I finally got my balls back!!!

[Roger B]

Still following along    I know a couple of modern Leonards who have got their control logic reversed. Thankfully there were emergency stop buttons by this time.

[Brian Rupnow]

Here we are, having fun with aluminum balls and a cardboard pointer.
<a href="https://www.youtube.com/watch?v=GB2ZG0HWQ_8" target="_blank">http://www.youtube.com/watch?v=GB2ZG0HWQ_8</a>

[Allen Smithee]

Forgime me if this is teaching people to suck eggs (again), but what we're looking at here are some fundamental aspects of control theory.

The physical throttle opening on an engine sets a *power* rather than a *speed*. Actually to be strictly accurate it sets a power *curve* because IC engines don't produce constant torque over a range or RPM.

So any governor that simply maps a physical throttle position to a speed will not deliver a true constant speed output. This is true of electrical/electronic governors and mechanical ones - they are (in control theory terminology) "direct-coupled".

There are two basic approaches to "fixing" this. Firstly we can make the coupling "indirect". This is easily in electronic systems (and even easier in software-controlled ones) but can also be achieved in mechanical systems with a bit of thought. The objective is to make a system that, instead of mapping a speed to a throttle position, maps a speed *error* to a *change speed command". One way of doing this would be to have a hydraulic throttle linkage, and have the governor control the valves so that an underspeed condition opens the "move to increase throttle" valve while an overspeed condition opens the "move to decrease speed" valve (and these operating points can be easily engineered to be variable whilst the engine is running). This approach can be nice and progressive if the valves have proportional opening, so that small speed errors only crack the valve open and bleed a small amount of fluid into the actuator cylinder whilst large speed errors whack the valve fully open. This kind of system is also inherently fairly well damped because they are based on the principle that the amount of feedback signal applied is proportional to the size of the error.

The second approach is simply to use very high gain in the system (this is how an op-amp works, for those of an electronic bent). The system is still directly coupled, but the gain is so high that you don't notice that your 1,000rpm governor has the throttle fully-shut at 1,010RPM and fully open at 990rpm so at high load the rpm might decay by 1% while the no-load idle speed is 1% higher than the nominal setting. Again, you can engineer it to have an operating point that is adjustable while the engine is running and this approach is much simpler to make (it's just a direct-coupled governor which is geared differently). The one thing you have to watch with this approach, especially if the load is very variable, is that high-gain feedback systems can be prone to oscillation if something else in the system happens to vary sympathetically with the throttle changes.

£0.03 supplied,

AS

[Brian Rupnow]

Last night, during a horrible case of insomnia, I worked out all of the linkage required between the lever arm on the governor and the throttle lever on a gas engine. The engine best suited to use with this governor, because of the way the carburetor faces and the throttle lever pivots, is my Fat Flywheeled Webster. Due to the physical shape of engine and the governor, (which evolved independently), the governor is going to have to set above the engine. This doesn't create a problem, although it may look a bit strange. I think I will build a 'bridge" over the top of the Webster, which the base of the governor will set on. This should line the pulley groove on the governor up with the pulley groove on the Webster, and put the governor lever more or less directly above the carburetor.

[Maryak]

One way of doing this would be to have a hydraulic throttle linkage, and have the governor control the valves so that an underspeed condition opens the "move to increase throttle" valve while an overspeed condition opens the "move to decrease speed" valve (and these operating points can be easily engineered to be variable whilst the engine is running). This approach can be nice and progressive if the valves have proportional opening, so that small speed errors only crack the valve open and bleed a small amount of fluid into the actuator cylinder whilst large speed errors whack the valve fully open. This kind of system is also inherently fairly well damped because they are based on the principle that the amount of feedback signal applied is proportional to the size of the error.

AS

Hi Peter,

The above was pretty well how all generator engines were controlled during my working life. Generators were allowed a maximum of 4% droop from no load to full load.

At one time I was on a visit to Maidenhead and the boffins there demonstrated a fluidic, (air operated), governor with no moving parts other than the output. The Hampton Court Maze had nothing on this baby. Seemed to work OK but never made it into the real world in my time.

Best Regards
Bob

[geoff_p]

When doing some work in the Lerwick (Shetland Islands) Power Station, I asked how they guarantee 50Hz from the Diesels.

They have two wall clocks, one is mains-powered (notional 50Hz) and the other by battery (quartz).  When the clocks differ by a few minutes, the engineer simply speeds or slows the engines via a manual linkage to the governors.

[Brian Rupnow]

This video shows a trial I made about 5 minutes ago to see if the Webster has enough power to spin the flywheel governor.
<a href="https://www.youtube.com/watch?v=BUhM5nUAPqA" target="_blank">http://www.youtube.com/watch?v=BUhM5nUAPqA</a>

[Brian Rupnow]

Now I have my thinking cap on. The carburetor throttle handle (which I have yet to build):eek: will be connected to the extension of the governor pivot arm (which I have yet to build.):eek: The governor pivot arm can exert a surprising amount of force, depending on how fast the governor balls are revolving. I will calculate the travel of both the throttle arm and the governor arm so they both travel roughly the same amount---but I don't want to throw a whole bunch of force into my carburetor and break it. So---we build a compensating connector link!! The link will be bent up from 2mm diameter uncoated welding wire. It takes VERY LITTLE force to move the throttle arm through its travel, as long as you don't "bottom it out" at either end of it's swing and then try to force it farther. The rather artisitic bends in the welding wire will let the wire deflect and take up any "over-travel", rather than break my pretty little Chuck Fellows carburetor.-

[Brian Rupnow]

Things are progressing well. I made an executive decision this morning that it was just too bitterly cold out in my main garage where I usually test my engines, so the governor and the Webster were moved in to my engineering reference table where it is toasty and warm. I thought it would be clever to make the governor arm from 1/8" diameter cold rolled steel, with a sliding adjustable block on it. That lets me change the length from the governor arm pivot out out whatever I need the linkage attachment point to be, very simply. Then I decided it would also be nice to be able to  adjust the length of the 1/16" diameter link between the governor arm and the carburetor arm, so another small round addition with set screw was added. I lied in the last post where I said the welding rod which I would make the link from was 2 mm. It isn't. Actually, it is 1/16" diameter. I thought it looked a bit "fat" in the solid model, but I didn't go out to the garage to check. Now, onward to build a new carburetor arm for the other end of the "link" to attach to.

[Perry]

Brian, watching all your updates with interest, aspecialy I like your super short "design to build" lead time - unlike mine which can be as long as few years uff..

Peter

[Brian Rupnow]

And here we reach a successful conclusion.
https://www.youtube.com/watch?v=CeyIhzOwhq0&feature=youtu.be

[Brian Rupnow]

And---If anyone wonders why I didn't use the fancy shaped connector link, it was deflecting too much and the movement of the governor lever was not getting through "cleanly" to the carburetor throttle lever because of that. The issue of different travels between the governor arm and the throttle arm was solved by having the easily adjustable connection point on the governor arm.

[Brian Rupnow]

It's quite amazing the number of different ways a mechanical governor can be configured, and it all has to do with the geometry of the linkage arms. As I said in one of my videos, not everything is scaleable 100%. If it was a larger more powerful engine, the throttle lever wouldn't have to move as much to bring the revs back up. As an interesting aside, I grew up in a sawmilling community, where we had a sawmill running a huge steam traction engine, two mills running Caterpillar diesel engines, and one mill running a water turbine. I don't remember too much about the water turbine mill, but I have read that flyball governors were used to control the sluice gates, letting more water through the turbine when higher amounts of power were required, and shutting the sluice gate part way when the higher power was not required. ( I never seen that). However, both the steam engine and the Caterpillar diesels ran on flyball governors hooked to the steam admission valve or whatever passes for a throttle on a diesel engine. Under no load conditions these engines ran along at a relatively low constant speed, but when a 20" white pine log was forced into the mainsaw, you could hear those engines "bark" from a mile away as the governors opened the throttles to whatever degree was required to cut those logs without dropping the rpm of the diesels or the steam engine. As soon as the log came out of the mainsaw, the engines would immediately return to their 'nominal' low speed setting. I walked past these mills every day on my way to public school, and can clearly remembering the loud "bark" of these mills when the saw bit into a log.

[Brian Rupnow]

At 4.38 minutes into this video, you will see a home brewed fan controlled governor on a model T powered buzz saw. I have never seen this done before except on small single cylinder engines, but it seems to be very effective.
<a href="https://www.youtube.com/watch?v=2LMv_Y4axYw" target="_blank">http://www.youtube.com/watch?v=2LMv_Y4axYw</a>