eCVT

[12p3phPMDC]

Inspired by the Miles' epicyclic retrodirect with the opposing spragues and
these picture examples, I want to explore the idea of eCVT using two motors
and two independent controllers.

The system would use the planet carrier as the output, the ring and sun gears
as inputs. It would then have two opposing spragues on the ring and sun gears
so that when torque is supplied from one motor, lets say on the ring, the sun sprague
locks and the carrier is driven. Vice versa, if the sun gear is fed torque from the other
motor, the ring sprague locks, and the planet carrier is driven.

You could probably build the system without spragues (see examples), but
It seems the stall torque and current would be high and cause heating problems.
Therefore the spragues would simplify the control at the expense
of regen. But you get dual redundant operation, i.e. if one motor fails
the other could still drive the system.

I have 62 tooth GT pulleys that would be perfect for a pressfit over
the ring gear of a 75mm Matex Planetary. Copying the overhead lift example,
I would then do a reduction off the ring drive motor to the 62 tooth GT pulley.
I would then do a direct drive out of one motor and feed the sun gear.
The reductions would need to be matched so
that the torque inputs from the two motors would be matched.

Sun to Carrier Reduction: Rs (fixed for given plantetary)
Ring to Carrier Reduction: Rr (fixed for a given planetary)
Motor to Ring Reduction: Rmr (adjustable from off the shelf pulleys)

The total ring reduction would be Rr*Rmr. Set to equal to the Sun to carrier reduction, Rs
and solve for Rmr. That way the torques from each motor relative to the carrier would be closely
matched at the same motor speed.

twin_motor_planetary.jpg (98.31 KiB) Viewed 2736 times

planetaryCVT.jpg (81.31 KiB) Viewed 2738 times

[Miles]

Did you see this?
http://www.endless-sphere.com/forums/viewtopic.php?f=28&t=7358
http://groups.google.com/group/nohassel/web/nohassel-faq-frequently-asked-questions?hl=en

[12p3phPMDC]

Thanks Miles.....

It pays to search....

[12p3phPMDC]

Whoa...you're too fast for me Miles...

pays to search...I posted on your thread, and it was gone it 1 second flat.

*** Since you live in the UK, wouldn't a better name be Kilometers??? j/k

(Back in Jazz Band, some guys used to call me that due to my Miles Davis influence, I was honored )

[Miles]

See also:

http://v3.espacenet.com/textdoc?DB=EPODOC&IDX=US5242335&F=0

http://v3.espacenet.com/textdoc?DB=EPODOC&IDX=US6286616&F=0

This is the system used on the iZip Express

[12p3phPMDC]

Cool!!

I am a flea on the shoulders of giants...

[Miles]

How about a throttle in each hand?

[12p3phPMDC]

That's exactly what I was thinking. It would be slick
and simple.

It would also be nice to have the torque mixed by a single
controller.

Man, that NoHassel drive is cool. For, now, something
that would work within the limitations of the off the shelf stuff
would be preferable and doable for me.

I am amazed at the amount of innovation that is occurring or has always
been there but never knew about it.

[12p3phPMDC]

Worked out some back of the envelope gearing calculations
based on OTS Matex planetaries.
Matex defines the reduction from the sun (input) to carrier (output).
i.e. 3:1, 4:1 etc.

using the overhead crane example:

r : reduction

Direct Drive Motor to Sun, rs = 1:1
Reduction Drive Motor to Ring, rr = ?

R: number ring teeth
S: number of sun teeth

Sun/Carrier reduction: rsc = 1 + R/S

Ring/Carrier reduction: rrc = 1 + S/R

If you want matched drives assuming identical motors, then

rsc = rrc * rr

rr = rsc/rrc

Sun/Carrier       R/S           Ring/Carrier         Ring Reduction    Pulleys ots
rsc                                       rrc                     rr                        
3:1                    2                 1.5   : 1             2:1                60:30
4:1                    3                 1.33 : 1             3:1                 60:20
5:1                    4                 1.2   : 1             4.17:1            72:17

The 5:1 starts to become less practical to feed the ring gear due to the large and small pulleys...
It would probably work, though. However a 5:1 gives you a nice initial reduction
on the sun gear. High powered ~70 kv motors would be nice. That would keep the gear speeds
from getting out of hand.

[12p3phPMDC]

bump for Ron's reference.

If one could take two outrunners back-two-back like
the airplane example, with the back motor driving the shaft
and the front motor driving the front bell....

The planetary gear will sum the torques supplied to
the sun and ring gears through the planet carrier.
The planet carrier becomes the output.

Then drive the sun gear with the shaft from the back motor and the ring gear
with the front motor bell.

You can have two throttles, one connected to each
motor. Then if you apply torque to say the front motor/ring gear,
then the planet carrier will be driven through the ring to carrier
reduction (if the sun is held). Vice versa...If you apply
torque through the sun gear, then the planet carrier
will be driven through the sun to carrier reduction (if the ring is held).
Now, supply torque to both and the planet carrier is driven by
both. A nice feature is that both inputs are reductions.

If you don't have some one way clutches, then the motor needs to be able
to provide stall torque continuously if *only* one motor is providing torque.
If both motors are providing torque, you may be able to
do away with the one ways entirely....
I'm not 100% clear on this yet...

I think a sophisticated controller with reverse capability can make this thing climb a mountain.
You can have extremely low ratios when you run one in reverse and the other forward,
and make the planet carrier creep along..
And you could run it in reverse....(eliminate the one way bearings though).

In the earlier post, I wanted matched torque inputs through gearing,
but, you might be able to match torques by selecting different motors..
i.e. 180kv for the large reduction and 130 kv for the small reduction.
There's nothing that says you have to match torques either...
I'm just square and symmetrical.

I'll post a pic when I get home tonight of the matex unit I have.
I'll need a higher torque unit when I pull this off. A design
is starting to solidify... I gotta finish the trike first.

Matex planetaries in a 75mm ring have a 12mm
splined input shaft.

Anyway, I going to pick up a small mill /lathe combination machine this weekend!!
I'll be able to do some lathe work and pull this off me thinks..

anyway, for other references, search for the Prius Hybrid synergy drive.
It's just a planetary with an extremely sophisticated controller, two motors, and an ICE.
It's a different layout, but similar in application. The motor/generator 2 is tied to the ring gear,
the ICE is tied to the planet carrier, and the motor/generator 1 is tied to the sun gear.

[amberwolf]

I am not sure if it is helpful, but I started a thread about this kind of thing, not realizing this thread was here (didn't find it in the brief search I did first).
viewtopic.php?f=2&t=15618&start=0

Basically thinking of taking an F&P motor cored out as the ring motor, whatever matching small gears I might have around as planets, (assuming I could make ring and sun teeth to match), and a modified outrunner/hub motor as the sun motor.

F&P axle removed from the cored out motor gets attached to planet carrier.

F&P rotor gets new section added *inside* the cored out stator, with ring gear on it's inside circumference. Bearings like little planets to ride the stator on.

Sun motor gets cored out, hollow tube axle installed with larger bearings to attach/support original hub casing back to it, with sun gear on outside of hub casing (between the flanges). Smaller bearings inside tube axle to ride on F&P axle.

Whole thing sandwiches together to be about the same width the F&P started out to be, maybe a little more. Same outer diameter, too.

Dumb idea?

[12p3phPMDC]

Dumb idea?

No, I think it is a great idea!! that's why I started this thread!!

I've slowly been aquiring materials to do this.
I got the planetary, the ring and sun pulleys, the one-way clutches, and
various support bearings, and some 4" aluminum rod for the housing.

I need to collect a few more bits for the mill lathe combo, i.e. boring bars (or make my own, thanks Thud!!),
and some cutting tool stock, etc...

Stock Drive products has a tutorial on planetary gear design. Miles also gives you the criteria in
his RD two speed planetary thread. I don't have them off the top of my head, but there
are rules to follow in terms of the number of teeth for the sun, planet, and ring gears for the whole thing
to spin without interference or binding (I think thats why).

Another interesting thing, is that Rohloff uses roller bearings on the planet carrier gears. This helps
to increase efficiency of the planetary. The matex modules are bushed. Miles said he was
planning on roller bearings as well.

I like the nested idea you've come up with, although I'm still trying to wrap my head around it...
I think that the large slower motors would be a great fit for the planetary, as speed decreases torque handling.

I've been searching for a reference for gearing diagrams....or skeleton diagrams....that let you see which bits are stationary relative
to the case, or axle, etc....but I keep coming up short! It'll help dump concepts to paper drawings that can
help get the point across..


anyway, I"ll come back to this again tonight to try and understand what you are proposing!

[Miles]

Miles also gives you the criteria in
his RD two speed planetary thread.

I'll post them here, too:

- The number of teeth on the Sun plus twice the number of teeth on the Planets equals the number of teeth on the Annulus.

- The number of teeth on the Annulus and Sun gears together, divided by two, must be an integer.

- To have equally spaced Planets, the number of teeth on the Annulus plus the Sun must be divisible by the number of Planets used.

[amberwolf]

I posted a bit more in the other thread in reply to stuff there, might help clear up some of the design fog I left here. Will make sketches as soon as I can get back to Lightwave3D (gurgle scatsup sucks for this, at least for me).

As to which bits move in relation to what, at least for what I intend, it's fairly simple:
--Planet carrier is fixed to the output shaft
--Ring (Annulus) fixed to outer motor (F&P)
--Sun fixed to inner motor.
--Rotating ring and Sun at same speed (opposite directions) gives zero output on planet carrier. (I think--I'm not great with planetary stuff yet, and I get confused when not actually touching/moving the parts). This is "idle"
--Increasing or decreasing speed of one of them relative to the other increases planet carrier speed.
--Both motors can at idle run at their optimal most efficient speed, taking as little current as possible.

One catch with a larger design like what I want, with nested motors, is that the planet carrier is going to be larger diameter, and must then handle more torque against it's attachment to the shaft. On the other hand, it also means it can develop more torque than a smaller diameter carrier, right?

The "idle" doesn't have to actually be running all the time, either. If both motors are spun down at the same rate, they'll still be in idle as far as moving the output shaft goes, and can consume less power (depending on where all the losses are in the system, like switching losses in the controller, etc, and where the "sweet spot" is with the motors' power consumption vs RPM). So it doesn't have to use up power all the time, only when you need to move it.

And even for very low output speeds, the motors themselves are still spinning at their more optimal fast RPMs, so that during the pretty common mode of starting up from a stop, you're not wasting all your power--instead you can use your power to drive the vehicle at speed. (yes, it will still take power to get the torque to startup from a stop, but it's not going to be huge currents doing it, causing overheating of the motor because it can't spin fast enough to keep BEMF high, and can't cool itself off by airflow).

All of that is assuming I am grasping how this thing should work, which is certainly not guaranteed. I can see it in my mind's eye, but sometimes I get splinters in it and don't see so well.

[Miles]

Will make sketches as soon as I can get back to Lightwave3D (gurgle scatsup sucks for this, at least for me).

Why don't you download the trial version of Alibre? You can continue to use it (with a limited number of parts) after the trial ends. Another alternative is the free version of One-Space-Modeller: http://www.ptc.com/offers/tryout/pe2.htm As you say, Sketch-Up is far from ideal, for this kind of thing.....

[12p3phPMDC]

Sounds like you have a good grasp of it all to me Amberwolf....

albeit, I'm still learning too...so I am definitely no expert.

What kind of motor would you nest inside the F&P motor? another smaller outrunner?

The complexities of reverse and the super high torque modes with the ring gear and sun gear spinning in opposite directions makes
me squirm a bit.

I think that you bring up some good points in regards to "idle" modes and current draw. I think that you can avoid huge stall currents at low rpm
by playing some tricks with the controllers. You would need a dead band possibly around low rpms where you would benefit from
reversing one motor and having the other in a forward direction to give you "crawl" gear. Then as you speed up, the other motor
would need to quickly change directions and go to a fast enough forward direction to avoid the dead band zone.

That's why I want to put some one ways in....to isolate the inputs, and to add redundant operation.
In a hybrid app, you could run the ring gear off a honda mini 4 stroke input and run the sun gear off the BLDC.
That gives you full "OR" operation, 1 or the other or both.

Then you have redundant operation via freewheels or oneways for the outputs too...You could have a tri-brid, Human Power, Electric, and Gas.
The one-ways mess with your ability to regen, but oh, well. you can't have it all unless you work at Toyota or Ford R&D
with the hybrid synergy drives. Hell, maybe we can figure it out?

I also think that there is a great potential for a Prius like hybrid drive here. It's pretty cool how the Prius uses two MGs (motor generators)
and an ICE. I think that they Prius has 7 possible modes of operation.. It doesn't have a dedicated starter...it uses one of the MGs to
start the ICE through the planetary....which is cool IMO.

This is from memory, I'll try to visit the prius sites and touch up my memory...

MG2 (Ring Gear drive coupled to differential)
ICE (tied to planet carrier)
MG1 (sun gear)

1. Full Speed Ahead, All three prime movers are contributing to forward motion and torque
2. MG2 only
3. MG2 and ICE providing forward torque
4. MG2 and ICE with MG1 generating
5. MG2 regen and MG1 regen
6. MG2 and MG1 full torque
7. Reverse

[Miles]

In the No-Hassel system, the speeds of the two motors are controlled automatically:

This is an overview of the control logic of the Nohassel propulsion unit´s two motors

The functionality of the two motors is controlled electronically. The basic function is quite logical and can be described as this control loop:


A) set the most powerful motor (LRK) to a speed that is within range of epicyclic gearing for the wheel RPM and optimized for LRK efficiency
B) set the speed of the smaller motor (HS) to a speed that matches LRK and wheel RPM (taken from a premade table in the master controllers program memory)
C) check if wheel speed is correct (based on sensor input from pedelec, brakes, throttle, current sensors, temperature sensors)
D) adjust target speed if needed based on input from C) and goto A)

[amberwolf]

Why don't you download the trial version of Alibre? You can continue to use it (with a limited number of parts) after the trial ends. Another alternative is the free version of One-Space-Modeller: http://www.ptc.com/offers/tryout/pe2.htm

Thanks! I will try both of them and see what they offer, and how intuitive their GUIs are (very important).

What kind of motor would you nest inside the F&P motor? another smaller outrunner?

Yes. Probably a rewound cieling fan motor, as they come in a range of small sizes and I happen to have a few.

The complexities of reverse and the super high torque modes with the ring gear and sun gear spinning in opposite directions makes me squirm a bit.

I don't really understand it enough to be squirming yet. After I play with ideas more I will probably begin to understand enough to get concerned. More likely I will have to break a prototype to grasp the problems.

I think that you bring up some good points in regards to "idle" modes and current draw. I think that you can avoid huge stall currents at low rpm by playing some tricks with the controllers. You would need a dead band possibly around low rpms where you would benefit from reversing one motor and having the other in a forward direction to give you "crawl" gear. Then as you speed up, the other motor would need to quickly change directions and go to a fast enough forward direction to avoid the dead band zone.

If I understand what you're saying, it seems easier to just keep them at higher RPM for the low-speed output, since it seems to me like the motors I've used so far (brushed) take less power at higher speeds (not including riding against wind resistance) than they do at lower speeds, given doing the same work. They don't get as hot, either.

That's why I want to put some one ways in....to isolate the inputs, and to add redundant operation.

I don't want to use anything like that because I'd like to have it capable of regen.

In a hybrid app, you could run the ring gear off a honda mini 4 stroke input and run the sun gear off the BLDC.
Then you have redundant operation via freewheels or oneways. Then you could have a tri-brid, Human Power, Electric, and Gas.

Interesting....

The one-ways mess with your ability to regen, but oh, well. you can't have it all unless you work at Toyota or Ford R&D
with the hybrid synergy drives.

Maybe use something like dog clutches to engage or disengage the drive of each input; then you could freewheel or regen, in any combination?


Regarding the Prius stuff, I don't pretend to understand how more than two inputs works. I haven't figured out a way to wrap my brain around it yet. It *sounds* great, but I just don't "get" it yet. I'll figure it out evnetually.

[12p3phPMDC]

What happened to the rest of your post Miles....?

I was trying to quote it...yikes...The fastest moderator on that side of the Atlantic strikes again....

So, as far as the NoHassel controller goes, it basically mixes the motor torques automatically to avoid
spinning the other side backwards...i.e. when torque is put on the ring gear and there is large intertia on the carrier (i.e. my bike and weight),
the sun wants to rotate backwards. For the design, I'm proposing, the freewheel would prevent that. For NoHassel and Solomon,
the motors would push back via control algorithms.

Its the special cases that were in your unedited post that mess with me......
I think two freewheels help to simplify these problems for me anyway...
You still may need really good starting torque from either motor if one motor is running and preventing the other
from starting due to back driven torque from the sun gear (or riing). This may require sensored solutions.
If you are feeding the planetary with reductions,it seems that the reductions will reduce the back drive torque at the motor.

I don't claim to understand all the Prius modes, I just know they exist.

[Miles]

What happened to the rest of your post Miles....?

Sorry..... It's all in the FAQ: http://groups.google.com/group/nohassel ... -questions

[amberwolf]

Well, it looks like while the type of unit I wanted to make (based on the Solomon unit) might be a *form* of IVT, it does not actually allow for different ratios for torque conversion. Meaning, it is indeed a variable-speed unit, but it is not variable-torque, in any traditional way.

Whatever the torque input is from the motors will be summed and presented at the output. So you'd still need high-torque motors on either end of the shaft (the sun and ring gears), to get good torque out of the planet carrier.

It does still allow you to spin both motors at very high speeds while keeping the output shaft at a very low speed, but that's all. I guess it would really be only good for if you had two motors that were half the torque you needed, you could connect them both with this device and drive one output. But there are other methods to do that.

Here's the original thread with the info, on DIYEC. Link starts at end of relevant part of thread, you can read upward from there if you want to see how the discussion reached that point.
http://www.diyelectriccar.com/forums/sh ... post164318

Also crossposted this to the other IVT thread.

[12p3phPMDC]

Amberwolf,

I think you bring up an excellent point here.

There is no dynamic gear ratio with this kind of system, only summing of torque from dynamic sources.

If you combine a planetary and a CVT, you can arrive at an IVT with one prime mover

http://cvt.com.sapo.pt/ivt/ivt.htm

Its pretty cool all different forms these things can take.

[Miles]

More split-torque technology:
http://nexxtdrive.com/

Specifically:

[Miles]

Article in AtoB on what has subsequently become NexxtDrive technology:
http://www.atob.org.uk/electric-bikes/e ... nsmission/

[Jeremy Harris]

That looks perilously close to the Toyota HSD principle to me, in fact so close that I'd be inclined to think that it might infringe on some of Toyota's patent claims.