Find IMU based on previous works

3.38K viewshardwareaccelerometer gyro hardware imu
0

Hi,
I’m building a quadrotor robot and my supervisor has asked me to make a report of the IMU modules used in previous works by teams or individuals, I couldn’t find such reports however. Does anyone know how to find such reports or how to find out what modules they’ve used? It’s really urgent. Thank you all.

Anonymous edited answer March 9, 2018
0
Anonymous 0 Comments

Great Post!

The IMU that you seek is the MPU 6000 as it is highly regarded as the very best IMU for Racing Drones…. Low noise, Great responsiveness, and good flying overall, it it can sample the gyro at 8kHz via an SPI bus, and also has great vibration tolerance.  These are a few attributes that come to mind.

This- however, has been subject to great debate in the FPV community.  For one, I think that a lot of it is a downplay of hype.

The MPU6000 for instance….. it’s an outdated, discontinued IMU (gyro) from invensense for quads and there is a limited supply of these units left in the world.

From the manufacturer, The MPU 6000 features a:

3-axis gyroscope, 3-axis accelerometer in a 4×4 package. The MPU-6000 is the world’s first 6-axis I²C MotionTracking device designed for the low power, low cost, and high performance requirements of smartphones, tablets and wearable sensors. The MPU-6000 is comprised of two parts with features summarized in the table below. To save space, the package size of the devices has been driven down to a footprint of 4x4x0.9mm (QFN).

 

The MPU6500 is newer than the above IMU sensors, but is not as popular. This IMU allows up to 32kHz updates from the gyro, and the gyro is more sensitive. However, most FPV pilots have found that they need to soft-mount this IMU to get better performance because it is too sensitive. Some have said that the MPU6500’s reputation for poor vibration tolerance is due to the flight controller not utilizing the full 32Khz gyro update rate-  So it’s a good idea when running an MPU6500 to use a flight controller that allows you to get readings at the full 32Khz sample rate.

This is probably due to the fact that it uses the I2C protocol, which isn’t usually used on flight controllers because I2C is a bit slow for flight controller sensor inputs and it doesn’t utilize the the gyro at a sample rate of 8Khz.

In Fact, The MPU 6000 IMU has been featured on very many flight controllers starting with the Naze32 rev 5.  It was actually regarded as a downgrade to move from the Naze32 Rev. 5 to the Naze32 Rev 6. due to the fact that it changed from the MPU 6000 to the MPU 6500.(regarded as an upgrade when biasing opinion from the spec sheet)

However, Looking again at the specs, a good engineer might throw out the MPU6500 for the newer, better? ICM20608 IMU.  This IMU was featured on the DTFc by OSDoge- A highly regarded FC from a highly regarded FC engineer….And a great flying board.

DTFc Flight Controller F3 6S Capable By DTF UHF dtfc

 

The DTFc features an upgraded ICM20608 IMU from Invensense https://www.invensense.com/products/motion-tracking/6-axis/icm-20608-2/

You would think this is the greatest thing since sliced bread because it has, compared to previous 6-axis devices, lower power consumption, lower noise, and a thinner package. The ICM-20608-G offers a duty-cycled operation mode for the gyroscope, that reduces gyroscope power consumption by half or more (depending on ODR) compared to earlier 6-axis devices. It also provides about 20% lower noise than previous devices, and about 17% thinner package compared to previous devices.

However, the FPV pilot simply hasn’t caught on to the newer tech.  There were some early opinions about the gyro on these newer FCs that really hurt these advanced Flight Controller products and kept the designers from advancing further.

Heck, even the newer Betaflight F4 board has an IMU6000 gyro-  This is OLDER TECH.

I guess this comes as a sore subject because even after writing the reply I am left to ask the question if it even really makes a difference, I mean the ICM20608 seems to fly fine over the MPU6000 at 32KHz IMO, but many pilots tend to disagree (could this be due to lack of understanding of the relationship between gyro update frequency and processor speed?)

The newer, ICM20602 seems to be the newer standard.

ICM-20602

 

The ICM20602 is one of the newest IMU’s from InvenSense suitable for use on FPV drones and it features much more accurate sensors with less noise. It can also read the gyro at 32Khz which makes it sutable for future use with STM processors F4-F7.    The Omnibus F4 V3, and SPF4 Evo flight controller feature this gyro and it’s a good indication that future Flight controllers will feature it as well as the other chips that are planned to be introduced with the latest smartphone releases in the coming times.

It’s an interesting note to point out that a lot of pilots have begun soft-mounting their FCs in an attempt to dull down the gyro and smooth out the performance. Could this mean that the newer tech is too responsive for the FPV pilot? I think so… Perhaps the newer IMU units show your true skills, and that’s why FPV pilots keep reverting to the MPU 6000- because it isn’t as precise, it isn’t new. and there is a bit of error in it’s output. After all, the MPU6000 is an old dog.

Regardless,  I think that you should start  with an MPU6000 , you might find that some of the newer IMUs work better for your application, but you will have to look at the spec sheets from Invensense and go from there.

One final thought, if the MPU 6000 is the best fit for the FPV Racing drone, could this mean an early demise for FPV racing drone technology as the Invensense IMU 6000 is no longer in production which means a limited supply remains.  Once they are gone, and the mobile industry continues building smaller, lower power, higher sensitivity Gyros that don’t work quite as well on racing drones, will the Racing drone pilot be left with the choice of using NOS (NEW-OLD-STOCK) to get the edge (or lack thereof) he is looking for? Perhaps a bit of software dampening will be required to help make up for the fact that humans are actually piloting these devices? Or maybe we simply need to ask Savitzky-Golay how to dampen our inputs.

Anonymous edited answer March 9, 2018

Question stats

  • Active
  • Views3375 times
  • Answers1 answer