# Mechanical mechanism that regulates gear's rotational velocity

Please do note that I am a civil engineer not a mechanical engineer.

I have a gear that rotates at a speed that varies with time. Usually, it rotates at speeds higher than the desired speed. Is there a mechanical mechanism that can regulate the output speed, i.e., a mechanism that can give a lower output speed without wasting power?

I am thinking about using a mechanism similar to the escape wheel in clocks but I am not sure if it can work with high torque and on a bigger scale.

Update: Thanks for all who contributed to answer my question. I have to add that the power that turning the gear (let's name it big gear) is coming from other small gears which gets there torque from twisted torsion springs. So I need the "big gear" to rotate in a constant speed so that all the torque coming from the small gears adds up. I hope the question is clearer this time.

• What have you searched for so far? What about the words "centrifugal governor"? Sep 22 at 19:09
• I searched for the "centrifugal governor" and it seems that it won't work if the system is turned upside down (which is a condition needed in my case). it also connected with a valve (steam engine). Sep 22 at 19:42
• You did not mention direction - vertical or otherwise in your question. May be an idea to improve your question then. And why couldn't you adapt the design to work with what you have instead of just assuming steam engines only? Sep 22 at 20:14
• Along similar lines, but without relying on gravity, I'd look into something like a centrifugal clutch, configured to act as a speed limiter? Sep 22 at 22:20
• Surely a calculated flywheel is the simplest? Sep 23 at 8:38

Mechanism you are looking for is something mechanical watchmakers have been searching for ages*. Because that is exactly what a mechanical wrirst watch does. The mechanism in question is called an escapement.

Now the classical escapement moves in pulses, but there is a sub variant known as a constant force escapement that moves at constant speed if the force usage is constant within some range. You can find several of those in patent databases. They are a bit finniky to set up. There is also the constant escapement which to my knowledge is still covered by a patent.

Escapements also often do not work upside down but watchmakers have allready thought about that.

* atleast about 700 years, its not untill recently that we have some good options.

• I don‘t believe the op wants to have high frequency discrete rotation…a constant force escapement isn‘t suddenly continuous because it now has a better regulation of force. Not to mention needlessly complicated when a flywheel could easily suffice. Sep 25 at 10:07
• @morbo the description of op is thus. But i agree, this just explains why its unikely for you to find such a continious regulation that has no active components without some sacrafice. We have been looking for such systems for a long time and thusfar we havent found a satisfying method of regulating the output of a spring, accurately or satisfactorily. OP does not describe the need very well so we can not answer the question very well. Today almost invariably you would use a electronic controllers for this. Sep 25 at 10:33

To my understanding the Hartnell governor is a variation that might be applicable to your needs.

Figure : HArtnell Governor source mechoholic

Essentially it moves the sleeve against a compression spring (instead of gravity), so - at least in theory - you could turn it upside down.

• what are you governing? These devices are usually attached to a valve that controls the speed. Sep 22 at 22:23
• @tigerguy To be honest, I agree with you that the best way is probably an electronic PID system. I was answering in the mechanical context set in the headline of the question, and from the constraints of the comments (governor was the first thing that sprung in mind). There is simply not enough information in the question to tell if there is a valve or not. Sep 22 at 22:44
• I thought I might be missing something obvious. Sep 22 at 22:48
• My estimation is that 99% of the use of the term "XY problem" is by people who think they know more than almost everyone else. The other 1% is from people who actually do. The problem is that the 99% erroneously think they are the 1%. ;) Sep 23 at 8:56
• @RockPaperLz-MaskitorCasket except in this case seems to be a bit true since the op has editted thequestion. Sep 24 at 5:13

There are multiple things going on here.

1. Your question indicates you have a variable input speed but want a constant output speed. The answer to this on its own would be a continuously variable transmission with a controller that sets the ratio according to the output speed. Any version that has a controllable ratio could be hooked to a simple pid controller with a setpoint equal to the desired output speed. Using examples from the wiki page this could be sliding a cone on its axis or changing the width and thus sheave diameter of a variable sheave transmission.

2. But then you talk about power. A transmission doesn't change power. The input power equals the output power, regardless of the ratios (subject to losses in the transmission of course). If your output speed is doing work, slowing down the input speed may not provide enough power to the output.

So we can keep an output running at a specified speed, but only if input power is enough to power the output.

1. The power used is determined by the load on the output, not by the speed of the input. So varying the output speed doesn't save or waste power. It just changes the rotational speed. I am worried this is an XY Problem; I think that understanding the total process would let us help you better.
• I edited the question to make it clearer Sep 23 at 11:05

Not sure if this exists off the shelf, but what if you installed in line with the spinning gear a device with two weights held in place by springs, which would move outward if the centrifugal force became too high?

I believe they also have devices that will declutch the engine if a speed is exceeded, but not sure that is a desired behavior.

• Sounds suspiciously like "centrifugal governor" as suggested above. Did you read the OP's objections? Sep 22 at 20:52

A very easy way is to attach it to a reducing gear, a gear with more teeth. You would choose the radius of the reducing gear like this.

If your output is usually 10% above desired speed, we choose

$$R_{final}=\frac{R_{initial}}{0.93}$$ Note we did not choose 0.90 because we want to hover near average desired output speed.

By just choosing 0.93 we get close to a speed that sometimes is the exact desired speed, sometimes a fraction more, sometimes a fraction less, but overall much closer to what we need.

By trial and error, we can bring the system to optimal speed.

Before using a gear first of all you have to know types of gears and their functions then you can easily use without any obstructions.