NJRCFLYER2

My Feedback: (42)

You're pretty close. During a collision, if the S/N ratio is good enough to discern the GUID, then yes, it will get interpreted properly and only the correct receiver will respond. However, if the S/N ratio is so poor as to prevent the GUID from being read, then information that was supposed to go with it is simply lost. Nothing happens. The receiver gets another chance on the next transmission.

aeajr

My Feedback: (2)

To whom is this addressed. Several people have expresed opinions on this subject.

I would point out that the process of detecting and dealing with colissions in a network are very well understood and widely used. IEEE 802.11, ethernet, is an excellent example. This is a CSMACD, Carrier Sensed Multiple Access Colision Detection network sharing scheme. The fact is that each computer reaches out to take control of the entire network, or sub network for some tiny fraction of a second. No one else can use it at that time. The point is that hundreds of computers, connected by hubs, can share an ethernet network and each one behaves like it is the only one on the network. Unless you have a few computers that is doing very heavy network work, the response time can be very good.

There were hundreds of papers written about how this can't work but for a few computers, yet today there are subnetworks with dozens of computers on them sharing the line. The point is that the methods for doing this are well known.

I have no interest in getting into an ethernet discussion, but simply point out that the idea of channel/frequency shraring may be new to the RC community, it is the standard way of working for most of the network world.

You must throw out all you know about requency sharing and frequency ownership as you learned it on 27, 50, 72 or 75 MHz. We have been operating in the dark ages while the rest of the world is in the modern age. Thank goodness we are finally getting into 1980/90s level technology of frequency sharing and colission detection and 2.4 GHz.

rmh

Thanks for the update - really.
My home computer is hooked up to fibre optics - My hometown was smart enough to vote in a fibre optics system - to everyone in town - and all I had to do was say "hook me up".
I watched the installation and the tech explained the basics to me - as I saw two tiny glass threads attached and then the noise level checked etc..
I now get phone and computer -and read info up and down as fast as any serve can throw it-- at better than a T-1 line speed.
The local phone guys who had a license to steal (QWest) went nuts -screamin "foul" . -as they have no way of reaching remotely same speeds--unless they go with fiber optics -
Just like the RC stuff - the change is dramatic.
Naysayers and Chicken Littles run around in circles --
I sold all my othe stuff- it worked just fine but the difference in what is happening with the DX 7 setup is just to big to passup -
servos which were deemed "ho hum" work extremely well with the new system .
oh yeh - my bill to the fibre optics guys is inc taxes $42.00 a month - and ALL the phone services -call forward - setup my own blacklists etc.. is just part of the package and I can do it all on the computer -
My various models on the 7000 Rx and 6100 rx work perfectly and response is unreal good
Color me " sold" on the new technologies .l

nonstoprc

My Feedback: (90)

Agreed 100%.

Also, I hope each vendor can publish the bandwitdh that their radio occupies. Because 2.4ghz can be used by Wi-Fi, WirelessUSB, wireless phones, etc, not like that our 72Mhz is dedicated to RC use, knowing the maximal operation conditions is warranted.

--qc

nonstoprc

My Feedback: (90)

I am not sure the comparison between Internet protocols with Spectrum technology is proper. Intenet transimission allows delay, while our remote control is pretty much real-time. Identified data packets (IDed) in FSSS setting guarantee that the receiver gets the correct ones, but can not guarantee that they arrive in real-time. For DSSS, the two channels reserved are all the bet. If a device is transimitting in those two channels (which should not following the FCC regulation), then interference happens.

Interference is real with either Spectrum and Futaba's 2.4ghz technologies, and that is why they have the fail-safe built into their receivers.


--qc

rmh

not comparing - just looking at new methods and old methods in "lectric stuff".

onewasp

{Quote}
Interference is real with either Spectrum and Futaba's 2.4ghz technologies, and that is why they have the fail-safe built into their receivers.
{Quote}

That is a guess on your part.
Do I know for certain?---- nope.
But then neither do you.
Logical sure------but --more logical--- is that many contests and many club fields now require the demonstration of throttle failsafe before access to the air is acceptable.

RFI-----no one in their right mind would ever say 'never'. BUT the most unlikely to have RFI of any system yet available or yet flown. Probably approaching death by lightning strike numbers---simply an expression of likelihood.

I don't EVER expect to see it.
To have it CLAIMED certainly---like failed switches and poorly maintained systems are CLAIMED as System failures.
So what else is new?

chrisF test pilot

I dont recall insulting anyone, and FYI I am a certified network admin and have been informally studying SS for a couple of years now so I am well aware of how it works. All RC SS systems fall under FCC Part 15.247, operating from 2400-2483.5Mhz. They are all sharing the same bandwidth, that is how SS works. And current RC SS systems do not employ collision detection/avoidance, so in this case isnt a good comparison to 802.11.
http://www.fcc.gov/oet/info/rules/pa...15-8-14-06.pdf

nonstoprc

My Feedback: (90)

I do not know 2.4ghz band interference is possible until this weekend, by googling on "FHSS vs DSSS" and read. Quite number of good papers and was convinced. To repeat, the band is an ISM (industrial, scientific and medicion), meaning a lot of vendors can jump onto it and design devices operating in that band. So the potential interfence sources are more than a few RC radios. Aniother critical point: FHSS can cause interruptions to DSSS and FHSS itself has the response degradation issue, under extreme situations (which is not controlled by RC radios).

So it is not a guess.

hilgert

My Feedback: (2)

Regarding 2.4GHz interference in the use of our RC SS systems please consider the following:

1) RC aircraft should be flow in a safe manner.

2) Flying over neighborhoods with houses is probably not flying in a safe manner.

3) Flying over open space, or unpopulated space, likely space with no structures, is more safe than flying over space occupied by structures and/or people.

4) Generally there are not 2.4GHz access points, cordless phones, etc., in such open spaces where safe flying is performed.

So, if one is flying safely, away from houses, offices, etc., then there are unlikely to be 2.4GHz devices that could cause interference between the pilot's TX and the aircraft RX. Not that such interference would even impact control over the aircraft mind you, but if it were to cause interference then I would propose that perhaps the aircraft is not being flown in a safe area anyway (since I don't typically see 2.4GHz WLANs and cordless phones out past the flight line at my local RC flying field).

Now, if you are flying off your local street and over housing areas, then it's "pilot beware". I do sometimes fly close-in off my cul-de-sac, and I have never had a problem.

Also, have done ground tests (see the "DX6 Actual Field Results" thread on RCGroups.com for some details on tests I did over a year ago) at thousands of feet, some directly in-line with 2.4GHz directional WLAN bridges (in other words, very high-gain directional antennas, so the maximum possible interference scenario) and found no problems. I also did some tests in my neighborhood on the ground with cordless phones on each end (TX end and RX end), as well as some 2.4GHz WLAN equipment, and still no problems.

The 2.4GHz band is SPECIFICALLY for devices to co-exist, and spread spectrum allows this to happen. I know it might seem very counter-intuitive to those familiar with existing narrow-band systems, but the RC world has been very much behind the curve in this regard. With spread spectrum it is possible to have hundreds or thousands of devices in an area co-existing, each completely oblivious to the other. In RC it's even better IF one is flying their aircraft in a safe area free of people and obstructions.

To me if one is flying safely, interference is not an issue as long as all the devices in the area are obeying the applicable rules (FCC rules in the US). Of course illegal 2.4GHz wireless video systems splattering all over the spectrum can disrupt this, but illegal devices would do the same in the narrow band world as well (especitally without the unique TX GUIDs and such).

-Hilgert

jpherit

Again, I am a little reluctant to talk about the Spektrum system or the Futaba aystem or any other because none of thes e folks have presented us with detailed descriptions of how they have implemented DSSS or FHSS. However, as you point out , the internet has a lot of good reading on the basics of these two techniques.

Before I go any farther, let us all understand that any interfering signal, or combination of interfering signals that is very powerful can and will defeat ANY wireless communiications link. However, DSSS can be designed to effectively minimize interference from a narrow band interferer as well as compliant DSSS interference. By compliant I mean observing power rules and through use of minimally interfering codes.

The key here is that the ISM band limits transmitter power and this makes the possiblity of high power interference problems very low.

If a FHSS signal of the type attributed by some to the Futaba system (ie 2ms dwell) and hops onto a DSSS channel for that brief moment, it is a narrow band interferer to the DSSS receiver for those 2 ms. This interfering signal is modulated by the DSSS receiver decoder and spread in frequency to a signal with peak strength reduced by the spreading factor while the desired spread signal is de-spread by the same factor. Now, if the data rate is ~10kHz and the spectrum is spread by the transmitter to 1 MHz, (which seems to be roughly what the Spektrum is doing) then this factor is about 100. When both signals, the desired narrowband signal, and the undesired spread interfering signal, are sent through a low pass filter then the interfering signal that makes it through the filter is 100 times weaker than the desired signal, all of whic hgets through the filter. If we assume they were about the same to begin with then the narrow band interferer is rejected. This factor is the famous "process gain" which gives DSSS its ability to reject interference.

A detailed discussion must include singal to noise ratio etc. etc; if you like you can see this development in mathematical detail in the book on spread spektrum by Dixon. To follow Dixon you will need to be familier with the Fourier transform and Correlation functions and their transforms, and some simple properties of various codes. Most engineers have studied these mathematical tools and should not find the development too difficult.

The key thing is the stipulation that transmitter power be limited. If so, then many simultaneous narrow-band nterferers can be rejected by DSSS. However, if the total narrow band power grows up towards, in our example, the factor of 100 then the demodulation will collapse and the error rate will climb quickly.

This means that if I am flying my parkie DSSS two miles away from an AMA site that has a very large number of 2.4 G Hz planes in the air at one, AND I fly my plane into this hornets nest of planes with my transmitter two miles distant, then there could be problems for me, (but not for anyone else).

In summary, a well designed DSSS can co-exist with power compliant, well designed FHSS systems except in quite unusual circumstances. There is no need for collision detection, requests to re-transmit, etc and all that stuff that works well in Ethernet networks but which would potentially introduce undesireable variable latency if used in RC.

I hope this helps a little.

NJRCFLYER2

My Feedback: (42)

The point is that collision detection involves the sending unit knowing that it happened and then re-sending the data in a timely way. In our case, that would be in time to form the next needed data frame. In the case of Spektrum and FASST, you have two different methods of sending data that has real time constraints associated with it. If the two manufacturers involved can explain how they each can successfully avoid and/or deal with collisions when they are both operating with a couple of dozen transmitters in simultaneous operation and do it within the necessary real time constraints that we have to operate within, I would be glad to know that. If they can back it up with actual testing, not anecdotal tales from a fly-in somewhere, that would be great too.

d_wheel

The method in which we use R/C seems as though we would never need to "re-send" any data. If we were sending data that would be stored and used at a later time, then resending bad packets would be necessary. However, the data being used by our aircraft is a real time event. By the time the receiver could tell the transmitter to resend a bad packet, it would no longer be needed.

Have I missed something here? Is there really a need for resending packets?

Later;

D.W.

rmh

anectdotal testing may seem like a "shade tree mechanic" approach to testing a system such as the SPEKTRUM and the FASST employ but-- (always the butt) I look at the results to date and can't find -in the SPEKTRUM , any verifiable examples of interferrence causing a crash. or servo problems etc..
I can say tho -that I have seen odd problems with the actual "binding" process. I can't pinpoint what the causes except to say it seems to be a particular 6100 rx equipped model I have - The rub is that it is an electric model --so I can't rule out the motor speed controller (BEC) as being the real culprit.
I really think I will have to isolate the rx from the ESCBEC -using a small battery pack -tho that is a pain --
The BEC if---for whatever reason-- - dips below the RX operating threshold-is bound to cause a problem and I can't tell for certain this is not what caused the problem.
The problem was that the model appeared to loose rudder control -so I landed it - and the rudder and then the elevator stayed in a hold -rudder straight and elev full up - no amount of stick moving could change this
But -(again) simply removing and replacing the battery connection to the ESC- instantly fixed it - elevator jumped back to neutral and the problem never re occurred. I flew another 10 minutes with no problem whatsoever. I di not rebind - I simply left tx stick at idle and "replugged" the ESC.
Definately not an inteferrence thing -

nonstoprc

My Feedback: (90)

Interesting.

What if there is only one fly field. One DSSS system turned on first and several FSSS ones turned on later. How many such FSSS radios does it take to completely diminish the "process gain" of the DSSS system?

carrellh

I don't think Futaba really thinks GS guys will be running in herds to buy an entry level six channel. But, the fact is it "can" control a lot of popular giant scale models (such as scale aerobats).

Only 5 channels are "required" in bobzilla's example:
1. Throttle
2. Aileron
3. Elevator
4. Rudder
5. Choke/Shutoff
You may not like the way you'd have to set it up (Y harnesses, sychronizers, etc) but it can do it and have a channel left over.

I think they really want to get a basic radio out there that is affordable so they can get some field exposure. They'll probably end up packaging this in the various RTF models. Once a person starts down the spread spectrum path, they're pretty much brand locked. When the person who buys an Avistar with Futaba SS goes for more planes, he has to buy Futaba receivers. If he decides to get an advanced radio, he's committed to Futaba because he has multiple receivers that will only work with a Futaba transmitter.

Spektrum did the same thing that it appears Futaba is doing. Brought out the DX6 to 'test the waters,' then brought out the DX7.

If I were marketing radios, I'd do the same thing. Try to hit the segment that generates the most sales volume first, then hit the more specialized segments.

Dreamer077

You can bet it has something to do with money. What they had was selling. Most likely the competion forced them to do it.

jpherit

The answer to your question requires full knowledge of the system details and a calculation of BER as a function of number and power of the other users. In your senario, the DSSS system is recieving a strong signal since its transmitter is near the other transmitters. This is a favorable condition. One then must develop the statistics of the interference. A simple example of this is the extreme, but unlikely case, when all N of the FHSS systems happened to hop at the same time onto the DSSS channel. So, in this extreme case if the receiver process gain was 100, and there were N=100 narrowband signals, each spread by 100, then the filtered signal would be equal to the desired signal. That means high error probabilty. Note that this is only for one 2 ms hop time (if we use the reported number for how long the FHSS signal stays on the channel of interest). The next hop would have most/many of the 100 hopped out of our channel. I think you can see the statistical nature here, for example, I assumed that all of the 100 of the FHSS signals happend to be all alligned in time. That is improbable too. In general there will be partial overlap, not complete overlap.

All I can say, in a general way regarding your question is that DSSS systems are quite robust against multiple user narrow-band interference.

Of course, one can imagine adding additional layers of protection on top of that provided by the processing gain. For example, forward error correction (FEC) can be (and is) added to digital transmission systems at the expense of some latency and additional hardware overhead. FEC is used, for example, in the CDMA cell phone uplink (remote to base). One can also imagine adding some form of error detection plus request to retransmit, but it seem to me that must be done with high speed so that a fram can be resent before the next frame arrives.
I have read some claims that Jim's Xtreme system and Nomadio use some form of return transmission, but I do not know what form either of these take so I can not comment on them. One could also imagine a hybrid system that uses both DSSS and FHSS. I would be very interested to learn more about their approach as well a more detailed info from Spektrum and Futaba.

chrisF test pilot

The answer was posted right before you asked the question. The answer being that they dont, neither of these systems are two way or involve any method of collision avoidance/detection or error correction as it is unnecessary ad would be detrimental to the application.

nonstoprc

My Feedback: (90)

Just wonder if FHSS and DSSS divides the 2.4 Ghz band the same way. If not, then is it possible that a FHSS channel can occupy more than one DSSS narrowband? This probably will make the situation worse.

chrisF test pilot

No they dont, they are completely different modulation schemes that use the available bandwidth differently. Not sure what you mean by the question, DSSS is not narrowband.

http://www.fcc.gov/oet/info/rules/pa...15-8-14-06.pdf 15.247 starts on page 100 and has the bandwidth and power requirements.

jpherit, the Nomadio stuff actually is FHDSS, I think the Futaba system may be also. Dont know a lot of info about Nomadio but do know they are using custom chips that they developed. XPS is using customized Maxstream chips. I do know from following it that with the speed of the processing with the Xtremelink Jim said they are using error correction and are able to send missing data before the next frame. Again not sure about what Nomadio is doing.

jpherit

I am not 100% sure I understand your question. Perhaps this will move in the right dierction for a common understanding.

A Spektrum system occupies two, ~ 1 MHz wide "channels" from a total of 80. So we can imagine that there could be 40 Specktrum units flying, taking all channels, when a FHSS system drives up and tries to fly. I assert that with a well designed system the FHSS system could fly. It would hop about among all 80 channels, each time the particular DSSS system that it hops into will reject the interference as I described above. (More precisely, the S/N ratio will decrease but the S/N ratio will stay well above what is necessary for rejection of signal). We could keep adding FHSS fliers until we have enough additional flyers that the S/N ratio degrades to the point that errors occur. When does that happen, I don't know precisely, but it could require a fairly large number of fliers above the original 40+1 I just described.

What about the other extreme? Suppose there are 40 FHSS (80?) systems flying and a DSSS user drives up and tries to fly. In principle he should be able to to so. What I don't know is how a Spectrum system will treat a channel with some noise at start up when it is searching for a "clean" CHANNEL. A reasonable design would accept some interference noise and it might chose the two channels with the least interference noise. It could then fly and it will not bother the FHSS because they jump out of the way quickly.

By the way, when I say narrow band, I am talking about the information rate/bandwidth which in RC is on the order of 10kHz. perhaps someone knows the precise bandwidth and can let me know). THis bandwidth is spread out to about 1 MHz. This IS spread spectrum!

Frequency hopping, in its simplest form, is a narrow band signal, again the same 10kHz, that bounces about in a prescribed pseudorandom fashion into (let's say) 80 channels. Both are spread spektrum and both have similar properties.

From the broadest perspective there are three ways, in general, that signals may be spearated. (provide multiple access without interference. They are: (1) Time division multiple access, (2) Frequency division multiple access, and (3) Code division multiple access. [TDMA, FDMA, and CDMA].
These, together with distance, separate users. We all understand FDMA, that is how the 72 MHz band manges access. One user per channel, but a lot of channels.
This is also how TV and radio provide access without interference. TDMA ia used in the network. Each user is assigned a its own time slot. Of course there can be contention (several users trying to use the same time slot) and there are various protocols to manage this situation. There is also the possibility of contention in FDMA, in the 72 MHz band the protocol used to handle it is the frequency board! It is not built into a system protocol.

Finally, there is CDMA, which is inherent a spread spectrum method. Here all users occupy the same time and frequency space simultaneously, but multiplex without interference using codes and their orthogonality properties. Note that even the Spektrum system is not using the multiple access properties of DSSS. They separate users by FDMA!! It is then a hybrid of CDMA and FDMA.

rmh

Fly em if you got em ---
I must be dense - All this seems to be much ado about relatively little on the practical side of things .
Please -- anyone one of the naysayers - please document one case of detectible frequency collision-with the DS7 system and the conditionsof the occurrance .
Just one -
Thanks

jpherit

I would be very interested to learn more about the Xtremelink approach. have followed the thread, but have not been able to discern just what he is doing.
As I mentioned, hybrid systems are indeed possible, and in fact hopping can be fast or slow and the frequency hop could be a multipleo of only a few 10's of kHz. In my example above I was presuming it was hopping between the 80 channels rather that within one channel. This is of course possible. I wonder about his down link. IF it is similar to the up link then it must have comparable RF power to the transmitter to have siimlar range. I hope he gets his system out soon so I can take a close look at it!

I am continually frustrated that I have to guess and infer what these folks are doing.

jpherit

My whole point is that "collisions" as you call them will happen when DSSS and FHSS are in the air together, but neither will be bothered by it. So I agree, you will find it very very hard to find a documented case of a "collision" causing a problem. That is the beauty of DSSS, contention is handled in an elegant way and retransmissions are not required. Here I assume that by collision you mean a FHSS system hopping into a spectral channel in use by a DSSS system.

I am not a naysayer, I am quite a big proponent of innovative use of spread spectrum concepts in our hobby. It is an exciting time.