Warning X8R RX - do not use with CDI on IC powered 'planes

[Andy J]

Interesting to note some of you use a second receiver with the X8's. Wondering if this is the sensible approach for all models not just those with ignition systems for the motor. My issue with the 8XR pro occurred in a standard 4 stroke glow, so don't think it is due to ignition noise but any noise on the supply lines. I am interested to know if a second receiver is used with the X8 range, which one would you select.

My tests of the 3A UBEC I had an issue with was inconclusive but will investigate further tomorrow now that I found some decent scope leads to prove the supply lines when connected to the Rx.

[Andy J]

Posted by Peter Christy on 08/05/2019 12:49:48:

Ron: Should have added - use a decent 4-cell pack, and plug it straight in to the receiver, not via a switch.

I'm trying to eliminate as many variables as possible!

--

Pete

Agree with Peter running a high voltage may be ok for the Rx but a lot of servos are not rated to operate above 6V. A fully charged 6v nimh battery will be 7.2v, LiFe will be even higher so unless you use high voltage servos this is putting a lot of strain on the servo electronics.

Ron... whilst I appreciate you picked up a bag of bits after the crash did you find each of the servos worked when you got home and tested each part of the electrical installation?

Briefly looked at a cheap powerbox but not entirely convinced this provides a solution to the problems we have here.

[Ron Gray]

The servos I use can handle the voltage, unlike the mistake I made a few months ago with some others!

From the totalled Sbach, the only servo that was damaged, written off, was the throttle one, all others have been thoroughly tested and are ok.

Regarding redundancy, my thoughts are that for a belt and braces situation I would use 2 R series RXs connected to each other through the Sbus ports. There is no point, imo, of connecting 2 X series RXs because if I can get any X series to fail, chances are they may both fail. But don’t forget that I’m talking about CDI equipped ‘planes, electric and glow I will still use my X series in.

Let’s not lose sight of the facts, there are 2 issues, firstly the X series RXs have poor protection against external interference and secondly they do not auto recover from the resultant lock out. It is highly likely that the cause of the interference is down to a poor connection in the HT lead, namely in the plug cap and is more likely to occur when using the smaller plugs as the plug caps lock on by pushing on and twisting.

Edited By Ron Gray on 08/05/2019 19:45:31

[Peter Robinson 9]

Hi

Just for info

I did some testing myself yesterday with the redundancy bus RB20 with xr8 pro and 6xr rx's. Also with twin 5cell nimah power supply. I swamped the x8r rx with the tx until it went to red led (rssi = 0), the system carried on working as normal. When the antenna was moved away the x8r reciever worked normally.

I tried to swamp the X6R rx with the tx but I could not turn it to red led even with the antenna touching the rx aerial. I got the impression the X6R is quite robust with signal reception.

**LINK**

Peter R

[Ron Gray]

Peter, good information however the circumstances are not the same! You managed to get the signal to fail but in my testing it is not the signal that fails, it is the interference that locks the RX, from which it doesn’t recover.

[Ron Gray]

Further to my previous tests here are some more:

[Geoff S]

Those are very interesting videos. I'm very grateful for the heads up because if I decided to convert my 2 petrol powered models from 35Mhz (Multiplex IPD receivers) to 2.4gHz then X8R would have been my choice.

Way back in the early 70s I was involved in designing an electronic fruit machine at Bell-Fruit using Diode Transistor Logic (DTL) chips for control (the reels were the old electromechanical types). Sadly, the main objective was to make them tamper proof so we conducted tests. (pub manager's often had no share in the machine profits and had hours to fiddle them). I brought in a motor cycle ignition coil and spark plug and proceeded to see how immune the logic (5v supply) was to a spark plug held close to the front of the case and succeeded in blowing most of the chips and emptying the coin hopper! So immunity to high voltage sparks near electronics is not a new problem. I don't know how the problem at Bell-Fruit was solved because I left shortly afterwards.

I do think Ron's test is quite severe with the plug so close to an unprotected PC board and I doubt if anything would survive unscathed. The main problem is the lockout and the failure to recover quickly (or even at all) rather than complete immunity.

btw I think Ron's very brave holding the plug lead! I've had a few ignition shocks in my time and whilst they're not fatal they're not enjoyable either

Geoff

Edited By Geoff Sleath on 10/05/2019 16:35:17

[Ron Gray]

Geoff, my tests have shown me that the R series RXs from FrSky appear to be fairly bullet proof regarding interference so I will be using those in future. Yes, I would agree that my test is severe but it is a worst case scenario so if a RX fails then there is a possibility of it failing in flight, I can vouch to it on 4 occasions! Regarding holding the lead, I’ve only had 1 shock and that was when I was also holding a sheet of litho plate as a test interference shield and allowed the spark plug to touch it, my reactions to the shock were, in my opinion, quite good! Not enjoyable though!

[Peter Christy]

OK, Ron, thanks for that. That eliminates the power supply from the problem.

My best guess is that the EMP from the spark is breaking straight into the microprocessor, as I theorised at first. I would guess that the "watchdog" is a piece of software in the processor, rather than an external kicker, so once the processor locks, so does the watchdog.

(A "watchdog" is a piece of circuitry that resets everything if it fails to detect activity for a set period.)

The interference rejection of the *receiver* (as opposed to the control circuitry) looks as if it holds up quite well. Its the controlling logic which appears to be the problem.

If I'm right - and this is all speculation - then it is difficult to see what can be done to improve things, short of a re-design, which is what they appear to have done.

It might be worth trying with some thin metal sheet wrapped around the receiver (not the aerials) and connected to the receiver ground. I would expect that to show a marked improvement. If it does, all that would be necessary would be to replace the case with something conductive and grounded. It doesn't need to be a metal case - a plastic one with a conductive paint would probably do, as long as the "shield" was grounded to the receiver.

--

Pete

[Ron Gray]

Pete, if you go back through this thread you will see that I have tried shielding the RX (I used litho plate) and that certainly works, however even with that shied in place if I hold the spark plug near the antennas the RX still locks out. Difficult to see how to shield the whole setup without affecting the signal reception. I had thought about a litho plate ‘firewall’ between the CDI and the RX but then it may also shield the RX from the TX!

[Peter Christy]

Ron, the trick is to connect the litho shield to the receiver -ve line.

A badly shielded spark plug - as in your case with the extra spark gap - produces an enormous EMP. You don't notice it so much these days, with TVs all digital, but back in the analogue days, a badly suppressed motor-bike could cause interference over a 100 yds away.

That's not to say that there isn't a problem here, just an illustration of the scale of the problem. I have to admit that its something that would never have occurred to me!

Many years ago,a club-mate (now well known in the large model fraternity) was having problems with a very well known and normally very reliable Japanese radio. The model would lock out at random and crash, yet when the wreckage was retrieved and plugged back in, everything worked faultlessly!

The service agents were unable to find a problem, so I got roped in to look at it. At the time, I had a very old, all valve EMI oscilloscope. I was prodding around the receiver, looking for anything odd, and when I changed ranges on the 'scope, the receiver locked up! The decoder chip got too hot to touch! Switching off and on again restored normal operation. Change range on the 'scope, and it all locked up again!

I came to the conclusion that the (proprietary) decoder chip had a faulty voltage regulator in it, which was latching up in the presence of the sparks generated internally in my 'scope when I changed ranges! Just like your X8R. I suspect one of the servos was noisy and providing the necessary glitch on the power rails to trigger the fault.

I sent it back to the official service agents, as they wouldn't sell me a decoder chip. They did replace it, and after that it behaved faultlessly, but it came back accompanied by a very snotty letter about letting "unauthorised personnel" tinker with "their" gear! And this was in the days of 27 MHz AM radios!

Seems some things never change!

At least FrSky do seem to respond to customer input!

--

Pete

 

Edited By Peter Christy on 10/05/2019 19:43:28

[Ron Gray]

OK, so the final couple of tests that I carried out following Pete's observations above:

[Andy J]

Ron

Don't think the problem is specific to spark ignition as we both have totally different set ups in terms in our models.

Did some testing with a RX8 Pro yesterday powering the Rx alone first with the 3A UBEC I had in the model that crashed. The noise was considerable as shown in the first picture with 260mV ptp.

Second test was with a 4.8V nimh pack and the supply is still not clean with 128mV of noise. So I think the issue is lack of smoothing on the Rx supply lines.

[Ron Gray]

Andy, the issues that I have seen are purely down to CDI interference, the videos quite clearly show that. That is the reason why I created this thread. May I suggest that if you have seen issues with your setup that can be attributed to other things then it would be better to create another thread for that. This way things will be a lot clearer for anyone visiting the FrSky part of the forum.

[Peter Christy]

First of all, thanks Ron, for your very thorough tests and demonstrations. I'm surprised that the shielding didn't work, but that - coupled with something else I saw - makes me suspect that the interference is breaking through via the power lines.

I noticed that in one of your demos, the receiver locked out while you were holding the spark plug over the battery, not the receiver. In your SBach, was the receiver battery close to the CDI or its battery, by any chance?

I don't know if you have the facilities to carry out this test, but it might be illuminating. You may recall that Spektrum, at one point, provided capacitors mounted on a servo plug, to help mitigate against the "brown-out" problem they faced on some receivers. I wonder if one of those would make a difference on the X8R?

A large electrolytic, like Spektrum used, is probably the wrong thing for RF suppression, but it might be worth a try in the first instance. I would have thought a smaller value - around 1nF - would work best for high frequency interference, but since this problem is proving devilishly difficult to cure, it might be worth trying the big electrolytic first. Also probably simpler, because you might know someone who has one!

The capacitor connects the positive and negative supply rails, and should provide a short circuit to high frequencies, whilst blocking DC. This *should* stop interference breaking through via the power supply.

Electrolytics are very good at suppressing low frequency AC, but are quite inductive and become less effective as the frequency rises. For high frequencies, a small value ceramic plate capacitor is probably better.

I can probably supply one of each, mounted in plugs ready to use, if you want to PM me your address.

Cheers,

--

Pete

[Ron Gray]

Peter - you are correct that in one test the RX locked out when I passed the spark plug over the battery, the interesting thing was that only happened after I re flashed it! In fact after the recent flash the RX was picking up more interference from further away! But in answer to the question regarding the battery placement in the Sbach, in order to get the CoG I had to place both the RX and CDI batteries under the turtle deck behind the canopy, so they were as far away from the CDI as they could be! I plan to re flash the RX with an older firmware to see what happens.

[Peter Christy]

Ron: PM sent.

It may be enough that the two batteries were in proximity for interference to "jump the gap". I used to make CDI units for cars back in the 70s - putting 400V through the coil instead of 12! Believe me, they never failed to start!

The pulse energy drawn from the battery was awesome, and suppressing the interference to a car radio was not easy! I'm thinking that having the two batteries close to each other could be inducing spikes in the Rx voltage supply, leading to the lock-up.

Having said that, my only gasser is a scale Bell 47 helicopter, with the batteries next to each other under the cockpit floor. That uses a Spektrum DSM-2 receiver. I have lots of problems with that heli, but mostly related to the vibration from the engine, rather than anything to do with the radio!

--

Pete

[Andy J]

Posted by Ron Gray on 11/05/2019 09:53:07:

Andy, the issues that I have seen are purely down to CDI interference, the videos quite clearly show that. That is the reason why I created this thread. May I suggest that if you have seen issues with your setup that can be attributed to other things then it would be better to create another thread for that. This way things will be a lot clearer for anyone visiting the FrSky part of the forum.

Ron

Think you have missed the point I have tried to make. This is that the only item we have in common to an identical mode of failure is the Rx. Therefore the investigation needs to be focussed on what mode of failure the Rx is prone too.

This belief that the problem is one of radiated susceptibility may well be correct, but electromagnetic compatibility problems can be caused by two means which are radiated or conducted means.

I appear to have highlighted the issue that the Rx8 Pro power lines are far from clean. So as subsequent posts refer, perhaps the use of appropriate filtering on the Rx supply is necessary to get rid of additional noise created by the ignition system.

[Andy J]

Peter PM sent.

[Andy J]

Peter PM sent.

[Ron Gray]

Andy, but with identical setups for power supplies to the RXs I get no problems with either the RX8R Pro or the R9 RXs however this thread is really about the X series, in particular the X8R.

[Ron Gray]

Thanks to PeteC I have now carried out further testing but with different sized capacitors on the power lines, unfortunately the X RXs still locked out. So having now replaced the Sbach 342 with another Sbach, this time a 300 I have installed the following:

R9 RX running 868 Mhz

RX8R Pro RX running 2.4Ghz

Redundancy Bus 10

2 x 3000MaH 2s LiFe batteries for RX

1 x 2500MaH NiMh battery for ignition.

So belt, braces, jock strap the lot!

[Alan Hilton]

Hi Ron

How do you run both frequency’s on a Horus 12? I’m away from home at the moment and can’t look at my tx . Can you run an external module and the internal at the same time

Alan

[Andy J]

Looks a nice installation Ron.

I also tried using capacitance on the supply lines but it did not improve the supply line noise on a X8R or the RX8R which both exhibited around 130mv of volt drop as it transmits telemetry.

As a precaution have now converted all my models to 6.2V nimh of a type identical to yours to try to keep the Rx supply as low an impedance as possible.

Model flown with the RX8R last weekend without issue using this set up. Second model ready to be flown with the X8R, so will be interested to see if any problems come up. Interestingly I still get the odd loss of telemetry at very close ranges to the Rx of ~1-2ft though but it does not seem to be an issue and I certainly don't get lock outs.

Like your idea of running two frequencies though.

[Peter Christy]

First of all, a big thanks to Ron for his painstaking approach to this problem. Even more remarkable is his forbearance in the light of his loss.

However, at this point, I'm out of ideas! It looks as if the cdi interference issue is inherent in the design - or components used - in the X8R receivers.

I have to say that I've been using X?R receiver in non-ignition models without any issues at all, so it does appear to be purely down to a susceptibility to ignition noise. And to be fair to FrSky, they have put out a re-designed receiver specially for this instance.

Sorry I couldn't be of more help, Ron.

--

Pete