM2/X and "Chinese Roulette"

Post by jhsa »

more channels?
My er9x/Ersky9x/eepskye Video Tutorials
https://www.youtube.com/playlist?list=PL5uJhoD7sAKidZmkhMpYpp_qcuIqJXhb9

Donate to Er9x/Ersky9x:
https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=YHX43JR3J7XGW
User avatar
rperkins
Posts: 1422
Joined: Sun Jan 08, 2012 12:51 pm
Country: -

Re: DSM2/X and "Chinese Roulette"

Post by rperkins »

I built a dx4e home brew. Works well for me. For the cost of procuring another donor and case plus the time involved for a home made result, adding a $30 module to my cart when purchasing an additional tx seemed appropriate. If only it wasnt orange :)

Btw -- a 32 bit number ,aka, 2 to the 32 power is = 4,294,967,295 so 4 billion UID's is not a stretch of the imagination :)
User avatar
rperkins
Posts: 1422
Joined: Sun Jan 08, 2012 12:51 pm
Country: -

Re: DSM2/X and "Chinese Roulette"

Post by rperkins »

And to be fair to the OP, the "chinese roulette' designation more than likely came from the magazine article he references. Hype sells magazines.
User avatar
Kilrah
Posts: 11108
Joined: Sat Feb 18, 2012 6:56 pm
Country: Switzerland

Re: DSM2/X and "Chinese Roulette"

Post by Kilrah »

pmullen503 wrote:But with the cost of a "real" Spektrum module from a donor radio so cheap (I got mine from a HP6DSM for $35), why use the OrangeRx module at all?
More channels, plug and play.
Helle
Posts: 577
Joined: Sat Jul 21, 2012 7:08 am
Country: -

Re: DSM2/X and "Chinese Roulette"

Post by Helle »

Hy,

according to the German DD8ED and the "Modell" magazin the problem are only with 2 and more orange DSM/DSMX moduls togehter are ON
and in DSMX Mode.

Because they say there are onyl 5 different GUID at the moduls.
and so someone can catch a flying modell when he switch his TX on too.

Not with original spectrum, because spectrum has unique GUID at each modul.
They use the GUID at the Cypress-chip to generate the hopping sequenze for their moduls.

So someone shoud test it at SPI interface to the HF-chip and with a Protokollanalyser

Or test it with two or three orange DSM/DSMx moduls and push the button

And here read this: http://www.rcgroups.com/forums/showthre ... 202&page=9

Helle
User avatar
rperkins
Posts: 1422
Joined: Sun Jan 08, 2012 12:51 pm
Country: -

Re: DSM2/X and "Chinese Roulette"

Post by rperkins »

Helle wrote:So someone shoud test it at SPI interface to the HF-chip and with a Protokollanalyser
I think sniffing the SPI is how they obtained the PN codes in the first place. The cypress chip is documented as far as the number of channels it can utilize

How about a statistical approach ?

I've read enough posts over the last year from this hammer22 dude to be able to say that he has a better idea than most about what is going on, so I post this with some level of confidence of its accuracy. Although this assumption assuming that all Orange modules use the same 10 frequency hopping sequences conflicts with the HK statement below of There are over 4 billion unique ID's assigned to OrangeRx modules.

http://www.rcgroups.com/forums/showpost ... tcount=118
hammer22 of rcg wrote:The DSMX frequency hopping algorithm appears to use the entire 32 bit GUID to calculate the hopping pattern.
http://www.hobbyking.com/hobbyking/stor ... ible_.html
HK says wrote:Module uses a unique Global ID for binding. In case of binding issues with other branded receivers press the "Change ID" button on module to change the Module ID to another unique value reserved for that module (5 ID's assigned to each module).There are over 4 billion unique ID's assigned to OrangeRx modules.
<scratches my head>

EDIT - i got to thinking by multiplying the 2 tx I was coming up with all possible GUID combinations of which more than one would match. I have corrected for that error. thanks

would that be (4,294,967,295/5 )* (4,294,967,295/5)chance of same GUID

oh.. not taking into consideration model match. how many models does average tx use. Lets forget about that and say each Tx only has one model. Once they get up to 5 models then you back up to the entire 32 bit space anyway.

hmm.. and by looking at the code at the same rcg post it appears that dsmx hops between 23 channels.
((4,294,967,295/5 ) * 23) * ((4,294,967,295/5 ) * 23) chance of same GUID on same channel at same time



that comes out to 1 in 3.903331044×10²⁰ 19,756,849,557 chance

I have no doubt there are flaws in my statistical ability. Maybe someone will come along and correct them. I still stand by the statement that 'Although dsm(x)/spektrum protocols apparently have their issues a GUID conflict caused by an OrangeRX TX module which causes an unrecoverable error in the rf link of the ISM band is very improbable'

I'm gonna try to leave you guys alone with this now :) Feel like I'm beating a dead horse.
Last edited by rperkins on Fri Mar 15, 2013 1:39 pm, edited 1 time in total.
User avatar
Kilrah
Posts: 11108
Joined: Sat Feb 18, 2012 6:56 pm
Country: Switzerland

Re: DSM2/X and "Chinese Roulette"

Post by Kilrah »

Helle wrote:Because they say there are onyl 5 different GUID at the moduls.
and so someone can catch a flying modell when he switch his TX on too.
This is an (IMO wrong) interpretation of what HK have put on their product page:
Module uses a unique Global ID for binding. In case of binding issues with other branded receivers press the "Change ID" button on module to change the Module ID to another unique value reserved for that module (5 ID's assigned to each module).
There are over 4 billion unique ID's assigned to OrangeRx modules.
I understand it as:
There is a 4 billion ID pool, and each module gets 5 unique ones of those. So each module has 5 IDs, but every ID is only assigned to one single module". E.g. Module 1 has 1,2,3,4,5, module 2 has 6,7,8,9,10, etc.

Some understand it as "every module has the same 5 IDs", and that's what caused all that "chinese roulette" thing.

With their understanding, you could take 2 ORX modules, bind one to a receiver, power it off, power the other on, press the Change ID button <5 times and you'd get control of the receiver. This has been tested and proven wrong.

With my understanding there would never be a collision possibility with 2 ORX modules. There could however be one with a Spektrum radio, as there is a chance one is using the same ID as that module. If this happened you'd press change ID, and as there might also be another Spektrum radio using it there are another 3 to choose from. Chances are tiny as only one Spektrum radio would use it, providing it even exists (no idea how Spektrum assigns their IDs, random or sequential). The thing is, with this system you should be able to bind a receiver to your module, press change ID once, lose control, press 4 more times, get control again, but AFAIK this has also been tested and doesn't work. So nobody understands what's really going on.

And we don't know if it only applies to DSMX, DSM2 or both... There was speculation that the first method (5 IDs, same for each module, or a different but limited number) would be used for DSMX due to the lack of knowledge of how to convert an ID to a channel table, requiring sniffing original modules, and the second (5 different on each module) for DSM2. Again, no proof it's correct nor evidence it's wrong AFAIK.

Again it would be easy for HK to clarify, if everything was right they could just explain in detail and the issue would be over. As the text is a bit vague, and every time someone asks the question the ticket gets closed without answer people tend to believe HK knows something is wrong but don't want to admit it. Maybe everything is OK but they're just stupid and doing bad marketing...

Who knows.
User avatar
Rob Thomson
Site Admin
Posts: 4543
Joined: Tue Dec 27, 2011 11:34 am
Country: United Kingdom
Location: Albury, Guildford
Contact:

Re: DSM2/X and "Chinese Roulette"

Post by Rob Thomson »

That is exactly how I understand it.

Sent from my GT-I9300 using Tapatalk 2
Slope Soaring, FPV, and pretty much anything 'high tech'
...........if you think it should be in the wiki.. ask me for wiki access, then go add it!
Helle
Posts: 577
Joined: Sat Jul 21, 2012 7:08 am
Country: -

Re: DSM2/X and "Chinese Roulette"

Post by Helle »

hy my friends,

here the link for the open source for the orange DSM2/DSMX software

sorry was wrong here its ok:
http://code.google.com/p/orange-tx-module-dsmx/

its for a ATmega with new features :P

Helle
Last edited by Helle on Sun Mar 17, 2013 7:54 am, edited 1 time in total.
User avatar
rperkins
Posts: 1422
Joined: Sun Jan 08, 2012 12:51 pm
Country: -

Re: DSM2/X and "Chinese Roulette"

Post by rperkins »

thanks Helle
mangled url above. maybe this one will work
http://code.google.com/p/orange-tx-module-dsmx/
User avatar
Rob Thomson
Site Admin
Posts: 4543
Joined: Tue Dec 27, 2011 11:34 am
Country: United Kingdom
Location: Albury, Guildford
Contact:

Re: DSM2/X and "Chinese Roulette"

Post by Rob Thomson »

So is this the actual code in the module?

Sent from my GT-I9300 using Tapatalk 2
Slope Soaring, FPV, and pretty much anything 'high tech'
...........if you think it should be in the wiki.. ask me for wiki access, then go add it!
User avatar
SR71
Posts: 109
Joined: Tue Feb 28, 2012 10:21 pm
Country: Italy
Location: Rome

Re: DSM2/X and "Chinese Roulette"

Post by SR71 »

Helle wrote:hy my friends,

here the link for the open source for the orange DSM2/DSMX software

http://code.google.com/p/orange-tx-m...mx/source/list
its for a ATmega with new features :P

Helle
Correct link is

http://code.google.com/p/orange-tx-modu ... ource/list
User avatar
rperkins
Posts: 1422
Joined: Sun Jan 08, 2012 12:51 pm
Country: -

Re: DSM2/X and "Chinese Roulette"

Post by rperkins »

doesnt look like it.

But it would be appealing to replace whatever comes on the module with open source code. The code on this site has some comments that appear to have gotten lost in a utf-8 shuffle but some of the russian is peeking through. Given that the link on rcgroups has a russian heritage I would say that this project is related to the those posts. hmm. wouldnt that be cool to use the orangerx tx module as a hardware basis for multiple protocols :) I have no idea if that is possible.

yes it is the same guy 'rw9uao'

hmm .. looks like there are 24 channels but I only accounted for 23 of them above in my pseudo math :)

Post Reply

Return to “er9x”