IMU Mount

The Rocketry Forum

Help Support The Rocketry Forum:

VultureSr

Member
Joined
Dec 7, 2021
Messages
13
Reaction score
1
Hi There!

Do you have experience with mounting IMUs in rockets so that no vibrations are transferred during the flight? Do you have any application examples?

Thanks in advance my friends!
 

OverTheTop

Well-Known Member
Joined
Jul 10, 2007
Messages
6,979
Reaction score
4,819
Location
Melbourne Australia
Have a look here. This is my Vertical Trajectory System:
FCresize.jpg

The mount is something that is available for consumer flight computers in model aircraft/drones.
 

OverTheTop

Well-Known Member
Joined
Jul 10, 2007
Messages
6,979
Reaction score
4,819
Location
Melbourne Australia
If I were to go the whole hog I would support it a little better. As it stands there are now more wires attached but they are spread around a bit more. I could also use thinner silicone insulated wires in lieu of the regular wires.

In a rocket there is more linear acceleration and I suspect a lot less constant vibration than is in an aircraft or drone.
 

VultureSr

Member
Joined
Dec 7, 2021
Messages
13
Reaction score
1
Thank you for link.
Do You use any vibration isolators? If so, how to choose not to suppress the signal from imu?
Doy You have any experience with imu mounting on flat syrface and and firmly fixed with screws?
Best regards
 
Last edited:

OverTheTop

Well-Known Member
Joined
Jul 10, 2007
Messages
6,979
Reaction score
4,819
Location
Melbourne Australia
The vibration mount has rubber grommets which are bolted to the deck. There is another frame that the grommets fit into that the flight computer mounts onto. The IMU mounts to the frame by double-sided tape and the grommets are bolted to the deck.

You could try fixing with screws directly to the deck but if there is sufficient high-frequency vibration it can degrade your measurements (and the results you get by integrating those). For the addition of some vibration mounts I wasn't going to take the chance.

I did an earlier version where the FC was mounted with thick double-sided tape that is what is normally used in drones for mounting the computers. I 3D printed a cage that mounted over that so any severe bumps would no tear the tape off the deck.
 

VultureSr

Member
Joined
Dec 7, 2021
Messages
13
Reaction score
1
Thank You for the answer.
Did You check measurements of IMU? What orientation angles did You receive in the flight?
 

neil_w

OpenRocketeer
Joined
Jul 14, 2015
Messages
13,690
Reaction score
6,928
Location
Northern NJ
In a rocket there is more linear acceleration and I suspect a lot less constant vibration than is in an aircraft or drone.
How much vibration is there in a typical rocket flight, anyway? I hadn't it thought about it but now I realize I have no idea. If there is vibration, what sort of frequencies?
 

VultureSr

Member
Joined
Dec 7, 2021
Messages
13
Reaction score
1
It seems to me that it depends on the type of rocket flight, speed, trajecetory and dynamics itself.
 

jderimig

Sponsor
TRF Sponsor
Joined
Jan 23, 2009
Messages
4,173
Reaction score
1,970
What vibration are you trying to suppress? The real vibration of the rocket or a false or independent vibration of the imu package?
 

OverTheTop

Well-Known Member
Joined
Jul 10, 2007
Messages
6,979
Reaction score
4,819
Location
Melbourne Australia
How much vibration is there in a typical rocket flight, anyway? I hadn't it thought about it but now I realize I have no idea. If there is vibration, what sort of frequencies?
I need to look at some old flight data and do the statistics on it. I don't have a feel for the levels although I suspect I know when they will happen based on various causes. Getting slammed by work currently so I might be able to have a look in January if I'm lucky.
 

jderimig

Sponsor
TRF Sponsor
Joined
Jan 23, 2009
Messages
4,173
Reaction score
1,970
Hi There!

Do you have experience with mounting IMUs in rockets so that no vibrations are transferred during the flight? Do you have any application examples?

Thanks in advance my friends!
  • If you have an ideal IMU you should mount it so the IMU has exactly the same motion as your rocket in flight, including vibrations. But not more motion, which means no slop in the mounting or the electronics bay or the sled etc.

  • However a real IMU has a finite sampling frequency and you do not want to pass any vibrations to the IMU greater than 1/2 the sampling frequency. Since most rockets have significant mass it is probably unlikely the rocket will have a very high frequency vibration.

  • If you think vibrations may be present in excess of half your IMU's sample rate then you can use an elastomer mount as shown in Over the Top's post to mechanically filter those high frequency vibrations out. However do not over-do that dampening because rocket vibration data can be useful to know.
 

Voyager1

Well-Known Member
Joined
Oct 23, 2015
Messages
816
Reaction score
297
Location
ACT, Australia
How much vibration is there in a typical rocket flight, anyway? I hadn't it thought about it but now I realize I have no idea. If there is vibration, what sort of frequencies?
It seems to me that it depends on the type of rocket flight, speed, trajecetory and dynamics itself.
I don't have any measurement data at the moment, but I do know that some hybrid motors can introduce significant on-axis resonance due to combustion instability (sounds like a bad case of flatulence!). This would need to be addressed if you were relying on IMU data when flying some hybrids. I do recall seeing this combustion resonance in accelerometer data in some of my earlier hybrid flights - I certainly heard it! Even some solid motors, e.g., sparkies, might display some combustion noise, too. There might also be aerodynamically-induced vibrations as well.
 

neil_w

OpenRocketeer
Joined
Jul 14, 2015
Messages
13,690
Reaction score
6,928
Location
Northern NJ
I don't have any measurement data at the moment, but I do know that some hybrid motors can introduce significant on-axis resonance due to combustion instability (sounds like a bad case of flatulence!). This would need to be addressed if you were relying on IMU data when flying some hybrids. I do recall seeing this combustion resonance in accelerometer data in some of my earlier hybrid flights - I certainly heard it! Even some solid motors, e.g., sparkies, might display some combustion noise, too. There might also be aerodynamically-induced vibrations as well.
What sort of frequencies are we talking about?
 

Voyager1

Well-Known Member
Joined
Oct 23, 2015
Messages
816
Reaction score
297
Location
ACT, Australia
What sort of frequencies are we talking about?
I don't recall the actual frequencies in my hybrid flights, but audibly they would have been anywhere from 50Hz to 100Hz or more, typically. It would vary from one hybrid motor to another. Even with the same motor hardware, a different grain could produce different results. Some documents do show spectral components of up to 1kHz.
 

DAllen

Well-Known Member
Joined
Jan 18, 2009
Messages
5,406
Reaction score
1,777
Location
SW Michigan
Soooooo what is the overall objective of said IMU? I find this interesting....
 

rocket_troy

Well-Known Member
Joined
Sep 9, 2013
Messages
417
Reaction score
214
Location
Melbourne, Australia
As I've mentioned in a hybrid thread, I purposefully run my hybrids unstable for the crowd-pleasing effect. Makes for a very ugly acceleration plot and whilst this is a non-useful extreme case for the purpose of this discussion, it's actually one of my smoother traces :)

acceleration plot.png


TP
 

Attachments

  • acceleration plot.png
    acceleration plot.png
    68.9 KB · Views: 10

jderimig

Sponsor
TRF Sponsor
Joined
Jan 23, 2009
Messages
4,173
Reaction score
1,970
As I've mentioned in a hybrid thread, I purposefully run my hybrids unstable for the crowd-pleasing effect. Makes for a very ugly acceleration plot and whilst this is a non-useful extreme case for the purpose of this discussion, it's actually one of my smoother traces :)

View attachment 493036

TP
A FFT of that data would be interesting....
 

OverTheTop

Well-Known Member
Joined
Jul 10, 2007
Messages
6,979
Reaction score
4,819
Location
Melbourne Australia
Even regular solid motors cause significant vibrations due to thrust instability. The last couple of shuttle flights were the most instrumented rocket ever flown due to additional logging on the SRBs. The ARES rocket had been designed to fly astronauts and they were worried the vibrations from the stretched SRB stage 1 would harm the astronauts.
 

UhClem

Well-Known Member
Joined
Feb 6, 2009
Messages
1,984
Reaction score
384
A FFT of that data would be interesting....
I did that once a while back. Digging through the archives I find plots for a K hybrid and M composite. They are like night and day. M energy concentrated at 0Hz but the hybrid spread out.
Kspectrum.gif Mspectrum.gif
Done so long ago that I barely remember it so details are lost. But since this data came from an RDAS with poor anti-aliasing filters...

Speaking of which, the RDAS let you configure the sample rate to 200SPS, 100SPS, or 50SPS. Unfortunately in addition to changing the rate at which data was stored it also changed the rate at which the apogee algorithm processed data. There was a notable hybrid flight way back when where the RDAS was set to 50SPS and the resulting aliasing caused it to have a very bad day.
 

rocket_troy

Well-Known Member
Joined
Sep 9, 2013
Messages
417
Reaction score
214
Location
Melbourne, Australia
I've never really bothered to FFT a hybrid either acoustically or via Cp Pc [I'm losing it] or via vehicle acceleration as acoustically at least, it's a function of not just internal chamber geometry and injector delta P, but also the flavour of the fuel.

TP
 
Last edited:

VultureSr

Member
Joined
Dec 7, 2021
Messages
13
Reaction score
1
  • If you have an ideal IMU you should mount it so the IMU has exactly the same motion as your rocket in flight, including vibrations. But not more motion, which means no slop in the mounting or the electronics bay or the sled etc.

  • However a real IMU has a finite sampling frequency and you do not want to pass any vibrations to the IMU greater than 1/2 the sampling frequency. Since most rockets have significant mass it is probably unlikely the rocket will have a very high frequency vibration.

  • If you think vibrations may be present in excess of half your IMU's sample rate then you can use an elastomer mount as shown in Over the Top's post to mechanically filter those high frequency vibrations out. However do not over-do that dampening because rocket vibration data can be useful to know.
  • I have a good mems sensor and magnetometer dedicated to flights
  • I am afraid that the vibrations of the rocket will be so large that they will be transferred to the sensor, which will disturb the measurement. This is very important to me because the INS algorithm is very sensitive to these vibrations and artificial rocket acceleration errors may arise due to high accelerometer noise. Wants to filter out the IMU most optimally
  • Are there any rules for selecting insulators to mechanically attach the imu to the rocket? Are they selected by trial and error?
 

VultureSr

Member
Joined
Dec 7, 2021
Messages
13
Reaction score
1
I don't have any measurement data at the moment, but I do know that some hybrid motors can introduce significant on-axis resonance due to combustion instability (sounds like a bad case of flatulence!). This would need to be addressed if you were relying on IMU data when flying some hybrids. I do recall seeing this combustion resonance in accelerometer data in some of my earlier hybrid flights - I certainly heard it! Even some solid motors, e.g., sparkies, might display some combustion noise, too. There might also be aerodynamically-induced vibrations as well.

Does this mean that if my IMU is at 400 Hz, it is best if the natural frequency of the rakirt is less than 200 Hz?
 

FredA

Well-Known Member
Joined
Jan 19, 2009
Messages
2,577
Reaction score
719
Does this mean that if my IMU is at 400 Hz, it is best if the natural frequency of the rakirt is less than 200 Hz?

In an IDEAL world that's all you get.....reality is less....read up on Nyquist.
 

jderimig

Sponsor
TRF Sponsor
Joined
Jan 23, 2009
Messages
4,173
Reaction score
1,970
  • I have a good mems sensor and magnetometer dedicated to flights

Which one?

  • I am afraid that the vibrations of the rocket will be so large that they will be transferred to the sensor, which will disturb the measurement. This is very important to me because the INS algorithm is very sensitive to these vibrations and artificial rocket acceleration errors may arise due to high accelerometer noise. Wants to filter out the IMU most optimally

The acceleration of the rocket launch is vibration, a big one. One cycle, broad frequency content. You want to capture that accurately, not over filter it. Any vibration frequency greater than the Nyquist frequency (1/2 sample rate) will be converted to garbage. Anything below is real response and should be preserved. I do not know what the INS algorithm is but real vibrations should be integrated correctly and not create error the processing the IMU. The integral of a sin wave over n periods is 0.

Most accelerometers have built in anti-aliasing (low pass) filters you can configure. An electrical filter and an mechanical filter are identical. I prefer the electrical filter because you know what you got. Below is an excerpt from the Analog ADXL372 datasheet.
In the absence of antialiasing filters, input signals whose frequency
is more than half the ODR alias or that fold into the measurement
bandwidth can lead to inaccurate measurements. To mitigate
this inaccuracy, a four-pole, low-pass filter is provided at the
input of the ADC. The filter bandwidth is user selectable, and
the default bandwidth is 200 Hz. The maximum bandwidth is
constrained to at most half of the ODR, to ensure that the
Nyquist criteria is not violated.

Check your IMU datasheet and configure it property and you should have no problems.
 
Last edited:

VultureSr

Member
Joined
Dec 7, 2021
Messages
13
Reaction score
1
Which one?



The acceleration of the rocket launch is vibration, a big one. One cycle, broad frequency content. You want to capture that accurately, not over filter it. Any vibration frequency greater than the Nyquist frequency (1/2 sample rate) will be converted to garbage. Anything below is real response and should be preserved. I do not know what the INS algorithm is but real vibrations should be integrated correctly and not create error the processing the IMU. The integral of a sin wave over n periods is 0.

Most accelerometers have built in anti-aliasing (low pass) filters you can configure. An electrical filter and an mechanical filter are identical. I prefer the electrical filter because you know what you got. Below is an excerpt from the Analog ADXL372 datasheet.


Check your IMU datasheet and configure it property and you should have no problems.

I have sensonor Stim300 with 80g range.
 

jderimig

Sponsor
TRF Sponsor
Joined
Jan 23, 2009
Messages
4,173
Reaction score
1,970
I have sensonor Stim300 with 80g range.
Looking at the datasheet I have a couple comments.

1. It looks like the module doesn't not have any anti-aliasing filter. So choose a high sample rate and maybe mount with an o-ring sandwich on the screws.

2. The maximum gyro rate of 400 deg/s is low for a rocket. If the rocket spins faster than 480 deg/s you will rail the gyro and lose ability to determine orientation of the rocket. Align your fins ahd hope your rocket doesn't spin.
 

VultureSr

Member
Joined
Dec 7, 2021
Messages
13
Reaction score
1
Looking at the datasheet I have a couple comments.

1. It looks like the module doesn't not have any anti-aliasing filter. So choose a high sample rate and maybe mount with an o-ring sandwich on the screws.

2. The maximum gyro rate of 400 deg/s is low for a rocket. If the rocket spins faster than 480 deg/s you will rail the gyro and lose ability to determine orientation of the rocket. Align your fins ahd hope your rocket doesn't spin.

Thank You,
I am trying to receive The data with frequency no less than 800Hz because this data are use in Kalman Filter to estimate orientation and next to estimate velocity and Position.
The rocket should be less spins than maximum range gyro.
 
Top