Independent tests prove lack of frequency hopping with XPS

[klhoard]

	Quote: Originally Posted by KrisW
	Okay. . give me one. . important. . logical. . all encompassing reason why I should NOT fly 72mhz. . and SHOULD fly 2.4 instead. <snip> Again .. ONE reason to switch.

.
.
Kris,

72Mhz is still working great for me too. I hate to see people still putting XPS in the same league as Futaba and JR, it's not.
.
.
If 72Mhz is working for you, then indeed stick with it. However, when you decide to consider making the jump, I wouldn't recommend basing your decision on the track record of XPS.
.
.

[sweetpea]

	Quote: Originally Posted by jonkoppisch
	Good question about the 13th transmitter. Also, Does the transmitter broadcast any kind of signal before it hears from the receiver? Ie, you turn on the transmitter (not broadcasting), turn on the receiver (broadcasts signal to trans), establishes link then the trans starts to broadcast 2.4 signal? I assume that the receiver is broadcasting a 2.4 signal so.. Another question. If 12 people are flying (the limit and fills the frequencies) and 3 more turn on, will the broadcasting receivers trying to link with the transmitters cause any interference as the channels are full?? According to posted specs though, 120 people can fly at the same time....

I can't remember why you can have 120 but there was a reason stated way back. If it is truly only 12......that does raise some questions at large events for the lucky #13.

From what I recall on several posts and talks with XPS.....the module does not transmit until it hears from the RX. The RX is the brains of the unit

	Quote: Originally Posted by klhoard
	. . Kris, 72Mhz is still working great for me too. I hate to see people still putting XPS in the same league as Futaba and JR, it's not. . . If 72Mhz is working for you, then indeed stick with it. However, when you decide to consider making the jump, I wouldn't recommend basing your decision on the track record of XPS. . .

Correction.......JR has yet to produce a 2.4 unit.........They use Spektrum 2.4 in their radios and Spektrum uses JR radios for their casings. But that is splitting hairs really.

I'm also with you......if your scared, unsure or anything else......stick with 72mhz.....you know the risks there. They have not changed in years.

2.4 is full of new risks that cause the same problem as 72mhz.....crashes. I would say you have to look at all the brands across the board for 2.4........(XPS. Spektrum, Futaba, ASSAN) then once you decide 2.4 is for you, look at specific brands that meet your needs/wants/price.

[jonkoppisch]

	Quote: Originally Posted by jonkoppisch
	Another question. If 12 people are flying (the limit and fills the frequencies) and 3 more turn on, will the broadcasting receivers trying to link with the transmitters cause any interference as the channels are full??

The more I think about this the more curious I am. All 12 channels are full, 3 more people decide to turn on. The receiver has to broadcast on 2.4 to establish the connection with the transmitter?? Something has to broadcast to something!! So now there are 3 more 2.4 sources at close range broadcasting looking for a connection. What frequency do they first broadcast on for the connection? etc.... Could they swamp 3 connected receivers?

Wait. Duh, it finally occured to me that the receiver will (or should) scan the frequencies first. If it sees that they are full then it should simply not broadcast for the connection? It would be interesting to have an oscilloscope (& know what the heck that I'm doing) and check this out...

[XJet]

	Quote: Originally Posted by jonkoppisch
	The more I think about this the more curious I am. All 12 channels are full, 3 more people decide to turn on.

This is one of the cool things about 2.4GHz, you *can* have multiple transmitters and receivers on the same frequency without problems.

Your standard 72MHz system is transmitting constantly -- that's because it has to work with a very narrow bandwidth and thus it needs to send data continuously to get enough information to the receiver at a rate that will allow smooth control.

A system such as XPS or Spektrum has much more bandwidth available to it (several megahertz rather than just 10 kilohertz) which means it can send the necessary data in a fraction the time. Think of it like a thin straw and a thick straw. If you have to suck up a gallon of liquid an hour, you'll have to suck continuously on the thin straw but with a thick straw you can drink the lot in just a few minutes then put your feet up and rest for a while.

Since the 2.4GHz system can send a whole frame of data in about 1/10th the time of your 72MHz system, it can transmit for just 10% of the time -- simply sitting idle for the remainder. This is called the duty-cycle (the percentage of available time that is used in every second)

During this other 90%, other transmitters (on the same frequency) can send their data.

Now all the receivers on the same frequency will receive all transmissions on that frequency but (and here's where "binding comes in") they will ignore the data that doesn't come from the transmitter they're bound to.

The receiver sits there waiting for the next lot of data to arrive. When a packet arrives it looks at it and either says "that's for me, and passes it onto the servos -- or says "not mine" and totally ignores it.

Since XPS appears to have a duty-cycle of about 10%, it means that in theory, you could have up to 10 transmitters all working simultaneously on the same channel. With a total of 12 channels available, that's how JD comes up with his "120 simultaneous users" number.

So what stops two transmitters from sending out their data at the same time and confusing the receivers?

Simple, the XBeePro modules always listen for other transmissions before they begin to broadcast their own signal. If a transmitter finds that the frequency is in use, it goes to sleep for a very short (random) time and then tries again.

This is called Carrier Sense Multiple Access (CSMA) and is a very common way for computer networks to all share a single cable on a large network. It's tried and tested -- but not 100% efficient.

What tends to happen is that as you get more and more transmitters vying for that frequency, they will each increasingly find that it's busy when they go to transmit. As a result, they go back to sleep for a moment and try again later.

The more transmitters are working on a single frequency, the more sleeping that is likely to occur between the time a transmitter is ready to send and the time it finds the channel unused by another.

It is not possible to get 100% utilization of the available frequency because, the more transmitters are contesting for the available time, the less available time there will be so the less likely the are to find a "quiet" period.

In the case of XPS, this situation is worsened a little because not only are all those transmitters trying to send data to their receivers, but those receivers are also trying to send an acknowledgment back to the transmitter (consuming more valuable time).

However, it is *true*, you can run 10 XPS systems on each channel, however there will be significant latency in the response of the system when you start approaching this theoretical maximum. Your servos will lag behind your control inputs and this could make flying very difficult.

I strongly suspect that this is a non-issue however, since I doubt many people would even dare take to the air if there were another 119 models flying :-)

Of course if I were just out to "knock" XPS (which despite the claims of some, I'm not) I'd come up with a reason why this is bad, but it's not -- it's pretty standard stuff for networking and 2.4GHz-based data links. It's nice, it degrades slowly under increasing load rather than failing without warning.

There's no problem here.

[BoneDoc]

XJet,

So what did you do on your testing? So far, you've been telling us of what XPS does NOT do, but not HOW. Curious minds want to know and get to the facts. If you have the time to make counterpoint arguments, surely you'll have the time to at least write about your methods.

I'm no fanboy as I do not own XPS. I have Spektrum right now, and I have issues as well from time to time (2 receivers died).

Please publish the facts.

[jonkoppisch]

Interesting. Thanks! I take it that the xbee modules are from wireless routers and such? What is the range of a typical 2.4 wireless router? Does it have the same strength as the xbee modules being used for rc? If so, with everyone getting wireless routers for their home networks, say 3 or 4 in the area, could this cause any interference? As I understand it, ground clutter would weaken or thin a lot of the signals but as you say, once you're in the air?

Added: Does a wireless router/network etc use more time on the frequency also? So instead of only using 10% it might use 30% or 70%??

[XJet]

	Quote: Originally Posted by BoneDoc
	XJet, So what did you do on your testing? So far, you've been telling us of what XPS does NOT do, but not HOW. Curious minds want to know and get to the facts. If you have the time to make counterpoint arguments, surely you'll have the time to at least write about your methods. I'm no fanboy as I do not own XPS. I have Spektrum right now, and I have issues as well from time to time (2 receivers died). Please publish the facts.

Because it becomes very easy to start lapsing into rather technical jargon (which benefits nobody except other tech-heads), I intend to make a video that will demonstrate the XPS frequency-hopping situation in a way that is easy to see.

I was first alerted to the problem when we were experimenting with 2.4GHz video-senders on a UAV wth an XBeePro -based telemetry system onboard.

If I turned the video sender on first, the XBeePro modules would work just fine, giving all the required range and no data-loss.

However, if I turned the XbeePro link on first, then turning on the video system would cause massive data-loss (over 99%). I realised this was because the XBee system is not frequency hopping and has no intrinsic ability to change frequencies when encountering interference.

I decided to see if the XPS modules had this frequency-change firmware so I repeated the experiment with an XPS receiver onboard.

As before, turning on the video system first produced a working link with acceptable range. The XPS worked fine.

However, when I turned on the XPS system first and checked it was working fine, it all stopped as soon as I turned on the video-sender.

A quick check showed that the video sender was operating on the same channel that XPS was choosing as its default at startup and that when exposed to this kind of sudden increase in noise, it would simply stop working.

This wasn't front-ened overloading or anything else , it was simply a case of the video-sender's transmission clashing with the XPS transmitter.

I then performed some other tests to see how far the video sender and XPS system had to be
separated before this didn't happen -- and found that even with 100 yards separation, the problem persisted.

The possibility that this XPS receiver was faulty had not been discounted but I now have another in my possession (thanks to a kind FG member) and I'll re-test with both units.

If I had a lot more time at my disposal, I'd set up a spectrum analyzer and invest a few days on this but I have to earn a crust too and we're on a tight timeline with an upcoming UAV tender so I'm already burning the candle at both ends.

However, I expect to produce something that will allow folks to see with their own eyes what's going on. If I get the chance I'll also upload some of the noise profiles from our own airfield which show the quite marked difference in 2.4GHz RF activity at ground-level versus altitude.

[bgold]

I should be patient

[XJet]

	Quote: Originally Posted by jonkoppisch
	Interesting. Thanks! I take it that the xbee modules are from wireless routers and such? What is the range of a typical 2.4 wireless router? Does it have the same strength as the xbee modules being used for rc? If so, with everyone getting wireless routers for their home networks, say 3 or 4 in the area, could this cause any interference? As I understand it, ground clutter would weaken or thin a lot of the signals but as you say, once you're in the air?

The XBeePro modules are designed primarily for the creation of data links and networks. The "range" of an XBeePro pair is "up to" a mile on the ground and more in the air -- that's not really an issue. Even with 10mW output (versus the 60mW standard output of XPS) you'll get far more range than you'll ever need on 2.4GHz.

Wireless internet *can* be an interference issue but most home routers are relatively low-powered and, because they're usually operated indoors, tend to emit only small amounts of RF beyond a hundred yards or so.

What can be more of an issue are things like powerful cordless phones (legal and illegal) operating in someone's back yard and thus clearly visible to a model and broadcast wireless internet. We have a commercial wireless internet service in the town where we test our UAVs and it has a profile on three of the XPS's 12 channels. It has three large radiation lobes, one of which comes very close to the airfield itself and you can observe the noise-levels rise significantly when you fly through it.

Even a BlueTooth mobile phone could potentially cause problems if it's activated in your pocket (or that of someone near you) while on the flightline. If it transmits on the frequency XPS is using, it could in theory obliterate the acknowledgment data sent back from the XPS receiver to the transmitter. If XPS doesn't get these acknowledgments it resends the data several times because it thinks the receiver never got it. This chews up more of the timeline. I don't know what happens if the timeline is almost full but I expect it *would* impact on other XPS systems also using that channel.

Other potentially worrying sources of interference are leaky microwave ovens and commercial 2.4GHz data-links -- or perhaps even a plane on 72MHz that's flying with a 600mW video transmitter onboard and is operating from the same field as you. I have a video sender here that completely obliterates the XPS system at significant ranges (I'll be testing this further in the New Year).

If you've got a 2.4GHz cordless phone or home network you've probably already seen just how much unwanted interference can occur between 2.4GHz devices. It's not uncommon for even a perfectly good microwave oven to kill the throughput on a home wireless network and likewise with the cordless phone.

If you've got diversity antennas or satellite receivers on your 2.4GHz gear then the effects of these interfering signals is likely to be much less -- since one of your antennas should be getting a strong signal from your transmitter at all times. With a single quarter-wave whip however, the chances of the valid signal being overwhelmed by an interfering one increases due to multi-pathing, polarization conflicts and dropouts caused by antenna null-points.

	Quote:
	Added: Does a wireless router/network etc use more time on the frequency also? So instead of only using 10% it might use 30% or 70%??

This depends on how much data is being transferred. If the network has (say) a 2Mbps maximum throughput and someone is sucking down a large file at full bore then that cbannel will be very highly utilized with very little space left on the timeline for anything else. That's one of the reasons that wireless internet connections tend to be so variable in their performance - as more people start using them they slow down a lot as the timeline becomes filled.

[aviti]

Some guys like to question X-jet because they think he has an axe to grind with Jim Drew but right or wrong, I'm glad someone that knows more than most of us is questioning this. There are others in the industry that know what's what with this stuff but they can't really comment on a competitor's system. It would be nice if we had a lot of non-biased experts like our own RC consumer reports but we just don't. I hope some other folks with the qualifications will also look at this.

[KrisW]

	Quote: Originally Posted by Kiwi
	Something I find a bit unnerving here is I remember clearly in the early days of the Web spamming on the benifits of XPS was the fantastic amount of channels available to this setup. I seem to recall it being stated that you could technically fly hundreds of models at the same time and get no interference between them. I am neither for nor against any manufacturers attempts to provide a reliable and robust RC system but I do find 12 channels a seriously small number of available frequencies. I know someone is going to say yeah but!!! You never have 12 planes flying at the same time but by god you can easily have more than 12 transmitters turned on in the pits if everyone thinks they are never ever going to shoot someone down or cause a problem. What happens when transmitter No 13 gets turned on.?? Does it just fail to link up or does it squeeze someone out with a weaker transmitter signal.?? Again I am not taking sides, nor do I want to open pandoras box but seeing as we have someone with a seriously indepth knowledge of 2.4 Ghz radio systems its an opportune time to ask I think??

Actually, the total numbr of frequencies would be about 6, to have totally clutter free, uninterrupted signal on BOTH channels in a pair. . . NOT GOOD.

Now, since an identifier string is encrypted, theoretically you can still have 78 pairs of frequency, with each half of the pair being stepped on by conflicting signals, but not matching the other half's PCM encryption, and so being ignored. . in THEORY....... and in theory keep flying . . ..

Theory often conflicts with reality .. and reality always wins.