Armchair Detective Required

[Andy J]

Dug the battery out today just to check what is was and it is indeed a 4 cell 2500mah pack.

 

Turned on the power and telemetry read 5.3V fore the battery volts which fell to 5.1v as I put resistance on the elevator servo arm. Found it impossible to stop the servo moving so don't think it is a dodgy servo or voltage issue.

[Andy J]

will check fail.

[Andy J]

Well its not a fail safe issue. Throttle goes to low and elevator and rudder stay at neutral. Will do a further check just to see if the elevator stays at the last commanded position.

[Andy J]

Nope, fail safe operates as intended with elevator and rudder returning to neutral on loss of signal.

 

Surprised no one has mentioned the mobile phone in my trouser pocket being an issue.

[kc]

I assumed you had the phone switched off at the flight line as recommended.   Did the phone ring at the crucial time?

 

A vintage model should just fly away if the signal is lost and the failsafe causing low throttle with rudder at neutral.   A nosedive sounds more like a clevis or structural failure if the Rx battery stands up to an extended test.                                      

[Andy J]

Certainly was not a failsafe issue, nor a clevis issue.

 

Could well have been a phone interference but would that have not been shown up on the telemetry feed? Certainly there are no incoming calls that day on the phone log. But as we all know the phone does periodically transmit data to the world and it could have been a message transmission.

 

Could the failure of the outer dihedral ribs which resulted in the TE joint be an issue?  Think this as unlikely as it would have resulted in a spiral and don't think the model turned in the dive of death.

 

[RottenRow]

20 hours ago, Andy J said:

Then without warning the model twitched and then immediately entered a vertical dive. At this point the throttle was closed and full up elevator was applied which appeared to have zero effect.

 

Are you saying that just the application of full up elevator had no effect, or that closing the throttle had no effect too?

 

Are the times shown in column B in the format minutes.seconds.tenths of seconds? Which makes the flight about 8 minutes long before the crash. Right?

 

Also, what is column E Alt(m)? I was thinking altitude but the model wouldn't have been flying at 343 metres.

 

Brian.

[Andy J]

It was the application of full up elevator had no effect. Assume throttle closed but cant be sure as the motor is very quiet and I am part deaf.

 

Time is in hours, minutes, seconds. Flight was probably a bit shorter as the time starts prior to engine start.

 

Think you can ignore Alt as the Rx does not have a pressure sensor.

[MattyB]

4 hours ago, Andy J said:

Dug the battery out today just to check what is was and it is indeed a 4 cell 2500mah pack.

 

Turned on the power and telemetry read 5.3V fore the battery volts which fell to 5.1v as I put resistance on the elevator servo arm. Found it impossible to stop the servo moving so don't think it is a dodgy servo or voltage issue.

 

IMHO that is not a test that proves a whole lot. You need to do a full monitored discharge cycle on that pack at a representative pack to see how it performs, only then can the battery itself be discounted as a potential root cause.

 

4 hours ago, Andy J said:

Surprised no one has mentioned the mobile phone in my trouser pocket being an issue.

 

3 hours ago, Andy J said:

Could well have been a phone interference but would that have not been shown up on the telemetry feed? Certainly there are no incoming calls that day on the phone log. But as we all know the phone does periodically transmit data to the world and it could have been a message transmission.

 

You have misunderstood how phones are (alleged) to impact a transmitter. They do not cause direct interference that is impacts the receiver; if they did flying on 2.4GHz would be impossible, as your RX "sees" hundreds of similar devices (phones, wifi routers etc) transmitting at a similar strength on the same band during the average flight. No, the supposed way they cause crashes is to permanently scramble the setup in your TX, altering things like servo directions or over-writing the bind information. If such an issue had occurred here you wouldn't have been able to re-establish control using the same model memory immediately afterwards, hence why we can rule out your phone 99.999%.

 

PS - The reason I use "alleged" in this post is that I am still yet to see any actual evidence of this happening; yes it's a BMFA recommendation, but in decades of flying in club environments I've never seen it actually happen in a way that can be directly linked to a phone. IMO the risk of flying powered models is far greater without a phone close by than with it, so I will continue to leave mine on in the pits thanks!

 

Another thread on this topic from a while back...

 

 

Edited February 16, 2023 by MattyB

[John Muir]

Certainly not the battery as telemetry reports the voltage at around 5V the whole time. The RSSI trace is very odd in that it starts out around 90 for a couple of minutes then increases to well over 100 when I assume the model was in flight, dropping back to what I would consider a more normal 60-70 after another four minutes or so. Then it drops a little more before the flight ends. There aren't any flat lines to indicate a loss of telemetry reception so the receiver was active.

Be interested to know what receiver you are using? What RF firmware is on the receiver? Is it properly matched to the RF firmware in transmitter (not the Open Tx version but the firmware from FrSky that you flash to the transmitter and receiver). Over 100 in flight seems strange to me but might depend on whether you are using ACCST or Access.

I think the mobile phone is unlikely to be an issue. I often fly alone and keep my phone on in my pocket while I fly for my own safety. It has never caused any problems in years of using FrSky gear. Or any gear for that matter.

 

John.

[kc]

Getting back to Andy Gates graph - does the red trace show the actual Tx output?   It falls off a lot just before the crash then switch is turned off and then when switched back on its increased power after a few minutes rest.  ( The Tx is then much nearer the Rx of course but wouldn't the Tx output be the same?)  To me that suggests Tx battery benefitted by the rest and is therefore a suspect.    Or have I got that all wrong?

 

A 'soak' test of the TX and Rx with servos worked  for same time, then turned off for a few minutes and on again after the rest would seem the next test to see what the telemetry traces are then.   Repeated without the rest for comparison.

Edited February 16, 2023 by kc

[Andy J]

Was using the Tx to test the aileron connections on a model and did notice the Tx voltage was only showing 2 bars out of 4 but don't think the Tx voltage is an issue, but it is strange that it increases from 7.5 to 7.6V in the log file.

 

Anyone know what A2(V) is?

[MattyB]

38 minutes ago, John Muir said:

Certainly not the battery as telemetry reports the voltage at around 5V the whole time. The RSSI trace is very odd in that it starts out around 90 for a couple of minutes then increases to well over 100 when I assume the model was in flight, dropping back to what I would consider a more normal 60-70 after another four minutes or so. Then it drops a little more before the flight ends. There aren't any flat lines to indicate a loss of telemetry reception so the receiver was active.

Be interested to know what receiver you are using? What RF firmware is on the receiver? Is it properly matched to the RF firmware in transmitter (not the Open Tx version but the firmware from FrSky that you flash to the transmitter and receiver). Over 100 in flight seems strange to me but might depend on whether you are using ACCST or Access.

 

RSSI only stays above 100 for around 15 seconds, so I suspect that is just the TX being at very close range initially after the model was initially powered up. However, I do agree the RSSI trace later in the flight when paired with the RX battery voltage (which does seem pretty steady) is rather weird - it does imply that (for some reason) the transmitted signal fromt he tX reduced over the course of the flight. I agree with kc that a "soak" test on TX and RX may be in order.

 

36 minutes ago, kc said:

Getting back to Andy Gates graph - does the red trace show the actual Tx output?   It falls off a lot just before the crash then switch is turned off and then when switched back on its increased power after a few minutes rest.  ( The Tx is then much nearer the Rx of course but wouldn't the Tx output be the same?)  To me that suggests Tx battery benefitted by the rest and is therefore a suspect.    Or have I got that all wrong?

 

Red = RSSI = Received signal strength indicator. Essentially it is the strength of the TX signal that the RX is seeing at any given point. 

 

13 minutes ago, Andy J said:

Was using the Tx to test the aileron connections on a model and did notice the Tx voltage was only showing 2 bars out of 4 but don't think the Tx voltage is an issue, but it is strange that it increases from 7.5 to 7.6V in the log file.

 

Not really - off load (i.e. after you switch the TX off to recover the bits) the voltage will recover a bit, so there is nothing in that particular part of the trace I would say is suspicious.

 

13 minutes ago, Andy J said:

Anyone know what A2(V) is?

 

Long answer - It's the old analogue telemetry port used in D series receivers to monitor (say) the voltage of your main lipo in an electric model. If nothing is plugged into it it will just give a semi random value (it was always rather complex to use as you needed to configure voltage dividers to get the right values - tbh I've forgotten how to do it, it was a right royal pain!).

 

Short answer - It's nothing that will really help with troubleshooting this particular issue.

Edited February 16, 2023 by MattyB

[Andy J]

Rx is a Rx6R. Just checked if I had anything in the Tx to programme a 6R but found nothing. Note I did download the file FW-RX4R-RX6R-c2.1.1 back in Sept 2020 on my PC but that does not stack up with versions listed on the FRSKy site for the Rx.

[Mike Blandford]

When the incident occurred, you turned the Tx off. After recovering the model you then turned the Tx back on.

Was the Rx powered all the time, and you then found the elevator was working correctly or did you power the Rx off and back on?

 

Mike

[Andy J]

Yes the Rx stayed powered Mike.

[RottenRow]

Certainly the variation in the RSSI level, apparently in three steps (seen on the graph), might be indicative of an issue, unless it can be explained by the model being flown significantly further away for some parts of the flight compared to others. However, if the received signal dropped to an insufficient level then I would have expected the failsafe to come into play, which Andy has verified works (throttle closed and controls at neutral). A Playboy type model would enter a reasonable glide at that point. Certainly not a vertical dive.

 

Also interesting are the dips in the receiver battery voltage, seen in the graph but even more obvious if you load the data for the first flight only into the Log Viewer in Companion, which expands the x-axis. I would suggest that this might indicate something limiting the current to the receiver... the battery voltage is measured there, not at the battery itself. The graph below shows battery volts in green and elevator movement in red. There seems to be an awful lot more elevator movement in the latter part of the flight, but note the dips in battery volts at around the same times. Perhaps the graph raises more questions than it answers...

 

 

I would inspect very closely the receiver battery connections, both those on the battery itself and the connector used to power the receiver. Perhaps power everything up and thoroughly wiggle all joints. Look into all of the connectors with a strong magnifier. Pull the wires in the connectors to check for poor crimping.

 

For model where the battery is hidden, I always write the battery type on the model in a convenient place, such as on the wing seat. That way I know what current I should charge it at (I always charge at constant current, 0.1C).

 

Other possibilites could be some kind of airframe failure, or significant weight shift within the fuselage. A Playboy, being a stable model, would be quite difficult to get to maintain a vertical dive I would have thought, unless something had broken.

 

Presumably, being a 3 channel model, the rudder servo is plugged into the aileron channel on the receiver, and the rudder Rx output is unused?

 

Can you recall how long after the model hit the ground you turned off the transmitter?

 

Brian.

Edited February 16, 2023 by RottenRow
Incorrect model name.

[Andy J]

Think from inspection of the log the actual model started its short taxi and take off run at around time 13:31:31. Impact was at around 13:34:09 when I let go of the up elevator demand. Tx was turned off at 13:34:24.

Rudder is indeed on the aileron channel.

Nothing came adrift in the fuselage to change the cog position. The only damage was to the outer dihedral breaks on both sides of the wing with the rear section ribs around that area completely destroyed. Main spar and LE remained attached though.

[john stones 1 - Moderator]

There any rudder (roll) when it did vertical nose down, feels very much like battery or loose plug in issue to me.

[Mike Blandford]

There have been a small number of NFL failures reported over some years (NFL = No Failsafe Lockout, there is a thread on RCGroups about them). Some of these have lasted a few seconds, some 20-30 seconds and some for a few minutes after which time control was recovered. There is some evidence that power cycling the Tx recovers the link.

As these have happened very rarely, there is insufficient information available to locate where any fault is.

There is also some evidence that telemetry continued to operate while there was no control.

 

I did manage a test using a XJT module in a Taranis (running erskyTx firmware) where I stopped the Tx sending data to the XJT, but kept the module powered. The module continued to send the last control information it had.

 

Mike

 

[Martin McIntosh]

This is strangely similar to what happened to me with very well flown and tested JR gear.  I throttled back my giant Limbo dancer at height soon after take off. It immediately put its nose down and I had no control whatsoever. Needless to say that no fault could be found with any of the gear which was subsequently flown again. Same a little later with, as it happened, my 1998 Ripmax LD which did the same thing but this time on full power. No failsafe operation and no faults found with the gear later.

In each case I stripped down the switch, Rx and all servos which were then examined under a high power magnifier. No battery or connector problems either.

My patch was known to have problems with a nearby gas pumping station maybe causing telemetry induced interference on 35mHz but not 2.4.

[Frank Skilbeck]

Just a thought, could it have been a transmitter hiccup which was cleared when you did the power cycle of the Tx. MPX had this issue with an update on the Profi, it only happened when certain telemetry sensors/values could cause the Tx to lock up, so it was very rare (but even so MPX issued a warning to stop using that release).

 

So the Tx program gets frozen and sends the wrong positions to the receiver, hence the vertical dive, but when you reboot the Tx it clears the fault.

[Brian Cooper]

A mobile phone is unlikely to cause a total Tx failure. . . . BUT for those who are sceptical about a phone causing issues, here is a little cautionary tale: 

 

I was doing the first flight with a Wot-4XL for a clubmate. 

It was flying nicely and only needed a few beeps of trim to get it flying straight and level. 

Prior to handing the Tx to my friend, I turned the model so it would be heading away from us.

I pushed the stick to the right...... and the model banked to the left. 

I pushed the stick a bit further and the model banked even further to the left. 

Uh-ohhhh....... 

 

Both aileron channels had reversed.

Fortunately, I got it back down without crashing it but, mentally, it was a challenge. 

During the urgency of getting it down, I was also hoping no other channels would go crazy. 

 

The only thing we could attribute it to was my buddy's phone which was in his pocket. 

We fixed the ailerons. . . . and put his phone back in his car. 

 

No more problems, but my buddy lost confidence in his radio. 

 

The radio was Multiplex.

 

Be careful with your phones, guys.  It seems they can create some mischief in our transmitters. 

[Martin McIntosh]

On the first day that evening electric r/c flight was allowed at the Nats I was first to take off with my Cap 21. Immediate aileron reversal! Very embarrassing, and I probably did not even have a mobile at the time. Gear was a 6 ch Spektrum on DSM1 and previously very reliable.

[John Muir]

Andy, if you are using ACCST the most up to date RX6R firmware is 2.1.1. If on Access it is 2.1.7. Version 1 firmware on ACCST had a rare problem that could lead to the sort of problem you just had, an uncommanded servo movement (the NFL that Mike mentions). Ver 2 was meant to fix this. You need ver 2 on both transmitter and receiver. The RSSI looks ok if you were actually next to the model when it was over 100. I assumed that was part of the flight.

I'm not sure the telemetry tells us anything very useful, other than the receiver was still sending it when the plane crashed, so it still had power.

Faulty servo? A plane like that doesn't just dive for the ground without the elevator going down (assuming no structural problem). I once had a problem with a plane where the rudder servo went bonkers when I operated the ailerons. Turned out the aileron servo was a dud which worked fine by itself but created feedback in the circuitry which made the rudder servo misbehave. Took ages to work that one out and luckily, while still on the ground. Cheap servos though and not well designed. The retailer withdrew them from sale.