IMU sensor failed on NEO

Hi,

today I have made the first flight to my Soxos DB7 without any problem, but when I was going to do the second flight, once I have powered the NEO during initialization the swashplate has begin to jump up and down several times, with NEO blinking in red and the VBCt has told me there is a IMU sensor failed and to check the log files.

I have tried to power up several times with the same result, even bind the NEO to another VBCt of a colleague and get the same error.

I have attached the logfiles of the first flight and the next time I have connected with the error. Rest of the logs of the mentioned power ups shows the same error as the second one.

A friend of mine told me the error dissapear changing the rescue initialization method from precise to standard, and once I have change, it shows the warning "IMU Sensor Z Axis hangs" but after that it finish the initialization ok, but the swashplate still jump up and down several times each time I power up the unit. Is that normal?

Btw, I have tried to do a rescue calibration and it has result in an error, but I have click the rescue switch and it seems to works ok.

What should I do?

Regards,
JC



1 mal bearbeitet. Zuletzt am 14.02.2021 20:44 von Arag_.

Quote:

this is a issue based on some helicopters (especially TDR2 with a very stiff frame) and thier microvibrations. They can't be monitored - and can't be dampened in a known way.

We have tested this effect extensively on a TDR2 - and it was nothing that could be solved permanantely (i talk about 200 flights in each variation, i have a local buddy here who did all the extensive testing on this heli) when using another NEO, dampening or mounting position. It resists with the heli and its mechanical things like gear, bearings, motor, stiffnes of the frame, tail drive, dampenings etc.

So - unfortunately we can't do anything to solve the initial trigger effect, just watch the sensor and if it hangs - give a message. The 15g z-axis sensor is driven to its end due those microvibs and stucks there. It could possibly also release again in next flight or ground handling as it is mechanical.

So - what we do - and why we can (and others not):

We have 9 sensors inside the NEO - 3 rate sensors as for all VBars already for flight control, then 3 accelerometers with 15g range, then 3 accelerometers with 200g range.

The 15g accelerometers have a factory calibration and are used during power up to calibrate the 200g ones which have no factory calibration. Both sets are used in flight to align the "virtual rescue model" together with the rate sensors (they are mainly used but have always a bit drift) from time to time when flying will permit this.

Now - knowing that a heli has some vibrations in flight - we measured them and found that even a Logo with plastic frame has 13g, a 600 nitro has 43g, a gasser up to 85g vibrations. So we have the 200g sensors in use in flight and fade over to them when the heli starts as the 15g were at the end of their measuring range.

So - if the high range sensors are once calibrated during power up - the rescue is safe.

And here we started to work. We added a "standard initialization" to the rescue panel which gives us the possibility to decide on power up based on a hanging 15g z-Axis sensor. If it hangs - we calculate the z-Value out of x and y (which is very close to a measured result, no noticeable difference in flight when using rescue) - and calibrate the 200g with this result. The swash will bump 3 times then as info for the pilot. If it does not hang - it will do the usual init and bumps 1 time as known.

The only disadvantage now is - you can't turn on the heli more than 45 deg (!!!) off from its "skids down" levelled position - but this won't care 99,999999% of all pilots as we think.

So - using this initilization all is working and no reason for warranty claim or repair - it would possibly shift, but not solve the initial problem.

Sidenote: if soemone has pro and enables the lightness - this sensors are used terefore too. If the Z-Axis error exists on helis with Pro or Pro rescue - do NOT use the lightness, set it to 0 umless you have a smooth running helicopter.

regards (EN) / Gruss (DE)

Rainer Vetter

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
Posting and answering - please use a PC and a standard keyboard. Answer all questions fully, take the needed time, don't forget probably requested attachments.

Für Posts und Antworten bitte einen PC und eine richtige Taststaur benutzen sowie alle Fragen vollständig abarbeiten. Nehmt Euch Zeit, beantwortet Rückfragen in Ruhe und vollständig.
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
Für Einzelkommunikation bitte ausschließlich Email verwenden! - PN, Messenger, Chat usw. werden nicht bearbeitet.
For single user communication please send a email only! PM, Messenger, Chat etc. won't be read.

Hi Rainer,

and what about the jumps the swashplate make during initialization?

In here you can see a video of the process: https://youtu.be/hMQtX9mmP1o

I guess that behaviour is not normal.

Regards,
JC

You mean the 3 jumps?

Quote:

The swash will bump 3 times then as info for the pilot. If it does not hang - it will do the usual init and bumps 1 time as known.

regards (EN) / Gruss (DE)

Rainer Vetter

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
Posting and answering - please use a PC and a standard keyboard. Answer all questions fully, take the needed time, don't forget probably requested attachments.

Für Posts und Antworten bitte einen PC und eine richtige Taststaur benutzen sowie alle Fragen vollständig abarbeiten. Nehmt Euch Zeit, beantwortet Rückfragen in Ruhe und vollständig.
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
Für Einzelkommunikation bitte ausschließlich Email verwenden! - PN, Messenger, Chat usw. werden nicht bearbeitet.
For single user communication please send a email only! PM, Messenger, Chat etc. won't be read.

Yep, usually it jumps just once.

Rainer, so, why it jumps 3 times instead 1 time?
I don't feel confident with this NEO right now.

Hi,

3 jumps inform that the alternative initialization has taken place. Flying is safe.

But if your model affects the accelerometers (Rescue sensors), I would not trust Rescue without having checked:

Do a couple of flights, leave the Rescue Status screen open, and go to a hover or land, to check if the ball re-centers (or almost), and that the orientation shown is correct (normal/inverted).

If this is the case during the whole length of a couple of flights (3D and everything), using Rescue is OK.
If not, Rescue would go where the ball is, worst case, south.

Can't be helped, there are models which have more issues than others, even of the same brand and type.
Cause are most likely high frequency vibrations (from gear box), hard, asymmetrical.
The usual shaft rpm vibration (main rotor, tail rotor, motor) are less likely, in that order.

Solution can be, mount the NEO differently (position, orientation, different tape), use the other case type, plastic or aluminum, some pilots have reported that changing made a difference, but there's no recommendation which is better, plastic or aluminum.

Kind regards

Eddi

Born to fly ...
forced to work.

Eddi E. aus G. Wrote:
-------------------------------------------------------
> Hi,
>
> 3 jumps inform that the alternative initialization
> has taken place. Flying is safe.

Ok, understood.

>
> But if your model affects the accelerometers
> (Rescue sensors), I would not trust Rescue without
> having checked:
>
> Do a couple of flights, leave the Rescue Status
> screen open, and go to a hover or land, to check
> if the ball re-centers (or almost), and that the
> orientation shown is correct (normal/inverted).
>
> If this is the case during the whole length of a
> couple of flights (3D and everything), using
> Rescue is OK.
> If not, Rescue would go where the ball is, worst
> case, south.
>
> Can't be helped, there are models which have more
> issues than others, even of the same brand and
> type.
> Cause are most likely high frequency vibrations
> (from gear box), hard, asymmetrical.
> The usual shaft rpm vibration (main rotor, tail
> rotor, motor) are less likely, in that order.
>
> Solution can be, mount the NEO differently
> (position, orientation, different tape), use the
> other case type, plastic or aluminum, some pilots
> have reported that changing made a difference, but
> there's no recommendation which is better, plastic
> or aluminum.
>

The point is that this heli has 43 flights with this NEO as you can see in the photo uploaded with no problems. Also, I have paid 100€ to use the rescue option, so if the rescue mode does not works correctly, I guess the unit is damaged and should be replaced by a new one as it is still under guarantee.
By the way, the rescue calibration it does not works: VBCt shows a calibration error message on the screen trying to do, but the x and x coords on the calibration status screen it shows 0,0, so it seems the calibration is ok.
Also, I have noticed with the calibration status screen opened, if I move the helicopter x and y coordinates change the values, but the z axis is kept with the same value, at the top of the bar.

Finally, I will try the procedure you have explained above in order to see if the rescue is working or not.

Regards,

JC

Hello JC,

we hear you.

But experience with the NEO for almost six years now, own and in customer support worldwide, tells us something different than blaming the NEO easily.

In the beginning when this occurred, also about five years ago, back then with the Henseleit TDR II, we replaced NEOs by the dozen, took the suspect ones back for inspection.
But the thing happened again, with the new units, on the same models, most likely sooner than later.
The same units that showed issues on this particular type of helicopter worked flawlessly on others, and vice versa.
Then, we found out about what caused the IMU Z-Axis to stick or hang, and we found a way to circumvent the initialization error, and initialize safely nonetheless (Standard Initialization, only possible because the VBar NEO as the only flybarless system so far has two sensors built in).
We also found that this does not seem to cause permanent damage to the sensors.

On the other hand, we found out (like all the other manufacturers of flybarless with Rescue), that it is not bullet proof.
Adverse vibration can and will affect the accelerometers and the gyros, and no different than countless years ago, with the first tail rotor gyros and tail drift back then.
Back then, you paid a couple of hundred Euros for a sophisticated tail gyro, and still had drift, still had to try to get rid of vibration or mount the gyro in a way so it was better decoupled from vibration, different pads, double pads, metal plate in between, velcro strap around.
The physics have not changed, although there as been advancement with the sensor technologies available.
Nowadays, the gyros can hardly be brought to the limit (still, I know of one particular helicopter which can do it almost at command).
But the accelerometers can, because they are more sensitive, and need to be.

They need to derive changes in gravity within +/– 1 G with decimals, while the running mechanics throw up to—hold fast—160+ Gs at them, all time record seen and documented.

If you want, I can add a special App to your VBar Control Touch, which will replace the standard vibration analysis.
There, you can see the G forces affecting the accelerometers.
I would just need the serial no. please.
I'd be curious, too, because we have no readings from a DB7 yet, but also no bad reports yet.

If you monitor the Rescue Status Screen, you will notice that the ball moves x/y if you move the model, and if you put the helicopter on it's side or inverted, the z-axis indicator will also change.

If the z-axis sensor of the IMU sensor (one of the two accelerometers inside the NEO) sticks, calibration is prohibited, and for a reason: if the sensors are not centering freely, learning the position (calibrating a wrong horizontal position) would be dangerous.
In this case, you can gently knock on the case, to release the sensors (use a dull tool, grip of a screwdriver, lighter, tap not too hard but also not too gently), then power-cycle the unit.
But horizontal calibration is usually not needed anyway, the factory calibration is good, and the sensors do not seem to change noticeably, over time.

By the way, other systems like Brain 2 seem to log the same thing, vertical sensor error, and bring up some sort of warning, you are not alone, we are not alone.

43 good flights are nice, but they say exactly ... nothing. I was enjoying my car for thousands of kilometers, until the left rear tire bust, and for no obvious or discernible reason ... di-dum, did a 360 on a busy highway.

On the other hand, different ambient conditions (temperature) can affect the whole thing, everything's even less flexible then, which in turn will make it a better conductor for vibration.

Best regards

Eddi

Born to fly ...
forced to work.

Here is the serial number of my vbct 11000100
Please, install the advanced vibration analyzer app you mentioned above.
I am going to test the heli tomorrow morning with the guidelines you have provided me. I will let you know the results.
Regards,
JC

Just do a updazte.

the app replaces the "usual" Vib screen at the Model status panel.

regards (EN) / Gruss (DE)

Rainer Vetter

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
Posting and answering - please use a PC and a standard keyboard. Answer all questions fully, take the needed time, don't forget probably requested attachments.

Für Posts und Antworten bitte einen PC und eine richtige Taststaur benutzen sowie alle Fragen vollständig abarbeiten. Nehmt Euch Zeit, beantwortet Rückfragen in Ruhe und vollständig.
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
Für Einzelkommunikation bitte ausschließlich Email verwenden! - PN, Messenger, Chat usw. werden nicht bearbeitet.
For single user communication please send a email only! PM, Messenger, Chat etc. won't be read.

RV Wrote:
-------------------------------------------------------
> Just do a updazte.
>
> the app replaces the "usual" Vib screen at the
> Model status panel.

Sorry, but I have read this message today, so yesterday I fly the heli with the guidelines Eddi has provided me and I did a monitor of the vibration with the standard app in the VBCt. The maximum level of vibrations during the flight was 4000, but I don't know if this value is too much or not. I perform horizontal and vertical tic tocs, funnels and some other figures. The value of the vibrations were similar in motor, main rotor and tail rotor.

Also, I check the rescue monitor screen during the flight on a hover, and x and y coordinates stays almost centered. I tried to do a rescue on a hover, and it works with any problems (the heli begins to gain altitude until i retake the sticks, decreasing pitch). Also, I do another rescue with the heli drifting to the right and getting lower and it works too in the same way.

I forgot to check the if the orientation of the heli is show inverted during the heli is flying inverted.

I just update the VBCt in order to get the advance vibration app installed, so I can make another try next weekend.

Regards,
JC

Hello JC,

the overall vibration reading doesn't tell much, taken by itself.

It's an average which takes all the vibration into account, and it doesn't tell the source, as well as it's not meaningful once you do aerobatics or 3D, where you create vibration on purpose.
Also, a bounce on the ground when landing, or ground resonance, can and will affect the overall reading.

To use it properly, run the model on the ground, on a soft surface, no blades, check the graph.
Or hover, check the graph.
Only there, it will show you a sensible readout.

With this in mind, the tool can help troubleshooting, or comparing if a particular change makes a difference.

Same goes for the extended vibration analysis tool.

If you have a buddy with you who can monitor or capture the readout, all the better.
Otherwise, the Screenshot App with a switch assigned can help, that's what it's meant for.

If it was OK the other day, what could have been different? Ambient temperatures (still a suspicion of mine)?

The original vibration analysis only monitors the gyros. Of course, extreme vibration levels will also affect the accelerometers, but you can only infer from that.
The extended tool also shows the noise on the accelerometers, and that's what would be of interest, in this particular case

Best,

Eddi

Born to fly ...
forced to work.



1 mal bearbeitet. Zuletzt am 22.02.2021 09:55 von Eddi E. aus G..

Hi Eddi

Eddi E. aus G. Wrote:
-------------------------------------------------------
> Hello JC,
>
> the overall vibration reading doesn't tell much,
> taken by itself.
>
> It's an average which takes all the vibration into
> account, and it doesn't tell the source, as well
> as it's not meaningful once you do aerobatics or
> 3D, where you create vibration on purpose.
> Also, a bounce on the ground when landing, or
> ground resonance, can and will affect the overall
> reading.
>
> To use it properly, run the model on the ground,
> on a soft surface, no blades, check the graph.
> Or hover, check the graph.
> Only there, it will show you a sensible readout.
Got it! I will try this next time.
>
> With this in mind, the tool can help
> troubleshooting, or comparing if a particular
> change makes a difference.
>
> Same goes for the extended vibration analysis
> tool.
>
> If you have a buddy with you who can monitor or
> capture the readout, all the better.
> Otherwise, the Screenshot App with a switch
> assigned can help, that's what it's meant for.
>
> If it was OK the other day, what could have been
> different? Ambient temperatures (still a suspicion
> of mine)?
Well, the failure of the sensor was in the same day (14/02). The first flight without sensor error was at 9:37 AM and the second flight when the error arises was made at 10:06 AM (half an hour later than the previous one). The ambient temperatures, wind conditions and other parameters was also the same.
Once I have changed the initialization of the rescue to standard instead precision, I flew the heli last weekend with similar conditions (temperature is about over 10ºC and almost no wind like the previous weekend).

>
> The original vibration analysis only monitors the
> gyros. Of course, extreme vibration levels will
> also affect the accelerometers, but you can only
> infer from that.
> The extended tool also shows the noise on the
> accelerometers, and that's what would be of
> interest, in this particular case

I will use the new app next time. I have seen it is possible to record the registers, right? In that case I guess is not neccesary to get screenshots, right?

Regards,

JC

>
> Best,
>
> Eddi

Hi again,

Quote:

Arag_
Well, the failure of the sensor was in the same day (14/02). The first flight without sensor error was at 9:37 AM and the second flight when the error arises was made at 10:06 AM (half an hour later than the previous one). The ambient temperatures, wind conditions and other parameters was also the same.
Once I have changed the initialization of the rescue to standard instead precision, I flew the heli last weekend with similar conditions (temperature is about over 10ºC and almost no wind like the previous weekend).

Thing is, the sensor element in question can stick and release, so whatever it is when you land, is probably still there when you power up again.
If it was sticking the flight before, and if you tap the NEO to release the sensor, it can release between now and then.

But there should be log file entries (hangs/released). If none, then all OK.

Quote:

Arag_
I will use the new app next time. I have seen it is possible to record the registers, right? In that case I guess is not neccesary to get screenshots, right?

I would still take screenshots, or photos with the smart phone, for your own records.
The logs will be saved to the cloud, but they can not be reviewed on the radio after powering down the model.

Kind regards

Eddi

Born to fly ...
forced to work.



1 mal bearbeitet. Zuletzt am 23.02.2021 08:50 von Eddi E. aus G..

>
> But there should be log file entries
> (hangs/released). If none, then all OK.

All the times I have power up the NEO, there has been the hangs warning during the initialization, and the swash always jumps 3 times instead once.
I will tap the NEO in order to see if the sensor begins to work again.

>
> I will use the new app next time. I have seen it
> is possible to record the registers, right? In
> that case I guess is not neccesary to get
> screenshots, right?
>
> I would still take screenshots, or photos with the
> smart phone, for your own records.
> The logs will be saved to the cloud, but they can
> not be reviewed on the radio after powering down
> the model.

Ok, I will take screenshots also.
>
> Kind regards
>
> Eddi


Regards,
JC

I just try to tap the NEO and the error of the z axis has gone, and now it just jump the swashplate once.

I have choose precision mode for rescue instead the standard and the heli, on the ground, seems to works ok.

Maybe the sensor was jammered and with the taps it is alive again.

Regards,
JC

Hi JC,

Standard basically only spares you the tapping and power-cycling if the swash plate jumps multiple times, at power up.
And the triple jump informs you.
So why go for precision instead I wonder?
The result is not different.
And if other errors would happen, the NEO would still warn, it does not simply disable the warnings.

We could also hide this effect from customers, but then you wouldn't know until it was maybe too late, you'd just be lulled into a sense of false security.
Would save us all a lot of explaining and cost us a shrug if something happened, like ... uh, guess it happens sometimes, but we don't know why ...

—Eddi

Born to fly ...
forced to work.

Hi,

Today I have been able to do 4 flights on the heli (all of the lipos) and it has worked without any issues.

I have take screenshots of the vibration graph during the flight (attached in the message). The first image in the zip correspond with a hover at 1900rpm in idle 1.

Also, I have write the log but don't know how to find it on the Cloud. I have record first and second flight. This screenshots corresponds to the second flight (on the first one I haven't realized to click on Vibration Graph button).

So, I guess the NEO is now working right without any problems.

If you see any other thing in the plots attached, please let me know.

Best regards,

JC

Hello JC,

what we can see is indeed 1-sided (not like a sinus) load on the z-axis, between 30 Gs and 50 Gs.
Aileron and elevator is less, and more symmetric.

The primary and more precise IMU Sensor (accelerometer) inside the VBar is good for +/– 16 Gs (all axis), while the secondary 'high range' sensor is good for +/– 200 Gs (less precise, though).
So the vibration load on this model is very well suitable to cause the z-axis sensor to stick (the initial issue you reported), and to release as quickly, with vibration.
Means, it _can_ happen that it sticks at power off and subsequently at power on (here: use Standard Initialization so you don't have to worry).

And it _could_ happen that the virtual horizon tilts, or worst case, flips, over time.

This is now where I would monitor/check the Rescue Status Screen a couple of times, during a couple of flights, and if the ball always returns to center, right side up, all should be well.

Not good, but not terrible – like they said in Chernobyl
And the DB7 has no record as far as I know, which is good in my opinion ... there are other helis which pop up on the radar more frequently.
It even happens on Logos, from time to time ...

The graph can not be shown on the cloud, this is why I recommended taking screenshots

Best,

Eddi

Born to fly ...
forced to work.

Eddi E. aus G. Wrote:
-------------------------------------------------------

> This is now where I would monitor/check the Rescue
> Status Screen a couple of times, during a couple
> of flights, and if the ball always returns to
> center, right side up, all should be well.

When you said the ball returns to center, do you mean x and y axis, right? Because z axis is top of the bar or bottom, right?

>
> Best,
>
> Eddi



1 mal bearbeitet. Zuletzt am 02.03.2021 13:37 von Arag_.

Hi,

right. The ball is for x/y-axis (you can see it move if you tilt or pitch the model, in your hand, or the NEO only).
Z-axis changes if you roll the model/NEO over to inverted.

—Eddi

Born to fly ...
forced to work.

Hi I am having the same issue here had 2 flights on an OXY 5, but to start with before the first flight I noticed it was showing cyclic ring active yet it had not been flown so i unplugged it an tried again it said the same thing in the log the third attempt and it then stopped showing cyclic ring active.

I had 2 flights everything seemed fine on the third flight I connected the battery and it would not initialize in the log it showed Init failed retrying then IMU sensor Z Axis Hangs then IMU Sensor fail, Have seen on the forums this can happen and to change the settings in rescue mode, I don't have rescue just the pro version so I am unable to to adjust anything.

I have looked back in the logs and have seen a lot of Init fail retrying which i haven't seen on my other model logs, it was suggested a flick on the neo can help a stuck sensor (I thought these sensors were all semiconductor and not mechanical ) so tried flicking the neo and to my disbelief it worked.

I had two more flights no issue and cyclic ring active only showed during a flight. I have checked the vibration with the blades off an its less than 100 in flight max 2500, I have seen in the logs init failed retrying yet the model works, not sure if I should still fly this neo can you give me your thoughts?
It is also has the latest firmware 6.4.30 VBar Pro

Many Thanks

Phil

Hi Phil,

for whatever reason, Oxy helis seem to more frequently show these effects these days. I can only make an educated guess and think it's related to the gearbox/gear mesh combined with the mount/mounting position of the NEO.

If you have Pro, and if you use Lightness/Anti Gravity (feature of 6.4) the accelerometers are now also in use.
There is no differentiation for initialization with this.
Only thing you can do is disable (zero out) Lightness in all banks, then, the sensors will not be queried at power on, because they will not be in use.

The sensors are actually micro mechanical things, there are moving parts inside. Even if it's German, check out this video beginning at approx. 42', there are also schematics and graphs in it which may be enlightening.
Sounds like pre-war Russian Jeep technology, if it does't start, treat the starter motor with a hammer, but it's the way it is.
The z-axis sensor is like a tongue lolling between a floor and a ceiling —= and if vibration becomes too excessive (more than 16 Gs, for that sensor), it can touch and occasionally hang/stick there until it's released (again) by vibration, or by a knock or flick.

The normal vibration analysis in the NEO is useful for tracking down rather low frequency (shaft rpm) vibration (e.g. 2,000 1/min / 33 Hz main rotor or 10,000 1/min / 167 Hz tail rotor or 18.333 1/min / 306 Hz motor).
It can not, does not show the kHz (2.64 in this case) 'teeth chatter' vibration frequencies e.g. of gears meshing.
And these are the worst, they are hard, steep-sloped, not even like a sinus but asymmetric, like getting fierce slaps only on your right cheek all the time.
This makes filtering a challenge, causes the most sensitive sensor (z-axis) to stick temporarily, and in result it can make Rescue (if installed) dangerous.
With Lightness / Anti Gravity, the model might only feel awkward (not lighter but heavier if reversed), or it might veer off a bit when transitioning knife edge, in flips or aileron rolls, as if center collective was off.

Kind regards

Eddi

Born to fly ...
forced to work.



1 mal bearbeitet. Zuletzt am 09.04.2021 09:09 von Eddi E. aus G..