Brushless DC Servo

Discussions relating to development and use of the OpenServo hardware.

Moderators: jharvey, Secondary Admin, Admins

robotjay
co-admin
Posts: 225
Joined: Tue Aug 01, 2006 3:58 am
Location: Nebraska, USA
Contact:

Post by robotjay » Thu Dec 31, 2009 7:52 am

I've just finished another schematic with two AVRs. As Kevin (aka "kbb") suggested, I've off-loaded most of the sensor readings to the Host/Master MCU, and the slave handles the actual motor commutation. Click the image below for a hi-res PDF version:

Image

I've drawn a hall-based current sensor from Allegro Microsystems, which looks like it'll fit the bill for what we need. It's a little expensive though (~$4), so I'm still open to cheaper suggestions. Let me know what you guys think. Talk to you soon.

-Jay
"Nothing is fool-proof; For we fools are ingenious and will find a way."

jharvey
co-admin
Posts: 362
Joined: Sun Mar 15, 2009 12:06 pm
Location: Maine USA
Contact:

Post by jharvey » Thu Dec 31, 2009 11:43 am

It appears you used one hall to get an overall system current draw. I think that what ginge meant was to get rid of the EMF and replace them with the hall chip. So you would need 3 hall chips.

robotjay
co-admin
Posts: 225
Joined: Tue Aug 01, 2006 3:58 am
Location: Nebraska, USA
Contact:

Post by robotjay » Fri Jan 01, 2010 7:44 am

jharvey wrote:I think that what ginge meant was to get rid of the EMF and replace them with the hall chip. So you would need 3 hall chips.
I hope ginge can weigh in on this himself, but I think he is under the impression that we would be using the sampled BEMF values to attain position, but we aren't. It's explained more clearly in the app note (section 2.1), but the BEMF values are only used in the timing of energizing each winding. Although the BEMF values from the OSv3 are noisy and mostly unusable, in this application, we don't need the signal to be all that clean for proper commutation. This is the same principle used in hobby ESC's. The thing that will make our board different from other ESC's is that we are closing the feedback loop with a master controller and outside position reference.

-Jay
"Nothing is fool-proof; For we fools are ingenious and will find a way."

lnielsen
Posts: 1
Joined: Mon Jan 17, 2011 3:25 pm

Post by lnielsen » Mon Jan 17, 2011 4:20 pm

This is my first post but I have been watching OpenServo for several years. Since I am new to the forum I can't post URLs so you will need to google some of the items I reference.

I am also looking to design a BLDC servo. I am very interested in building a DARwin-OP (on sourceforge) robot but I am hoping to lower cost by using my own servos. Robotis has not released too much information yet on their new Dynamixel RX-28M servo that DARwin-OP uses but it has a high quality brushed DC motor with a magnetic encoder and a 4.5 Mbps serial bus. I figured if I was going to do the effort I would try to upgrade to a BLDC motor.

That said, in my research, I found several references that argued against a sensorless design because of the complexity of finding the initial location when you stop and start which a servo does often. I also was leaning towards a sinusoidal drive to lower noise. It does not sound like the Atmel USART can run at 4.5 Mbps which could be another issue for me. I was looking at a dsPIC design based on the dsPIC33F. They have a $25 Microstick (on microchip) board that would make a prototype easy.

You may not be interested but I thought I would share my thoughts.

robotjay
co-admin
Posts: 225
Joined: Tue Aug 01, 2006 3:58 am
Location: Nebraska, USA
Contact:

Post by robotjay » Mon Jan 17, 2011 8:08 pm

lnielsen,

DarwIn-OP is a perfect candidate for OpenServo mods. Dynamixel is using a magnetic encoder, huh? Jared Harvey has been talking about his OpenEncoder for years now, which features a magnetic encoder. Sounds like a certain company may be watching these boards... ;) The next OpenServo version will have out-of-the-box support for the OpenEncoder.

The design uses a chip that CAN detect initial location, without needing to store current location in some sort of memory. It's truly elegant in its simplicity, although the chip itself is kind of pricey. (You may be able to get enough sample quantity to do what you need, though.)

The biggest hurdle for doing something like this is finding some sort of BLDC motor that comes with a gearbox attached OR by finding a BLDC that will fit into a TowerPro MG995 (or whatever the currently cheapest hobby servo is.) I gave up my search and shelved this project because I wasn't able to find an inexpensive solution. PLEASE share with us your research findings. We'd be glad to offer any advice you need. Talk to you soon.

-Jay
"Nothing is fool-proof; For we fools are ingenious and will find a way."

lourencohen
Posts: 3
Joined: Thu Feb 20, 2014 6:16 am

Post by lourencohen » Thu Feb 20, 2014 7:16 am

I was also not aware much about the AREF pin on AVRs and wanted to know more. Your discussions really helped me to understand about the aspects and features of the same in a much simple way. Thanks for sharing the same.





Thanks
Lourene
windows8-problems.com

Post Reply

Who is online

Users browsing this forum: No registered users and 0 guests