Benefits of a two stage transmission?

[Cozz]

Hey Extrapilot,

I'm not sure I understand some of your response, I'll try to clarify...

Of course I get to proclaim something a good design (I don't get the should be part), it's my opinion, and I'm entitled to it. I'm not familiar with the 425?

It is true that the radial component of the force does try the separate the gears. But, I'd say that deformation in the gear itself would be negligible and the concern would be that force displacing (bending) the shaft, to separate the gears. Thats why I like the design of a support both above and below the gear. The more points and further spread, at which you resist this force, the smaller and lighter the supports need be.

I'm not certain how you derive simple to be generally heavy or bulky? I'd argue quite the opposite.

I'm not certain you understand the cost that I'm referring to. The cost to design and manufacture, in order to generate a profit. This also spills over to "simplicity", as a component that is simpler, is usually less intensive to manufacture, and cheaper... 99% of products designed, are "designed for manufacturing", and compromise needs to take place for this... Comparing two parts, that meet (not exceed) the required performance (strength, weight, durability, etc.), the simpler design is always wins.

Hope that helps. If it's just taste/opinion that separates us here, then so be it.

[Cozz]

...and to bring this back to the context of a single-stage vs multi-stage transmission, you would need to assess the gain in RPM vs any associated losses (weight, complexity, frictional/inertial losses, etc.)

A perfect example would be calculating climb rate (could easily be done in excel), using the Protos 380 and Goblin 380. These two are the same "size" heli, and can be compared given all other things are equal (motor and battery). I might try this 'n report back...

[extrapilot]

This isnt art. We are not sharing opinions about color schemes. We are talking about your proclamation that ‘flexing under load should not be an issue.’ It is an issue. I don’t know if you are generalizing (i.e. a well-designed gear/frame ‘should not’ deform to a degree that performance suffers), or if you are implying this as fact (that this is the case- though you haven’t measured it). That is the ‘should’ problem.

The main issue for these helical single stage setups with a plastic main gear, as was explained earlier in the thread, is axial gear deformation.

Regarding mass/bulk and your ‘simple is ALWAYS better’ proclamation, find two power supplies with equivalent current and voltage ratings- one switching, one conventional- and look at mass and volume. Oops. This extends to most systems- it is just less obvious for many of them.

What about a single part? To stick with our niche, take a 3D heli main blade. Do you actually believe that a simple blade- a 2D NACA0012 extrusion can compete with a well designed CFRP 3D-contoured blade, for a given mass/span/ average chord? The complex blade will outperform the simple one in every way that matters, at a lower mass and drag for equal lift. If you try to dumb this down with your caveats about ‘meets but does not exceed,’ the advanced blade will simply be lower in mass and have reduced volume and drag for the same lift at a given pitch and RPM.

You want a system example? DFC. In your world, it is simple due to low part count, and no need for a complex scissors swash drive. And, it meets the specs for the application (can provide pitch control to blades via a swash etc etc). But, to do it, the arms have to be heavier and have a larger cross section. Because flap leads to high bending forces on the arms, the grips need to be stronger, which implies mass and/or volume increase. And, none of that is beneficial- you want the least moving mass in a cyclic process- and you want the lowest cross section possible to minimize drag.

This just goes on and on. You may try to caveat this to the point where you have ‘same profile, same mass, same volume, same material, same same same,’ but I hope you can see that then you end up with identical parts, and the question is moot.




Regarding this test you mention in #22 here- even if you could calculate climb rates in Excel (you cant), you have 10s of variables here that matter hugely in ways you cannot know without advanced testing. And even if those variables didn’t exist, what do you hope to gain? Some universal conclusion about transmissions, where they can only apply to a particular design in a particular frame anyway?

[Cozz]

Oh boy, this has gotten a bit on the argumentative side...

No this isn’t art. But, like art there is no right/wrong answer, there are just different solutions (to the same problem). As the original question asked, is there a benefit to a two stage transmission? Short answer is yes, there’s higher blade RPM and potentially a narrower frame. Where it gets interesting is asking, at what consequence? And, is it worth it? AGAIN, there’s no real tangible right and wrong.

Now, flexing... I am most certainly not suggesting there isn’t any, there HAS to be, that’s physics 101. But I dispute that there is enough to cause the gears to be pushed out of mesh. Firstly, the tangential force is the (far) larger component, and not responsible for pushing the gears apart. Secondly, the radial force is “reacted” by the stiffness of the gear itself (and they’re incredibly stiff plastic. Plus, the larger the diameter, the more material there is to be stressed, and “absorb” this force), the bending stiffness of the shaft and the shaft supports built into the airframe. All of these components work together to resist the separating (axial) force, preventing them from being “pushed” out of mesh.

Sorry, you’ve completely lost me regarding; mass/bulk, power supplies switching and conventional...

Regarding my comments of simplicity in design, I think you’ve moved into a purely argumentative space. The concept of simplicity can only be used when comparing apples with apples, not apples with watermelons (like welding machines to formula one cars)... Again, I’m sorry you’ve lost me regarding blades. All main blades I’ve seen are a symmetrical naca profile(?), unless your referring to tip designs? This is a design evolution, seeking increased performance, an entirely different argument... You can compare differing head designs, that being; one with, one without (DFC) followers. But, you’ve misunderstood my context of “specification”, as I mean performance specifications; as stiff as, as fast as, etc. (I don’t fully understand the in’s and out’s of head design, so unfortunately I can’t comment)

Of course you can model (calculate) and compare climb rates using excel! What you’re referring to is ACCURACY. A common misconception is that a model (I’m talking mathematical) isn’t useful unless it’s deathly accurate, which is complete BS!... You can roughly model the heli (a “point-mass” analogy), roughly estimate lift at various RPM, and thus (roughly) investigate the effect of each variable change. It simply requires a ‘lil engineering/mathematical/physics common sense and sense of what’s reasonable.

[Cozz]

Just started to build up the calc's in Excel, and it didn't take long to debunk this topic (which has probably already been discussed elsewhere, and teaches me not to believe what I read in forums without validating).

Firstly, the Golbin 380 isn't a "compound" geartrain, it's a "simple" configuration. Whoops!

Secondly, I moved on to compare the 500 with the Gaui X5, and looked up ratios to find the simple geartrain favours higher RPM. I then investigated further, to see what gears were used at the motor.

So, seeking higher RPM is not the goal of heli’s with compound geartrains (that I looked at), they’re simply seeking higher torque... All Goblins have a larger gear/pully/whatever at the motor (and smaller main gears, in the case of the 380), and use motors with lower Kv ratings... There are two ways to approach performance; seek higher RPM at a lower torque, or seek higher torque at lower RPM... think two bicycles side-by-side, travelling at the same speed, one rider pedalling faster, the other pedalling slower... it also explains the greater gear-face size.

[Block137]

Originally Posted by Cozz

....

Now, flexing... I am most certainly not suggesting there isn’t any, there HAS to be, that’s physics 101. But I dispute that there is enough to cause the gears to be pushed out of mesh. Firstly, the tangential force is the (far) larger component, and not responsible for pushing the gears apart. Secondly, the radial force is “reacted” by the stiffness of the gear itself (and they’re incredibly stiff plastic. Plus, the larger the diameter, the more material there is to be stressed, and “absorb” this force), the bending stiffness of the shaft and the shaft supports built into the airframe. All of these components work together to resist the separating (axial) force, preventing them from being “pushed” out of mesh.

I think he meant the up-down force caused by helical(slant teeth) gears.
Which in this case, acts perpendicularly to gear plane so it is torque bending the gear not pushing apart.

[Cozz]

Hey Block

...put simply, this component is of little concern. Helical gears are cut at a very slight angle.

When the gears "mesh", they do so at a point, and that point "slides" down the gear face as the gears turn... the force components are illustrated well here:

https://www.controleng.ca/servosoft/...urce/Gears.htm

Basically, the gears in any geartrain (as in an RC heli) won't come out of mesh through normal operation. Not unless they're incompletely designed, or badly neglected to a point that wear intervenes.

[learnedthehardway]

From owning a Chase (3 belts) and a old Protos (1 belt) the only real benefit I see is reducing stress put on each belt by dividing it across 2. Can tolerate higher torque loads that beat up the belt on a single stage setup.

Other than that it's just more parts and more drag. At least on the two designs mentioned the motor pulley/pinion is the only means of changing the gear ratio so no difference there.

Chase has short flight times and is a PITA to work on compared to my Mini Protos. Vastly prefer the Protos for those 2 reasons.

If you fly hard smack 3d the Chase would probably be a better choice.

[Cozz]

Sorry, I hate to be that guy... The stresses aren't distributed across the number of belts present, they're passed from gear-to-gear through the belt, or mesh... bigger power value, bigger belt.

If you assume there's no losses (there obviously is, through friction, inertia, etc. but you assume there isn't for ease of investigation sake), the motor has a power value "at the shaft", that equates to a torque (force at the gears outer edge, dependant on gear radius/diameter), and it's passed from shaft-to-shaft. The power from the motor shaft gets passed onto the main shaft. Small diameter gears spin faster, with less torque, big diameter gears spin slower with more torque, but both have the same power value.

You can pass along bigger power values, more gradually (stress wise), using more stages, and that can lead back to what you were saying... and also, like you mentioned, leads to more weight and losses, and complexity. That equates to being a PITA for maintenance (especially when packaging in a small airframe).

You can see this design philosophy taking shape with various heli's; the goblin 380 is one stage, along with the new protos 380. I think the threshold for two-stage geartrains is within a 500 sized airframe.

[extrapilot]

Originally Posted by Cozz

Hey Block

...put simply, this component is of little concern. Helical gears are cut at a very slight angle.

When the gears "mesh", they do so at a point, and that point "slides" down the gear face as the gears turn... the force components are illustrated well here:

https://www.controleng.ca/servosoft/...urce/Gears.htm

Basically, the gears in any geartrain (as in an RC heli) won't come out of mesh through normal operation. Not unless they're incompletely designed, or badly neglected to a point that wear intervenes.

There you go again. You don’t get to proclaim some self-implying statement as anything but that. ‘A gear that is properly designed to not deform won’t deform.’ Yea- great insight.

Back to reality-
It is very typical to see 5-8hp peak at the motor for 700-class machines now. At 8hp at 20,000RPM at the pinion, with a 20deg (typical for our setups) helical angle, and a 0.7” pinion dia, you will see 13lbf axial force.

Now- you go and hang 13lb from the periphery of your standard main gear, and if you have a real dial indicator- indicate the deflection. Oops. Try at half that mass. Oops. Then, heat the gear to operating temperature at 2min into the flight and measure it. Oops^2

You are a guy who- instead of putting a motor/transmission on a dyno to read efficiency- wants to instead test the climb rate of different machines using incredibly complex rotor systems and infer from that powertrain efficiency. To complicate that more, you are going to estimate climb rate in Excel. I don’t know if you realize you absurd this is- but it is entertaining.

[Cozz]

Haha, there you go again, trying to start an argument.

Nothing I'm putting forward here is an opinion, just First principles of basic physics and machine dynamics, that is learned in any mechanical engineering degree... don't take my word on it, look it up in a text book!

Your the only one here talking about a 700. But, the same applies.

Standby, I'll do the math on these gear forces, and prove the numbers.

(you clearly don't realise how powerful/useful excel is, but that's beside the point. You don't seem to be able to grasp how modelling works, and how/why details are "neglected" within reason... look up "discrete time-step modelling" and "quasi static assumption")

[Cozz]

Didn't need to... see attached worked example (note:at twice the helical angle measured on my X3 main gear, of 15deg)

...from this data you just need to back calculate bending stress in the shaft, and gear itself (depending on shape factors and material)

[Cozz]

...and here, read this:

https://www.nar.org/wp-content/uploa...-Report_v2.pdf

(after a 2 minute search on google... there are many other examples of "discrete numerical models", all carried out within spreadsheet software)

And attached, my quick calves, before seeing what was going on...

[Cozz]

The calc's for deflection of a Gaul X3 main gear... which I believe is made from GF-30 Glass Fiber reinforced Nylon (E of 12x10^3 N/m^2)... as a 2D "slice" of its shape is a muddle, I chose sections of varying thickness from 6mm (it's MAX) to 3mm (it's MIN). In reality it would be somewhere in the middle.

It's a rough estimate, but you can see it takes 512kg (5kN) to deflect this material 1.2mm at a length of 33mm (the resisting length of the main gear)!

People very very VERY much underestimate "plastic" parts. Basically, it'll break before it bends, so unless your snapping main gears, you have no issues!

[guyguerra]

For me none of this really matters. I have Logo's and Whiplashes with one main large gear, and I have Goblins and Mostro's that are two stage. I've flown these hundreds of times, many times in the same day. I can not tell the difference in the way they perform. All of these Helis have a good réputation and Logo's verses Goblin's is really just a personal choice. If you notice a difference it's probably the electronic set up and not the main gears. I like them all.

[Cozz]

I completely agree

(...and within my original post before this all got sidetracked)

[extrapilot]

Originally Posted by Cozz

Haha, there you go again, trying to start an argument.

Nothing I'm putting forward here is an opinion, just First principles of basic physics and machine dynamics, that is learned in any mechanical engineering degree... don't take my word on it, look it up in a text book!

Your the only one here talking about a 700. But, the same applies.

Standby, I'll do the math on these gear forces, and prove the numbers.

(you clearly don't realise how powerful/useful excel is, but that's beside the point. You don't seem to be able to grasp how modelling works, and how/why details are "neglected" within reason... look up "discrete time-step modelling" and "quasi static assumption")

Look it up in a text book? Some of us do this for a living, using real tools- you know- Ansys, CFX, Catia, etc.

Your rough estimate suggests that an X3 gear will deflect 1.2mm at its periphery under an axial (tangent) force of 512kg?

Attached is FEA for a 700 gear. It is under a tangent force of 20.4lbf at a small section of its periphery, which is what it will see at 8hp with a 30deg helical angle on a stock pinion.

The material is a Delrin type, as commonly used in the this industry. At this loading, there is a deflection of very nearly 1.5mm.

Why do you think that better manufacturers mill herringbone gears from billet or rough cast Delrin, if axial loads and deformation are not an issue? Just to add cost and complexity to the manufacture?

What in the world are you talking about with this ‘break before it bends’ nonsense? Even if this were CFRP- it will bend- a lot- before it breaks, in anything like this type of geometry.

Anyone can take their 600-700-class machine, bolt its frame securely to a jig, and hang a 20lb bowling ball from the periphery of the main gear. Watch what happens.

[Cozz]

Congratulations, you use CFD software, so therefore you'd know in the background of your software, runs code using the same equations anyone uses in a spreadsheet (with half a clue), and any result can be replicated in a spreadsheet, only slower, and without a pretty interface... you've missed the point entirely! I'm talking about rough calc's, "back-of-the-envelope" calculations to investigate what's going on.

Probably(?), some plastics are "brittle", some plastics are "ductile" ...the numbers suggest that it can handle any force it'll ever see in this heli. THAT's my point ...aaaand I don't care how commonly Delrin is used in the hobby, it's not in my heli, nor did I use that materials' data, your comparing apples to oranges again.

Okay, you've investigated your ULTRA perfectly (haha) accurate numbers to be 1.5mm... what does that mean? Apply common sense to your findings? Apply theory to your numbers? my guess is that that deflection amounts to very little, it's negligible, apart from a brief increase in friction and decrease inefficiency.




Anyways... I think that'll do it for me. It's been a pleasure Extrapilot

[Cozz]

*FEA... CFD... same thing.

[Weule]

I'm guessing the 2-stage design also reduces the gyroscopic effect of having a large main gear.