Drivetrain Power Loss

HomeBuiltAirplanes.com

Help Support HomeBuiltAirplanes.com:

wsimpso1

Super Moderator
Staff member
Log Member
Joined
Oct 18, 2003
Messages
7,836
Location
Saline Michigan
Bike shaft drives and extensions have a lot of moving parts between the crankshaft and the end wheel. A lot of bearings, gears, differentials, clutches, and all these add up. Look at the linked web page, it is a very realistic approximation and allows school students to calculate the losses and efficiency. There are even specs for every single mechanical piece how much efficiency is lost on each depending on the type.
Most of the parts between engine and rear wheel are there in both systems. Clutch, primary drive, countershafts, gears, dog clutches, most of the bearings, and so on. They have to be there anyway. There are no differentials in a motorcycle. The chain final drive has two pair of bearings while the shaft drive needs four pair. The shaft and CV joint add some pounds, and that is probably most significant on a racer. We are not talking race bikes, we are talking airplanes that kill us a significant fraction of the time when power is lost. So I vote for reliability on maintaining power...

Every loss will also depend on speed, torque, temperature and a lot of other parameters. Short story: look at the calculating model. When you'll get a chance to disassemble your bike shaft drive for service or oil change, note all the parts inside. Write them down. Look for their specs and efficiency loss. Add it up and you got the difference between crankshaft power and rear wheel power. Then seek an XS1100 chain drive.

Check out our tough Yamaha XS1100 Chain Drive Conversion
Where are the specs and efficiencies listed for the internals? Where is the model? None are on the website you provided.

You see, people raced with your bike more than 40 years ago, because they understood the information I wrote about here and the thought of engineers. Some ditched the shaft drive and switched to chain drives to reduce losses and improve efficiency. Extra weight was reduced, even a lighter swingarm could be swapped.
Or maybe they ditched the shaft drive because it was not up to the increased power of the engine once prepared for racing and the lightness was a bonus. The seller's of the kit do not cite advantages, just that they have it and it won a race 40 years ago.

Billski
 

Vigilant1

Well-Known Member
Lifetime Supporter
Joined
Jan 24, 2011
Messages
6,043
Location
US
Or maybe they ditched the shaft drive because it was not up to the increased power of the engine once prepared for racing and the lightness was a bonus. The seller's of the kit do not cite advantages, just that they have it and it won a race 40 years ago.
On a race bike, decreasing unsprung weight (really, inertia) is key to keeping the tires in contact with the pavement despite any bumps, etc. I can see that the designer of such a bike might choose a lighter chain drive over a shaft regardless of any slight differences in efficiency. And they would be slight.
I owned a car that had the rear disk brakes mounted on the inner end of the half-shafts, right beside the differential. This did a good job of reducing unsprung inertia, but made brake maintenance a real chore. You'd read the micrometer in poor light so you could convince yourself that the rotors were still within limits rather than drop the rear end to change them.
 
Last edited:

Toobuilder

Well-Known Member
HBA Supporter
Log Member
Joined
Jan 19, 2010
Messages
5,120
Location
Mojave, Ca
Yep, much of the asserions concerning drivetrain losses and "work" are laughable. The idea that there is a huge disparity in power output between the GSXR1100 and the GSX 1100G due to the shaft drive is also laughable. I have owned a few 1000 CC "Superbikes" in my day from the major manufacturers including a 88 Yamaha FZR 1000, 90 Honda CBR 1000, 91 Kawasaki ZX11, and yes, even the 92 GSXR1100 mentioned earlier. ALL of these high strung sportbikes had their basic engines used in more pedestrian models. But that is not to say the engines were a "direct" transplant - In every instance Im aware of, the cams, carbs, CR, exhaust, and induction were different. Yes, the core engine architecture may be the same, but the tuning is significantly different. Beware of making an apples to apples comparison just because the engines have the same bore/stroke and look the same.
 

Toobuilder

Well-Known Member
HBA Supporter
Log Member
Joined
Jan 19, 2010
Messages
5,120
Location
Mojave, Ca
On a race bike, decreasing unsprung weight (really, inertia) is key to keeping the tires in contact with the pavement. I can see that the designer of such a bike might choose a lighter chain drive over a shaft regardless of any slight differences in efficiency. And they would be slight.
I owned a car that had the rear disk brakes mounted on the inner end of the half-shafts, right beside the differential. This did a good job of reducing unsprung inertia, but made brake maintenance a real chore.
Unsprung weight is a factor, but so is the torque reaction of the shaft as it tries to climb the ring gear on accel and decel. The "pogo" effect on the shaft drive bikes is quite unsettling. Unless you use a fairly complex multi link and driveshaft like BMW Paralever system, that is. And there again, lots of weight to solve a problem that does not exist with a chain and simplex swingarm.
 

DangerZone

Well-Known Member
HBA Supporter
Joined
Sep 5, 2011
Messages
2,183
Location
Zagreb HR
I just schemed up a bevel gear drive that might stand a 7500 rpm 100 hp engine for a long safe life and much higher loads for shorter times, sort like a motorcycle operation. 1-3/4” pitch diameter, ¾” shaft diameter, bearings 1” and 4” from the pitch line on the input shaft. Put that horsepower at that speed on the gears and you get 960 pounds at the pitch line, and lose 4.8 pounds to drag (assuming the high end of the rolling drag). That also puts 320 pounds on one bearing and 1280 on the other, giving a total on 1600 pounds load and 8 pounds drag:

(4.8 lb * 0.875 in + 8 lb * 0.375 in) * (1 ft/12 in) = 0.60 ft-lb of drag torque or 0.857 hp at 7500 rpm. Looks like that worst case friction gear set is below 1% friction losses when running 100 hp to me. Done correctly, it can be under 0.4%, and is regularly done in automotive powertrain hardware.

We get similar results in CV joints, parallel shaft straight cut gears, etc.

Billski
Ok, is that the method you calculate a drive efficiency? Am I understanding correctly, that you evaluated this 100HP engine with a shaft drive having less than 1% efficiency loss from crankshaft to rear wheel, meaning having 99HP at the rear wheel?

I'll just repost the link to show what we were taught how to calculate efficiency by calculating first each component, and then calculating the overall efficiency.


This is a huge difference, because this calculation would show only around 84HP at the rear wheel. After calculating, we would check the original values 100HP at the crankshaft and 84HP at the rear wheel with a dyno. The measurement usually confirms the calculus up to 1% deviation. How do you verify your results of crankshaft (input) and rear wheel (output) difference?

Then there is the GSXR1100, a faired sport bike requiring racing posture, and not really intended for more than one lucky individual at a time. Engine makes 156 hp at 10,000 rpm. To do that with the same basic engine means cams, induction, and exhaust tuning all aimed toward maximizing specific output, and usually drives torque way up the rpm band. Yeah, a peaky powerplant. Is this intended for easy riding? Nope. All out performance? Yep. Then the rest of the bike... much more compact set up, clip on bars, racing style fairings, yeah, a place is provided where a second person could ride along, but that is not what this bike is for. Looks more like a road racer than anything else. And you know what, the same spec sheet that says 156 hp engine also says 138 hp at the rear wheel. Somehow it is supposed to lose 18 hp or about 12% getting the power to the road.

So let's do what in the engineering world we used to call the laugh check.

Billski
Ok, I never heard of this laugh test but we usually do a reality check. After calculating every single component they are all added up - and then verified and measured by an accurate instrument. The engine power is rated at the crankshaft, it needs certification that it has the power declared by the manufacturer.n Otherwise it does not pass the tests and wouldn't be allowed to ride or drive in the streets. The crankshaft power can also be measured by the same dyno and other instruments (they are massive). Interestingly, the results always add up to be exact.

This is quite common that a motorcycle of that era (1990s) had losses of 18HP or 12% getting the power to the road or rear tire. I am trying to understand what you are doing differently that you get a loss of only 1%, which is so far away from reality comfirmed by both motorcycle manufacturers and independent measurement. Could you please elaborate how do you verify that your calculation is confirmed by actual measurement of power at the crankshaft and afterwards at the rear wheel?
 

Vigilant1

Well-Known Member
Lifetime Supporter
Joined
Jan 24, 2011
Messages
6,043
Location
US
This is quite common that a motorcycle of that era (1990s) had losses of 18HP or 12% getting the power to the road or rear tire. I am trying to understand what you are doing differently that you get a loss of only 1%, which is so far away from reality comfirmed by both motorcycle manufacturers and independent measurement.
Here's your claim:
Bear in mind also that chains&sprockets lose only 2 to 3% efficiency, while belts 10-20% and shafts 20-35% efficiency. Hence if you prefer some KISS solution, it might be wise to see if you can use some simple and proven concepts.
You say chains are 97 to 98% efficient, shafts are 65% to 80% efficient.
Are you talking about the entire drive line and tire efficiency in the case of shafts, but just the chain and sprockets in the case of chains?
In any case, there's no way that any motorcycle has an engine-to-road efficiency as low as 65% unless the gearbox and tires are full of oatmeal. And the road has 3inches of sand on top.
 

DangerZone

Well-Known Member
HBA Supporter
Joined
Sep 5, 2011
Messages
2,183
Location
Zagreb HR
Most of the parts between engine and rear wheel are there in both systems. Clutch, primary drive, countershafts, gears, dog clutches, most of the bearings, and so on. They have to be there anyway. There are no differentials in a motorcycle. The chain final drive has two pair of bearings while the shaft drive needs four pair. The shaft and CV joint add some pounds, and that is probably most significant on a racer. We are not talking race bikes, we are talking airplanes that kill us a significant fraction of the time when power is lost. So I vote for reliability on maintaining power...



Where are the specs and efficiencies listed for the internals? Where is the model? None are on the website you provided.

...

Billski
No differentials? Ok, so what would you call the rear shaft drive differentials that are sold on ebay when a shaft drive differential breaks due to hard abuse? It looks like this:

1997-2001 BMW K1200LT K1200 NON-ABS Rear Wheel Drive Axle Differential Gear Case | eBay

1975 MOTO GUZZI 850 T 850T REAR DIFFERENTIAL | eBay

1991 SUZUKI KATANA GSX1100 G GSX 1100g FINAL DRIVE REAR DIFFERENTIAL 89 90 91 | eBay

The specs and efficiencies of every bearing or joint is provided by the manufacturer. Most list them in catalogs, good quality parts have them no matter if a bearing is SKF or NSK. Poor quality bearings usually do not have that, some Chinese bearings have nothing.

I'm really trying to understand how come there is so much difference in approach...
 
Last edited:

DangerZone

Well-Known Member
HBA Supporter
Joined
Sep 5, 2011
Messages
2,183
Location
Zagreb HR
Here's your claim:

You say chains are 97 to 98% efficient, shafts are 65% to 80% efficient.
Are you talking about the entire drive line and tire efficiency in the case of shafts, but just the chain and sprockets in the case of chains?
In any case, there's no way that any motorcycle has an engine-to-road efficiency as low as 65% unless the gearbox and tires are full of oatmeal. And the road has 3inches of sand on top.
Please read the link to how efficiency is calculated. Every component is calculated separately, and then the whole system is calculated by adding all those losses together. At least that's the way it has been done in EUrope for the last 40 or so years. I am now curious about how it is done in the USA and hopefully Billski will provde some answers beacsue I believe the guy knows stuff.

Drivetrain losses (efficiency) – x-engineer.org

If a chain driven GSXR1100 loses power from 156HP at the crankshaft (before the gearbox) and provides 138HP at the rear wheel, then the chain is only a small part of the overall 12% losses. Between the crankshaft and rear wheel are the gearbox, the clutch, and a few other things which reduce this efficiency. The same engine with a shaft drive will not have only 12% losses, it will have much more.
 
Last edited:

Vigilant1

Well-Known Member
Lifetime Supporter
Joined
Jan 24, 2011
Messages
6,043
Location
US
Please read the link to how efficiency is calculated. Every component is calculated separately...
So, again, in the earlier comparison (max 99% efficiency for chains, max 80% efficiency for shafts) are the asserted efficiencies for the whole driveline including the tire? Just a few components? It should be consistent to be useful.
 

TiPi

Well-Known Member
Log Member
Joined
Aug 25, 2019
Messages
408
Location
Peeramon (AUS)
There is no point in trying to educate DZ, he can’t see that these “much higher losses in a drive shaft setup” would generate heat and we all know that this excessive heat is not there.
He is also relying on ebay for the definition of a differential!!!
 

Dan Thomas

Well-Known Member
Joined
Sep 17, 2008
Messages
5,819
Nonsense. I specifically wrote in previous posts that all losses add up in the end, including heat losses and friction losses, among other losses.
"Friction" losses are just another way of generating heat. They're not a loss category of their own. All of the heat generated in a drivetrain is due to friction. Even the noise generated creates heating of the atmosphere. Stirring oil heats it. Bending metal heats it. Friction.

An argument like that is like the old "torque is most important" when discussing power. But torque is useless without RPM when discussing power. Torque times no RPM is no power.
 
  • Like
Reactions: BJC

DangerZone

Well-Known Member
HBA Supporter
Joined
Sep 5, 2011
Messages
2,183
Location
Zagreb HR
So, again, in the earlier comparison (max 99% efficiency for chains, max 80% efficiency for shafts) are the asserted efficiencies for the whole driveline including the tire? Just a few components? It should be consistent to be useful.
Wrong. I wrote:
Bear in mind also that chains&sprockets lose only 2 to 3% efficiency, while belts 10-20% and shafts 20-35% efficiency. Hence if you prefer some KISS solution, it might be wise to see if you can use some simple and proven concepts.
This means a motorcycle engine with a final chain drive could have 12% efficiency loss (as an example 10% internals and 2% chain) while this same bike with a shaft drive could have 30% efficiency losses (as in 10% internally and 20% in the shaft drive). One influences the other, too.
 

DangerZone

Well-Known Member
HBA Supporter
Joined
Sep 5, 2011
Messages
2,183
Location
Zagreb HR
There is no point in trying to educate DZ, he can’t see that these “much higher losses in a drive shaft setup” would generate heat and we all know that this excessive heat is not there.
He is also relying on ebay for the definition of a differential!!!
Excellent, feel free to educate me, how do you verify your power loss or power gain measurements? How do you call a motorcycle shaft drive differential? Or you don't call it anything because you know nothing about it?
 

TiPi

Well-Known Member
Log Member
Joined
Aug 25, 2019
Messages
408
Location
Peeramon (AUS)
Excellent, feel free to educate me, how do you verify your power loss or power gain measurements? How do you call a motorcycle shaft drive differential? Or you don't call it anything because you know nothing about it?
Start here: Differential (mechanical device) - Wikipedia
the motor bike is lacking a key ingredient to qualify as a differential (only 1 output). They are just 90deg angle drives.
 

Toobuilder

Well-Known Member
HBA Supporter
Log Member
Joined
Jan 19, 2010
Messages
5,120
Location
Mojave, Ca
Perhaps there is a bit of a language barrier here, but a "differential" is a means to split power equally between at least two driving wheels. If there is but one wheel, there is no differential.

Also, the Superbike era of the late 90's was NOTORIOUS for elevated claims of "crankshaft" power - just like the American Musclecar HP wars of the late 1960's. This is a lot more marketing hype than engineering units.
 
Last edited:

eether54

Member
Joined
Aug 11, 2019
Messages
5
Excellent, feel free to educate me, how do you verify your power loss or power gain measurements? How do you call a motorcycle shaft drive differential? Or you don't call it anything because you know nothing about it?
I'm not much for posting, I mainly just read everyday and this thread has proven to be an interesting one.

I don't have anything scientific to add, just lots of first hand experience with all three drive systems on motorcycles.
I can tell you that on a 11000 mile trip the modern fuel injected 955cc chain drive motorcycle got 9-10% worse fuel mileage than the 30 year old 850cc pushrod air-cooled v twin with a shaft drive. Oh, and the shaft drive bike was had more drag and a heavier load.
All things things being equal, under consistent load (ie running 75 mph down the highway) the differences between the three drive systems are negligible.

The idea that shaft drive is as inefficient as claimed is just not correct.
Check out what Dr John did with moto guzzis vs the big Japanese manufacturers. It's interesting reading if nothing else.
 
Last edited:

Dan Thomas

Well-Known Member
Joined
Sep 17, 2008
Messages
5,819
Also, the Superbike era of the late 90's was NOTORIOUS for elevated claims of "crankshaft" power - just like the American Musclecar HP wars of the late 1960's. This is a lot more marketing hype than engineering units.
I remember that, and the feds, in about 1973, had to mandate marketing HP numbers to be determined by taking the HP at the rear of the transmission, with all engine accessories installed and operating. Water pump, alternator, emissions stuff--all had to be there and working for the HP measurements. HP numbers dropped by about a third.

And that wasn't all friction and accessory load losses. Not nearly. I'm pretty sure those automakers were also fudging their numbers not only by measuring flywheel HP but by multiplying redline RPM by peak torque, which occurs well below redline. That will give an impressive but completely false number. The US outlawed that, but did the Japanese?

It's a bit like the famous Craftsman HP ratings on their compressors and vacuum cleaners and so on; they were claiming stuff like a 3.5 HP motor on their vacuum cleaners, which a 110VAC 15-amp circuit could never supply. 3 HP is 2238 watts, which is over 20 amps at 110V. They were cheating by measuring the starting inrush current, which lasts about half a second, and the thermal breaker will tolerate that. At normal running load the HP would be maybe 1.5 HP at most. The inductive resistance of a spinning armature will limit current flow once speed has been achieved.
 

Protech Racing

Well-Known Member
Joined
Jul 10, 2020
Messages
417
I raced 100cc bikes a lot as a kid. The larger the difference in gear sizes made a large impact on drive line loss. Just like they tellyou in Mech engineering,
For example a 100cc Hodaka in 4 th gear ,driving a certain , rear gear sprocket, made 12hp. at the wheel .
In 5th gear, driving an larger sprocket for the same IPR( inches per revolution) the bike put down 11HP at the same wheel and speed.
Keeping the gears as close to the same size as possible will produce the most power .
Keeping the gears straight cut will also help out put by reducing side loads and slipping loads. Less heat = more power out.
FWIW, my chassis dyno allows me to measure coast down drag. Some gear sets are 22 HP on the FWD VW, that I race and tune. That is my upper limit. I rebuild them with loose fit bearings etc at 22HP.
On the same note, i found that running in 3 rd gear gives more power then 5 th gear . similar to the Hodaka 30 yrs agao. I set up the best cars to spend as much time in 3 rd gear as possible now. The 3rd gear sizes are large and closest to the same size, as is the final drive gear set.
 
Last edited:

wsimpso1

Super Moderator
Staff member
Log Member
Joined
Oct 18, 2003
Messages
7,836
Location
Saline Michigan
Ok, is that the method you calculate a drive efficiency? Am I understanding correctly, that you evaluated this 100HP engine with a shaft drive having less than 1% efficiency loss from crankshaft to rear wheel, meaning having 99HP at the rear wheel?
No, I postulated a robust 90 degree bevel gear set that would likely be durable, calculated the pitch line forces for 100 hp and 7500 rpm, then calculated the likely losses by the gear mesh and bearings on the shafts. I did the calcs using a rather conservative 0.005 coefficient of drag for steel rolling along on steel. That is how gears and roller bearings are modeled by folks who design these systems, and usually the folks in automatic transmissions assume even smaller drag, because, quite frankly, automotive transmission gears and bearings do better than this. The method can be expanded to cover real gear sets with known dimensions, but it is common for these gear sets to do modest fractions of 1% loss. Oh, and this is way more info on modeling than that the pack of... info... you have been supplying.

I'll just repost the link to show what we were taught how to calculate efficiency by calculating first each component, and then calculating the overall efficiency.
I feel it is necessary to point out that the example run through in the article is illustrative of the method, but is also quite conservative and does not make any attempts to calculate the friction, then torque lost, then energy losses of real systems. As I mentioned in an earlier post, in the automatic tranny world, we are accustomed to whole gearbox efficiencies around 97%, and that is with around half of the clutches open and dragging as well as spinning quad compound planetary, transfer, and final drive gears. The example used provided no basis shown for each of the numbers. The drag in individual gearsets was quite high compared to anything used in modern light truck or car powertrains since at least 1990 MY, where my professional involvement in these machines stretched started.

Ok, I never heard of this laugh test but we usually do a reality check.
Same thing...

After calculating every single component they are all added up - and then verified and measured by an accurate instrument. The engine power is rated at the crankshaft, it needs certification that it has the power declared by the manufacturer.
Says who? Certification in the US specifies the start conditions, dwell before driving, then the speed vs time trace for city and highway cycles. On the test, measured is actual speed vs time over the cycle, fuel used and emissions produced. The rest of the cert is that the evaporative emissions control system works, and the whole thing has to be demonstrated that it can be expected to go 100,000 miles without going smoggy. Once the vehicles are being sold, the fleet is allowed a very small number of emissions related failures before a target mileage number or a recall to fix the systems is triggered. The Euro and Asian cert process were modeled on the EPA processes, but with a little more rigor in finding a test cycle that reflects actual usage a little more than the EPA cycle did. In 1970, the data collection and sorting was less sophisticated than it is now. Story is the EPA lab instrumented a car, had ten employees drive them home and back, and picked the secretary's trace as looking closest to the middle of their range. That cycle is still used 50 years later.

Internally, both of the automakers I was technical leadership within measured whole transmission efficiency aas part of proveout to internal customers that the tranny was a good thing to have in the product. Calibrated dynos put energy in at the flexplate and take it out at the prop shaft or half-shafts. We did this both with running transmissions (pumps, controls, etc all running) and with pinned clutches to show that the clutches were holding. But that was for internal use and as part of data sharing agreements with the EPA and other regulatory outfits for policy analysis.

Otherwise it does not pass the tests and wouldn't be allowed to ride or drive in the streets. The crankshaft power can also be measured by the same dyno and other instruments (they are massive). Interestingly, the results always add up to be exact.
Again, says who? You are assuming way more government oversight than is actual. The emissions and fuel use are what are tracked. If the car is a dog, they do not care. Oh, there are lots of safety checks in FMVSS, but measured power? Nope. And most of it is self certified too.

This is quite common that a motorcycle of that era (1990s) had losses of 18HP or 12% getting the power to the road or rear tire. I am trying to understand what you are doing differently that you get a loss of only 1%, which is so far away from reality comfirmed by both motorcycle manufacturers and independent measurement. Could you please elaborate how do you verify that your calculation is confirmed by actual measurement of power at the crankshaft and afterwards at the rear wheel?
Again, I suspect exaggerated power output claims, poor calibration of dynos, and very likely, a fair amount of slip from the drive wheel to the roller - typical automotive tires peak in load carrying capacity at around 7% slip. A 100% powertrain can not do better than the speed ratio between tire and rollers. Chassis dynos are great for durability tests though.

The only components I can begin to feature that have anything like the losses you are citing are torque converters when operating open. At zero speed ratio, they make zero efficiency, efficiencies at high speed and modest power can be around 90%. and that requires lots of oil flow and substantial coolers. That is also why torque converters have had bypass clutches that mechanically couple the engine to the transmission, with efficiencies around 99.5%. Open converters at high power and low speed ratios will overwhelm all traditional cooling systems. On dynos for checking torque converter function at speed ratios of 0.5 and less, cold oil in and big water flow directly over the converters is used to prevent temps that would otherwise fry the clutch linings, seals, soften the springs and other heat treated components, and generally cook the device and the oil flowing through it. Let' just say I know about power loss inside devices. Without torque converters, modern automatics only barely need cooling. Put the gadget in to allow launch, and you need cooling...

Billski
 
Top