2 stroke thread (with occasional F1 relevance!)

[Tommy Cookers]

these 2-wheeler CVTs seem to have inherently poor mechanical efficiency ie 90% at best ?

presumably this applied to the DAF/Volvo type (at the time I assumed otherwise) that was quite a good seller
ie no-one mentioned how the 10% emerged as heat

a fixed ratio (ie flat or toothed belt) bicycle transmission is about 98% efficient


these posts about manual slipping of the clutch (to give the engine more time at its peak power) for better acceleration !
the Burgman is another CVT whose manual mode makes the engine spend less time (than auto mode) at peak power and so gives worse acceleration
though it's fully compatible with cultural preconceptions and those who feed them

[manolis]

Hello Tommy Cookers.

Regarding the efficiency of the V-belt CVT’s,
here is a plot for the Suzuki Burgman650 SECVT:



The plot has been made based on the data / lab tests presented in the:

“Implementation of a Suzuki Electronically controlled Continuously Variable Transmission in a Formula Student racecar,
Eindhoven University of Technology Department Mechanical Engineering Dynamics and Control Technology
Group Eindhoven,
October, 2008)”,

at http://www.pattakon.com/PatBox/SECVT.pdf

At heavier loads / long gear ratios the CVT of the Burgman 650 operates at some 95% efficiency.



Quote from http://www.triketech.com/Drivetrain/n360.html :



The chart gives a pretty accurate snapshot of efficiency levels found in the N360, Internal Gear Hubs (IGH), and a 10 speed derailleur for comparative purposes. No two test dynos ever give the same results on the same parts so take it with a grain of salt.

What's important to consider is how often will you be riding at either end of an extreme of the range? Chances are over 50% of your pedaling is done at a ratio of .08 ~ 1.2. Contrary to what most folks claim, the N360 is pretty efficient in that range; more efficient then the typical IGH.

Derailleurs are the most efficient however as sprockets get smaller and chainline moves off center it causes erratic changes in the efficiencies for each gear.

IGH's suffer from drag as each gear is still meshing, just not engaged to the output shaft. More gears is more meshing losses. Ground gears can slightly reduce friction but comes at considerable cost.

End of Quote



Here is the NuVinci 360:



Here is the Alfine11 IGH (Internal Gear Hubs) of Shimano (not real CVT):



And here is a 10 speed derailleur of Shimano:




According the plot, a 10 speed derailleur has an efficiency that varies with the transmission ratio, with an average of “only” 95.5%, or so, and a maximum below 97%.



Contrary to the NuVinci hub CVT:



and to the IGH Hub transmission:



(both utilizing a primary transmission from the pedals to the rear wheel; the primary transmission comprising a toothed belt or chain and a pair of sprockets, and relative friction loss)

the PatBox:



needs not a “primary transmission” and saves this way some 2% of friction loss.

Thanks
Manolis Pattakos

[Tommy Cookers]

well nothing seems to show a bicycle CVT with the usual belt and the usual non-enthusiast 50 Watt customer would be over 90% efficient

the Burgman example is interesting, a high power expensive device with a fancy metal-link belt and its efficiency is poor at low powers
amusingly it also has a train of 5 gears to get the power from this so-called transmission to the rear wheel
though this waste of power matters little when the powerful engine is running heavily throttled in town traffic

afaik there the world has never seen a full set of hub gear efficiency data in 120 years
very early gears had ball-bearings on the planet spindles, these were soon deleted and so efficiencies then fell
and about 40 years ago oil lubrication was universally replaced by grease, efficiencies have plummeted eg in the 5 and 7 speed versions

the non-enthusiast might be expected to buy this reassuring device ie the expected 3 speed or similar
(as I did, this and NuVinci models were heavily discounted in January - btw your source doesn't show the NuVinci as efficient at low powers)

and/or - why wouldn't the conventional CVT be reassuring to and actually meet the needs of the non-enthusiast bicycle customer ?
(if the ratio didn't 'hunt' eg won't hard pedalling (impulses) lower the ratio in synchrony with impulses and impair user confidence ?)

the enthusiast customer will be hard to satisfy - good luck with that !

[manolis]

Hello Tommy Cookers

You write:
“well nothing seems to show a bicycle CVT with the usual belt and the usual non-enthusiast 50 Watt customer would be over 90% efficient”



In a bicycle with the PatBox CVT like:



the conventional chain is replaced by a V-belt,
the conventional sprockets are replaced by two conical pulleys (each having its own spring),
while a thin plain auxiliary belt and a pair of rollers/tensioners s are added for the control of the transmission ratio.

The thin-plain auxiliary belt and the two rollers/tensioners add an insignificant amount of friction (just think the way they function).

Relative to a typical bicycle with chain and sprockets, the efficiency in a PatBox CVT bicycle drops, and this drop of efficiency is dominated by the friction of the V-belt.


So, the question turns to “what is the efficiency of a V-belt”?


According the official tests of the Carlisle,

(more at http://www.pattakon.com/PatBox/energy_l ... ciency.pdf )

the efficiency of power transmission belts is not as low as most people used to think.














The efficiency of an industrial cogged V-belt at rated power plays around 96% (with the maximum being well above it).


The V-belt for the PatBox CVT has not significant reasons (EDIT: others than the narrower angle of the conical pulley) for being less efficient than the industrial V-belts of the above mentioned tests of the Carlisle.


The simple-minded approach is that using a cogged V-belt having something like 95% efficiency (see the above plots and tables),
and avoiding other friction losses (like the friction loss in the “primary transmission” required in the HUB transmission systems for bicycles (IGH like and Nuvinci CVT like)),
the PatBox CVT besides being a true and easily controllable CVT, besides being a lightweight, simple and cheap to make transmission for bicycles, it seems capable for higher efficiencies, too.


The average bicycler is reluctant to change habits, or to test new solutions.

However, when other bicyclers will overtake him easily and finish tireless the same race while he finishes exhausted, he cannot help wondering what is going on and ask for a test drive.


Thanks
Manolis Pattakos

[Pinger]

(if the ratio didn't 'hunt' eg won't hard pedalling (impulses) lower the ratio in synchrony with impulses and impair user confidence ?)

For the above reason, CVT that has torque sensitivity is a non starter on a bicycle. Maybe on a tandem with pedal/cranks phased 90 degrees apart and both riders pedalling, but for solo rider with only two legs, nah.

[Pinger]

''@ Pinger
so can't we have a CVT type bicycle transmission with 'master' V pulley controlled ie rigid in each of only eg 5 rider-selectable widths ?
the other pulley would still be sprung of course
(or rigid with a seperate jockey-type belt tensioner)
(or rigid and behaviour-matched with the master pulley)

we could call those rider-selectable positions gears''

That would knock the Continuously Variable out of CVT....
The pinch forces to be manipulated are considerable. Achievable via a thumb operated lever or are we requiring both hands on a three foot lever to effect a ratio change?
I'd surmise derailleur is here for a while longer.

(What needs to be remembered is that ratio change on CVT is effected by closing a pulley and forcing the other open. Merely opening a pulley will cause belt slippage).

[manolis]

Hello Tommy Cookers.

You write:
"so can't we have a CVT type bicycle transmission with 'master' V pulley controlled ie rigid in each of only eg 5 rider-selectable widths ?"


I think you mean something like this:





The lever of the PatBox can lock at a number of positions, with each position corresponding to a specific transmission ratio (like having a manual gearbox with, say, six ratios).


The same can be done by having some slots (or stops) wherein the control lever (beige) of the PatBox bicycle:



to lock, providing a number of discrete transmission ratios:


The question is: when you can have an infinity of transmission ratios, why to limit yourself in just 6, or 10, or 20?

Thanks
Manolis Pattakos

[Tommy Cookers]

@ Manolis
haven't you highlighted the efficiencies of various belts that are in different ways unuseable in CVT ie notably the cog types ?
the CVT-suitable wrap type belts seem to be 90-91% efficient

I apologise for deleting a post when others were in the process of responding

[manolis]

Hello Tommy Cookers.

You write:
"haven't you ringed in red efficiencies of types of belt and modes of use ?? ('cog') that can't be used in CVT ?"


I think there is a missunderstanding here: the cogged V-belts are different than the toothed belts. In a cogged belt the "cuts" are not teeth; they are threre to make the belt more flexible.


Cogged V-belts are typical in conventional scooter (and sled) CVT's.

Here is a photo of the cogged V-belt used in the CVT of the Suzuki Burgman 400 (it is a normal CVT controlled by a conventional "centrifugal" variator / governot; it is different than the electronically controlled SECVTof the Suzuki Burgman 650).



It can be regarded as a “double cogged” belt (the cuts are at both sides (internal and external) of the belt).


Quote from the pdf of Carlisle :

"Bending hysteresis is the principal factor determining power loss comparisons between cross sections.
Consequently, due to increased flexibility over plain base belts industrial V-belt cogged constructions require the least energy and run at lower temperatures under no load.
Reduced cogged hysteresis is reflected by the lower temperature, although enhanced heat transfer from tooth
turbulence and greater convective area is an additional factor."



By the way, here it is shown schematically the complete transmission of the Burgman 650 SECVT



Before the SECVT unit, there is the primary transmission (reduction 1.5).

After the CVT there is the final transmission (another reduction of 1.58) comprising 7 (seven) passages of the power from gearwheel to gearwheel .



Supposing a 2% friction loss (98% efficiency) at every one of the eight steps the power passes (excluding the energy loss inside the SECVT), the total loss of energy is 15% plus the enrgy loss inside the CVT!

Compare to the "direct transmission" in the PatBox bicycle.

Thanks
Manolis Pattakos

[Tommy Cookers]

[quote=manolis]
Tommy Cookers.You write:"haven't you ringed in red efficiencies of types of belt and modes of use ?? ('cog') that can't be used in CVT ?"

I think there is a missunderstanding here: the cogged V-belts are different than the toothed belts. In a cogged belt the "cuts" are not teeth; they are threre to make the belt more flexible
Cogged V-belts are typical in conventional scooter (and sled) CVT's.

Here is a photo of the cogged V-belt used in the CVT of the Suzuki Burgman 400 (it is a normal CVT controlled by a conventional "centrifugal" variator / governot; it is different than the electronically controlled SECVTof the Suzuki Burgman 650).
[url]http://burgmanusa.com/gallery/d/40204-2/IMG_8242.JPG[/url]
It can be regarded as a “double cogged” belt (the cuts are at both sides (internal and external) of the belt).
Quote from the pdf of Carlisle :
"Bending hysteresis is the principal factor determining power loss comparisons between cross sections.
Consequently, due to increased flexibility over plain base belts industrial V-belt cogged constructions require the least energy and run at lower temperatures under no load.
Reduced cogged hysteresis is reflected by the lower temperature, although enhanced heat transfer from tooth
turbulence and greater convective area is an additional factor.[/quote]

this point had crossed my mind
though partly because other sources use the approved term 'notched' for what is here being called 'cogged'
and your catalogue source showing cogged belts as having reinforcing cords in the teeth ie as a belt for synchronous use
and some of what you ringed seems to be rather wide ie 'joined' belts that cannot be used here
my interpretation went the way it did

EDIT ok notched/cogged is typically claimed to be worth about 2% efficiency over wrapped
but your high efficiencies shown come from having large pulley size, high speed, and high power (ie using the belt at full capacity)
so this application won't yield these high efficiencies to the non-ehthusiast customer (the same reason I didn't buy a NuVinci)
with chains and sprockets the low-load/no-load losses are insignificant, not so with a belt and pulleys
(this difference was clear in pushing around my 'highly-efficient' (synchronous) belt drive Kawasaki 250)

[manolis]

Hello Tommy Cookers.

You write:
“but your high efficiencies shown come from having large pulley size, high speed, and high power (ie using the belt at full capacity)
so this application won't yield these high efficiencies to the non-ehthusiast customer (the same reason I didn't buy a NuVinci)
with chains and sprockets the low-load/no-load losses are insignificant, not so with a belt and pulleys”


Besides the “transmission efficiency” there other things that matter:


The interruption of the power flow during a gear shift is a significant limitation of the chain / sprockets (derailleur) design, especially for new bicyclers (or aged bicyclers having some issues with their bodies).

Isn’t the interruption of power flow a performance limitation, too?


The easiness of the gear-shifts; the majority of the bicyclers have ten or more gear-ratios, however they use one or two; a wrong gear shift in an uphill is anything but pleasant for the average bicycler.


The automatic control over a CVT can keep the pedalling at the ideal rhythm, and the bicycler tireless (the average bicycler has more interesting /pleasant things to do than dealing with a “strange”, at least, gear shifting mechanism.


The overall efficiency.
As in a scooter, similarly in a bicycle the gear train should focus on the optimization of the efficiency of the power source (the bicycler), because there is the most gains.
If the bicycler has to align with the limitations and tricky handling of the bicycle’s transmission, the overall efficiency drops

Simpleminded question: why only 21 gear ratios and not 121 gear ratios?

With 121 gear ratios one can choose a ratio even closer to the ideal one.
However, the side effects are more than the advantages.

With an easy to handle CVT, one can have not 121 but 1021 ratios easily selectable (with the automatic control being the next reasonable step.


So, even with a, say, 5% lower average transmission efficiency than a derailleur system (with the average efficiency of a good derailleur being around 95.5%, the above 5% appears too much), a PatBox CVT can still provide a better “overall efficiency” to the average bicycler.



Another significant issue is the cost.

Here is the Alfine IGH of Shimano (which is not a CVT) :





Its complication (just imagine the work the shaft in the lower photo needs) justifies the 400, or so, Euros of its price (no matter how strange it appears a bicycle having a HUB gearbox costing 400 Euros!)

And what it offers? Just 11 gear ratios, and an average efficiency of some 92%.


In comparison,

I don’t see a reason for a PatBox CVT bicycle (having an infinity of easily selectable gear ratios) to be (in mass production) more expensive than a similar conventional bicycle.


Thanks
Manolis Pattakos

[Tommy Cookers]

[quote Manolis] .......The automatic control over a CVT can keep the pedalling at the ideal rhythm ......[/quote]

what automatic control would that be ?

[manolis]

.......The automatic control over a CVT can keep the pedalling at the ideal rhythm ......

what automatic control would that be ?

Hello Tommy Cookers.

The PatBox is a quite simple system.
The displacement (angular or linear) of a roller/tensioner defines the transmission ratio.

There are many ways to automatically control the PatBox CVT in a bicycle.

For instance, by an electronic controller and a servo motor, as in the NuVinci Harmony:

Quote from http://www.bikeradar.com/us/commuting/g ... -13-47420/

NuVinci Harmony shifting system review
Changes gear automatically



. . .
They have allied this novel hub gear with a small servo motor that replaces the cable on the manual version, and gone one step further by adding a microchip that enables you to set your preferred cadence and changes gear automatically to keep it constant.
. . .

Switching to fully automatic
. . .
In practice, we tended to use a higher cadence for steeper hill climbing and a lower one for flatter riding; this meant a fairly infrequent twist of the grip shifters.

The overall effect is miraculous, leaving you free to forget about gears and concentrate on enjoying the ride.

End of Quote


In a few words, what is necessary fot the automatic control of the PatCVT in a bycycle is a servomotor (not necessarily electric) and a controller / selector (to put the desired cadence). The servomotor, under the control of the controller, varies the transmission ratio so that the cadence to return to the selected one.


PS.

In the previous posts they were mentioned the NuVinci Hub CVT and the Shimano Alfine 11 IGH.

Here is the 14 gear ratios Rohloff SpeedHub:



which is regarded as the king of the IGH's.

Its price?
Above 1000Euros in the eBay.

What if offers?
14 discrete gear ratios (actually 2*7 gear ratios) in a wide range.

And they claim a high efficiency:



Quote from https://www.rohloff.de/en/experience/te ... fficiency/:

"The overall average frictional loss within the Rohloff SPEEDHUB 500/14 ranges between 1.5 - 5%."

End of Quote.

As the other IGH's, the Rohloff SpeeHub also needs a primary transmission (which adds some 2% friction loss).

Thanks
Manolis Pattakos

[uniflow]

The Traia engine suffered drive failure in early testing due to the ratio being 2 to 1 exactly. If you look at one you will see the drive has a very large and carefully designed rubber damper to try and control the torsional problems. Unhappy torsional resonance were set up in the drive system at certain RPM outputs. A bad idea to run exact divisible reduction numbers
In the gear reductions (PSRU) I build for experimental aircraft use (Subaru, Honda, Suzuki, Chevrolet) I use a hunting tooth setup where the the tooth numbers are not divisible exactly so the two gears are constantly changing their contacts with each other.
The reason aviation engines need to have a cylinder removed from time to time is their very design. Air cooled engines are just too hard to control without thermal issues, hence cracking and burnt valves. Look at any glider tow plane, always cracking cylinders. Extreme heat load up to release height then quench cooling on the way down. The water cooled engines we use have a certain amount of insulation via the thermostat. Subaru engines use has become wide spread in the autogyro, they are a reliable / durable power plant. Air cooled aviation engine just can not compete.

Sorry seems I've come in to this thread several pages back. I'll shut up.

[manolis]

Hello Uniflow.

A way to reduce the torsional problems, even with 1:2 primary reduction, is to reduce or eliminate the inertia part of the torsional loads the gearwheels transmit.

Here is an unconventional Variable Ratio gearwheel:



that reduces, or eliminates, the inertia torque loading the transmission.

Instead of pitch circles, the cooperating gearwheels have properly calculated pitch curves (like the B0 and B1 in the drawing).

The gearwheel with the uneven teeth is secured on the crankshaft.

With such a primary transmission, the conventional even-firing straight-four 4-stroke engines (those with the flat crankshaft) release the gearbox (and the rear tire) from the inertia torsional loads and improve the drivability (to do the same, Yamaha changed from flat to cross-plane crankshaft and made motorcycles like the famous M1 and R1, wherein for the sake of the lower inertia loads it was sacrificed the even firing).

The system works for any number of cylinders (for a single cylinder, the shape of the crankshaft gearwheel is slightly different than in the drawing above).

Thanks
Manolis Pattakos