PhoenixOne

Hello all,

I am finally ready to upgrade to another jet and at the same time I am thinking of switching my radio. I am currently using a JR10X which is PCM. I would like a really nice scale jet, (Tam Superhornet is what I'm looking at right now.) I heard that he is using the Multiplex Royal Evo 12 which is IPD receiver not PCM. What is the difference? After asking countless sources, nobody seems to be able to give me a straight answer as to what happens if you get hit. I hear that this is what most use in Europe and it is better than PCM. Is this true? And I guess I might as well start a poll or something. I assume JR is better than Futaba? I just want the best possible radio security that I can have at the moment.

Please help I am confused!

Thanks

Mark

Gordon Mc

My Feedback: (11)

Yeah - I've seen a lot of people push IPD without being able to explain it. "It's better", they say, but myself and another guy who have some DSP knowledge and had some questions about IPD tried for a long time to get explanations, to no avail. Then suddently the best explanation I ever saw (much better than any official Hitec people had been able to provide) came from a guy called Harry Curzon - RCU id "HarryC". If he doesn't pop in here to answer your questions, you might want to search through his prior posts, and / or drop him a PM asking for help. He knows his stuff, and will give you the straight dope.

unknown

http://www.multiplexusa.com/Support/..._DS_IPD_GB.PDF

Lets hope is last crash with the SUPER F-18 was not because of this new IPD system

PhoenixOne

No I heard through several people that Tam's Superhornet prototype that crashed in Arizona had a JR10x. That is what started my curiosity in this whole PCM, IPD thing. From what I hear, if he had been using a multiplex with IPD that probably would not have happened. Im just curious, thats all. I'm sure we all want to avoid any chance of a lockout with the best technology we have, especially with the investments involved with these jets.

Mark

unknown

I wish these RC/RADIO manufacturers would try out dual frequency transmitter SYSTEM. For example your radio transmitting on both Ch 17/20 and your receiver receiving both channel.

gregmax

How about a Digital Spectrum Modulation system like in the link below?

http://forums.radiocontrolzone.com/s...5&page=1&pp=25

It is also a 2 way system so telemetry can be sent back!

They are apparently going to produce a system for model aircraft.

The Spectrum system is available now but only a 3 channel system for surface use (model cars), and as it is just
a module system is should just plug into your compatible transmitter and then just change receivers over.

Or another similar system under development,

http://forums.radiocontrolzone.com/s...d.php?t=165549

S_Ellzey

I think I know the answer to this one, so let me take a shot at it.

PPM – Pulse Position Modulation - The technique that has been in use for over 30 years in R/C. The transmitter creates a series of pulse by switching on and off. A very long sync pulse is generated followed by a series of short pulses, one per channel. It is the width of each of these short pulses that commands a position. After the receiver sees the sync pulse the next pulse it sees is sent to the first servo, the next pulse after that goes to the second servo and so on. Each servo moves to the commanded position defined by the width of the pulse it receives. You can actually see the pulses on an oscilloscope.

PCM – Pulse Code Modulation - The transmitter sends out a digital stream, a bunch of ones and zeros. The receiver sends this data to a computer that then creates a series of pulses that are sent to the servos, just like PPM. The receiver’s computer has logic in it that handles missing data, hold and failsafe.

IPD – Intelligent Pulse Decoding – The transmitter is standard PPM, but now the receiver is sending pulses to the computer, which then times the length of the pulse, and then creates an identical pulse to send to the servo. If the computer can not discern a pulse, its logic will create one to fill in the void, hold and failsafe like PCM.

The advantage of IPD appears to be that if a PCM system has just one bad bit in a data frame the whole frame is thrown out; thus PCM must have a solid link to work well. With IPD, the system only has to see the pulses, which is probably a bit easier to do, so you are probably able to get a little more at the extreme edge of your range (just how good are your eyes?), and it probably is a bit more tolerant of noise.

Personally, I have never had a PCM problem. The JetCat and Ram ECU’s have only reported a couple missing frames over a years time.

Steven

JohnMac

Steven,
Your description is quite accurate. IPD to all intents and purposes does the same as PCM. There are some pros and cons but they do not amount to much except for the speed. The IPD has the same speed as PPM, not the slight delay of PCM. However, I flew my first jets with PCM and the slightl PCM lag didn't really bother me. I have been flying IPD recievers for about 5 years. I think I can detect a slightly better servo response from the IPD but its a moot point for me.
The other benefit of IPD is it is not brand specific; it can be used with any PPM Tx.
Finally, Multiplex had already developed a 12 channel Rx. It was not going to be easy to convert this into PCM at that time but developing it into an IPD Rx was quite possible. Right now I am using the 12 & 9 Channel IPD synth Rx,s in my jets. Solid as a rock!
Regards,

John

JohnMac

Hi Mark,
Just hold on there son
I am a true blue Multiplex flyer, but I think to claim that a Multiplex IPD Rx would have perfromed better than a JR PCM may just be stretching things a little. None of the sysytems we use actually provide what the name claims - failsafe! If any Rx goes into failsafe, and stays there, the model will crash. Does not matter which brand. Theres nothing very safe about this failure mode. PCM/IPD is better than PPM in my opinion, but only by degree.
Will one brand be more resitant to interference than another? Perhaps. But will that matter in the end, on that particular day, at a particular site, witha particular model. Too many variables. It's anybody's guess.
So unless you have some very specific eveidence about a particular crash, specvulating that another brand of Rx might have prevented that crash, is I suspect just that, speculation.
Regards,

John.

HarryC

My Feedback: (1)

Very unlikely to be true. If interference caused PCM to go into failsafe then the IPD would have done so as well.

IPD is just a method for PPM to get error detection and programmable failsafe so that it appears to the the user to behave like a PCM failsafe. The rx is cheaper smaller and lighter than a PCM rx, and works with any brand Tx unlike PCM which is proprietary. PPM systems usually update a little faster than PCM systems but would you notice? Those are IPD's advantages over PCM, I do not believe it has any technical advantage over PCM.

PPM error detection is a well tried routine, and has been available as a channel add-on, and built into electric flight ESCs for many years. Multiplex even built it into their digital servos a decade ago so each servo had individual failsafe. PPM error detection does not look for precise errors like PCM can, rather it looks for servo pulses that are shorter or longer than the possible min and max of a servo pulse. Hence it checks that a pulse is valid, rather than correct. Multiplex then bolted on the fully programmable failsafe feature of PCM. If a pulse is valid it is sent to the servo, and copied to the memory. If a pulse is not valid it is discarded and the last good pulse sent from memory to the servo. Thus it does a "hold" rather then a glitch. If it has been on hold for 1/2 second and the pulses are still not valid it then sends the servos to the positions which you have programmed into it, such as throttle closed and all cother controls at neutral, or whatever you have chosen. As soon as a valid packet starts again it comes out of failsafe. I can verify that it does work, I have been using IPD for almost everything for about 5 years now, have had 3 failsafe events, two of which went into and then back out of failsafe in the blink of an eye - when you are doing a harrier with a Synergy it is rather obvious that it has gone into then back out of failsafe!

The obvious question is that if it can only check for validity and not errors, won't it get fooled by pulses that are longer than they should be but still within the time limit allowed? Only two pulses, the first and last, can be fooled like this and that would require some very precise interference which is so unlikely it can be discounted. If a pulse is extended in its ON period by interference then it runs into and merges with the next pulse. If the interference then stops, the two pulses together are longer than the max valid pulse and so are rejected, the strict sequence is also now broken so the hold settings have to be used for all the rest of the packet until the next packet starts. The downside is that if just the synchro pulse gets hit then the whole packet is lost even if it is all good data, whereas PCM tends to split the data set into two smaller packets and if one gets hit the other will still be ok.

Programming an IPD is done on the rx. You switch the radio on and set the tx to the failsafe settings, then push in and remove the programming plug in the program socket on the rx.

The new synthesiser IPD rx have the facility to switch the failsafe off when you are doing a range test, so that the error detection and failsafe does not mask range problems, the synth rx also have error counters which count the total of data erorrs, low voltages etc.

H

XJet

For those that are interested, I've almost completed a review/shootout of six receivers (PPM/FM, IPD/DSP/TSR and PCM) in which I've compared the performance of the individual units and the technologies they're using.

As a previous poster has stated, there are pro's and con's to IPD when compared to PCM and neither is perfect -- but the results of the tests I undertook may surprise some.

Included in the test are two fully synthesized receivers (the Polk Seeker and Sombra Shadow).

All receivers were subjected to a range of adverse RF environments including high levels of impulse noise and burst noise (both radiated and on the servo/battery leads) as well as test for intermod/cross-mod resistance, sensitivity, selectivity, recovery from lockout and failsafe, the effectiveness of extra features (such as TSR) etc.

Right now I'm sorting through a *huge* pile of data produced as a result of this testing and once I've knocked it into some delightfully crafted prose I'll publish it on the Web.

EddieWeeks

I have run a RC-Car with a camera all over the hood, in people houses where ranges
of 100m from inside my house to inside another house was commen..

The IPD would run out of range at some of the furthest houses but I never lost contol..
It just got jettery.. When I moved the sticks I could see the wheels move just slowly and
I had the ability to drive back in range..

I have seen some PCM system range check and it just quits working...

Eddie Weeks

EddieWeeks

Man I want to hear all about this... Keep us updated please..

Eddie

BMT

Fact remains that any post rf demodulation sceme that has more processing than simple PPM will be better/safer. The quality of the firmware has a much bigger bearing on which commercial brand is better. At the end of the day the differences are all subject to the quality of the RF link and not much else. In the event of signal path degradation the differences become more or less acedemic since it happens suddendly and hardly ever progressivly. In my experience with Hitec/JR and Multiplex(of years ago admittedly) JR comes out on top for design quality AND build quality. I have no experience with Futaba unfortunately. But I have changed my mind about JR servos recently....
The way forward is surely Spread spectrum but the current unlicenced bands offer frequencies which are rather directional and may not be as fit and fly as the current sub 100MHz systems.
Cheers
Andre

HarryC

My Feedback: (1)

I suspect that Multiplex developed the IPD range for commercial rather than technical reasons. Multiplex has long been a very innovative company and I believe that it was the first to use PCM on a model radio. But they never updated it and were still using the original version with a low bit count and slow update rate. They desperately needed to design an up to date version with 9 or 10 bits and a better update rate and, as John mentioned, they needed a 12 channel rx as well. Mpx was faced with an R&D cost, for what is a very small sales volume compared to Futaba and JR. This was before the Royal Evo whose sales volumes must feel like winning the lottery for Mpx. Therefore I guess they felt that the cost to develop a better failsafe rx and have a 12 channel version would be lower if they developed IPD and potentially have greater sales since anyone can use it with any brand of radio, even with non-computer Tx. Therefore they stopped making their PCM rx in favour of the IPD. Having gone over to PPM, and no new PCM rx being made, their next Tx the Evo did not need PCM just plain PPM so that must have cut the costs of that by a little bit too.

Interestingly, Graupner/JR also makes 12 channel radios, and in 12 channel mode they are PPM only, not PCM. Graupner/JR make a 12 channel rx but it has no error detection or failsafe which is odd since the sort of models that use 12 channels tend to be expensive, large and fast, and in most need of failsafe! If you want to use a Graupner/JR MC24 or MX22 in 12 channel mode and have failsafe, you have to buy a Multiplex IPD rx.

PPM and IPD has limits though, and 12 channels is probably close to the limit, since every additional channel adds to the length of time the data packet requires and thus reduces the refresh rate. Mpx uses compression in 12 channel mode to reduce the packet time and thus improve the refresh rate, but Graupner does not and takes the full hit of time that 12 channels requires. The compression can only be taken so far before it starts to have an impact on accuracy.

The only way to get more and more channels yet not reduce the refresh rate is to use PCM with an increasingly fast bit rate. I have been curious as to why PCM systems have been slower than PPM, but Futaba is showing the way with the new 14MZ which has not only squeezed in more channels but has really fired up the refresh rate as well. If Mpx ever wants to make a radio with more than 12 channels then it will have to consider coming back to PCM, but the market for radios of more than 12 channels will be very small indeed and with the other big players already establishing themselves there the sales volumes would probably not justify it.

The makers of PPM DSR rx make great claims for it, so I will be very interested to see the results of XJet’s tests.

H

basimpsn

My Feedback: (5)

Sorry wrong post

mr_matt

My Feedback: (10)

I think, until the 14MZ, that all PCM radios use normal Frequency Shift Key (FSK) modulation. My guess is that the 14MZ is the first radio to use some sort of direct digital modulation, like QPSK or something. That seems to be the only way they could get that much information down that small (10 kHz) channel.

XJet

Don't forget also that as CPU power goes up, the options for implementing more sophisticated lossless compression algorithms also increase.

Bearing in mind that most of the time the servos in a plane move very little, a great deal of the bandwidth used by systems without compression (or with minimal compression) is wasted.

The tricky bit is (of course) dealing with the fact that any number of frames may not be delivered to the receiver so delta-frame compression can only be used sparingly -- but I still think it quite possible to fit 14 channels into the existing bandwidth using FSK, espeicially if the primary channels were prioritized over secondary or tertiary channels where increased latency of response was of less importance.

unknown

Now this is a system i like [8D]. Anyone using this system?? or know anyone that is using this system (DSL receiver)

http://www.acteurope.de/html/dsl_receiver_system.html

HarryC

My Feedback: (1)

That's not a valid strategy once mixing is allowed because you never know what the primary channels are. A 4 servo wing on a glider will have in effect, 4 aileron channels, but if the tx only gives priority to channels numbered 1-4 then all the aileron servos on higher channels will not match the response.

H

HarryC

My Feedback: (1)

I may be completely misunderstanding the nature of data frequency and available bandwidth, but.......

We do not have a 10khz bandwidth, we have 10khz spacing. The data we send is not an anlogue sine wave such as sound waves with all the infinite variety that they contain across their spectrum and thus requiring enough bandwidth to do it. We are only sending frequency X or freqency X +3khz and we only measure if they are on or off, we do not measure any data modulated within the carrier such as sound wave radio has to do. Surely then we can toggle between X and X+3khz at phenomenal frequency because there is no sine wave data modulated into the carrier frequency itself, we merely need to measure "is frequency X on or off, is X+3kz on or off?" If I am sending frequency X to you, and you have extremely sensitive and fast reacting receiving equipment, I could switch off the transmission for a tiny tiny fraction of a second and your receiving equipment will detect that momentary off. I don't understand how bandwidth affects that. If I combine switching off frequency X with switching on frequency X+3khz for that tiny tiny fraction of a second, your receiver will see that momentary on. That's all we ask our model radios, both PPM and PCM, to do.

Maybe the receiver needs a whole wave to arrive in order to know whether it is actually getting X or X+3khz, but in that case our data frequency is limited by the carrier frequency, e.g. 35mhz or 72mhz and not 3khz.

XJet

Unfortunately life isn't that simple :-)

But to simplify as much as possible...

Whenever you change the frequency of a radio signal you generate sidebands.

In fact, the faster you change the signal, the more sideband frequencies you generate and hence the more bandwidth is consumed.

In the case of our digital RC transmitters, the square-looking signal that is used to perform the modulate is actually not quite square. If you look carefull at the vertical lines that represent the transition from a high signal level to a low signal level (and v/v), you'll see that they're slightly sloped -- that is to say that the voltage doesn't go from zero to maximum or maximum to zero instantly.

The slope on this vertical section of the wave is called the rise-time and it is the angle of this slope that determines the maximum modulation frequency that is actually being used. The steeper the slope, the higher the effective maximum modulation frequency being seen by the transmitter.

In the case of FM (and FSK), things get a little complicated due to the introduction of something known as the modulation index -- which is effectively a ratio of this maximum modulation frequency to the amount of frequency change produced in the transmitter's signal.

Without going into the math and boring the snot out of everyone, it suffices to say that the bandwidth of an FM or FSK modulated carrier is primarily determined by the rise-time of the modulation signal fed into it.

Now you can make the pulses and the inter-pulse gaps shorter, thus allowing more data to be transferred -- but unless you reduce the rise-time (increase the steepness of those vertical lines) you find yourself reaching a point where there's no flat bits left on the top or the bottom because the pulse/space has simply become too short for the available bandwidth.

In theory, zero-point FSK could allow the use of much shorter rise-times without producing more sideband frequencies -- but then you start getting into the more sophisticated modulations schemes anyway.

HarryC

My Feedback: (1)

Xjet, fantastic info, thanks very much for that. I knew it couldn't be simple otherwise the model radio makers would have been able to do it long ago but with my very limited knowledge of waves and their side effects I couldn't see what the problems are.

To satisfy my curiosity, have you a rough idea what the max bit rate is for basic FSK PCM in a model radio with typical model radio slopes? I am just wondering is it in the region of 500bps or 5000bps etc?

At the risk of boring others, (but hey they don't have to read this bit!) what are sidebands, how are they created, how does their frequency relate to the carrier or the modulation frequency, and why should we be bothered about their existence?

H

Gazzer

I'm not bored just acutely aware of my personal limitations!!!!!!

This is fantastic information and it seems there is a tremendous amount of knowledge around.

If people can distill this to us simpletons, then it will be really valuable.

Call me old fashioned, but do I read the early summaries, as,

Differing systems, all capable, all as reliable as they can be, some slight variance in the way they achieve failsafe, but the probable results very similar?

I guess I ought to concentrate on a first class installation, with good contacts, fresh efficient batteries, clean aerials, properly set aerials and i should be OK!!!!!!

PCM10 X by the way!!


Keep the thread going though!!!

Gazzer

BMT

X-jet explained it as best as anyone can here. The fact is that servos have a resolution of at best 1024 bits and a min transit time of 100ms for 60 deg. So it follows that when you send a new "dataframe" you only need to send the most significant bits to get the servo going and send the least significant bits before it is even relavant for the final target position on the servo. For a 5 deg move you only need to send a delta of a fraction of the full frame anyway. I think the Futaba engineers may have used this strategy which merely requires a bit of lateral thinking, but took a bit long to get there....
IPD gives an alternative to this solution since PPM has infinite resolution (more or less). It cannot be as secure a PCM but under real conditions of signal blockout its acedemic anyhow.

Cheers
Andre