K&N RU-3100 comparison to Amsoil eaau6065

Amsoil 6065 filter installed on the Cadillac STS-V Spectre intake

Amsoil 6065 filter installed on the Cadillac STS-V Spectre intake

I recently replaced the Spectre filter in my 2008 Cadillac STS-V (supercharged 4.4L V8) with Spectre CAI with a K&N RU-3100 due to delayed availability of the Spectre replacement.  I also picked up an Amsoil eaau6065 filter.  These are both the same form factor — 6″ mouth, 6″ long, 7.5″ to 5″ cone shaped.  The K&N is oiled gauze, and the Amsoil is a microfiber dry filter.

The K&N was $39.86 ($43 delivered via Amazon) and the Amsoil $60.05 (just over $72 delivered).

The test runs were done using a similar method but on different days; 88F for the K&N, 91F for the Amsoil test.  To test the two filters I used HPTuners to capture both lb/min output from the MAF during a wide open throttle acceleration run in 2nd gear, and vacuum pressure at the supercharger inlet in PSI.

This was a heat-soaked test comparison, meaning that the engine of the STS-V, intercooler, and air inlet were hot, and the day was hot.  This results in lower performance, but was similar across the two tests.

MAF and SC Inlet Vacuum

MAF and SC Inlet Vacuum

A higher MAF reading is better, and a lower SC inlet pressure reading is better.

What the test showed is that the 2 filters recorded equivalent MAF readings, but the SC inlet vacuum for the K&N was lower than the Amsoil air filter by 0.0-0.15 psi, or 0-4 inches of water.

Comparison of hot runs K&N and Amsoil

Comparison of hot runs K&N and Amsoil

I also ran a virtual dyno which appeared to show equivalent performance with test to test variances.  The Amsoil result had equivalent peak hp (366 hp vs 367 hp) and appeared to make more torque at lower RPM, but my impression is this is virtual dyno variance because we don’t see a variance in the MAF readings at similar points.

Hp Tuners Delivered Torque Comparison

Hp Tuners Delivered Torque Comparison

The PCM’s evaluation of delivered torque from each run appears to be very equivalent.  Not surprising as this depends on MAF input, but persuasive that output performance between the two is similar.

Conclusion

The K&N appears to flow slightly better — less vacuum is pulled — but the resulting air flow through the MAF in lb/min and Delivered Torque calculations look very equivalent between the two.

Testing Toward MBT — New Max Torque

Today I continued my testing toward MBT by adding timing across the rev range to move the commanded advance closer to the MBT vs RPM vs Cyl Air table.  1 degree of advance was added from 3000-3600 RPM, and 0.5 degrees from 3800-6000 RPM.

The V was cold, in cool air, almost out of fuel.

I went to a less densely populated HPTuners table to gauge if that improved the data capture.  My impression is this was useful, but that we are capturing info as fast or faster than the related PID are updating anyway.  I would prefer to have the more diverse info points than oversample a smaller set for future tests.

More timing makes more Torque

More timing makes more Torque

This graphic shows today’s delivered Torque in blue, and the previous delivered torque in yellow.  The current total advance in red, and previous in green.  This shows that by touching up the timing the V gained ~5 lb ft of torque throughout the range.

HPTuners data delivered torque & calculated hp comparison to previous

HPTuners data delivered torque & calculated hp comparison to previous

This graphic is the same info but with calculated HP compared current and previous.  The actual high points were:

Hp RPM Torque RPM
Previous 541.65 6585 453.0 3663
Current 555.78 6710 471.0 3478
Current 545.6 6573

The previous test run topped at 6585 rpm where the 2-3 shift hit; today’s run was for 545.6 hp at a similar 6573 RPM.

HP Tuners at Max Torque

HP Tuners at Max Torque

This is a hptuners snapshot at max torque point

Hp Tuners at Max HP

Hp Tuners at Max HP

and here is a similar shot at max hp point.  I forgot to tell the car it was okay to shift to it redlined in 2nd for this test.

Virtual Dyno comparison of previous capture and today's run

Virtual Dyno comparison of previous capture and today’s run

The virtual dyno run shows the low torque spike but only a 386 hp peak measured.  Today’s run in red, previous odd run in blue.

MBT retune 0-60 chart

A 0-60 run measured from the HPtuners log took 5.17 sec without rollout.  The PerformanceBox unfortunately didn’t properly capture the run so I don’t have a good figure for what it would have been with rollout.

MPH Speed
0
10 0.98
20 1.70
30 2.22
40 3.02
50 4.14
60 5.17

I really wanted to capture MAF lb/min but that was one of the things I did but should not have removed when lightening the HpTuners data table.

Summary

  • Tuning the engine across the RPM range did improve the power output as expected
  • The power gained by these slight improvements of ~5 lb ft shows that the engine is already operating close to optimal
  • Some KR is observed at 3385 RPM and 4000 RPM, suggesting that the OEM tune is close to the knock limited max torque

My conclusion is that there some gain with additional timing, but that the OEM approach is very close to optimal and provides some margin for knock / detonation.

Update:

I reduced 0.5 degrees of timing advance everywhere there was a hint of KR.  A follow-up run hot (more likely to knock) showed no KR in the tuned area.

Virtual Dyno comparison same day cold & light vs hot and heavy

Virtual Dyno comparison same day cold & light vs hot and heavy

To run the KR checks I also captured a 2nd gear test run.  Although the runs are adjusted per SAE for temp and pressure the adjustments don’t have as much impact as temps do on the STS-V.  The V runs ~22hp less power hot after adjustments, and 30+ hp missing without adjustments.

While getting my tables fixed I also found an S/C inlet pressure pid.  This sounds like pressure at the inlet of the supercharger, and is in absolute pressure.  So if we subtract this value from the barometric pressure in psi we can get vacuum at the supercharger inlet.

SC inlet pressure vs boost and RPM

SC inlet pressure vs boost and RPM

The SC inlet pressure goes from 0 psi at low RPM up to 1.30 psi at redline.  This figure appears to be ideal for planned air filter and intake testing.

Testing Toward MBT — 6400 RPM at 23.5 Degrees

Today’s test is a consideration of if a Cadillac STS-V supercharged VVT DOHC 4.4L V8 will make more power or less power with slightly less timing at 6400 RPM.  I am using Hp Tuners to data log and for one graphic, and Virtual Dyno for the other graphic.

Cadillac’s Max Torque vs RPM vs Timing table suggests that 23.5 degrees of advance is ideal at 6400 RPM, but the PCM commands more.  So we test and see which is more right.

Torque and HP vs Timing

Torque and HP vs Timing

This graphic shows Delivered Torque in blue, Calc Hp in green, and timing in red.  During this run I had also made a small adjustment to 2-3 shift points so that we get closer to redline in 2nd before the shift.

Test: Our focus is on 6000-6700 RPM — does the HP drop off with less timing, or does it stay on an up-slope?  My observation is that it stays on an up-slope.

Virtual Dyno runs

Virtual Dyno runs

I also captured 2 virtual dyno runs, on the same road, same day, same direction.  The first was with the intercooler temps lower, and the 2nd just after the 1st so with intercooler temps higher.  Clearly the results are sensitive to intercooler temps, as well as normal variances in test method, etc.

Although the first run in red slows an early spike that over-shadows the high RPM performance, in neither run does the performance fall off a cliff past 6000 rpm, which is what we are looking for (above 400 whp is a good result in this range).

HpTuners gauge snapshot at peak RPM 2nd gear

HpTuners gauge snapshot at peak RPM 2nd gear

Conclusion

The reduction in timing to 23.5 degrees appears to make power similar to the higher timing point — no disadvantage from 6000-6600 RPM.  We want to run the least timing that produces equivalent power, so this seems a good result.

The slight adjustment in 2-3 shift points was successful at getting better high-rpm data in 2nd for testing.