Being dragged screaming and kicking towards 2.4!

[Frank Skilbeck]

Simon you quote "As DSM2 does not frequency hop, it is more susceptible to signal fading and blocking. Hence why you need satellite receivers and antenna installation is more critical"

Why is frequency hopping less susceptible to signal fading and blocking?

Not all Spektrum Rx's use satellites, the AR600 for instance just has two aerials, plus I've found the smaller 6110 series with just small aeriels very good.

As a user of FrSky, M-Link and DSM2 I must admit I've yet to experience problems with any of the systems.

[Simon Chambers]

Posted by Frank Skilbeck on 21/08/2012 22:15:37:

Simon you quote "As DSM2 does not frequency hop, it is more susceptible to signal fading and blocking. Hence why you need satellite receivers and antenna installation is more critical"

Why is frequency hopping less susceptible to signal fading and blocking?

Not all Spektrum Rx's use satellites, the AR600 for instance just has two aerials, plus I've found the smaller 6110 series with just small aeriels very good.

As a user of FrSky, M-Link and DSM2 I must admit I've yet to experience problems with any of the systems.

Yup and notice that the AR600 is a new receiver that supports DSMX. Likewise the AR400 is new receiver that supports DSMX but no satellite receivers. Spektrum are finally stopping peddling satellite receivers as something to fix all ills. Also both have two antenna's that need to be positioned correctly (90 degrees apart) and away from each other. I would imagine Spektrum are probably assuming most users are now on DSMX anyway which is frequency hopping.

Spektrum have traditionally specified a satellite receiver system is full range, and otherwise its parkflyer. Hence the 6110 is parkflyer range.

Before explaining the benefit of frequency hopping, a quick, brief explanation on signal fading. RF waves, like sound waves bounce off all objects. Signal Fading is when two of these waves (of which have bounced off two different surfaces, thus like an echo you get in sound waves), one the inverse of the other, cancel each other out. You can do the same thing with sound waves - often done school physics experiment.

Frequency hopping, if you get two waves that cancel each other out, in a couple of milliseconds time, the frequency changes. When the frequency changes, the wavelength changes, hence the signal will no longer cancel out.

Signal blocking is when a RF absorbent/reflective object prevents the signal getting to the receiver. Now to combat this, if you have two antenna's in separate polarisations (i.e. 90 degrees apart), you can not only receive signal that are polarised differently (e.g. bounced off a couple of surfaces which changes its polarisation), but also as the antenna is (or should be) further apart than the other, it is less likely to be blocked too.

Satellite receivers are a solution to a problem that can also be overcome with frequency hopping.

Cheers,

Si.

[Peter Christy]

Sorry Simon, I don't know where you got your information from, but a lot of it is just plain wrong!

DSM-2 DOES frequency hop - but only between two channels (selected at random). IF by chance the system selects two adjacent channels - which I've never seen mine do - then a broadband analogue video transmitter could cause it problems. BUT the very nature of the spreading technique used mostly enables the system to overcome this. A DSM-2 system can recover a signal from BELOW the noise floor - something the FrSky and similar systems cannot.

The Spektrum system was originally fully frequency hopping - so the theory about the chips not being able to manage it just doesn't hold water. It was switched to a direct sequence system for production because this provides a measurable improvement in the path budget compared to full frequency hopping systems. Now you might argue that a 3dB advantage in a path budget of around 130dB is minimal, but it still represents a doubling of the system's immunity.

The satellite receivers are not essential, but do provide an improvement in the diversity of the reception. I have a thermal soarer which uses an early DSM-2 receiver without a satellite, and I've never experienced any range or interference issues with it. But signals at 2.4 GHz are much more susceptible to physical blocking and aerial orientation issues than at lower frequencies, so anything that can be done to minimise these problems is going to improve reception reliability. I know about this only too well, having run into similar issues on 459MHz in the past! Full frequency hopping is no substitute for the added diversity provided by a satellite receiver, though since both systems are exceptionally robust, there are economic reasons for not providing satellite receivers. But these are *economic* reasons, not necessarily technical ones!

The bandwidth of a FrSky system is about 100 KHz. The bandwidth of a Spektrum is about 2 MHz. I know. I've measured both!

JR DSX-9 at 1 MHz per cm (DSM-2 system):

FrSky at 1 MHz per cm:

(Wish there was a preview button so I could check whether these links work or not before posting!!)

As you can see, the Spektrum (as indeed does FASST) occupies a substantially larger bandwidth than FrSky, and it is this, along with the subsequent de-spreading of the signal, that provides its high immunity. (And you have no idea how difficult it was to capture those images!!!)

I currently run both DSM-2 and FrSky systems, and have complete faith in both. My son currently runs Futaba FASST, and that has also proven to be a robust system.

But please lets end this nonsense about DSM-2 not hopping. It DOES, and I can see it quite clearly on my spectrum analyser. It just doesn't use as many channels as other systems, but for very good reasons of its own!

Cheers,

--

Pete

 

 

 

Edited By Peter Christy on 22/08/2012 00:11:03

Edited By Peter Christy on 22/08/2012 00:14:51

Edited By Peter Christy on 22/08/2012 00:15:49

Edited By Peter Christy on 22/08/2012 00:20:55

[bouncebounce crunch]

if thats your granny's heartbeat she must be cranky.

[Mark Powell 2]

Simon (Chambers),

Misinfomation.

'Low power and short range??!!'. I said that, it is not misinformation at all. Most of it is short range, a range of 100 yards or less is common in its major uses, WiFi , medical, and the like. And it is limited to 10mw or something (not sure what that limit is - re-read, I accept your 100 mw). My ancient, long thrown away, 27 MHz Kraft system was one watt. So were others. We 'get away' with an apparent long range because of techniques such as dual aerials, multiple receivers, high sensitivity, and so on. But 2.4 transmissions are essentially 'local'. Except for the illegal ones of course.

Fasst?

A quote from a RipMax advert. 'Both the Acro Wot Foam-E FTR aand RTF versions can be operated with a Futaba S-FHSS transmitter AND ALL FUTURE FUTABA TRANSMITTERS' (my emphasis). So they are all going to incorporate S-FHSS,  both low  and high end, obviously. Does not bode well for Fasst, does it? And future S-FHSS can incorporate telemetry simly enough. So despite Futaba's marketing blurb I an suspicious. I suspect RipMax possibly letting the game away was accidental.

 

 

 

 

 

Edited By Mark Powell 2 on 22/08/2012 05:38:20

[bouncebounce crunch]

lets get ready to rrrrrumbLLLLe.

[Peter Christy]

Posted by Mark Powell 2 on 22/08/2012 05:08:48:

My ancient, long thrown away, 27 MHz Kraft system was one watt. So were others.

True, but it would only have been outputting around 100mW erp (Effective Radiated Power). At 27 MHz - and indeed 35 MHz - a hand-held box with a very short (electronically, not physically) aerial is not a recipe for efficiency! I designed Mick Wilshere's Talisman transmitters, and when we submitted them for testing we were also putting in around 1 watt DC - but the measured output was just under 100mW erp!

Conversely, at 2.4 GHz, where the wavelength is only a few centimeters, and hand held box can make a reasonable ground-plane, and the aerials can actually have gain instead of loss! All the standard 2.4 GHz aerials have a gain of 2dB. The transmitters have an output of around 60 mW, which when boosted by the (slightly directional) aerial gives 100 mW erp.

In a similar fashion, the very nature of a direct sequence spread spectrum signal leads to a degree of "coding gain". And the more you spread the signal, the more gain you get from this. So in theory, the DSM signal illustrated in my earlier post will enjoy a much higher coding gain than the FrSky one, as it is employing over 10 times the bandwidth!

One thing I should have added was a screen-shot of a FASST signal for comparison:

Again, a desperately tricky thing to grab, but you can clearly see that the bandwidth is even greater than DSM-2, (pic is again at 1MHz per cm) which will help in the absence of a satellite receiver, as of course will the extended aerials used by non-satellite receivers.

It wil be interesting to have a look at the new Futaba and JR systems when I can manage to get my hands on them for comparison!

--

Pete

[Mark Powell 2]

Peter,

I remember the Talisman. A very small, neat transmitter. Although I know the basis of how these things work I am not qualified to judge the electronics in detail. I did not think that 27/35 aerials were very efficient, and in some ways it is reassuring to know that the ERP of the 2.4 transmitters is much the same. Didn't know the Fasst system was fairly similar to the Spektrum. Had a quick search a while ago to see but didn't really find anything. Futaba seemed a touch reticent even in their 'technical notes'.

Thanks for your clarifications.

[Peter Christy]

Mark,

Most of these systems are "hybrid" systems, being neither pure hopping nor pure direct sequence. The Spektrum and Futaba probably have the most in common, using a very wide "spread" along with hopping, though the Futaba uses more hopping than even DSM-X.

One final point, and then I'll shut up for a bit

The Spektrum DSM system was the first practical system to use 2.4 GHz. As such, it was somewhat over-engineered - as indeed were the early digital systems back in the 60's! Back then, most proportional systems came with a fail-safe, but as the systems got refined, the over-engineering was slowly removed, leading to smaller, lighter and cheaper systems. I suspect the same is happening on 2.4 GHz, where as manufacturers become more comfortable with its properties, elements that are seen as no longer necessary will be removed. Satellite receivers may well be such a case of unnecessary over-engineering, but its a little early to pass judgement just yet!

Spektrum caught all the Japanese manufacturers with their pants around their ankles, and if it wasn't for Paul Beard's efforts. we would still be dependent entirely on 35 MHz. The system he devised has proven to be robust and reliable and if, as a result of competition, the systems get simplified and drop in price, then we will all benefit. But just because a system is updated and modified doesn't necessarily imply there was anything wrong with the original. It may just mean it was initially over-engineered - and that is not a bad thing!

--

Pete

[Simon Chambers]

Firstly, I really don't want this to end up as another DSM2 vs other systems debate. There are enough of these around as it is.

> DSM-2 DOES frequency hop - but only between two channels (selected at random). IF by chance the system selects two adjacent channels - which I've never seen mine do - then a broadband analogue video transmitter could cause it problems. BUT the very nature of the spreading technique used mostly enables the system to overcome this. A DSM-2 system can recover a signal from BELOW the noise floor - something the FrSky and similar systems cannot.

Ok, ok, changing between 2 channels is just about frequency hopping - however I would call that a very loose definition of frequency hopping! In fact in the documents accepted and published on the FCC OET site, regards DSM2 as a DSSS system - not a hybrid FHSS/DSSS or a FHSS, just a DSSS.

So yes DSSS, compared to say GFSK, will have better sensitivity when a given section of spectrum is congested (especially when proven in the lab enviroment!). However thats not of the point of frequency hopping in the real world. By hopping channels quickly enough, you can avoid those areas of interference completely. Given modern 2.4GHz beasts such as 802.11b/g (22MHz b/w) and 802.11n (40MHz b/w) taking up huge chunks of the spectrum, a couple of them in a nearby local area could swamp DSM2 if sitting on nearby frequencies. By hopping you can transmit between this interference when others transmitters are using that bandwidth and then use this bandwidth during periods its not.

Even though DSM2 will pick 2 clean channels (not always far apart as many others have shown), if it picks two channels that overlap with, say a couple of 802.11n AP, the airborne receiver could end up losing packets when flying towards the source of interference, due to not enough transmitter signal propagating through the interference. This is achievable if the transmitter is turned on, while for example still in a protective metal case, on the floor, etc and it has performed its initial and only clear channel assessment. So it could pick two channels that, once moved away from the shielded area, are actually not optimal.

Now for the most part, in R/C use down a field or a slope, 2.4GHz interference is far enough away not to be an issue. However down at the local park with a few nearby large housing estates, this becomes much more of an issue. Especially when the airborne receiver moves away from the transmitter and nearer to the sources of interference (i.e. AP in houses).

I'm certainly not saying that DSM2 isn't suitable for R/C use (even though some individuals do, I certainly do not), in the same regards 35MHz is perfectly suitable for R/C use too. However in the real world R/C use, pure DSSS is really not superior when there are localised sources of interference.

Now of course you could say all this is baloney. However I'll pose a question. If DSSS is so superior than a FHSS system to receiving through interference - why does both Bluetooth and ANT+ use GFSK modulation with frequency hopping, if DSSS is superior?

Even Zigbee, which is DSSS, has a Frequency Agility system built into its protocol when the channels it is communicating on have become too congested.

The groups that produced those specifications certainly wouldn't have chosen their modulation formats lightly!

I would also point out, that its not just me who believes this, the R/C manufacturers do too. None of the fairly recent new airborne R/C systems out there use pure DSSS. If they do use DSSS (such as JR's DMSS), they incorporate frequency hopping into it to make it a hybrid system.

Incidentally Futaba in their new S-FHSS system has their RF chipset configured as 2-FSK modulation with a bandwidth of 232KHz. No DSSS involved.

[Simon Chambers]

> The Spektrum system was originally fully frequency hopping - so the theory about the chips not being able to manage it just doesn't hold water. It was switched to a direct sequence system for production because this provides a measurable improvement in the path budget compared to full frequency hopping systems. Now you might argue that a 3dB advantage in a path budget of around 130dB is minimal, but it still represents a doubling of the system's immunity.

I understood, from speaking to a former spek engineer, that to take advantage of benefits of DSSS and full frequency hopping, it would require the different channels to have variations of the PN codes to prevent interference with other DSM transmitters. However this would require a large NV storage area and due to the speed of reading NV memory, a large shadow cache area for it. Pushing up the memory requirements of the uC (and changing chosen uC at the time) and thus cost. Of course this is only what I was told by him and I don't have another source to verify it against.

>The satellite receivers are not essential, but do provide an improvement in the diversity of the reception. I have a thermal soarer which uses an early DSM-2 receiver without a satellite, and I've never experienced any range or interference issues with it. But signals at 2.4 GHz are much more susceptible to physical blocking and aerial orientation issues than at lower frequencies, so anything that can be done to minimise these problems is going to improve reception reliability. I know about this only too well, having run into similar issues on 459MHz in the past! Full frequency hopping is no substitute for the added diversity provided by a satellite receiver, though since both systems are exceptionally robust, there are economic reasons for not providing satellite receivers. But these are *economic* reasons, not necessarily technical ones!

I do not see the benefit of having an additional receiving circuit, when the same results can be achieved with a RF switch and dual antennas. If needs be, the data can be sent twice on a channel, with the receiving end switching between its alternative antenna's based on a fixed time cycle - as Hitec does it. Futaba S-FHSS switches antenna's on channel hops - but has a similar packet transmission rate and more channel hops than Hitec. It’s not as if R/C protocols transmit a lot of data - most only send a handful of bytes per packet taking very short period. There is plenty of time to transmit repeat packets between the 18-25ms servo update rate. Indeed, from examining a number of R/C systems, this is what they do.

> As you can see, the Spektrum (as indeed does FASST) occupies a substantially larger bandwidth than FrSky, and it is this, along with the subsequent de-spreading of the signal, that provides its high immunity.

Yup I agree, on a given portion of the frequency band. However as above, by hopping around sources of interference, can be just as effective.

> (And you have no idea how difficult it was to capture those images!!!)

I don't doubt it! It’s been a long time since I've had to take screenshots on a CRT scope/analyser/etc. I've unfortunately no longer have access to a proper Spectrum & Network Analyser anymore (well not to justify borrow one to use for examining R/C protocols!) after ditching my job with a well-equipped lab to work for a startup.

Cheers,

Si.

[Simon Chambers]

Posted by Peter Christy on 22/08/2012 13:25:26:

The Spektrum DSM system was the first practical system to use 2.4 GHz. As such, it was somewhat over-engineered - as indeed were the early digital systems back in the 60's! Back then, most proportional systems came with a fail-safe, but as the systems got refined, the over-engineering was slowly removed, leading to smaller, lighter and cheaper systems. I suspect the same is happening on 2.4 GHz, where as manufacturers become more comfortable with its properties, elements that are seen as no longer necessary will be removed. Satellite receivers may well be such a case of unnecessary over-engineering, but its a little early to pass judgement just yet!

I certainly agree that DSM was a brilliant system, and came out at a time when 2.4G band hadn't be thought about for R/C use. Also the 2.4G spectrum back then was a lot less congested than it is on. Designing a system from a blank sheet with nothing to base on is certainly not easy.

Satellite receivers are a case of over-engineering - nothing wrong with that, but I thoroughly dislike Spektrum's marketing to say that Satellite receivers are superior. For all intents and purposes a switched dual antenna setup is functionally identical.

If you look at any modern system, out of all the chipsets they can use, they very few use one with DSSS capability now. Even if they do, they don't rely solely on DSSS to provide a interference free, strong link, they also frequency hop.

I still belive adamantly that FHSS is superior to pure DSSS in R/C use for a stronger link with coexistence with other 2.4GHz transmitters and other 2.4GHz traffic/interference/noise.

However like I've already said, down the flying field or slope site, with very little 2.4Ghz traffic its a bit of a mute point. It's not as if DSM2 planes are dropping out of the sky all the time!

Of course there is the whole other subject (which again, really has been done to death) of the relatively high brownout voltage in Spektrum receivers that has been reckoned have caused the majority of unexplained accidents - due to inadequate power systems, etc, etc, etc.......

Cheers,

Si.

[Mark Powell 2]

Simon, I am learning a lot. Just for fun, at this very moment, this box I am typing on is transmitting, via 802.11n a very large file to another box, wirelessly. My model plane is close to the boxes, on the floor, to give it the worst condition possible, and is satisfactorily responding to the 'servo monitor' signal being tranmitted from the bottom of my large garden with the transmitter in short range 'test' mode, again, down on the grass, with its aerial deliberately 'worst condition' pointing near enough directly at the plane

Why's it working fine?

[FunnyFlyer]

Great thread guys - the Forum at its best!

[Simon Chambers]

A simple answer, one 802.11n AP doesn't completely swamp the 2.4G band. Whatever R/C system your using has probably either chosen a part of the band away from where the AP is transmitting on or is channel hopping around it. Also unless your running a mixed 802.11b/g network, it could quite possibly be transmitting in the 5GHz band.

If you had a couple of 802.11n in 2.4GHz mode set to 40MHz bandwidth, you're likely to see different results.

Cheers,

Si.

[Mark Powell 2]

Simon,

Thank you.

Funny Flyer,

Have a look at the 'Ducted Fan in Theory and Practice' thread if you have not already done so. Amusing thing about that one, I looked at a couple of professional jet airliner pilots forums and some of the threads are asking the same questions, space rockets and all, and getting the same assorted answers and disagreements. The 1850s theorist, Froude, is currently in danger on ours..

[Peter Christy]

Simon,

<Sigh!> I said in my last post that I'd shut up for a bit, and I certainly don't want to get involved in a point by point refutation of your posts - some of which I agree with, by the way! But I do feel that a couple of the issues you have raised need refutation!

Firstly, the point about a wideband WiFi (or similar) signal completely blocking a DSM signal. This completely ignores the effect that de-spreading the wanted signal - and consequently spreading the unwanted one - has. See my analogy of Excalibur rising from the lake! The only thing that I'm aware of that could cause a DSM link to fail is a high power analogue video signal. Anything that uses a spread spectrum technique of any kind may cause some degree of packet loss, with a consequent slowing of the system, but should not cause a total blockage of the signal. That is assuming it is designed in compliance with the regulations, because in layman's terms, that how the regs say they MUST operate!

Moving on to the satellite receiver issue. I hinted earlier that I had come across this issue on 459 MHz. When I initially designed my 459 MHz gear, one of the biggets problems I came across was getting it to work in a helicopter. It works fine in my old "Lark" (mostly wood), but getting it to work in Shuttle (mostly plastic) proved a real challenge. I did eventually get it to work satisfactorily in a Shuttle, with careful aerial routing, but it took a lot of trial and error.

The problem was simply that there was so much metalwork around that it was very difficult to ensure a "line-of-sight" path between transmitter and receiver under all orientations.

Of course at 2.4 GHz - with an even shorter wavelength - the problem is exacerbated. Clever modulation methods, as employed by 2.4 GHz gear, can do much to overcome reflection and fade, but at the end of the day, if the signal doesn't get through, then its game over!

The problem is even worse in jets. Again, you have a large area of metal that can easily get between transmitter and receiver, casting a shadow over the receiver, with the added safety concerns of weight and speed.

With the standard - even extended - aerials used on single receivers, it is extremely difficult to make an arrangement where it is not possible to "screen" the receiver from the transmitter.

But, with a tiny satellite receiver, you can put the main receiver in the fuselage, a satellite in the wing - or even both wings - and maybe even one in the tail. That way there is ALWAYS one receiver that can "see" the transmitter, regardless of orientation!

And don't believe that rubbish about 2.4 GHz being immune to screening. It is no more or less susceptible than any other system. I well remember an early adopter of a full frequnecy hopping system putting it in a carbon-fibre hulled speedboat, having been given assurances by the manufacturers representatives that it would work fine. It did - until he put the carbon-fibre hatch in place, at which point the system simply locked out instantly!

All my helicopters (except the Lark!) use satellite receivers, and will continue to do so for the time being - because I KNOW that it works!

--

Pete

[Simon Chaddock]

Just to get back to my cheap and nasty Skyartec 706!

Whilst it has no 'on board' model memory as it can only be programmed from its PC programme any number of individual configurations stored and called up as required but you have to have a lap top at the field to make any adjustments!

It is obviously intended for their own RTF planes as it has distinct settings - standard, delta wing and 3 helicopter modes (CCPM-, CCPM+ & Flybarless).

Channels 5,6 & 7 (marked gain, pit & gear) are only available in the helicopter modes so as yet are untested.

Nevertheless in 'standard' mode (4 channel) it does work as it should with adjustable (but only via a PC!) dual rate and exponential features. As you might expect it is free from glitches and has quite a range as well.

It does have one surprise up its sleeve though! In the rather brief manual it says 'gear - flame out switch' but nothing else.

It took me a while to discover what it meant but it is exactly that - in Standard and Delta Wing mode the TX 'gear' switch cuts the motor power but leaves everything else on!

I have added to this post simply in case somebody feels tempted to do what I did and buy a cheap 2.4 radio. It will probably work fine but don't expect too much!

[Codename-John]

Posted by Simon Chaddock on 24/09/2012 11:19:08:

It does have one surprise up its sleeve though! In the rather brief manual it says 'gear - flame out switch' but nothing else.

It took me a while to discover what it meant but it is exactly that - in Standard and Delta Wing mode the TX 'gear' switch cuts the motor power but leaves everything else on!

All radio`s come with something similar, you pull it all the way back to cut power to the motor yet everything else still works the same, its called a throttle stick lol

[Simon Chaddock]

To my surprise the Skyartec 2.4 installed in my 3 channel all Depron "trainer" has so far has performed faultlessly although without meaningful instructions programming the exponential feature has proved 'interesting' but it works well enough.

I have also iscovered that the Skyartec 2.4 system is quite unique so is not compatible with anything else! Not really very surprising.