View Full Version : [Question(s)] Servo Control Circut

06-01-2009, 03:31 AM
Hello everyone,

I am fairly new to this board and do mostly reading on the subjects that I am currntly interested in atm. So please do not be harsh in my nubbish questions. :veryhappy:

Okay, nubbish question #1.

Is there a "complete" schematic diagram out there for the servo control circutry such as that used in standard hobby and robotic servos ( PWM )? I have googled and yahoo'd till I find the screen blurry and I go cross-eyed and can only find sites such as: http://mysite.du.edu/~etuttle/electron/elect32.htm . I know that I may be reinventing the wheel but I am researching to take things to "Really Large Scale" if you get my drift. :veryhappy: I would like something, if possible, that I can replace the H-Bridge with my own parts that fit my power requirements.

Nubbish question #2.

Is the signal line of a servo seperate from the voltage source for that servo? The reason I am asking this question is "IF" I built my own servo and that servo's power requirements happen to be let's say 24VDC, would I be able to run the SSC32 as the controller for that homebrew servo "IF" the control circutry was identical.

Thank you all for any help and for putting up with the NUBB.


06-01-2009, 09:46 AM
I doubt you will find a "complete schematic" for your hitec/futaba servos since they are proprietary and any who reverse engineered and posted a schematic would probably open themselves to a lawsuit....

However, have you seen the open servo project? http://www.openservo.com/ An open source digital servo based on an AVR microcontroller.... this is probably what you'll want to look into using.


06-02-2009, 01:21 AM
So, how big of a motor would you like to drive to move how much based on a servo signal input? Here is one answer (http://secure.oatleyelectronics.com//product_info.php?products_id=204&osCsid=bfa1228ffda8158b1cef35bd1332d417), for driving a little bit bigger motors. There are the RobotEQ controllers that are supposed to take feedback to drive a bit bigger motors. The datasheet for the NJM2611 (http://semicon.njr.co.jp/njr/hp/fileDownloadMedia.do?_mediaId=417) has a servo control circuit at the end, not sure what brand servo it might have been used in. Another possible is the MC33030 datasheet (http://www.onsemi.com/pub_link/Collateral/MC33030-D.PDF).

Question 2, no, the signal line is not seperate, as the signal has to reference a ground. Otherwise the control eletronics inside have not basis for which to read the control signal. As to the actual question, yes, you can have servos that run on 24 volts be controlled by 5 volt electronics, as long as the grounds (reference) is tied together. It would be possible to completely isolate one side from the other, but there will be the need for a ground on each side.

06-02-2009, 10:05 AM
Thank you both for the great links, I am still amazed that I could not find anything like what you both pointed me to. I am not looking for any "brand" of circut per se, all it has to do is act like a "brand" so that I may use it with an SSC32 Controller. I like the idea of that kit, should be easy to replace the transistors/mosfets (whatever they are) with bigger beefier ones to handle something like: http://www.allelectronics.com/make-a-store/item/DCM-1353/24VDC-350W-MOTOR-11-TOOTH-24-WIRE/1.html . Then all I would have to do is worry about gearing everything down to the speeds I would need.

Also, I see where I made my mistake on #2, I think I meant to ask if the decoder circut can be seperate from the H-Bridge circut as long as both circuts have the same ground line to the power source? I will try to draw up a small *jist* picture later tonight at work.

I did write out this post at 4:30 am at work, LOL. Like I said NUBB hehe. Well not a complete nubb, I did graduate the Navy Basic Electronics and Electricity school way back in '88, which means I know just enough to be dangerous. :rolleyes: Unfortunately, my skill set goes more toward programming and networking then electronics.

Again, thank you for the links, I will study everything more closely at work tonight when I have time to think. LOL

06-02-2009, 01:37 PM
What are you going to use for positional feedback?

I paid a lot more than $20 for those motors. [sigh]

06-03-2009, 01:55 AM
Well, if I go with the kit from Oatley then it would be a pot connected to the output shaft. Of course I would have to figure out some gearing scheme for that.

Now correct me if I am wrong on this, when you gear down (let's say 2500 rpm to 60 rpm) you increase the torque that the output shaft produces. Would it be better to take a lot of steps down or would it be better to take as few as possible to get the above numbers correct with as much torque as possible?

I am working on a scetch drawing but have reports to do in a few, will definately get it up when I get home.

TY and goodnight.

06-03-2009, 02:15 AM
I'm by no means a mechanical expert, but I would expect that a single stage reduction would have less friction loss than say a 3 stage reduction, however it's obviously a lot larger.

I do think there's a whole lot of difference in the method that's being used for the reduction though.
You can accomplish a big speed reduction /torque increase by using a worm wheel, but there's a lot of friction in one of those, and as such it isn't very efficient.

06-03-2009, 12:14 PM
As someone that has designed AP-level Physics exams, I may not be a mechanics expert, but I think I can chime in on the physics:

Worm gearing has a very high range (generally between 4:1 and 75:1) but has a very low efficiency range (generally 20-70%)

Most of the rest of the traditional gearing types range from 92-99% [with the exception of Planetary gear boxes which are also high reduction but very very high loss].

In order to calculate a gear train, you multiple the efficiencies of each in the line together, so if we had three gears of 94% efficiency in a train: 0.94 * 0.94 * 0.94 = 0.943 = 0.83 = 83% efficient.

So if we had eleven gears of even 98% [that's efficiency is tough to do at the hobby level]: 0.9811 = 80% efficiency, but at say 94%: 0.9411 = 50.6% efficiency (!)

Of course, that's excessive, but you get the idea.

Now, gear speed also plays a part here, and that part is called "churning losses", and is based on the speed through the lubricant. Very difficult to calculate, it's primarily derived from experience and WAG. The speed the gear turns through the lubricant will increase churning losses as it goes up.

So the upshot: Don't use more gears than you have to. Gears are tremendously efficient by most standards, but if you get silly with 'em, they'll get silly back with you.

06-03-2009, 02:31 PM
Lynn, thanks for that lil gem of info there.
Despite my love of mechanics I can't seem to force myself to actually read up on it, it's the little bits and pieces of info like this that make this forum stand out from all the others!

06-04-2009, 12:06 AM
Thanks, ScuD! I was going to go all out and post all the math, but I've been sick and just didn't feel like it by the time I hit reply. :(

Maybe I'll do a quickie tutorial with all the math as soon as I'm better. That just seems like one of those tutorials that had oughta be done here. :) Calculating planetary gearbox efficiencies is eye opening, imho. ;)