That's right, Nintendo's Motion Plus addon is out, and inside is a 3 axis, 16 bit gryo (well, a 2 axis gyro and a 1 axis gyro). Which is way better than the 2 axis, 10 bit one I was expecting. Communicates via I2C and costs less than $20. This could be pretty revolutionary for hobby IMUs. Coupled with a Nunchuck, its a 6DOF all digital sensor for $40!Details on the innards and protocol here.For the moment, it seems that you need to switch between the Motion Plus and the Nunchuck via a somewhat complicated method. Can't just read them both from different addresses. Hopefully that will be resolved when people better understand what's going on inside it.There's also some mention of it being able to run at two different sensitives toggled by some i2c command. And possible temperature compensation (they did this in software, only using the gyros?)Eventually I'll write some arduino code for the Motion Plus, but there are probably lots of you out there who could do it better/faster, so by all means give it a go!PS: Anyone know about progress on adapting the DCM method to ardupilot/arduino?
BTW..most A/D converters we have been using with autopilots are only 10-12 bit. 14 bit is plenty and I agree, offloading to the i2c device is better. What needs to be done (beside software) is a little board to mount the nunchuck and motion plus guts. Just 4 wires to it. i2c and +3.3 power.What a bargain for $35 and a small board to mount to for a 6DOF that's better than the Sparkfun 6DOF $125 unit IMO.
Xander: good point. We're pretty handy with Eagle and PCB fabbing at this point, so we tend to think around custom PCBs. But for most people doing one-off prototypes I agree that the Wii boards are a good alternative to breakout boards and breadboards.
You make a good point about the epson, hadnt realized it was so low speed. Though its not exactly the same part # as the one in the m+....so that might not be the case.
Hmm, the documentation previously said 16-bit. Looks like that wiki has changed it to 14-bit now. Still, not really sure why they'd bother with that if the last 2 or 4 bits would be useless.
Ok, for me, with far more mechanical skills than electrical, it's easier to integrate 2 boards than build my own. It's trivial to mount the boards, but I have no experience soldering smd chips. And I'm not really sure what advantage you get out of building your own beyond a slight weight/size reduction. Previously I would have likely used breakout boards, leaving me with the IMU split across 3 boards and costing near $100. This solution may not be the best for finished products, but for prototyping I still think it's great.
Yes, I would agree. The Wii controllers are fun for one-off hacking, but if you're trying to build an autopilot it's cheaper, simpler and much smaller and neater in the end to just put those sensors on your own board.
1. EPSON X3500W is 100 deg/s yaw gyro. Maybe that is adequate.
2. I2C messages from Motion Plus indicate 14-bit, not 16-bit, and in any case, it's highly unlikely that you are seeing more than 10-12 bits of useful data, depending on the quality of their A/D voltage reference.
I don't argue that this is a cheap way to get a IDG600. Beyond that, I think you end up with a kludge of 2 circuit boards for 6 DOF, so "ease" wouldn't be part of my description.
@Howard:
1) This is a 3 axis gyro already (i assume) quite orthogonal. I would have difficulty creating something as orthogonal myself.
2) Yeah, $30 for an IDG300 (which may not be fast enough), plus $30 or so for that final axis. You're now at 3x the cost of this unit.
3) With a low powered uC, high frequency analog reads can start to eat up a lot of processor time. I'm only too happy to offload the process if I can. And why would I want to go through the trouble of buying and integrating an AD7998 when I already have 16-bit ADC on this board?
I use Parallax Propeller micro that has i2c built in. No analog inputs tho. So i2c is the best interface for me. Doesn't the arduio micro have i2c also ?
I love I2C (uBlox5 has an I2C interface), but this hardly seems worth the trouble, other than a cheap source of IDG600 (if you have Jack's heat gun). You can buy a 2-axis IDG300 for $30. If you already have A/D, there's no additional cost. If you want I2C A/D, add an Analog Devices AD7998 which gives you 8-channel 12-bit for $8.
Comments
Hmm, the documentation previously said 16-bit. Looks like that wiki has changed it to 14-bit now. Still, not really sure why they'd bother with that if the last 2 or 4 bits would be useless.
Ok, for me, with far more mechanical skills than electrical, it's easier to integrate 2 boards than build my own. It's trivial to mount the boards, but I have no experience soldering smd chips. And I'm not really sure what advantage you get out of building your own beyond a slight weight/size reduction. Previously I would have likely used breakout boards, leaving me with the IMU split across 3 boards and costing near $100. This solution may not be the best for finished products, but for prototyping I still think it's great.
2. I2C messages from Motion Plus indicate 14-bit, not 16-bit, and in any case, it's highly unlikely that you are seeing more than 10-12 bits of useful data, depending on the quality of their A/D voltage reference.
I don't argue that this is a cheap way to get a IDG600. Beyond that, I think you end up with a kludge of 2 circuit boards for 6 DOF, so "ease" wouldn't be part of my description.
@Howard:
1) This is a 3 axis gyro already (i assume) quite orthogonal. I would have difficulty creating something as orthogonal myself.
2) Yeah, $30 for an IDG300 (which may not be fast enough), plus $30 or so for that final axis. You're now at 3x the cost of this unit.
3) With a low powered uC, high frequency analog reads can start to eat up a lot of processor time. I'm only too happy to offload the process if I can. And why would I want to go through the trouble of buying and integrating an AD7998 when I already have 16-bit ADC on this board?
Bottom line: Can't beat the price or the ease.
-
1
-
2
-
3
of 3 Next