Replace a Servo with a DC Motor to reduce battery drain
I am considering using a 12V DC motor geared down to 3 RPM. It will use a MOSFET to switch +on/-on for direction and off to put it into a holding position. To develop position information Ill use a printed encoder disk or drilled holes and an IR detector to read until it needs to stop as needed. Part of the proof of concept will be to reverse torque the motor shaft to determine the stationary friction to overcome in order to change the stopped position. My hope is to use this as an adequate brake and eliminate the need for the typical holding current of a servo. If that works it should be a simple, inexensive way to go. If it doesn't brake I might try a low voltage holding current as a position holder. Also, the encoder would be watching for undesired movement.
I have the impression servos or steppers are more commonly recommended for this type of application. But the motor will only be on for short periods and so the elimination of holding current would be a big plus. Is there some reason this isn't done?
Comments welcome.
Dom..
I have the impression servos or steppers are more commonly recommended for this type of application. But the motor will only be on for short periods and so the elimination of holding current would be a big plus. Is there some reason this isn't done?
Comments welcome.
Dom..
Comments
What RPM is the 12 volt motor running at when the output shaft is running at 3 RPM? In other words, when your planned IR detector signals a change in one position exactly how many revolutions of the motor have gone by?
Secondly, if the motor is driving a load CCW and suddenly requires changing direction to CW, how many motor shaft revolutions will be required before the output shaft actually changes direction? (AKA backlash in the gears)
I don't actually expect you to answer these question here ... they are just food for thought. As, I myself made a servo from a small DC motor w/gear box that was designed to turn a radar antenna on a boat. It operated at 12V and turned apx 12 RPM.
But it had a 'coasting' element to contend with, every time you turned it off. It did not come to an abrupt stop. So that proved to be quite difficult to eliminate the 'hunt-n-seek' introduced by the over correction in each direction. Also, it was constructed using flip/flops, timers and gates ... rather than a micro controller (quite a few years back)
Even when its job was reduced to pointing lights around to expose terrain at night, delivery of the analog command signal to the remote location was difficult ... it worked as advertised on the bench ... but, failed in the field.
Obviously your not talking about a worm drive here. And I doubt you have an issue with that when geared to 3 RPM. Now, if you were talking 3 rev/sec ... that would be different.
And, one final thought inspired by your subject line 'Replace a Servo with a DC Motor to reduce battery drain':
BTW ... I see from your location, that we are just 100 miles away from one another ... howdy neighbor!
... Tim
Is there a reason why a worm drive isn't being considered?
As Tim mentioned, if you stop sending pulses to a servo the current used is greatly reduced.
Controlling a normal DC motor as if it were a servo is not as easy as it first sounds. I've done it, but it was a lot harder to get it to work than I had anticipated.
You mention using a MOSFET to turn the motor on and off by I assume you want to control the motor in two directions so you'll need some sort of h-bridge circuit which complicates the electronics.
Nice to know there are enthusiast nearby.
Your question have gotten me to rethink some things.
Really got my attention, so I went back to the Parallax servo datasheet and it specs a quiescent current of 7 ma. I can manually turn the servo shaft. Therefore any torque will move it unless it is meant with an equal and opposite push from the servo motor. So it seems holding it in position will require energy from the battery above the quiescent.
http://www.parallax.com/sites/default/files/downloads/900-00005-Standard-Servo-Product-Documentation-v2.2.pdf
Don't know the motor RPM, so assuming it's 5,000 or more. The encoder will be moving at 3RPM rather than the motor speed. I think the Prop can characterize and compensate for the overshoot by possibly ramping down the voltage as it approaches it's new setting. I'm thinking the 1,000:1 gear reduction will have so much friction that I won't be able to turn the output shaft and therefore no voltage applied will act as a brake.
On another note it might be possible to use something like the 1.5ms - 20ms driving a MOSFET to switch the 12V into a cap for smoothing. The final voltage applied to the motor would depend on the duty cycle of the control. Might need to up the frequency but, if it worked, that would give me a programmable ramp, maybe.
Hmm, I wonder if the coasting was because of the inertia of driven mass or the motor spinning after power was removed? I can see now that a hard on-off will smack things around a bit and rattle placement.
Thank You Very Much for your food for thought.
Dom...
Thanks for looking at this and posing questions.
I don’t know the pros and cons of a worm drive versus other types. So I punted.
Some general criteria is to minimize mechanical parts, use off the shelf and keep one-off parts to a minimum. The motor under consideration has the power, reduction, shaft size, mounting holes and fits with other available parts like bearings, shaft cap and misc.
Also I’ll be pressing a 5mm bearing onto a 5mm D shaft --- not kosher but should work?? I haven’t worked with H-Bridges yet but it should be interesting. With this project I’m pretty sure I’ll need to make some circuit boards anyway or use Parallax Dual Motor Driver MC33296. That might make it much easier and quicker to prototype.
Erco’s thread about “Make Your Own Encoders” was really helpful, he did a pretty good job pointing out the pitfalls. The other user comments were great.
Sounds like you learned a lot when you did this. I’d be really interested in hearing about some of the problems you ran into. I’m getting ready to hit the “Buy it Now” button but not if this won't work.
Thanks a lot for your comments, greatly appreciated.
Dom..
I'm only used worm gears with Lego sets myself.
BTW, I think the MC33296 is a good h-bridge. Parallax sells the double chip board, but if I were to purchase these again, I'd purchase them as individual units. The "inverse" pins on the double chip board have been tied together which pretty much makes them (the inverse features) useless. Pololu has said in hindsight they wish they hadn't done this.
By using the inverse pin, the chips can be driven in pulse/coast mode with only two control pins. IMO, the two control pin option presented on both Pololu's and Parallax's websites aren't a good way to control these chips since it causes the motor to brake between pulses (pulse/brake). The pulse/coast vs pulse/brake issue is even more important when one wants to reduce the current load from the motors. In my experience pulse/coast uses much less current than pulse/brake.
This is not true at all. I have several that can.
I thought I was stating that too simplistically.
I think it's save to say it's harder to drive a worm gear in reverse than a "normal" gear box.
It may be that the worm drive gearboxes I have been using are not the cheap kind, they are rather nice. They cannot be relied upon to hold position if under anything but the slightest load.
There is a specific criteria to determine if a worm drive can be back-driven. I can't get my hands on my gearing docs at the moment to quote it, but it depends on the ratio and some other tooth geometry as well as the coefficient of friction. A general rule-of-thumb is 100:1 or higher will not back drive, but even that isn't a guarantee.
-Russ
Hi Russ,
I'm sure the ratio is ball park about 1000:1. So I'm depending on the gear friction plus ratio to work as a brake. I intend to have a position encoder and the Prop monitoring, so the motor could be tweaked back or held in position if it starts to back drive. I'll be ordering parts today. "The proof of the pudding is in the eating."
Thanks for your insight,....
Dom...
I did some digging around and I found that old motor I was referring to. This video should require little explaining.
[video=youtube_share;gyp2ilnAm2g]
... Tim