OpenServo.com

ginge · Posted: Tue Jul 04, 2006 4:42 pm Post subject: OpenServo High Current controller

Hi all,

I have just finished and build a controller for a large 6A motor using OpenServo MCU and reference design. It can supply up to 12A continuous, and 15A peak. This is a large board, but has loads of connectors and outputs.

It's a little rough around the edges, and you could consider this a version 0.001 pre-pre-alpha. It does work though!

Check it out here
http://www.headfuzz.co.uk/?q=OpenServoBig

I update with a photo when I'm back form my holidays (2 weeks)

Barry

mpthompson · Joined: 02 Jan 2006 Posts: 650 Location: San Carlos, CA

Looks great Barry. Have a good vacation.

coallen · Joined: 09 May 2007 Posts: 8

Hi Barry,

I am interested in making use of your PCB design for a high current driver. I have some motors that I suspect to be in the 10-12 amp range that I would like to use in a robot arm design. Could you tell me if the PCBs on your web site are the latest version and if they can be re-used?

Thanks in advance,
Chris

ginge · Posted: Wed May 09, 2007 10:06 am Post subject:

Hi Chris,

Thanks for the interest.

The designs I have on my website are slightly dated now, but are functional.

I created a slightly newer version of this design here:
http://www.headfuzz.co.uk/?q=node/49

which uses a hex display instead of LEDs. This fixed up a couple of problems.

The amount of current your motors require is rather large. I can't see my boards being able to drive your motors if that is the average current. The L6203 driver I used can support continuous current of 6-7A with a absolute max peak of about 12A. The figures I posted above were slightly too high, and misrepresented. I apologise, it was my mistake reading the spec sheets.

I would recommend you look at creating a discrete H-Bridge using banks of High current parallel FETs to meet your demands.

Something I also later added to my design was a small PTC fuse to one of the motor terminals. This stops the bridge blowing out if there is a sustained stall on the motor. I would recommend doing this on any design you use.

In summary:- I don't think my driver boards are going to work out too well for you.

Let me know if there is anything I can do to help out with a suitable design. I will shortly be needing a very high current driver much like you. Hopefully I can work though the design and create something that is suitable.

Barry
_________________
http://www.headfuzz.co.uk/
http://www.robotfuzz.co.uk/

coallen · Joined: 09 May 2007 Posts: 8

Actually, 6-7 amps is what I suspect the average current to be with max draws in the 10-12 amp range. Though I'm just speculating. I need to test them out and see what they really are.

I like your design because it's also got a current sensor built in. From what I've been reading, OpenServo 3.0 might have that as well. What are your thoughts on driving heavier duty MOSFETS from the output of the stock OpenServo design?

Thanks again for your help. And I would just like to say you've got a very nice site.

Chris

ginge · Posted: Wed May 09, 2007 12:56 pm Post subject:

Chris, Thanks Smile

OpenServo v2.x also has the current sensor. It has been a feature since the beginning.

coallen · Joined: 09 May 2007 Posts: 8

If I were to hookup larger MOSFETs to the output of the standard OpenServo board, would there be any other changes I would need to make? What about the current sensor? Would I be in danger of frying it?

Also, what are your thoughts on using an absolute encoder instead of a potentiometer? Is there really that much benefit? The encoders such as the ones from US Digital are nice, but expensive (~$50) compare to a cheap pot.

Thanks!
Chris

ginge · Posted: Thu May 17, 2007 2:13 pm Post subject:

Hi Chris,

If you use larger FETs on the standard OpenServo, you could run into several problems.

The first problem will be the PCB trace current capacity. You will end up burning those out before you fry the current sense resistor.
The traces onboard will not carry more than a couple of amps without damage.

The next problem, as you have already identified, is the current sense resistor. The wattage of this component will need to be uprated by a large margin. As the wattage of the resistor increases, so does its package size. You can't expect to fit a much larger resistor in place of the one already onboard.

There will be other less defines problems due to the inductance and the PWM driving. You will need to tweak the PWM settings. You will also run into problems with AVR brownouts where the motor is saturating the board. Adding some large protection caps should help here.

As for the encoder. This is an idea that keeps getting thrown around. I personally coded up some basic encoder support that worked well for the modified boards, but maxed out the AVR's processing capability at 2500-3000 ppr. If you want to use a stock 8MHZ OpenServo board to do the counting of the pulses, expect to have to code it up in super optimised fashion, or suffer from horrible PID and PWM performance.

I think we may have more luck with the 20MHZ version 3 that is currently in the works, but this is entirely wishful thinking at this point, as it is all a work in progress.

On a side note, I am currently working on a new design for my robot boards that will sink a continuous 10A current and 15A peak. The board is obviously a little larger than an OpenServo, but provides much more power than the previous modifications I have posted up. I expect this to be posted up on my website in the next couple of weeks, time permitting.

Barry
_________________
http://www.headfuzz.co.uk/
http://www.robotfuzz.co.uk/

coallen · Joined: 09 May 2007 Posts: 8

Hi Barry,

Thanks for the quick response.

ginge · Posted: Thu May 17, 2007 3:06 pm Post subject:

Hi,