Independent tests prove lack of frequency hopping with XPS

[tadawson]

	Quote: Originally Posted by socalsal
	Question> can a system be made to "hop" just by re-writing the software?? or is it a hardware issue? My thoughts are if it is a software issue then would it not be a simple case of updateing the software to enable a system to do this? If it is the software what is preventing someone from writing a "update" and providing it to others?? are there some type of laws against it??

According to most here, it *does* hop, just not under all the conditions that they think it should . . . . Jim is supposedly rewriting the code to add those other conditions . . . and only time will tell . . .

- Tim

[GBR2]

JD has recently said the following:

"The current hopping code (v1) is done only when there are no communication issues between the units. This is a predictive technique that sees a potential issue and will make a change before a problem arises. A byproduct of this technique is extended range. The new hopping code (v2) adds a reset of the frequency allocation when a complete (not partial) saturation occurs. A third version of the code (v3) will add the ability to use as many backup channels as the user likes, including all of them to provide a psuedo-full time frequency hopping. Individual frequencies will also be able to be assigned to the remote receivers. Since the majority of our hardware has been in-field flashable since October of last year, and all future hardware will be in-field flashable, we have the capability to provide new firmware as these features become available."

Although I believe there might be some issues in actually doing it based on the hardware being used by XPS. XJet would be able to address the question about software/hardware better. Personally, even if XPS came out with new code, it would really have to be demonstrated that it actually does what XPS says it will do before I'd commit any XPS equipment into anything but a foamie. I'd also would like to see some antenna diversity, which XPS said they were going to do but that hasn't materialized yet either.

[Woketman]

	Quote: Originally Posted by socalsal
	Wow.you are really mad at this JD... I can understand this. However why would you call anyone such names??? All this does (name calling) is cheapen your image in the eyes of others. You have the right to say anything you want . However don't you feel that letting this subject degrade to name calling is going a bit far?

Whatever. The word that I used to describe JD is accurate and a perfect name to call a lier. Call a spade a spade. He DEFINITELY deserves it!!!

[capgains]

What is antenna diversity, and what does it do for you? I mean, I understand the way the Spektrum has the 2 pieces of one, and that FASST has 2 whips, but do you really need it, or is it more of that warm fuzzy type stuff? Personally, I don't care if it has no antenna if it works.

[Toumal]

	Quote: Originally Posted by capgains
	What is antenna diversity, and what does it do for you? I mean, I understand the way the Spektrum has the 2 pieces of one, and that FASST has 2 whips, but do you really need it, or is it more of that warm fuzzy type stuff? Personally, I don't care if it has no antenna if it works.

Well, consider this: A monopole antenna (such as what you have on your receiver) exhibits different reception "sensitivity" depending on whether the antenna points straight at the TX, or is at a 90 degree angle to it. The best reception occurs when both antennas are parallel to each other. The worst reception happens when both point straight at each other. Depending on circumstances, you can have quite severe signal dampening just because of antenna orientation.

If you have antenna diversity, you obviously have two antennas. By arranging them in a way that has them at a 90 degree angle to each other, it is impossible for the airplane/helo to be in a position where both antennas point straight at the transmitter. If one antenna does point at the TX, the other will be at a 90 degree angle and thus have a much better reception. The receiver simply uses whatever antenna delivers the strongest signal level at any given moment.

Spektrum's DSM2 has both antenna and frequency diversity, which means they use two frequencies at the same time, and they have two antennas per frequency. Futaba doesn't have frequency diversity, but instead hops to a new channel many times per second.

Antenna diversity is no magic bullet against bad receiver location, like putting it all into an RF-tight fuselage. But in real life, when your model is in a bad position and maybe the engine block further blocks off the radio signal, having a second antenna with momentarily better reception can mean the difference between a safe landing and a crash.

I wouldn't call it a "fuzzy-feeling-feature", rather it's definitely a safety improvement.

[XJet]

	Quote: Originally Posted by GBR2
	JD has recently said the following: "The current hopping code (v1) is done only when there are no communication issues between the units. This is a predictive technique that sees a potential issue and will make a change before a problem arises.

Yes, the way JD has tried to make this work is by monitoring the number of retries required to send a packet. If the noise floor *gradually* increases, the retry count will also gradually increase. Providing the increase is slow enough, you'll have time to negotiate a new frequency and hop.

Unfortunately, this just isn't a real-world situation.

When an interfering signal appears in the 2.4GHz world it tends to just "appear" from nowhere -- not signal its impending arrival by slowly increasing in intensity as is required for this algorithm.

Such signals usually appear suddenly when an interfering device is just switched on (instant appearance) or when the interfering device becomes "visible" as you fly into its signal (ie: if it was previously shielded by terrain, buildings or other features).

So this is absolutely *useless* as a method of dodging interference -- as Kiwi's tests clearly proved.

If it really were able to "predict" interference and jump before it arrived -- why didn't it see that signal which slowly crept up the band towards the XPS operating frequency?

No, JD has blatantly lied on this one when he claimed that XPS would hop away from an interfering signal. It doesn't, it can't and it won't. And, to his credit, even he then backed away from that initial claim (but only after the proof was published) to say that it was simply a range-improvement strategy.

	Quote:
	The new hopping code (v2) adds a reset of the frequency allocation when a complete (not partial) saturation occurs.

There he goes with "saturation" again (sigh!).

I suspect that what he's going to do here is negotiate two frequencies (a primary and a fall-back) at turn-on time. The system will use the primary until such time as it is unable to sustain communications, at which point it will "fall back" to the secondary.

Now this is "better than nothing" but not without a few "gotchas" of its own.

Assuming that he does an energy scan to establish which channels are quietest and then picks the one with the lowest noise as primary and the one with second-lowest as secondary, a problem could exist when several XPS systems are being flown simultaneously and noise hits part of the band.

If they've all chosen the same secondary channel (quite probable) then you could wind up with the whole lot of them trying to use the same secondary channel when their primaries get hit - and remember that XPS only has a choice of 12 channels (unlike Spektrum and Futaba). It's better than the existing option but still not perfect -- because that secondary channel could also be in the part of the band that's hit by the noise -- but that's true of Spektrum also.

There are also some *very* subtle issues in respect to the handshaking required to decide when/if to fall back to the secondary channel and whether to jump back to the primary if that channel is also lost to noise. At least with Spektrum the receiver is always listening on both channels -- the XPS receiver can only listen to one at a time so it really needs to be sure it's listening to the same one that the transmitter is broadcasting on.

	Quote:
	A third version of the code (v3) will add the ability to use as many backup channels as the user likes, including all of them to provide a psuedo-full time frequency hopping.

That's never going to happen. Using the XBeePro with its reliable transport layer there's just insufficient determinism in the link to support FHSS.

[capgains]

Okay, but is this a short range type of problem, or an out of LOS problem? I've always been told to avoid pointing the antenna of the TX at the airplane, but never seen problems if I did... on 72 Mhz. The signal might not be as strong, but still well in reception to maintain apearantly normal control.

If I'm not mistaken, 2.4 is a flatter wave. Does that make it more critical regarding antenna orientation?

[Toumal]

	Quote: Originally Posted by capgains
	If I'm not mistaken, 2.4 is a flatter wave. Does that make it more critical regarding antenna orientation?

"Flatter"? I'd say that the higher the frequency, the more severe the attenuation effect near the antenna's pole. There's a lot of difference between 72MHz and 2.4GHz in terms of general propagation behavior.

As one of my RF tech teachers would've said: Never look into a metal pipe's open end. Either someone's flushing their toilet or there's microwaves coming out

[BZFrank]

	Quote:
	As one of my RF tech teachers would've said: Never look into a metal pipe's open end. Either someone's flushing their toilet or there's microwaves coming out

LOL!

BTW and FYI It seems IFS/XPS is now banned from the Speedcup plyon/speed competition and very likely more competitions this year, due to a series of crashes / flyaways (lockouts and failure to activate failsafe) here in Germany. It is likely other contests here will follow these rules too until the situation is clarified.

http://translate.google.de/translate...hl=de&ie=UTF-8

[jonkoppisch]

[Toumal]

Individual incidents should not be taken as definite proof for anything. However, one should keep in mind the number of incidents with XPS/IFS, and the number of failures involving FASST. I'm going out on a limb here, but I'd say it's a safe bet that even if you count every instance where the exact failure reason cannot be determined, there's far less incidents with FASST than XPS.

As spectacular as it may look, these images alone don't prove an XPS failure. Oh I myself do believe that this was the case, but as far as concrete, legally-sound proof is concerned, no.

What we need is a guy/gal losing his/her plane due to XPS lockout, with pre-crash installation pictures and a flight logger logging all voltages and servo movements. And possibly a 2.4GHz spectrum analyzer log showing that XPS was shot down mid-flight by interference. And a sworn statement.

(And even THEN you will find people who point out that all this doesn't mean anything because, well, they didn't have problems with XPS.)



For me, the glaring design deficiencies are much easier to prove. I dare JD or Graupner sue me for stating that their frequency hopping solution is inferior to true FHSS, and that their single-antenna solution is inferior to antenna diversity. They know damn well that they'd lose, because every single engineer with a bit of an idea about these things will tell them the same thing. I would, however, be very careful with statements that this or that crash was caused by XPS. Unless you have definite proof of that, stating such things is dangerous (and potentially wrong too!)

[jonkoppisch]

This picture was taken from this post... It's a result from the pylon races..

	Quote: Originally Posted by BZFrank
	LOL! BTW and FYI It seems IFS/XPS is now banned from the Speedcup plyon/speed competition and very likely more competitions this year, due to a series of crashes / flyaways (lockouts and failure to activate failsafe) here in Germany. It is likely other contests here will follow these rules too until the situation is clarified. http://translate.google.de/translate...hl=de&ie=UTF-8