Virtual Dyno, Wheel HP, Calculated Engine HP, and You #Motorama

My Cadillac STS-V as most modern cars keeps up with how much torque the powertrain is delivering.  It uses this info to make adjustments in power delivery if needed.

Knowing the Delivered Torque, we can derive the Calculated HP, since we know the standard definition HP = (Torque x RPM)/5252

Engine RPM (SAE) rpm Delivered Engine Torque ft.lb Calc Engine HP
3564 423 287.05

Virtual Dyno focuses on lovely HP/Torque graphs of power at the wheels, but of course is based on data tables which one can read from the graph node points.

Virtual Dyno WHP
RPM WHP
4489 272
4600 274
4693 278
4836 284

How do the Virtual Dyno wheel hp values relate to the engine’s calculated hp values for the same data run?

Delivered Torque to Virtual Dyno discussion

The Red line in this chart is the HPTuners Calculated HP.  I have applied smoothing 3, which is to say I have averaged the 3 prior and 3 following values at each point in order to smooth the graph.  The Yellow line is the Virtual Dyno wheel HP values for the exact same test run.

These two measure different things — the calculated hp is power at the crank; the Virtual Dyno is hp at the wheels.  We might expect them to differ by a standard dyno adjustment for transmission losses for automatic transmission vehicles, or 20%.  The Green line is the Virtual Dyno adjusted upwards to account for 20% transmission losses. Still not on top of the engine hp line.

The adjustment needed in this case is around 26%, which is the purple line (between the green and the red lines).  At high RPM this line overlays almost perfectly with the calculate engine HP line, and is closer at very low RPM, although not at mid-range values.

There are multiple sources of error and adjustment variables among these lines.

  • The engine uses a model to determine what makes engine power, which may or may not directly relate to more power at the wheels. I suspect it does relate closely to more power at the crank.
  • My test area is at 635 feet elevation, which is a 3% adder due to air density
  • In Dynojet mode, Virtual Dyno adds 9% to emulate a Dynojet result  — as a side note, I enjoyed this story on Dynojet development
  • Virtual Dyno uses car weight, driver weight, and weather entries.  If I am off on my entry it would effect the calculation.
  • Virtual Dyno is adjusting to SAE standard
  • It also uses vehicle coefficient of drag, and frontal area to adjust for wind resistance.  If I am off on these entries, or have a head-wind or tail wind, it would effect the calculation.

Now all these are normal issues to be dealt with, but the amazing thing I suppose is that it works at all — that we get meaningful data to gauge by.  My purpose for exploring these issues in this article was to get a handle on how the Virtual Dyno result differed from the Delivered Torque.

A short summary might be the STS-V has transmission losses of 22-24% and I am at the equivalent of 3% elevation, so 25-27% losses and that is what we see in the data.  However, one purpose of the thumbrule of average transmission losses for an automatic of 20% and manual of 15% is to be able to compare dissimilar cars — and how much WHP the car is putting down is what matters for how the car moves.  So the long and the short of it is, the car is putting around 400 whp and this car may have to make 540+ hp to do that, but an ‘average’ competitor will be able to do that with 500 hp.

I tend to use both to gauge mods — if the delivered torque / hp go up, that’s good, and I confirm the same trend in Virtual Dyno to determine the benefit of a mod.

What do you think?  Any advice on how to remove any variables or better use the tools?

 

 

More Air Filter Testing – K&N on Spectre CAI Cold/Full

I am gathering more test information using HPTuners and Virtual Dyno on my 2008 Cadillac STS-V across different modifications.  Today’s runs with a full tank of gas and nice weather after an overnight rain.  HPTuners shows the baro that the engine perceives, so I am using that for baro adjustments.

KN cold 2013-05-25 2 run comparison with weird midtorque run adjusted

This morning’s run in red vs the previous odd run in blue.  Today’s run showed the expected shapes — high torque at low RPM, high HP at high RPM.  Huge torque reading at low RPM today.  Not sure what additional variables the automatic transmission might be tossing in — these runs are in TUTD (touch up, touch down) mode with 2nd gear manually selected.  I have also adjusted yesterday’s run to reflect that it was with lower fuel before I filled up today.

hptuners for kn cole 2013-05-25 1300203

The Virtual Dyno is estimating wheel horsepower.  400 WHP with 20% drivetrain losses for an average automatic would equate to 500 hp at the crank.  HP Tuners estimated the engine was making 543+ hp at the crank for an instant at 6651 RPM, and above 520 from 6300 RPM on; 1- 400/520 hp  which would be a 23% drivetrain loss.  My goal for the STS-V was to get to a solid 400 WHP, which was reached with exhaust and CAI. Now I would like to get to a solid 425 WHP…

After the huge mid-RPM torque and low high HP run in the comparison above, I made an adjustment to ensure the engine wasn’t limiting boost.  In this theory, if the supercharger were breathing better and hit a higher boost point earlier than expected, the PCM might pull boost at high RPM.

Today’s run showed the engine is pulling and building boost all the way to fuel cutoff.  So either the previous odd run was just that, an odd dataset, or the new adjustments are working.

Here is a comparison of last night’s hot run vs today’s cold run:

kn cold 2013-05-25 comparison to kn hot 2013-05-24

Both of these are adjusted for weight, barometer, and temperature per the SAE standard, so they should read identically (within the error of the method used).  However, my STS-V is effected by heat to a greater degree than the standard adjustments correct so even with adjustments hot reflects lower results than cold.  This has added a dimension to testing.

 

Update:

Today’s graph had some data noise at the beginning of the run that caused the high torque spike.  With that removed the updated graph below is correct for today:

KN 2013-05-25 Correction