Building a better crankshaft (crank and slider)

[paradisePhysicist]

I've read through this thread as well as watched the video and I also don't 'get' it. The crank mechanism such as the one in the video works as it does because top and bottom dead center are undefined with respect to push or pull on the slider. The flywheel is a remedy to allow us to do what needs to be done while still leaving these two areas undefined. There is no way around it without adding complexity. It can be a flywheel, more sliders with timing that determines when to push, etc. But you cannot get around it by simplifying what we already have.

Looking back at my original post, its not 100% perfect, in retrospect I'd have rephrased one of the lines above the gif, I didn't anticipate the minor mistake was going to result in stress and drama.

Anyway, I am willing to add complexity as long as the complexity is robust and doesn't require a flywheel, I am looking to be able to just push the slider and get full rotations without having to add a bunch of mass and decrease the acceleration.

 

[Averagesupernova]

You do realize you will be adding a new set of problems right? I guess if you want to find that out, go ahead.
-
Make your slider engage two gears with a ratchet pawl mechanism. They will each turn the opposite direction but only half the time. Stop and reverse the direction of the slider at will. Connect the two opposite turning gears together with more gears to get the rotation constant and in the same direction. No flywheel. New set of problems include more reciprocating mass as compared to previously, a ratchet mechanism that will wear quickly, more parts in general. Looks like you haven't built a better mousetrap, but by golly that pesky flywheel is gone.

 

[paradisePhysicist]

You do realize you will be adding a new set of problems right? I guess if you want to find that out, go ahead.
-
Make your slider engage two gears with a ratchet pawl mechanism. They will each turn the opposite direction but only half the time. Stop and reverse the direction of the slider at will. Connect the two opposite turning gears together with more gears to get the rotation constant and in the same direction. No flywheel. New set of problems include more reciprocating mass as compared to previously, a ratchet mechanism that will wear quickly, more parts in general. Looks like you haven't built a better mousetrap, but by golly that pesky flywheel is gone.

I am generally leery of rachet type mechanisms due to the reasons you mentioned. Currently the best solution on the table so far is the half-gear mechanism, however that has some problems as well. I will link to the video of the half-gear mechanism here, it is similar but somewhat different than the gif posted by Lnewqban. Here it is:
https://www.ytprivate.com/watch?v=DfDPyWdlfz0

So far this one is my favorite but I am on the lookout for other options.

 

[hmmm27]

magnets.

 

[paradisePhysicist]

magnets.

Hmm I thought of this yesterday, but haven't yet worked out the specifics. The idea was going to be something involving the crank turning another gear, and that both gears had a magnet attached that boosted the rotation, in such a way that it also didn't interfere with rotation on the stall angles. Haven't really explored the idea yet because of all the sweltering heat, AC is on yet still been sweating all day. Though I don't have Lego magnets, I am willing to cross that line and use magnets, breaking legos is simply not my policy.

 

[some bloke]

Just so that I am on the right page here...

When you say a dual input piston, are you maning one of the pneumatic pistons the Lego makes, and that it is going to both push and pull the crankshaft, and that the problem is getting from the end of the push to the start of the pull? Is that about the summary of it?

I've made a piston engine out of lego technic before using the pneumatics, but it was 4 pistons and a flywheel to make it run, all pistons pushing and pulling. They were 90° offset to allow 2 of them to be pushing/pulling whilst the other 2 were switching.

I am mulling this over at the moment. One thing I am considering is if you were t owork out a mechanism where the start and end of the stroke is translated by 90° to push the crank up or down, passing the problematic point of the 179-181° changeover. Alternatively, a crank-in-a-crank could do something similar...

I shall ponder. Maybe even crack out the lego later!

 

[paradisePhysicist]

Just so that I am on the right page here...

When you say a dual input piston, are you maning one of the pneumatic pistons the Lego makes, and that it is going to both push and pull the crankshaft, and that the problem is getting from the end of the push to the start of the pull? Is that about the summary of it?

I've made a piston engine out of lego technic before using the pneumatics, but it was 4 pistons and a flywheel to make it run, all pistons pushing and pulling. They were 90° offset to allow 2 of them to be pushing/pulling whilst the other 2 were switching.

I am mulling this over at the moment. One thing I am considering is if you were t owork out a mechanism where the start and end of the stroke is translated by 90° to push the crank up or down, passing the problematic point of the 179-181° changeover. Alternatively, a crank-in-a-crank could do something similar...

I shall ponder. Maybe even crack out the lego later!

Interesting, what is a crank in a crank? Are you referring to adding a gear underneath the crank?

 

[some bloke]

Interesting, what is a crank in a crank? Are you referring to adding a gear underneath the crank?

I am pondering exactly that right now.

The thing I am considering is whether you could have the crankshaft such that where the conrod would connect, there would instead be a second crank which is turned relative to the original crank. This would be such that (in a horizontally laid out engine, crank on the left and piston on the right) the main crank would be at 3:00 (180°, no power from the piston) whilst the secondary crank it at 12:00. The piston would push on the second, smaller crank, turn it, which in turn pushes the main crank beyond the 180° mark and then when the piston pulls back, it would be pulling the secondary crank at 3:00 (so have no control over which way it turns) but the main crank at beyond 3:00, meaning the net forces will make the secondary crank turn in the correct direction.

It's all in my head right now, I'll see if I can put it into some doodles later!

 

[some bloke]

The doodle is complete but I fear that without animating it in some way it is going to appear as utter nonsense...

I will try and make a gif for it later if I get the time!

I will also say that this will likely make a very poor engine as I think it will result in the speed juddering all over the place... but it could negate the need for a flywheel, anyway!

 

[paradisePhysicist]

Interesting and this gives me some ideas, think I'm going to try them in the simulation.

 

[some bloke]

The basic thinking behind it is that if you have two cranks, both of which are turned by the same piston, then when one is at it's maximum/minimum then the other is not and the net result is that one will turn and the other will not resist, so the whole assembly will turn. No idea if it'll work!

 

[paradisePhysicist]

The basic thinking behind it is that if you have two cranks, both of which are turned by the same piston, then when one is at it's maximum/minimum then the other is not and the net result is that one will turn and the other will not resist, so the whole assembly will turn. No idea if it'll work!

Hmm, based on this I tried to put two gears in the same spot in the simulation. Then a rod at then minimum location to gear 1, then a rod at 90° location to gear 2, both attached to the same piston. It will not work and how do I explain this... I used springs to give it some leeway incase the rotations and positions were not exactly precise, still didn't work, I think the reason is because when gear 2 reaches the maximum position, it wants the piston to go in reverse, but at the same time, gear 1 wants the piston to still go forward, therefore it can't rotate. I will draw a picture to make it clearer.

Actually I think I could have just used 1 gear for this actually, but in the simulation I used 2 gears stacked on top of each other.

I haven't tried your doodle yet but I will try that next.

 

[hmmm27]

Sorry if I missed the post explaining this, but is it a powered upstroke ? And, the only problem is that you can't seem to go through 0 or 180 without rotational momentum... ?

 

[paradisePhysicist]

Sorry if I missed the post explaining this, but is it a powered upstroke ? And, the only problem is that you can't seem to go through 0 or 180 without rotational momentum... ?

Yes.

 

[hmmm27]

okay... so, you've got

- a big circle, featuring
- a little circle near the edge.

Make the little circle a slot : same placement, same width but say a length of 3x, the diameter of the original. Angled, as measured at BDC(or TDC), at 45deg.

Works fine in 2d ; probably too messy in 3d, but might give you something to work off of.

 

[paradisePhysicist]

Make the little circle a slot : same placement, same width but say a length of 3x, the diameter of the original. Angled, as measured at BDC(or TDC), at 45deg.

Works fine in 2d ; probably too messy in 3d, but might give you something to work off of.

Hmm, can you draw me a rough sketch, I'm not quite sure what the visualization of this is. Are you saying to cut a slot into the gear, put the conrod into the slot, or something else?

 

[hmmm27]

Something like...

The first one is normal, it's the one you have issue with. The second one uses a slot, not a circular hole, to hold the crankpin.

 

[Averagesupernova]

Something like...

View attachment 285613The first one is normal, it's the one you have issue with. The second one uses a slot, not a circular hole, to hold the crankpin.

What happens at the top of the stroke?

 

[hmmm27]

What happens at the top of the stroke?

Same as what happens at the bottom... except upside-down. 45deg before tdc/bdc, the pin switches from the inside end of the slot to the outside.

 

[paradisePhysicist]

Something like...

View attachment 285613The first one is normal, it's the one you have issue with. The second one uses a slot, not a circular hole, to hold the crankpin.

I will try running this in the simulation.

I tried this idea (post #21) into the sim, and it cannot work for linear input, no matter how many teeth are in the circular gear (the simulation is very forgiving).
https://grabcad.com/library/reciprocating-mechanism-1

I attempted to put post #44 into the sim but the instructions seem unclear, is the small gear anchored to the world space or supposed to slide around the big circle, I am not sure, I have a lot of questions about that drawing.

 

[hmmm27]

I will try running this in the simulation.

#55[edit: #52] * idea does rely on inertia ; if the piston-arm is comparatively too heavy, you'll still be stuck.

* so, how does one link to a specific post in a thread ?

 

[paradisePhysicist]

#55 * idea does rely on inertia ; if the piston-arm is comparatively too heavy, you'll still be stuck.

I tried post #52 it in the sim, it works if the friction is low, but if there is a decent amount of friction it won't work, gets stuck when the diagonal is in the outer region. Post #55 contains different concepts so idk which one you are referring to.

* so, how does one link to a specific post in a thread ?

You can use the quote button.

Also, I tried the idea I said was my favorite (post #38) and it sucks for linear input, back to the drawing board ig.

 

[hmmm27]

I tried post #52 it in the sim, it works if the friction is low, but if there is a decent amount of friction it won't work, gets stuck when the diagonal is in the outer region. Post #55 contains different concepts so idk which one you are referring to.

Sry, meant #52, not #55. post amended. Given length of slot was for an example, not a calculation.

 

[some bloke]

Hmm, based on this I tried to put two gears in the same spot in the simulation. Then a rod at then minimum location to gear 1, then a rod at 90° location to gear 2, both attached to the same piston. It will not work and how do I explain this... I used springs to give it some leeway incase the rotations and positions were not exactly precise, still didn't work, I think the reason is because when gear 2 reaches the maximum position, it wants the piston to go in reverse, but at the same time, gear 1 wants the piston to still go forward, therefore it can't rotate. I will draw a picture to make it clearer.
View attachment 285607
Actually I think I could have just used 1 gear for this actually, but in the simulation I used 2 gears stacked on top of each other.

I haven't tried your doodle yet but I will try that next.

Yes, this was similar to my first thought before I dismissed it for the same reasons!

The trick is finding a way to have 2 cranks which can be turned by a single con-rod. That's where my design came in, though I am struggling to find a way to make it on any online gear generators - none of them assume a gear is fixed and the rest rotate around it!

 

[paradisePhysicist]

Sry, meant #52, not #55. post amended. Given length of slot was for an example, not a calculation.

I will post gif of experiment, I don't think it will work at any length or configuration without assistance from rotational inertia (ie., a heavier gear or flywheel needs to be added.)

I am looking into the one cylinder locomotive, the only website I can find is this website, which says the diagram is wrong, so I haven't figured out how they built it yet. http://www.douglas-self.com/MUSEUM/LOCOLOCO/onecylinder/onecylinder.htm

Yes, this was similar to my first thought before I dismissed it for the same reasons! The trick is finding a way to have 2 cranks which can be turned by a single con-rod. That's where my design came in, though I am struggling to find a way to make it on any online gear generators - none of them assume a gear is fixed and the rest rotate around it!

Hmm can you edit your diagram because I'm a bit confused about it.

 

[some bloke]

I will post gif of experiment, I don't think it will work at any length or configuration without assistance from rotational inertia (ie., a heavier gear or flywheel needs to be added.)
View attachment 285646I am looking into the one cylinder locomotive, the only website I can find is this website, which says the diagram is wrong, so I haven't figured out how they built it yet. http://www.douglas-self.com/MUSEUM/LOCOLOCO/onecylinder/onecylinder.htm
Hmm can you edit your diagram because I'm a bit confused about it.

I think you have too long a slot on the gear there. I would have the slot at 45° from the radius, and significantly shorter than that!

The goal of the motion there is that when the crank stops turning from the forward motion, when the piston pulls back the connection slides along the slot and downwards, so that it is below the crankshaft axis and therefore will now pull the crank in the correct direction.

The issue there is that the crank will be turning & stopping throughout the cycle as the con rod stops driving it and instead slides through it.

I will try to come up with a clearer way to convey my idea, but it will probably involve making an animation to explain it as it's a bit hard to picture the way everything moves! This is an adaptation of a mechanism I designed years ago to create a long linear motion from rotary motion within a confined space. It effectively doubles the piston motion for the crank diameter, and also creates exact linear motion rather than rotary for the con rod. I'm adjusting it with an offset in the idea that it would create the dual crank effect I'm trying to come up with for you!

Can you explain which design you're trying to figure out on the link, as there are loads on there, and all of them seem to involve flywheels!

Can you also explain why a flywheel is not wanted if you are trying to create an engine which will turn continuously in one direction?

Ignore that, I re-read the thread and you're goal seems to be about creating a reversible cam & slider mechanism without relying on a flywheel. I wonder if this is one of those things which may prove mathematically impossible...

I'll look forward to trying to find out...

 

[gmax137]

Can you also explain why a flywheel is not wanted if you are trying to create an engine which will turn continuously in one direction?

I think everyone reading this thread would like to know this! Especially in light of this:

The issue there is that the crank will be turning & stopping throughout the cycle as the con rod stops driving it and instead slides through it.

That seems a fatal flaw for any useful slider crank mechanism.

 

[paradisePhysicist]

I think you have too long a slot on the gear there. I would have the slot at 45° from the radius, and significantly shorter than that!

What is 45° from the radius lol, a radius is a scalar with no angle.

The goal of the motion there is that when the crank stops turning from the forward motion, when the piston pulls back the connection slides along the slot and downwards, so that it is below the crankshaft axis and therefore will now pull the crank in the correct direction.

I will try another configuration of that but I don't think it will work.

The issue there is that the crank will be turning & stopping throughout the cycle as the con rod stops driving it and instead slides through it.

Yes, some efficiency will be lost with that as well.

I will try to come up with a clearer way to convey my idea, but it will probably involve making an animation to explain it as it's a bit hard to picture the way everything moves! This is an adaptation of a mechanism I designed years ago to create a long linear motion from rotary motion within a confined space. It effectively doubles the piston motion for the crank diameter, and also creates exact linear motion rather than rotary for the con rod. I'm adjusting it with an offset in the idea that it would create the dual crank effect I'm trying to come up with for you!

Ok cool. Looking forward to the animation but also in the meantime you could add more labels with the diagram, with the current diagram I'm not quite sure how to build it.

Can you explain which design you're trying to figure out on the link, as there are loads on there, and all of them seem to involve flywheels!

Oh the design that was suggested is the Neilson one-cylinder locomotive. The only problem is that seems to be the only website I've found of it online, and the book about it doesn't seem to exist.

Can you also explain why a flywheel is not wanted if you are trying to create an engine which will turn continuously in one direction?

Ignore that, I re-read the thread and you're goal seems to be about creating a reversible cam & slider mechanism without relying on a flywheel. I wonder if this is one of those things which may prove mathematically impossible...

Yes. I don't think it is mathematically impossible, but would be interesting to see some kind of mathematical theorem or axiom saying it is impossible.

I'll look forward to trying to find out...

Thanks.

I think everyone reading this thread would like to know this! Especially in light of this:

Flywheels reduce acceleration and require more force to move. Also would be nice to just have a device that converts linear to full rotational motion inherently.

That seems a fatal flaw for any useful slider crank mechanism.

True but if the slide is short enough maybe the efficiency lost is not so bad.

 

[hmmm27]

seems a fatal flaw for any useful slider crank mechanism.

Ball bearing in a tube crossing the crankshaft axis at the appropriate angle.

----------

Heheh, the actual(ish) solution is to use the concept in the DWFTTABCWHATEVER thread (but not the propeller implementation) to scoot a weight around the outside of the crankarm at twice the crank angular velocity,

Easy peasy.

 

[paradisePhysicist]

Ball bearing in a tube crossing the crankshaft axis at the appropriate angle.

Sort of a diagonal 3d angle or is the angle purely 2d?

Heheh, the actual(ish) solution is to use the concept in the DWFTTABCWHATEVER thread (but not the propeller implementation) to scoot a weight around the outside of the crankarm at twice the crank angular velocity,

Hmmm what thread is this? Also adding a weight seems to be a bit like a flywheel lol...

The issue there is that the crank will be turning & stopping throughout the cycle as the con rod stops driving it and instead slides through it.

I tested this in a new configuration as suggested it still locks up. The simulation is very forgiving so at first this isn't apparent, but if you test it by letting go of the throttle and only applying it periodically in bursts, then it will lock up at the 2 stagnant regions.

I think I can now define it as an axiom or more clear definition, an inherent linear-to-rotary mechanism can do full rotations without the use of momentum, a non-inherent linear-to-rotary mechanism must be boosted by momentum to do full rotations. Full rotations is kinda vague also cause you can just attach a little gear and get full rotations on the little gear, I think maybe the phrase "continuous rotations" is the ticket.

Anyway, as predicted, after looking at the graphs, the crank with the groove in it is somewhat less efficient, still does useful work though. The standard model rotates for about 53 seconds the crank with groove rotates for only 48 seconds when both start with the same velocity. Also the graph curves are not as smooth.

 

[gmax137]

Flywheels reduce acceleration and require more force to move. Also would be nice to just have a device that converts linear to full rotational motion inherently.

OK.

I think you always have some flywheel effect: the crankpin and the big end bearing are orbiting the crank centerline, plus there has to be a connection between the crank and the crankpin. This may not "look" like a flywheel but there will be some rotational inertia. How much is needed to get the crank to rotate past the top & bottom dead centers? I concede, there is not always enough without a flywheel (eg, the gas model airplane engines won't run unless the propeller is on the shaft).

 

[hmmm27]

Sort of a diagonal 3d angle or is the angle purely 2d?

2d : at two points in a full crank-rotation, the bb rolls to the other side, providing weight to push the crank rotationally.

[note: the next bit of the conversation regards a different mechanism]

Hmmm what thread is this?

Something about an unpowered vehicle moving faster straight downwind than the wind itself. Unintuitive... and nifty.

Here's the relevant vid.

Using that as a reference, imagine two concentric circles: a static one on the outside, a rotating one on the inside, with a little go-buggy in-between, whizzing around (exactly)twice as fast as the inside circle : for each one revolution of the circle, the go-buggy completes two revolutions.

Also adding a weight seems to be a bit like a flywheel lol...

Like everything else in the system which has a rotational component, it acts as a flywheel. Which has nothing to do with how it gets the crank through both tdc and bdc.

Basically, just like the ball bearing rolls over to the other side to add its weight at the right placement to push the crank through both tdc and bdc, the go-buggy manages to be in the right place at the right time twice during a crank revolution.

I think I can now define it as an axiom or more clear definition, an inherent linear-to-rotary mechanism can do full rotations without the use of momentum, a non-inherent linear-to-rotary mechanism must be boosted by momentum to do full rotations. Full rotations is kinda vague also cause you can just attach a little gear and get full rotations on the little gear, I think maybe the phrase "continuous rotations" is the ticket.

And you mean what by "inherent" and "non-inherent" ?

 

[paradisePhysicist]

2d : at two points in a full crank-rotation, the bb rolls to the other side, providing weight to push the crank rotationally.

Hmm I will try various configurations and see if it works, but I'm suspecting it will either help the device get past the 179 angle and hinder it getting past the 359 angle, or vice versa.

[note: the next bit of the conversation regards a different mechanism]

Something about an unpowered vehicle moving faster straight downwind than the wind itself. Unintuitive... and nifty.

Lol.

Here's the relevant vid.

Cool vid, seen it the other day. Didn't study it to 100% see how it works but I shall study this.

Using that as a reference, imagine two concentric circles: a static one on the outside, a rotating one on the inside, with a little go-buggy in-between, whizzing around (exactly)twice as fast as the inside circle : for each one revolution of the circle, the go-buggy completes two revolutions.

Like everything else in the system which has a rotational component, it acts as a flywheel. Which has nothing to do with how it gets the crank through both tdc and bdc.

Looking for non flywheel solutions but this idea does seem interesting for experimentations.

Basically, just like the ball bearing rolls over to the other side to add its weight at the right placement to push the crank through both tdc and bdc, the go-buggy manages to be in the right place at the right time twice during a crank revolution.

And you mean what by "inherent" and "non-inherent" ?

inherent means the system works without an external boost. A working inherent system gives continuous rotations even with heavy friction and no momentum boost. The flywheel acts kind of like a boost mechanism to get past the limitations of the system. An loose example is a car that goes 190 kph topspeed, does not go 200 inherently, but you can use a nitro boost to go to 200 kph.

 

[hmmm27]

Looking for non flywheel solutions

So, since it isn't actually a flywheel, beyond having a rotational component, like many other bits and pieces of the system like the crankshaft, crankarm, piston-rod, etc...

inherent means the system works without an external boost. A working inherent system gives continuous rotations even with heavy friction and no momentum boost. The flywheel acts kind of like a boost mechanism to get past the limitations of the system. An loose example is a car that goes 190 kph topspeed, does not go 200 inherently, but you can use a nitro boost to go to 200 kph.

Without rotational momentum, a basic system (piston->piston-rod->crank-arm->crankshaft) is never getting past tdc/bdc, regardless of how friction-free the bearings are.

 

[some bloke]

Here's a mechanism which I think would work. You could have the central oval cam (which doesn't rotate) rock back & forth to change direction.

I am imagining the piston to the right of the wheel, pushing the pin to the left. The pin interferes with the fixed cam, so cannot move past it. The pin is in a slider. The image is the 4 stages of half a stroke, and it behaves the same on the way back.

The idea is to have a mechanical movement where the first parts of the pistons stroke are converted into non-linear motion - in this case, the pin is deflected sideways by the central cam, and starts the wheel rotating anti-clockwise. At the start of each rotation, the pin slides along the cam and starts the wheel going in the correct direction.

If you're interested, this was inspired by the mechanism inside of a clicky pen, which converts repeated reciprocating motion to rotary in a similar way to what you are requesting - have a look at a transparent clicky pen and you'll see the mechanism spins as you click it!