Drive train loss riddle. (Page 3/4)
Will AUG 31, 10:39 AM

quote
Originally posted by FieroWannaBe:

Provided everything is properly accounted for and maintained by the operator, you can expect a reasonable repeatability run per run if the acceleration rate is controlled.



Well... this is the issue, isn't it? Did you hear about the pre-production C8 Corvette that a magazine took to a dyno shop? They laid down an unreasonably high number and there was a whole lot of "C8 Corvette SERIOUSLY UNDERRATED!!!!!!!one!11" headlines. Turns out the dyno wasn't tuned up right for the Corvette and gave bad data. Of course there wasn't NEARLY as much publicity for the correction as there was for the initial article.

Here's my grenade:

With a known MoI of M, a known test duration of S, test start RPM of RPM1, test finish RPM of RPM2, the average power delivered to the drum during the test time is M*(RPM2^2 - RPM1^2)/S.

Any device that calculates a different number for the same test is WRONG.

[This message has been edited by Will (edited 08-31-2020).]

pmbrunelle AUG 31, 12:15 PM

quote
Originally posted by Will:


Do all hub dynos include DUAL hub units?

A single hub dyno sounds like a recipe for blown up diffs...



There's a hub unit for the left side; it's just hidden behind the car. At 0:22 in the video, as the camera approaches the computer screen, the top of the LH hub unit comes into view.

Plus, I've been in that room a few times; I saw boxes for both sides with my own eyes.
FieroWannaBe AUG 31, 02:56 PM

quote
Originally posted by Will:


Well... this is the issue, isn't it? Did you hear about the pre-production C8 Corvette that a magazine took to a dyno shop? They laid down an unreasonably high number and there was a whole lot of "C8 Corvette SERIOUSLY UNDERRATED!!!!!!!one!11" headlines. Turns out the dyno wasn't tuned up right for the Corvette and gave bad data. Of course there wasn't NEARLY as much publicity for the correction as there was for the initial article.

Here's my grenade:

With a known MoI of M, a known test duration of S, test start RPM of RPM1, test finish RPM of RPM2, the average power delivered to the drum during the test time is M*(RPM2^2 - RPM1^2)/S.

Any device that calculates a different number for the same test is WRONG.




If I remember it had something to do with AWD vs RWD setup, in combination with not using the road load measurement for the car that wasn't ON THE DYNO. A rookie oversight, and hard for me take at face value...


The math isn't proprietary, its standard Newtonian physics, What different systems do to overcome control system variables, numerical analysis, data collection and hardware configuration, and data processing becomes proprietary. Some companies patent their methods.

I'll just leave it as this:
If the math says:
M*(RPM2^2 - RPM1^2)/S=300HP
and the device reports 301.5 HP is it wrong?
Or is it reporting with results within .5% accuracy?

Your statement is leaves out the balance of forces and moments on the shaft. That's where a controlled rate test will differ from a pure inertial dyno test. It measures more that just average power, The dyno actively measures torque on the shaft needed to control the rate of acceleration, in addition the average power produced to achieved the determined rate of acceleration.

At the end of the day, are you there to get a certified accurate weights and measures with number that can be traced back to how it correlates to some gauges in France? Or are you there to see how a change affects the control and power production of your engine? They are tools, and if they provide repeatable results, there is no reason not to utilize it for a before after picture.

[This message has been edited by FieroWannaBe (edited 08-31-2020).]

FieroWannaBe AUG 31, 03:04 PM
See above

[This message has been edited by FieroWannaBe (edited 08-31-2020).]

pmbrunelle AUG 31, 07:53 PM
An instrument is wrong if it reports a value outside of its promised accuracy.

As far as traceability to artifacts in France, that ceased to be true since the 2019 redefinition of SI units. No longer is a documented chain of comparisons to the IPK needed. Now that the SI system units are derived from natural phenomena measurable by all, the system has become "free", no longer chained to a central authority (well, besides the standardization defined by the BIPM).

As for dynos and my car, if it can report in standard units, that's a plus for the end-user. In an ideal world, I don't want to be tied to any particular dyno.

When I use a mic to measure the diameter of a shaft, I can use any brand of mic and obtain a similar result (in terms of the SI metre, which is reproducible anywhere). This interchangeability of measures is part of what allows our civilization to exist today. I understand that the dyno industry has not reached that level of maturity yet, but we should keep our eyes on where we want to head towards.
FieroWannaBe AUG 31, 09:36 PM

quote
Originally posted by pmbrunelle:

An instrument is wrong if it reports a value outside of its promised accuracy.

As far as traceability to artifacts in France, that ceased to be true since the 2019 redefinition of SI units. No longer is a documented chain of comparisons to the IPK needed. Now that the SI system units are derived from natural phenomena measurable by all, the system has become "free", no longer chained to a central authority (well, besides the standardization defined by the BIPM).

As for dynos and my car, if it can report in standard units, that's a plus for the end-user. In an ideal world, I don't want to be tied to any particular dyno.

When I use a mic to measure the diameter of a shaft, I can use any brand of mic and obtain a similar result (in terms of the SI metre, which is reproducible anywhere). This interchangeability of measures is part of what allows our civilization to exist today. I understand that the dyno industry has not reached that level of maturity yet, but we should keep our eyes on where we want to head towards.



My point is don't find yourself chasing a point of diminishing returns. Yes you can use a inside micrometer to frame a doorway in your garage, but to what end, would a tape measure suffice? Would you pay a precision machinist to cut the trim inside your house? There is a time and place for exacting precision and tight accuracy targets, and there are costs associated with such.

Then the price goes from 120$ and hour for dyno time to over $500 an hour, as an example. You can be the one customer of 100 that wants that level of precision, traceability and accuracy, but does that mean that one garage is going to shell out the extra capitol to get to that point. The dyno industry has been able to achieve high levels of precision, but not for the price hobby and sportsman level shops pay. That is just the truth of the matter. The software isn't necessarily less capable, the math doesn't change, but the level of accuracy, precision for the sensing hardware and capabilities and automation of the data logging software is definitely going to vary by price of entry. But that doesn't invalidate the equipment shops are using today, nor make the data they provide wrong.

I am not saying dyno's are inaccurate, but I am saying they aren't all built to the exact set of measurement standards one would expect at an R&D unit built for a major manufacturer or F1 team, the cost comparisons are over a factor of 10s. Like you said a chain of money exists, which will include a bottom line. You can't expect businesses to go bankrupt trying to please your impulsive desire for data OCD. Make sure the shop you use is competent and understands proper maintenance and calibration practices, and you will receive results with a level of repeatability that will help tune your engine management to a level that is sufficient for almost all users. Championship winning race engine are tested on chassis dynos from these American manufacturers. Automotive OEMs buy chassis and engine dynos from these American manufacturers. They do work, and they do report report quality data.

Measuring the diameter of a shaft is a single gauge, with an easy to baseline standard. When you dyno you car, you take on many measurements, and they all have compounding effects on the reported result, which inevitably involves calculation, which includes a level of error propagation. Even engine manufacturers go back and forth and don't always use or agree on a defined testing standard.

It will probably scare you when you find out the last time your local hot rod garage had their torque wrench calibrated by an outside service...

As far as the actual parasitic loss of the Fiero drive train. Yes it is going to go up when power levels go up, that is why us enthusiasts always like to associate it as a percentage of power. You can find yourself a dyno that can estimate it by measuring the decel side of a dyno run, but of the few that have dyno'ed their cars, fewer have done that. And few Fiero owners want to find out. What value do we gain? Regardless of the power the engine makes, its never going to put more down to the tires, since we are almost always going to find ourselves with a transmission after the flywheel, and most of us treat it as a black box of gear ratios.
FieroWannaBe SEP 01, 08:49 AM
I have been laying in bed thinking about all the comments in this thread about parasitic loss of Fiero drivetrains, and I think people are going to read it and think that Dynos aren't accurate tools at measuring horsepower.
So I thought it would help to say none of them are necessarily wrong, but they go about measuring in different ways that can influence the number you see at the rollers.

For chassis dynos, there are 3 basic ways chassis dynos here in the US operate.

Inertia- A known mass roll is spun up to a set speed (RPM1 to RPM2) which will happen in a measured t.
Steady state- A shaft is spun at a set RPM, as engine power is applied the dyno applies torque to resist any acceleration of the dyno shaft.
Constant Acceleration - A set acceleration rate is maintained by the dyno applying torque to its shaft to resist the engines power to accelerate it faster.

As it was stated to earlier, It takes more power to speed up an amount of inertia in a faster amount of time. So as an inertia dyno, such as DynoJet, reads the power from a run with the same engine at two different power levels, the amount of power the cars driveline absorbs will be different from the two test due to this difference in accel rate.
In contrast to a constant acceleration run, the amount of power to accelerate the inertia of the car and dyno will be near the same for the two tests, so the reported number at the wheels will be different than in an inertia dyno, because at a controlled rate the driveline inertial losses shouldn't change run to run.

Two different brand inertia dynos, will have different baseline inertias and different acceleration rates, even then the absorbed power of the vehicles powertrain differs test to test because of rate differences.

If the speed of the car and dyno shaft remains constant the torque measured is the torque it takes to keep that speed constant by resisting engine power. Do this at multiple speed steps and you can create a plot of torque vs speed. This takes a lot of careful work tuning the control system to be stable with the car attached. If the speed isn't very stable, all the instability will lead to small accelerations up and down at all the data capture points and that can influence reported numbers. The smaller and faster you make these steps, the closer this test starts to become a controlled rate test. Generally steady state tests are used for R&D, pinpointing problem areas of an engine, or for powertrains with very heavy inertias. As inertias get smaller, tuning takes more effort and faster systems. Power train resonance becomes a durability factor as well.

As you move from a pure inertia based test, to a steady state test, the affects of the powertrain (engine to tire) inertia is diminished in power reporting. As the acceleration is slowed the power it takes to accelerate that driveline is reduced.
That is the biggest reason a dyno jet reads such a different number from mustang dyno, the test methods are not the same. Some other dynos can operate in blended modes of all three, and give the operator a lot of control.

All of this is outside of the environmental variables, the issues with tires, and traction on curved rolls.

Yes there is difference in precision and accuracy between a $25,000 wireless rotary torque transducer with a 12 point calibration, and a $500 linearly calibrated bending beam load cell, but it is not going to invalidate any results you achieve on the chassis dyno, they are going to report values very close to each other. There is a lot that goes into moving from the manual valve operated point to point test of 30-40 years ago, to a fully automated recording of a WOT test, and the electronics hardware needed to do so can also affect the way readings are recorded. Just compare it to the different ways fuel can be metered into an engine, and how different ways engine management can be accomplished, there definitely isn't a standardized way.

In addition, when comparing engine dyno vs wheel dyno numbers, just pay close attention to magazines advertised numbers on engine dynos, they usually run open long tube headers, that wouldn't fit in any car, no accessories, no air filters or even big air metering trumpets, and an usually an electric water pump, or even an external water pump with very cold operating temps. And no one knows if those are raw or corrected power numbers, and they rarely state test conditions. Automakers conduct tests to a higher level of standards, but there are still installation variables, although that is tightening up as time progresses. If someone wants to setup their engine on an engine dyno, with all accessories as they exist in the car, you are probably going to have to pay for that extra set up time, and even then, your car's exhaust system, probably isn't going to fit in the cell.
Will SEP 01, 09:07 AM

quote
Originally posted by FieroWannaBe:

I'll just leave it as this:
If the math says:
M*(RPM2^2 - RPM1^2)/S=300HP
and the device reports 301.5 HP is it wrong?
Or is it reporting with results within .5% accuracy?




Of course there's a tolerance to any physical system...

But when the DynoJet says 301.5 HP and the Mustang says 265 HP...

One has a tolerance, and one is wrong.

A man who has a watch knows what time it is. A man with two is never really sure.


quote
Originally posted by pmbrunelle:

There's a hub unit for the left side; it's just hidden behind the car. At 0:22 in the video, as the camera approaches the computer screen, the top of the LH hub unit comes into view.

Plus, I've been in that room a few times; I saw boxes for both sides with my own eyes.



I saw the second hub unit.

I haven't really looked into hub dynos very much. They're probably fine as tuning tools as long as they're consistent, but the visco-elastic power consumption of a loaded tire is non-trivial and the MoI of the wheels & tires isn't part of the result. If you want to measure how much power is available for accelerating your car, the hub dyno is going to overestimate.

[This message has been edited by Will (edited 09-01-2020).]

Will SEP 01, 09:19 AM

quote
Originally posted by FieroWannaBe:

Would you pay a precision machinist to cut the trim inside your house? There is a time and place for exacting precision and tight accuracy targets, and there are costs associated with such.




Well aaaaaaaackshully...

I added a vanity nook to the downstairs bathroom in my house before I moved out in 2016 to go to Afghanistan. The nook had to accept a 48" wide vanity. The vanity was built on a machine, so it would be 48" and square. I spent possibly more time than I should have plumbing, squaring and measuring the top and bottom plates for the new walls to make sure that the finished dimension of the nook was 48 1/16" across the required height and depth of the vanity, as well as square enough that the vanity would just slide in. When the contractor installed the studs and finished the walls, his jaw dropped when the vanity slid right in with zero problems or "adjustments".
FieroWannaBe SEP 01, 09:30 AM

quote
Originally posted by Will:


Well aaaaaaaackshully...

I added a vanity nook to the downstairs bathroom in my house before I moved out in 2016 to go to Afghanistan. The nook had to accept a 48" wide vanity. The vanity was built on a machine, so it would be 48" and square. I spent possibly more time than I should have plumbing, squaring and measuring the top and bottom plates for the new walls to make sure that the finished dimension of the nook was 48 1/16" across the required height and depth of the vanity, as well as square enough that the vanity would just slide in. When the contractor installed the studs and finished the walls, his jaw dropped when the vanity slid right in with zero problems or "adjustments".



It all depends on how you value your time.