johnls

My regular flying site is notoriously "noisy' in terms of radio signals. There are microwave links going across the site, generators operating and government agencies testing "equipment", the likes of which I am not privvy to. Approximately 2 1/2 years ago I lost three high value aircraft ( 2 jets and a large gasser) to radio problems in a six week period, whilst operating 35mhz. I was about to quit the hobby when 2.4ghz became readily available and to this day I had not had an issue and I have felt confident in the equipment Today however the weather was kind, but the site wet and muddy from recent rains so I decided to go fly a couple of glow powered planes. Both have Spektrum AR7000 receivers powered from 2200mah lipos, using powerbox switch/regulators etc. and using a JR pcm12x tx. Whilst I was flying a collegue was flying an electric plane using a new JR pcm 9x tx and AR500 rx , when he had a lock out and crashed his plane. I then flew my aircraft over the same spot and my controls locked in the neutral position before I regained full control a couple of seconds later. I flew back into the same area and the same happened to me again. I am now extremely concerned at flying my jets or large 35%-48% size aircraft on this site. Probably more intriguiging is what is happening?, my confidence in this technology guiding expensive aircraft has been shattered.

J.F

dunsfold?

GrayUK

Does anyone have the same problem with FASST on the site?

Paul

flyinfool1

My Feedback: (2)

While the 2.4 technology that we are currently using is in most cases the best link available to us today, it is still not bullet proof.
We have a fairly low power output, The RX can simply be overwhelmed by a very powerful signal so that it can not even see the signal from your TX that it is looking for.
It is very possible to have a noise level that is just to high for the RX to sift out the signal that it is looking for.

First thing I would try is to see if you can reproduce the "hit" by having someone hold the TX at a pilot station and then walk the plane to the area where it was hit.
Look around and see it you just happen to be flying right in front of a high power antenna at that location. Or possibly it is a spot where the beams of 2 different directional antennas are crossing.

If it is a high power signal that is originating from nearby, your only options would be to find out if there is time when that signal should not be present, to find a safer field to fly at, or roll the dice and fly.

johnls

Jeff
I think you have a valid point. remember also, that our transmitters, in the UK, operate to a lower power output level than in the States.

Paul, Nobody is using the FAAST system on this site. It would be interesting to know how it coped.

JF,
Yes, its at Dunsfold. Home of Top Gear, film sets , military UAV's etc

Anybody know of a flying site in the South East of the UK?

John

simonmiller

John,

Probably the best flying site in the South East - Junction 22 on the M25. Lots of jets fly from this site:-

http://www.phoenixmfc.org/website/index.php

Regards

Simon

gruntled

Just to help rule out other issues:

You are certain you guys are not pointing the antenna directly at the plane when it passes this particular spot?

jason

My Feedback: (1)

As Paul said, it would be interesting to know if FASST had the same problem on that site. I don't profess to be an expert on 2.4 technologies but if a Spektrum set picks two channels which are free at the start of the flight and half way through your flight you get interference on them channels for a long period of time surly it's going to cause a problem whereas if you have a FASST system which selects free channels throughout the flight you might be ok?? As I say, I don't really know a whole lot about the way it works but Futaba's system does make sense to me.

Jason

BaldEagel

My post was inspired by a bottle of Claret and was not worthy, so I have deleted it.

Mike

ww2birds

My Feedback: (14)

Jason,

I wondered the same thing about the Spektrum sets picking and holding two frequencies, assuming this would be inferior to FASST technique that spreads the energy over the entire band. You'd think the hopping would just have to be better - more robust. But I dug into it a bit more and learned that a well-designed DSM system does various forms of modulation that spread out the RF energy around the two selected frequencies, and actually does much better than you would think than if it were just "parked" on two frequencies.

The details depend on the specific design choices, so we'd have to get one of the JR/Spektrum design engineers to comment .. but it is an interesting topic, and not much in the way of facts or real engineering data available to us.

It would be fascinating to see a real-world site where the design choices of DSM2 and FASST could be compared with actual aircraft flying, I don't know of any such experiments or incidents. Mostly what we hear (and observe ourselves) is that they both work fine almost all the time.

There was some discussion of powerlines taking down some JR stuff at a California jet event last year, but I don't think it was ever concluded what was really happening, and given the market share JR has at the moment in the US Jet market, I don't there there were many (or any?) FASST systems to try .. and even if there had been, the risk of flying in an areas that took out someone else's plane might make people chicken out!

Dave

Molar mender

I find it interesting that you would mention "Today however the weather was kind, but the site wet and muddy from recent rains so I decided to go fly..." key words being wet, muddy, rain. I was doing research for the antenna arrangement for my new project using 2.4ghz and ran into some interesting information concerning the attenuation of radio waves. Attenuation is when a signals power is reduced as it is being transmitted. One source states, "Attenuation increases with a rise in frequency or distance. Also when a signal collides with an obstacle, the attenuation depends strongly on which type of material the obstacle is made of. Metal obstacles tend to reflect a signal, while water absorbs it." Another source (http://www.inpathwireless.com/faq.php) states "Even though rain itself does not cause major propagation problems, rain will collect on the leaves of trees and will produce attenuation until it evaporates. " Third site has "Water. That is why microwaves — which are 2.45 GHz — cook food. They excite water molecules because the water absorbs the energy and otherwise inhibits propagation." (http://www.isp-planet.com/technology/water.html)

Most of the research is from the high powered communications systems. Our transmitters only put out 500mw which is not a strong signal compared to the communications transmitters. I never had problems with 72mhz on wet mornings because the wavelength is so long but the same condition with my foamy 2.4ghz and it was all over the place. I don't usually fly 2.4ghz when it is noticeably wet.

So the place where your airframe had the lockout could have been in an area that could not get enough signal due to the attenuation of the signal by the wet conditions. Just something to think about and do further research on.

Roy

George

My Feedback: (57)

I'm starting to hear more occurrences of this happening. I JUST got off the phone with a buddy of mine back home and he was telling they are suddenly losing an abnormally large number Spekrum-guided aircraft (all shapes and sizes) in the same general location/area. This area just happens to be the same exact area I had the lockouts when I was testing the first XPS system. This cannot be coincidence, something IS affecting the systems, we just don't know what!

R/
George

gunradd

My Feedback: (9)

I herd some kind of new police radar gun is on 2.4 and it hit the whole band. I hope its not true.

gruntled

It could not get FCC approval if it did. All 2.4 in the USA has to play nice together.

Robrow

John, do you know if your setup tried using another of the installed satelite receivers? It's unfortunate that you don't have the data logging system that Weatronic systems have, although I'm not saying Weatronic 2-way signal would survive the same environment any better, you would have a record of which channels had the problem.

Rob.

c/f

My Feedback: (1)

OK here is a thread on Spektrum interferrence immunity, lots of good dialogue with a RF transmision proffesor (Mihia) at a university and John Adams of Horizon. John sent this guy a NIB DX7 to perform an interference test here is the test result post as well.

original thread. (http://www.rcgroups.com/forums/showt...ktrum+problems)

In a nut shell the interference immunity we enjoy comes more from the transmitters collision avoidance feature when booting up, versus the rx immunity to outside sources. The test show the vulnerability of current crop of 2.4G when the TX collison avoidance feature is curcumvented. Test also show Spektrum dual freq choices are very close to one another less than 3 apart, and only took one of the two being interferred with will lock out the rx...........



Test done



OK, I tested the DX7 this morning in my lab and this is how it went: I
first started the transmitter and then the receivers in close
proximity - they got lock really fast and without any problems. Then I
used a spectrum analyzer to find the two frequencies of the link. It
happened that both frequencies were close to channel 10 (actually one
was on 9 and one on 11). Then I started a single wireless card on
channel 9 transmitting to another one also on channel 9. Although the
wireless cards were supposed to go to 400mW, the software prevented
them for going that high - they only went to 63mW (18dBm). That's
fine, the interference would be far worse at 400mW, but 63mW should be
enough. At first nothing bad happened (i.e., good link - both of
them) even with the receiver very close to the transmitting
antenna. Then, looking at the throughput of the wireless cards I
figured out what was happening - the wireless cards were sensing the
transmissions from the DX7 and were yielding to these, i.e., the
wireless cards were careful not to transmit over those transmissions
from the DX7. That is easy to understand as 802.11 uses a CSMA/CA
algorithm, where CS stands for "carrier sense", i.e., for each packet
it's first checking if there is another transmission going on and if
it is, it waits until that is done rather than interfering.

However, this polite behavior was happening while the transmitter was
in the same room with the transmitter-receiver pair. In reality, the
transmitter would likely be farther away from the offending
interference source. Therefore I started to walk away with the
transmitter and at about 40 yards from the wireless cards the DSM2
link died down completely. At that point the signal from the
transmitter was considered too weak by the wireless card and it just
started sending packets without any problems. At that point I stopped
the traffic flow between the two computers and the link restored
almost immediately. Started it again and it died. Repeated a couple of
times to make sure. It was clearly a case of interference, where
starting traffic on 802.11 was able to shut down the DX7 link at
will. Thus, I think that this conclusively shows that DSSS can
interference from wireless cards, in some cases even from a single
card at a relatively low power (Cisco access points, popular in
universities and enterprises usually go to 100mW).

At that point with the clear case of interference, the transmitter was
about 40 yards from the receiver, the receiver was about 2 yards from
the transmitting computer, the receiving computer was about 5 yards
away and was receiving at full speed. The receiver (AR7000) was
plugged into the 1100mAh NiCd battery that came with the transmitter
and that was charged for about 10 hours last night and used for about
1 hour this morning. The receiver also had a digital servo on it, but
I stopped using movement of the stick as an indication of a good link
almost immediately, as I figured out that the lights are a far better
indication of the link quality than the servo (they start blinking as
the link goes bad and go completely off when the link is completely
off). The transmitter battery was also charged for the same 10 hours
as the receiver battery, and in the beginning showed absolutely full,
and at the end just a tad under completely full (5-10% less than full
or so).

Although I was able to shut the link down with a single wireless card,
in general, it takes two of them to shut it down. After this test I
started the transmitter and receiver two more times to see where it
chooses its frequencies. The second time it also chose two nearby
frequencies, that could be shot down with a single card (on channel 7
this time), and the last time it chose the frequencies at the ends of
the band (requiring channel 11 and channel 1 to shut it down).

So... what next? I considered trying to see how much interference it
takes to shut it down, that is to start moving the receiver away from
the source of interference and see how far away it can get before the
link restored. However, I figured out that this experiment would be of
little value, as indoor propagation (where all this experiment took
place - in my lab) is significantly different from outdoor propagation
and it would be difficult to translate any results from indoor
outdoor. At the same time it's rather inconvenient to do this
experiment outdoor - the PCs are desktops, the spectrum analyzer is
not hand-held either, etc. It would most likely also take longer and
require at least one helper to get sufficient results for a good
model.

So, now that there is no question in my mind that external
interference (even FCC legal interference) can shut down a DX7, what
are the chances of a shot down? As I pointed out before, there are too
many variables to give a general answer. However, there are a few
points that can be considered:

- choosing two independent channels at arbitrary frequencies can be
risky with respect to protection from wide-band sources of
interference, like 802.11. If the user is unlucky enough to get those
two frequencies close to each other, a single access point can bring
down a plane by shutting down both frequencies. What is the chance of
this happening? If we assume a bandwidth for an access point of 20MHz,
at a field with a single access point on a given channel X, then the
chances of both frequencies being on the channel of the access point
are about 1/16. Basically one in 16 times, if the plane is relatively
close to the access point (as I pointed out, today's experiment does
not clarify how close the plane has to be to the AP) and the access
point becomes active for a few seconds, the plane goes down.

- at the same time, choosing the frequencies at a fixed 40MHz
spacing increases the chances of collisions with other spectrum
users. I wrote two small scripts to figure out what are the chances of
two users to choose the same frequency given a bunch of users and it's
clear that two independent frequencies are far better than fixed
spacing (I included here the results - the black line considers only
frequency collisions, while the red line tries to take into account
the duty cycle - assumed at 12%). The probabilities shown are that at
least one user out of all the users can be shot down because
other users chose his frequencies - it's clear that at very few users
the chances are very small, but get up fast as the numbers of users
increase. Also, this probabilities should be considered in conjunction
with the chances of a plane to be closer to a source of interference
rather than from its own pilot - if each plane flies closer to its
pilot than of any of the sources of interference, there should be no
problems.

- also looking at those probabilities, it's clear that even with
four transmitters (the one from John Adams and the three that Diggs
has), the chances that one of the transmitters would have both
frequencies chosen by two others are really small (smaller than one in
one thousand). Thus, the experiment with two transmitter is not
feasible using stock transmitters - if we'd have access to the
firmware, that would be easy.

- probably the easiest fix that Spektrum may consider implementing
is imposing a minimum of 20MHz between the two frequencies, while
keeping the choice of the channels independent of each other. This
would probably not require any changes to the receivers and only a
minor firmware upgrade to the transmitter. This should fix the problem
with interference from wireless access points, potentially fixing
Diggs problem at our field - I am not sure if access points were the
problem - I didn't do any measurements at our field, but if
interference is the problem, this would fix it. If it's a BEC, of
course it would not fix it. This is a very simple fix with almost no
downsides.

- this would not fix, however, the problem with many users - for
that there should either be a limit on the number of simultaneous
users, or more significant changes to the system (a DSM3 or
such). Alternatively, requiring the users not to fly close to other
users (as potential sources of interference) would work too, but it
may be impractical.

Please let me know if you have any questions about this (I'm sure
there will be at least a few), as well as suggestions for further
testing - I'd be glad to do them as long as they:
- take a reasonably short time
- ideally not require outdoor testing, as I'd have to carry a lot of
hardware out and buy batteries and inverters
- may result in a conclusive and useful outcome.

Best,
Mihai
Attached Thumbnails Probability that at least one user gets its two frequencies killed by another one when using fixed spacing. 98.2 KB · Views: 99
Probability that at least one user gets its two frequencies killed by two other users when using an independent choice of frequencies. 103.6 KB · Views: 51

gunradd

My Feedback: (9)

I use a DX7 and was hit and my receiver did not go into fail safe because the motor was set to turn off in fail safe and plane fly straight. When I lost it the plane was flying all over the place with the motor running until it finaly almost flew out of site then into the ground. It took over a week just to find it. I have since switched to the AR9100 and have had no problems with 4 remote recievers. I have seen XPS and futaba also get hit at our feild and it always seems to be in the same spot. I just thought it was always user error because 2.4 has always been rock solid but now I am not so sure. I got to where I didnt even do range checks anymore I was so spoiled but now I do them every flight. The part that really bothers me is when you get locked out it should go into failsafe and mine was doing uncommanded turns and just going all over. I was helpless watching it fly away I even had time to try and cycle the power to see if I would get the link back.

johnls

Now I am asking around at the flying site it appears I am not alone, its just that the other guys didn't tell anyone or put it down to one of those things!! Helicopters have been having problems and crashing after experiencing lockouts on both JR and Futaba 2.4 set ups. Several aircraft have also crashed or had a problem from which they recovered. The orientation of the ariel is not the issue, unfortunately these size machines haven,t got a data logger.
Simon, thanks for the advice on the Phoenix club. I put my application in today, they seem to have a waiting list but fingers crossed.
I'm not trusting anything expensive at Dunsfold, so other than next weeks winter jet meet I'm staying grounded until I get a new site.

John

NUNU

data logger wont help when the aircraft crashes and burns..I lost confidence a long time ago on 2.4....wont go into it I will say the marketing on 2.4 is so good that no matter what happens the balme will be placed on the pilot or the antenna..2.4 is still too new to trust on expensive crafts I feel

rhklenke

My Feedback: (24)

The simple fix to this problem? Use FHSS - like FASST!

Military systems that use spread spectrum to avoid active interference, i.e., jamming, use FHSS. The reason being that in order to jam a FHSS system completely, you have to fill the entire band with noise at a sufficient level to block the transmission - much harder than simply blocking two fixed frequencies. Also, SS systems that use a single frequency, (like 802.11) use a long sequence in their DSSS transmission to spread the energy out fairly wide across the bands. If you look at a DSM2 transmission, the channels are fairly narrow - i.e., not spread out very far by the DSSS part of the transmission, and thus are fairly easy to interfere with.

I've said this before and I'll say it again, the Spektrum guys are to be admired because they were first and they were a small company that developed their own equipment, but their contention that they selected the "2 fixed channel" method over FHSS because it was "better" even though it was "harder" to do than FHSS is *pure* marketing BS. They were the first to come up with this method and if it was in fact "better," then FHSS then, like everything else in engineering, one has to ask why nobody else did it that way before - especially since spread spectrum radios have been around for 20+ years...

I still think that sometime in the future you will see JR move to a FHSS system - just like everybody else - because its superior to DSM2.


Bob

GrayUK

Interesting and informative posts.
I have and will continue to be a supporter of FHSS, just makes sense to me.
You say no one fly’s FASST at that site, i would be happy to do so as a trial.

Paul

blhollo2

My Feedback: (278)

I dont think that was a good idea even though you were investigating the lockout but however reasonable..

johnls

Paul, Are you at C/Jets next Friday? If so I'll speak to you then.

John

simonmiller

John,

I believe that pheonix has about 10 spaces available. Moreover, our membership renewal is coming up in January. If you need any help do not hesitate to contact me and I will see what i can do.

Regards

Simon

gruntled

Definitely something to keep in mind if I ever fly my planes in a small room with wireless routers in it, which is . . . . . never.

The 3+ years of trouble free DSM and DSM2 operation (DX6, DX7, 9303) me and my buddies have experienced speaks for itself as far as I am concerned.