Let's compare MPU-6000 and L3GD20

1,153 views
Skip to first unread message

Nikolay Arsov

unread,
Apr 16, 2012, 3:41:36 PM4/16/12
to uavdevb...@googlegroups.com
Hi all,

Has anyone compared the MPU-6000/6100 ( Invensense ) and L3GD20 ( ST ). Aside from the 6000 motion processor, it seems ST gyro is not so bad...could say it is good enough. More, it's MRF is too close to 6000 one...>24kHz.

Dear developers, please place here your PRO and CONs.

Best regards

Nick
L3GD20.pdf

Peter Hollands

unread,
Apr 16, 2012, 4:11:40 PM4/16/12
to uavdevb...@googlegroups.com
Nick, I suppose we should also enclose the 6000 specification for easy comparison.

Each document is > 40 pages, so a lot to think about.

The Digital Motion Processor is irrelevant for our application, as we cannot program it.

Best wishes, Pete
PS-MPU-6000A.pdf

Nikolay Arsov

unread,
Apr 16, 2012, 4:22:19 PM4/16/12
to uavdevb...@googlegroups.com
Hi Pete,

Thanks for placing the MPU-6000 datasheet. I'd just forgot to. About the DMP....I'm not interested in and suggest not to discuss it.
I'm interested in comparing the two 3 axial gyros as I'm still digging into that field....

Best regards
Nick

William Premerlani

unread,
Apr 16, 2012, 7:00:11 PM4/16/12
to uavdevb...@googlegroups.com
Hi Nick,
I am desperately looking for analog gyros. Do you have any news?
Best regards,
Bill

Nikolay Arsov

unread,
Apr 17, 2012, 1:50:09 AM4/17/12
to uavdevb...@googlegroups.com
Hi Bill,

There are some good and some bad news about the analog gyros as follow:

A/ Invensense - just YAW axis - ISZ-505.....bad news;

B/ ST - L3G3250A - good news - will be available...bad news - do not know when. I have talked to their technical manager....he doesn't know too.. :-(

C/ Sensor Dynamics ( now Maxim ) - SD740 - a good gyro with analog, I2C, SPI output. ( datasheet, schematics, sample code ) attached. I talked to them ( Maxim ) too, but they do not know whether and when they will reestablish the manufacture of this gyro.

In other words, bad news are more than good ones. So sorry.....I'm in a black hole too.

So, I'm waiting just one word from you - "DIGITAL" , which will take me out of the black hole.... :-)

Best regards

Nick
SEN0095.zip

William Premerlani

unread,
Apr 18, 2012, 3:59:22 PM4/18/12
to uavdevb...@googlegroups.com
Hi Nick,
It looks like the handwriting is on the wall...the diydrones UAV world is moving to digital gyros. There is a good chance that the next version of the UDB will have the digital gyro/accelerometer from Invensense, with the SPI interface.
Best regards,
Bill

Nikolay Arsov

unread,
Apr 19, 2012, 1:27:00 AM4/19/12
to uavdevb...@googlegroups.com
Hi Bill,

I'll finish the preliminary design of AUAV2 and post the schematics in the thread - AUAV2(TINY UDB5) discussion.

Best regards

Nick

Claudio Carbone

unread,
Apr 22, 2012, 7:00:28 AM4/22/12
to uavdevb...@googlegroups.com
Il giorno mercoledì 18 aprile 2012 21:59:22 UTC+2, William Premerlani ha scritto:
Hi Nick,
It looks like the handwriting is on the wall...the diydrones UAV world is moving to digital gyros. There is a good chance that the next version of the UDB will have the digital gyro/accelerometer from Invensense, with the SPI interface.
Best regards,
Bill
 

Been doing a couple calcs regarding the above mentioned question.
To read each sensor 8000 times a second, you have to:
write the address, write the register, read 2 bytes for every acc axis
write the address, write the register, read 2 bytes for every gyro axis.
 - This is assuming the 6 registers holding high-byte and low-byte values for each accelerometer are consecutive as well as those of the gyroscopes.
In the case of the Invensense parts, that is true.
-

This means each read effectively is 16 bytes long.
Now assuming you can chain these without delays, this accounts for a complessive read throughput of
16 * 8 * 8000 = 1024000 bits/second
way above the fast I2C specs and near the maximum speed of SPI!

So at the moment there is no Invensense part capable of sustaining this thourghput (both MPU6050 and MPU9150 have I2C at 400KHz) apart fom maybe the MPU6000 which nobody has tested.
Moreover it is also a problem of interface: nearing the real megabit of throughput represents a taxing environment for any low-level MCU interface.
Most SPIs are rated at this speed as maximum, while there are almost no parts supporting megabit I2C (apart from I2S parts maybe).

Anyone got a look at the ST LY4150 pitch/yaw?

Claudio



Nikolay Arsov

unread,
Apr 22, 2012, 10:46:21 AM4/22/12
to uavdevb...@googlegroups.com
Hi Claudio,

Yes we had considered all ST gyros with Bill, but their common problem is low mechanical resonant frequency....as low as 4kHz....bad. The only two interesting gyro with high MRF is L3G3250, but the problem is it is not in production and ST doesn't know if it will be.

Best regards

Nick
Reply all
Reply to author
Forward
0 new messages