cancel
Showing results for 
Search instead for 
Did you mean: 

LSM9DS0 Accelerometer noise

Erich Neumann
Associate II
Posted on January 17, 2017 at 19:16

Hi.

I am using the LSM9DS0, and have observed a lot of noise on the z-Axis of the Accelerometer Data. The attached picture is a plot showing the noise while operating at 100Hz in the +-2g range. Is this a normal level of noise for this device, or am I doing something wrong, or is the device defective?

kind regards.0690X000006064IQAQ.png

Erich Neumann

#lsm9ds0 #accelerometer #noise
9 REPLIES 9
Oliver Beirne
Senior
Posted on January 17, 2017 at 23:00

Hi

neumann.erich

I have moved your thread to the 

https://community.st.com/community/mems-sensors-community/mems-sensors-forum?sr=search&ampsearchId=7c0ad53c-0ba8-464a-8ec1-0ec9a544bd8e&ampsearchIndex=1

where someone should be able to assist you.

Thanks

Oli

Miroslav BATEK
ST Employee
Posted on January 18, 2017 at 10:57

Hello Erich,

can you please share the values in text form or attach an excel file, so I can check it.

Posted on January 18, 2017 at 11:20

Hello Miroslav,

sure, no problem. here is the values in text form:

0,9606323242

0,9723510742

0,9658203125

0,9698486328

0,9688110352

0,9644775391

0,9722900391

0,9704589844

0,9642944336

0,9791259766

0,9736328125

0,9665527344

0,9619140625

0,9720458984

0,9658203125

0,9647216797

0,9718017578

0,9605102539

0,9632568359

0,9747314453

0,9681396484

0,9627685547

0,9709472656

0,9575195313

0,9515380859

0,9769287109

0,967956543

0,9680175781

0,9744262695

0,9681396484

0,9664916992

0,971496582

0,9672851563

0,9697265625

0,9570922852

0,973449707

0,9660644531

0,9713134766

0,9621582031

0,9642944336

0,9688110352

0,9682617188

0,9688110352

0,9709472656

0,9696044922

0,9600830078

0,9566040039

0,9677734375

0,97265625

0,9761962891

0,9678955078

0,9717407227

0,9627685547

0,9721069336

0,9733276367

0,9577026367

0,9709472656

0,9569091797

0,9683227539

0,9697875977

0,9760131836

0,9555664063

0,9730224609

0,9582519531

0,9653930664

0,9691162109

0,9626464844

0,9694824219

0,9642944336

0,9615478516

0,9675292969

0,96875

0,9627075195

0,9727172852

0,9649658203

0,9642333984

0,9641113281

0,9597167969

0,9580078125

0,9566040039

0,9693603516

0,9730224609

0,9674682617

0,963684082

0,9678955078

0,9581298828

0,9624633789

0,9723510742

0,9647216797

0,9611816406

0,9743041992

0,9681396484

0,9609985352

0,9708251953

0,9694213867
Nejc Okorn
Associate II
Posted on January 22, 2017 at 13:41

Hi

oliverbeirne

‌,

Batek.Miroslav

‌,

I have the same issue as Erich, in fact I can confirm, that some of the devices using this sensor does not have any noise or should I say there is expected noise and the other devices have the same noise as presented by Erich. Ihave tried 10 different sensors and at least every second device has this noise issue.

Pictures from the one without the noise and the one with noise below.

0690X000006066OQAQ.png

0690X000006066nQAA.png

Unfortunately we discovered this issue right before the release date of the device, because we were only testing it on one device...

I hope someone will be able to help us, because we really do not want to switch the sensor before the release.

Best regards,

Nejc

Miroslav BATEK
ST Employee
Posted on January 24, 2017 at 11:43

Hello Erich, Nejc,

I received information that the typical accelerometer noise density is 150 ug/sqrt(Hz) 

@FS=2g and ODR=100Hz.

Your values seems to be much higher.

Here are a few hints:
  • Check power supply and blocking capacitor, noisy Vdd has affect to the output values

  • Read data only if DRDY interrupt is trigerred, do not poll status register for new data

  • Please check there are really no surrounding vibration when you measure the noise in still state

Posted on January 24, 2017 at 18:35

Hello

thanks for the replay.

I have to say that we already did everything we could for first and last suggestion you made to get the clearest input voltage as possible and I was measuring the input voltage myself, but could not see any noise.

There is no vibration and the screen shotsI provided in previous post were taken at the same time and the devices were on the same surface.

I don't see why the second suggestion would be any different? But Iwill try this and will report back.

Thanks,

Nejc

EDIT:

The results are the same, but I did get a lot better results using +/- 2G scale option.

Anyway the results on +/- 2G are still not as good as on +/- 8G measuring outcome onthe device that works fine(picture from the previous post with good results)

________________

Attachments :

Screenshot from 2017-01-25 17-47-33.png : https://st--c.eu10.content.force.com/sfc/dist/version/download/?oid=00Db0000000YtG6&ids=0680X000006Hyut&d=%2Fa%2F0X0000000bEf%2FPNYc3VDpBHujylJjkxsieUj2jTZ7ierM0qX6hB9VpFo&asPdf=false
Posted on February 01, 2017 at 13:41

Hello,

I have checked the Supply voltage and blocking capacitor, have switched to using the data_ready to trigger the readout of the chip and checked for vibration being mistaken as noise. these measures did not help.

then we replaced the chip with a new one, and the noise is much lower, as seen in this image:0690X00000606CHQAY.png

I also observed that the noise level is highly dependent on the supply voltage, as seen in the following image. is this normal?

0690X00000606CMQAY.png
Posted on February 02, 2017 at 11:06

Hello Erich,

thank you for the data, I discussed the issue with my colleagues, unfortunately they never had this kind of problem with LSM9DS0.

I got only one advice to check that the Block Data Updated (BDU) is enabled (in CTRL_REG1_XM register).
Posted on February 02, 2017 at 22:40

Batek.Miroslav

‌ Looks like we might get this chip to work properly. We took your advice and set BDU on

CTRL_REG1_XM and you can see the improvement on the image below.

neumann.erich

‌ this is something you should also try.

0690X00000606DoQAI.png

I have some other questions if you can help me out.

We took our device out on the field to test it out. The device purpose is to be exposed to different altitudes from around 300m ASL to 3500m ASL. We are suspecting that there is different accelerometer and gyroscope offset in x,y,z. Because the device is not behaving properly when altitude rises. Unfortunately we can not say that for sure, because we will have to take our notebook with us to make some measurements(device is displayless).

Is it common for offset to change on different altitude? Does pressure affect accelerometer offset?

I am unable to find in datasheet or anywhere on web information whether accelerometer in

LSM9DS0

iscapacitive orpiezoelectric?

Thanks,

Nejc Okorn