Computing hp loss on a chassis dyno
Race Director
Joined: Apr 2007
Posts: 11,152
Likes: 890
From: South Western Ontario
The simple fact is that there are 3 types of losses that consume engine HP.
#1 - rpm based losses. These vary with rpm but are the same for any rpm no matter what power level is being transferred. You have windage losses which are simply caused by the parts spinning in the grease, oil or air. The pump in an auto falls into this category. At 5000rpm it's producing say 200psi and requires the same power to drive it no matter how much HP is passing through the transmission.
#2 - Variable losses. These change depending on the power level and rpm. You have the gear losses due to the gears meshing and sliding over each other. Since the transmission is typically in the 1:1 ratio with the input and output shaft locked together these only apply to rear-ends on the dyno. You can have the torque converter losses, which vary depending on how much it's "slipping".
#3 - inertia losses. The faster you accelerate the drive train parts, the more power goes into accelerating any component that is rotating. These aren't a loss in the conventional sense. You're really just storing energy. The power used to accelerate the parts doesn't produce heat. However, it gets lost as heat the moment you apply the brakes. If you coast, the stored energy will allow you to coast further as it gets absorbed by the windage losses and the vehicle drag losses. If you don't clutch in or neutral shift, some also gets absorbed by the engine braking effect.
Overall, saying it's fixed or a percentage are both wrong. Since #2 and #3 both increase as the HP increase, the percentage is likely the closer approximation in most cases. So, it's an OK approximation to use. That is all. I have seen posts with 2 different engine combo numbers and the data supported the fact that the losses aren't a fixed or percent value. The loss in power went up but the loss in percent went down.
I won't even get into how accurate each dyno might be or why. Add errors there and the numbers can be skewed either way. An inertia dyno DOES NOT measure the HP or torque. It measures the acceleration rate of a drum and calculates everything else.
The rambings about how the gear losses in the transmission change the power it consumes sounds rather odd to me since a dyno run typically uses the 1:1 gear ratio where the input and output shafts are simply coupled together and no gears are being used. How could any gears cause an extra loss when no power is being transmitted through the gears?
If you want this to be any help then pull all 4 numbers off the curves at the same rpm. This type of data isn't useful to this discussion since each number is at a different rpm.
#1 - rpm based losses. These vary with rpm but are the same for any rpm no matter what power level is being transferred. You have windage losses which are simply caused by the parts spinning in the grease, oil or air. The pump in an auto falls into this category. At 5000rpm it's producing say 200psi and requires the same power to drive it no matter how much HP is passing through the transmission.
#2 - Variable losses. These change depending on the power level and rpm. You have the gear losses due to the gears meshing and sliding over each other. Since the transmission is typically in the 1:1 ratio with the input and output shaft locked together these only apply to rear-ends on the dyno. You can have the torque converter losses, which vary depending on how much it's "slipping".
#3 - inertia losses. The faster you accelerate the drive train parts, the more power goes into accelerating any component that is rotating. These aren't a loss in the conventional sense. You're really just storing energy. The power used to accelerate the parts doesn't produce heat. However, it gets lost as heat the moment you apply the brakes. If you coast, the stored energy will allow you to coast further as it gets absorbed by the windage losses and the vehicle drag losses. If you don't clutch in or neutral shift, some also gets absorbed by the engine braking effect.
Overall, saying it's fixed or a percentage are both wrong. Since #2 and #3 both increase as the HP increase, the percentage is likely the closer approximation in most cases. So, it's an OK approximation to use. That is all. I have seen posts with 2 different engine combo numbers and the data supported the fact that the losses aren't a fixed or percent value. The loss in power went up but the loss in percent went down.
I won't even get into how accurate each dyno might be or why. Add errors there and the numbers can be skewed either way. An inertia dyno DOES NOT measure the HP or torque. It measures the acceleration rate of a drum and calculates everything else.
The rambings about how the gear losses in the transmission change the power it consumes sounds rather odd to me since a dyno run typically uses the 1:1 gear ratio where the input and output shafts are simply coupled together and no gears are being used. How could any gears cause an extra loss when no power is being transmitted through the gears?
If you want this to be any help then pull all 4 numbers off the curves at the same rpm. This type of data isn't useful to this discussion since each number is at a different rpm.
Last edited by lionelhutz; Oct 24, 2013 at 09:14 PM.
Safety Car
Joined: Oct 2004
Posts: 3,926
Likes: 739
From: Youngstown Ohio
Ok I get what you're saying. The HP loss is greater but the % loss remains constant. Sure I agree that is true mathematically speaking.
But mike said;
"Another key is understanding that losses are neither a fixed amount nor fixed percentage- they rise exponentially (non-linear) with RPM and torque."
So according to that statement power loss would rise % wise as well.
Your percentage of loss was less at a higher HP level than Revi's at a lower power level. Agreed?
Also it was stated that a manual transmission looses less power percent than an automatic. Again this did not prove to be true between your loss % wise on a th400 vs revi's on an M20 trans, 22% VS 28%.
As far as multiple dyno sessions is concerned, I'm not referring multiple sessions on one car, I'm referring to multiple dyno examples from different cars in order to get a larger sampling of engine dyno's VS chassis dynos of the same car for various transmission types.
In any case I have found out what I needed to know and I appreciate everyone's input.
Even Mike managed to not insult while teaching
you'll never make je** of the year doing that mike, might wanna rephrase your responses to keep up appearances.
But mike said;
"Another key is understanding that losses are neither a fixed amount nor fixed percentage- they rise exponentially (non-linear) with RPM and torque."
So according to that statement power loss would rise % wise as well.
Your percentage of loss was less at a higher HP level than Revi's at a lower power level. Agreed?
Also it was stated that a manual transmission looses less power percent than an automatic. Again this did not prove to be true between your loss % wise on a th400 vs revi's on an M20 trans, 22% VS 28%.
As far as multiple dyno sessions is concerned, I'm not referring multiple sessions on one car, I'm referring to multiple dyno examples from different cars in order to get a larger sampling of engine dyno's VS chassis dynos of the same car for various transmission types.
In any case I have found out what I needed to know and I appreciate everyone's input.
Even Mike managed to not insult while teaching
you'll never make je** of the year doing that mike, might wanna rephrase your responses to keep up appearances.
Racer
Joined: Apr 2011
Posts: 367
Likes: 7
From: Congupna Victoria
Maybe an alternate look at this might be that with ANY gear there are losses...
Try to remember your High School Maths/Physics of a force acting an angle...that force can be broken up into a force in the X-axis and the Y-axis....
The same happens with gears - either in the transmission and/or in the diff.
The force is NOT transferred at 100% efficiency..in the case of the Crown wheel and pinion in the diff, some of the force is trying to push the gears apart, exerting force on the casing/housing.
This is loss.
The greater the angle of gear cut, the greater the loss (that's why racers like straight cut gears in gearboxes/transmissions) - noisy, but less power loss.
To accurately mathematically model WHAT the actual loss is for an individual combination, would not be hard, but it sure ISN'T a constant and it ISN'T a simple percentage.
HOWEVER, as most people can understand percentages, it's usually simpler to use percentages and in most cases, it's a reasonable approximation.
Imagine a person trying to push a car down the road.
If he pushes directly from behind in the direction he wants to go ALL of his effort (force/power/hp) gets translated into moving that car in that direction (minus frictional losses in tires etc)
BUT if he pushes at a 45 degree angle to the direction he wants to go only 0.7071 (1 divided by the square root of 2) of his effort gets translated into a force pushing the car down the road - the rest is resisted by the cars resistance to be pushed at right angles to the direction that the wheels are facing.
NOW if TWO guys push at the same 45 degree angle, they both only translate 0.7071 of each of their efforts, BUT the car accelerates faster.
They both have lost some of their effort - that's why the loss CANNOT be a constant and in this simple example would be a %.
BUT with so many different cars/engine/transmission/differential etc combination - there is no simple EXACT way to predict, just an estimate.
And the simplest Estimate is a % and it is widely accepted that is is between 15-25%.
Try to remember your High School Maths/Physics of a force acting an angle...that force can be broken up into a force in the X-axis and the Y-axis....
The same happens with gears - either in the transmission and/or in the diff.
The force is NOT transferred at 100% efficiency..in the case of the Crown wheel and pinion in the diff, some of the force is trying to push the gears apart, exerting force on the casing/housing.
This is loss.
The greater the angle of gear cut, the greater the loss (that's why racers like straight cut gears in gearboxes/transmissions) - noisy, but less power loss.
To accurately mathematically model WHAT the actual loss is for an individual combination, would not be hard, but it sure ISN'T a constant and it ISN'T a simple percentage.
HOWEVER, as most people can understand percentages, it's usually simpler to use percentages and in most cases, it's a reasonable approximation.
Imagine a person trying to push a car down the road.
If he pushes directly from behind in the direction he wants to go ALL of his effort (force/power/hp) gets translated into moving that car in that direction (minus frictional losses in tires etc)
BUT if he pushes at a 45 degree angle to the direction he wants to go only 0.7071 (1 divided by the square root of 2) of his effort gets translated into a force pushing the car down the road - the rest is resisted by the cars resistance to be pushed at right angles to the direction that the wheels are facing.
NOW if TWO guys push at the same 45 degree angle, they both only translate 0.7071 of each of their efforts, BUT the car accelerates faster.
They both have lost some of their effort - that's why the loss CANNOT be a constant and in this simple example would be a %.
BUT with so many different cars/engine/transmission/differential etc combination - there is no simple EXACT way to predict, just an estimate.
And the simplest Estimate is a % and it is widely accepted that is is between 15-25%.
Drifting
Joined: Oct 2005
Posts: 1,615
Likes: 150
From: Forth Worth TX
Looks like you're trying to compare apples & oranges.
There are three HP ratings, Gross, Net, and RWHP. My data shows Gross and RWHP. The "standard" 15% loss is a ballpark number for when you compare Net HP to RWHP, not Gross to RWHP. My car has 341 gross/300 net(est)/ 244 RWHP. If you use those numbers I would lose 18% from Net to RWHP.
You can't use my data to compare against C3 Stroker's data. His results are with a different dyno and only show the GROSS HP to RWHP loss (22%), which is good, but isn't the complete formula that we are trying to address in this post. For C3 Stroker to complete his formula he would need to build/run a second engine (with more or less HP) on the same engine dyno as the first, then again on the same chassis dyno (same exhaust, same everything) and see if he gets a 22% loss or something different.
The OP wants to know if you can assign a specific HP loss number to a transmission or a percentage. In order to find that answer you need a formula.
You need to run two engines on the same engine dyno to find GROSS HP. (I did that. The results were 285hp and 341hp.)
Next you would need to install each engine in the same car and make chassis dyno runs on the same dyno. (I did that. The 285hp engine lost 23.85% and the 341hp engine lost 28.44%.) Those are the facts, not guesses, conjecture or theory.
With the only difference being the power output of the engine itself, we have to ask where did the additional 5% loss come from? My opinion/belief is that the more power you try to put thru the drivetrain, the more drag/restriction/loss you will get.
The important part of this formula is keeping all variables the same except for engine output.
As an example of skewing percentage numbers;
My chassis dyno runs were made with OEM exhaust manifolds, 2.5" exhaust pipes, and mufflers installed. If I had made the chassis dyno runs with headers installed I would have lost less RWHP when comparing to the original GROSS HP. That inturn would have made the total drivetrain loss percentages look to be less.
(Lets assign 25hp as the increase for the mythical headers and see what happens.)
285hp@4900/336tq@3400 Gross HP on engine dyno.
217hp@5000/282tq@3250 RWHP on chassis dyno. This was a 23.85% loss.
242hp@5000 (with headers) RWHP. This would only be a 15.08% loss. This ends up being better than C3 Strokers results. (I presume C3 Stroker is running headers on his.)
341hp@5600/387tq@3400 Gross HP on engine dyno.
244hp@5000/304tq@3600 RWHP on a chassis dyno. This was a 28.44% loss.
269hp@5000 (with headers) RWHP. This would only be a 21.11% loss.
We can play with numbers and percentages all day long, but the fact remains my car lost an additional 5% (comparitively) by only changing engine output.
There are three HP ratings, Gross, Net, and RWHP. My data shows Gross and RWHP. The "standard" 15% loss is a ballpark number for when you compare Net HP to RWHP, not Gross to RWHP. My car has 341 gross/300 net(est)/ 244 RWHP. If you use those numbers I would lose 18% from Net to RWHP.
You can't use my data to compare against C3 Stroker's data. His results are with a different dyno and only show the GROSS HP to RWHP loss (22%), which is good, but isn't the complete formula that we are trying to address in this post. For C3 Stroker to complete his formula he would need to build/run a second engine (with more or less HP) on the same engine dyno as the first, then again on the same chassis dyno (same exhaust, same everything) and see if he gets a 22% loss or something different.
The OP wants to know if you can assign a specific HP loss number to a transmission or a percentage. In order to find that answer you need a formula.
You need to run two engines on the same engine dyno to find GROSS HP. (I did that. The results were 285hp and 341hp.)
Next you would need to install each engine in the same car and make chassis dyno runs on the same dyno. (I did that. The 285hp engine lost 23.85% and the 341hp engine lost 28.44%.) Those are the facts, not guesses, conjecture or theory.
With the only difference being the power output of the engine itself, we have to ask where did the additional 5% loss come from? My opinion/belief is that the more power you try to put thru the drivetrain, the more drag/restriction/loss you will get.
The important part of this formula is keeping all variables the same except for engine output.
As an example of skewing percentage numbers;
My chassis dyno runs were made with OEM exhaust manifolds, 2.5" exhaust pipes, and mufflers installed. If I had made the chassis dyno runs with headers installed I would have lost less RWHP when comparing to the original GROSS HP. That inturn would have made the total drivetrain loss percentages look to be less.
(Lets assign 25hp as the increase for the mythical headers and see what happens.)
285hp@4900/336tq@3400 Gross HP on engine dyno.
217hp@5000/282tq@3250 RWHP on chassis dyno. This was a 23.85% loss.
242hp@5000 (with headers) RWHP. This would only be a 15.08% loss. This ends up being better than C3 Strokers results. (I presume C3 Stroker is running headers on his.)
341hp@5600/387tq@3400 Gross HP on engine dyno.
244hp@5000/304tq@3600 RWHP on a chassis dyno. This was a 28.44% loss.
269hp@5000 (with headers) RWHP. This would only be a 21.11% loss.
We can play with numbers and percentages all day long, but the fact remains my car lost an additional 5% (comparitively) by only changing engine output.
Last edited by Revi; Oct 25, 2013 at 09:37 AM.
Thread Starter
Le Mans Master
Joined: Apr 2011
Posts: 6,284
Likes: 1,171
From: Hermosa
There are three HP ratings, Gross, Net, and RWHP. My data shows Gross and RWHP. The "standard" 15% loss is a ballpark number for when you compare Net HP to RWHP, not Gross to RWHP. My car has 341 gross/300 net(est)/ 244 RWHP. If you use those numbers I would lose 18% from Net to RWHP.
Are most engine dynos a gross HP value? The ones I see on youtube look to be running no accesories, not even the water pump.
My chassis dyno runs were made with OEM exhaust manifolds, 2.5" exhaust pipes, and mufflers installed.
Melting Slicks
Joined: Mar 2008
Posts: 2,666
Likes: 289
From: Spring Park MN
As I stated earlier, the dynos themselves vary from one to another. I did this testing more than once. Test on one dyno and then go straight to another on the same day with everything else the same and came out with quite different numbers.
I did the same kind of testing with 6 different flow benches. They are worse than dynos. The numbers were all over the place.
Like I said before, they should be used as testing equipment to check your work. Use the same dyno (or flow bench) all the time with the same test setup and procedures. In this way they are a powerful tool.
So if you test an engine on a brake dyno that can vary from other brake dynos and then put the car on an inertia dyno that can vary from other inertia dynos and will show quite a bit of difference from a brake dyno, and what do you have?
If you had both an engine dyno and a chassis dyno, you could establish a trend of power loss in the chassis after doing a lot of engine to chassis dyno tests.. This would be good for your own testing and you could then pick out excess parasitic losses on individual engine/car combinations, but the information would only apply to your dynos!
I did the same kind of testing with 6 different flow benches. They are worse than dynos. The numbers were all over the place.
Like I said before, they should be used as testing equipment to check your work. Use the same dyno (or flow bench) all the time with the same test setup and procedures. In this way they are a powerful tool.
So if you test an engine on a brake dyno that can vary from other brake dynos and then put the car on an inertia dyno that can vary from other inertia dynos and will show quite a bit of difference from a brake dyno, and what do you have?
If you had both an engine dyno and a chassis dyno, you could establish a trend of power loss in the chassis after doing a lot of engine to chassis dyno tests.. This would be good for your own testing and you could then pick out excess parasitic losses on individual engine/car combinations, but the information would only apply to your dynos!
