A New Method (?) to Dynamically Change Cam Timing when Engine is Running

I was doing some calcs for the gear train for the 2 stroke diesel I am building and was thinking about the possibility of adjusting the cam timing when the engine was running. Say, for changing the injection advance and retard curve. Here is a method that is new to me, but might be old news to some. I don't really know. But I do know that "real" new ideas are hard to come by.

Here is the plan for changing the cam (or any gear) timing. You start with 3 gears installed in a straight line. The drive gear, an idler, and the driven gear. In the example, all gears are 36 tooth mod1 metric. They are installed such that there is almost no backlash. The center idler is mounted on a small cross slide so that it can be pushed off the straight line axis +/- 1MM.
In the top picture you can see that all 3 gears are in line and at a 36mm center distance. You can see the zero point of one tooth on the drive gear and one tooth on the driven. While the engine is running if the cross slide of the idler is pushed 1 mm off center, the zero tooth on the output gear will rotate 6 degrees. Also, if you pull the idler to the other side of the centerline by 1mm, the gear will rotate a total of 12 degrees to allow for a +/- 6 degree timing change. The movement of the idler gear could be done with a governor, or by hand, for starting for example.

The change in the backlash of the gear train was a concern, but calculations show that moving the idler 1mm off center changes the center distance only .014 mm, and the backlash of the entire gear train only changes by
2 x (the change in center distance) x the tangent of the pressure angle of the gear.
Twenty (20) degrees is a common pressure angle. The math is shown in the picture.

For the example shown, the driven gear can rotate +/- 6 degrees (+/- 2mm on the circumference), but the backlash only varies from approximately zero to .010mm (.0004"), which I think is acceptable for most applications.

Thanks for taking a look and let me know if I missed something or if you have seen this method before.
Thanks,
Lloyd

Seems to make sense to me.

I have been planning to attempt the same thing with an externally driven differential gear - having one output fixed but adjustable and the other driving a cam, so rotation of the fixed output is actually an input, adding or subtracting from the rotation of the cam driving output.

I used something similar to make a mechanical computer from lego, eons ago, to seperate the X and Y outputs from a single joystick.

My idea for timing advance was to use a chain drive with a spring loaded idler on the slack side and an adjustable idler on the other side of the chain. Moving the adjustable idler would change how much chain was between the driving and the driven gear, which would change their phase.

There was a steam valve gear that used two cams per valve, where one was fixed and one was adjustable. The valve would only open when both cams were "extended" so by moving one cam relative to the other you could change the cutoff.

On the dockstrader valve gear software, it was the "skinner uniflow".

A way to change timing while the engine is running is by moving the camshaft longitudinally, the cams being wider with a progressive change in profile in their thickness, then a different profil comes into play.

Interesting idea! Seems viable to me ... I look forward to hearing how it works out.

Gedeon Spilett said:

A way to change timing while the engine is running is by moving the camshaft longitudinally, the cams being wider with a progressive change in profile in their thickness, then a different profil comes into play.

Click to expand...

On my printed engine I had a stack of cams, but instead of moving the camshaft, I moved the follower/pushrod to the desired cam.

rolphill said:

On my printed engine I had a stack of cams, but instead of moving the camshaft, I moved the follower/pushrod to the desired cam.

Click to expand...

Any pictures of this set up, perchance?

The Ignoble Troll said:

Any pictures of this set up, perchance?

Click to expand...

Halfway through the video, it shows the cam setup. The pushrods are constrained inside guides that I can slide back and forth to "select" different cams. You can see the small threaded rod that sticks out the side is attached to the two guides. Moving that rod moves the guides. The "tappets" that the pushrods push on are large enough to cover all positions.

This method means you don't have to try to slide the whole camshaft and thus deal with sliding gears/sprockets.

rolphill said:

Halfway through the video, it shows the cam setup. The pushrods are constrained inside guides that I can slide back and forth to "select" different cams. You can see the small threaded rod that sticks out the side is attached to the two guides. Moving that rod moves the guides. The "tappets" that the pushrods push on are large enough to cover all positions.

This method means you don't have to try to slide the whole camshaft and thus deal with sliding gears/sprockets.

Click to expand...

Very cool, thank you for sharing!

rolphill said:

Halfway through the video, it shows the cam setup. The pushrods are constrained inside guides that I can slide back and forth to "select" different cams. You can see the small threaded rod that sticks out the side is attached to the two guides. Moving that rod moves the guides. The "tappets" that the pushrods push on are large enough to cover all positions.

This method means you don't have to try to slide the whole camshaft and thus deal with sliding gears/sprockets.

Click to expand...

Sweet! So many approaches to the same problem.

rolphill said:

My idea for timing advance was to use a chain drive with a spring loaded idler on the slack side and an adjustable idler on the other side of the chain. Moving the adjustable idler would change how much chain was between the driving and the driven gear, which would change their phase.

There was a steam valve gear that used two cams per valve, where one was fixed and one was adjustable. The valve would only open when both cams were "extended" so by moving one cam relative to the other you could change the cutoff.

On the dockstrader valve gear software, it was the "skinner uniflow".

Click to expand...

I really like this idea, but maybe with a timing belt instead of a chain.
Actually, I like it better than my idea of shifting a gear off-axis. Pushing an idler roller against a belt ought to be easier, I think.
Lloyd

Have you looked at the Maudsley steam engine valve gear? It used a parallelogram of gears to adjust the phasing of the crankshaft and eccentric shaft to adjust cutoff and change direction.

This may be a more explanatory picture from the Holzapfel catalogue.

Similar to post above - here's a design I came up with for changing the timing between two driven shafts for an industrial application.

The two main shafts are also slightly adjustable in the pitch between axles - hence the floating idler gears each constrained in its own orbit about its sun gear by arms (black) and the two idlers to each others' pitch by further arms (turquoise). This arm is jacked from side to side to change timing.
By making the idlers have a different number of teeth, when the larger of the two idlers is displaced it induces a greater angular change to the smaller idler - hence a change in timing between the two main driven shafts.

This was on a machine that produces wrinkled steel wire and can be offset with the machine at full power.

I also repurposed this design to endurance test CV joints under torque and rpm which without being able to induce powerless torque would require insane power to perform. The torque being dialed in by offsetting the driver and driven ends to the required torque.

Just a thought - Regards, Ken I