S4 ignition timing vs intake plenum side at WOT
12-29-2014, 01:28 PM
#61
Inventor
Rennlist Member
Rennlist Member
Thread Starter
Yet more data with long and short grouping now.
Oddly, it appears the left plenum's central long cylinders 2/3 are the ones that knock the least? (The smooth areas for the left past 4800 rpm are where I am retesting the limits, but it is too cold here 40F/4C to do much knocking.) It could be I have the bits crossed and I have long and short reversed, but I'm fairly certain on the logic.
Whatever the physics it pulls like a freight train and is whisper quiet. Pretty much I can drive one gear higher than I could before all this - 3rd pulls like 2nd used to, etc. I stopped logging in the rain after I spun the tires in 3rd at 4000 a couple of times...2nd, even with the A/C on, is a lost cause.
The average of the right side is even closer to the stock curve. Nicht wahr?
5/8 = left short, 2/3 = left long, 1/4 = right short, 6/7 = right long
Oddly, it appears the left plenum's central long cylinders 2/3 are the ones that knock the least? (The smooth areas for the left past 4800 rpm are where I am retesting the limits, but it is too cold here 40F/4C to do much knocking.) It could be I have the bits crossed and I have long and short reversed, but I'm fairly certain on the logic.
Whatever the physics it pulls like a freight train and is whisper quiet. Pretty much I can drive one gear higher than I could before all this - 3rd pulls like 2nd used to, etc. I stopped logging in the rain after I spun the tires in 3rd at 4000 a couple of times...2nd, even with the A/C on, is a lost cause.
The average of the right side is even closer to the stock curve. Nicht wahr?
5/8 = left short, 2/3 = left long, 1/4 = right short, 6/7 = right long
Last edited by PorKen; 12-29-2014 at 02:20 PM.
12-29-2014, 02:03 PM
#62
Nordschleife Master
Yet more data with long and short grouping now.
Oddly, it appears the left plenum long cylinders 2/3 are the ones that knock the least? (The smooth areas for the left at high rpm are where I am retesting the limits, but it is too cold here 40F/4C to do much knocking.) It could be I have the bits crossed and I have long and short reversed, but I'm fairly certain on the logic.
Whatever the physics it pulls like a freight train. Pretty much I can do things one gear higher than I could before all this - 3rd pulls like 2nd used to, etc. I stopped logging in the rain after I spun the tires in 3rd at 4000 a couple of times...2nd, even with the A/C on, is a lost cause.
The average of the right side is even closer to the stock curve. Nicht wahr?
Oddly, it appears the left plenum long cylinders 2/3 are the ones that knock the least? (The smooth areas for the left at high rpm are where I am retesting the limits, but it is too cold here 40F/4C to do much knocking.) It could be I have the bits crossed and I have long and short reversed, but I'm fairly certain on the logic.
Whatever the physics it pulls like a freight train. Pretty much I can do things one gear higher than I could before all this - 3rd pulls like 2nd used to, etc. I stopped logging in the rain after I spun the tires in 3rd at 4000 a couple of times...2nd, even with the A/C on, is a lost cause.
The average of the right side is even closer to the stock curve. Nicht wahr?
I wouldn't discount the possibility that there's some issue somewhere. Like one of the knock sensors having a bad connection or something.
However, assuming away any possible problems and taking is as given that these data are correct, the exhaust explanation seems the most logical one. I am not saying it's likely, just that it's not inconsistent with your data.
There are two cylinders that go into overlap soon after another cylinder has fired in the same bank. #3 fires 90 degrees after #1 and #5 fires 90 degrees after #6. It may be that the blowdown from #3 stuffs #1 at overlap and blowdown from #5 stuffs #6 at overlap. This would lead to cylinders #1 and #6 to knock a lot.
If I understand your numbering correctly, #1 is in the RS group and #6 is in RL group. Those are the groups that knock a lot. Did I get this right? [Edit: Ken's new labeling seems to indicate I did in fact get this one little piece right.] If I got that right, then it's possible (I won't say probable) that the knocks are coming not from the intake manifold but from the exhaust manifold.
One (painful) way to test this would be to install a long-tube headers and see if the knock disparity goes away.
Another way would be to change your cylinder grouping a bit to 1/6, 2/8, and 3/4/5/7. The first group has the same bank blowdown from another cylinder after 90 degrees, the second group has the same bank blowdown after 270 degrees, and the last group has the same bank blowdown after 180 degrees. If you find that only 1/6 knocks easily and 2/8 don't knock much at all, then that result wouldn't allow us to reject the exhaust theory.
I didn't come up with all of this on my own, of course. You can read more in the following SAE paper (2011-01-0337):
Blowdown Interference on a V8 Twin-Turbocharged Engine
Apoorv Agarwal, Hosuk Jung and Kevin Byrd -- Ford Motor Co.
Robert A. Stein, Ahmad Kassem and Paul Whitaker -- AVL Powertrain Engineering Inc.
Christian Spanner -- AVL List GmbH
Here's the most relevant section. I've changed the cylinder numbers to match the 928 cylinder numbering and firing order:Apoorv Agarwal, Hosuk Jung and Kevin Byrd -- Ford Motor Co.
Robert A. Stein, Ahmad Kassem and Paul Whitaker -- AVL Powertrain Engineering Inc.
Christian Spanner -- AVL List GmbH
V8 BLOWDOWN INTERFERENCE AT HIGHER ENGINE SPEED. As engine speed is increased, the pressure wave travel time in the exhaust manifold becomes significant in terms of crank angle degrees. This causes the blowdown interference pulses to arrive later at the cylinder of interest, and paired cylinders for each bank no longer behave identically.
Cylinder pressure pumping loops for the right bank of the V8 engine at 4750 rpm are shown in Figure 11. As previously explained, cylinders 2 & 8 do not suffer from blowdown interference. For cylinders 4 & 7, the 180° interference pulse from cylinders 1 & 6 arrives after exhaust valve closing at 4750 rpm and consequently the pumping loop is very similar to that of cylinders 2 & 8. ...
Of primary interest at high speed is the effect of the 90° interference on cylinders 1 & 6. Because the 90° pulse arrives later in the cycle of cylinders 1 & 6, it not only increases exhaust stroke pumping work, but also increases the exhaust pressure during overlap, increasing residual fraction. As a consequence, cylinders 1 & 6 are more knock limited at high engine speed than the other 6 cylinders.
Cylinder pressure pumping loops for the right bank of the V8 engine at 4750 rpm are shown in Figure 11. As previously explained, cylinders 2 & 8 do not suffer from blowdown interference. For cylinders 4 & 7, the 180° interference pulse from cylinders 1 & 6 arrives after exhaust valve closing at 4750 rpm and consequently the pumping loop is very similar to that of cylinders 2 & 8. ...
Of primary interest at high speed is the effect of the 90° interference on cylinders 1 & 6. Because the 90° pulse arrives later in the cycle of cylinders 1 & 6, it not only increases exhaust stroke pumping work, but also increases the exhaust pressure during overlap, increasing residual fraction. As a consequence, cylinders 1 & 6 are more knock limited at high engine speed than the other 6 cylinders.
Last edited by ptuomov; 12-30-2014 at 10:41 AM. Reason: Added the reference
12-29-2014, 02:19 PM
#63
Inventor
Rennlist Member
Rennlist Member
Thread Starter
12-30-2014, 04:17 AM
#64
Inventor
Rennlist Member
Rennlist Member
Thread Starter
Hmmm...forgot about the advance rate per spark...will have to program around this at WOT. The left side may still very well be higher, just not that high.
The EZ retards immediately but IIRC normally only advances a few degrees per each spark, so the left side may be much lower - it never reaches that high before the next cylinder. (I'd better lower that waaay down before trying it after recoding.
)
The EZ retards immediately but IIRC normally only advances a few degrees per each spark, so the left side may be much lower - it never reaches that high before the next cylinder. (I'd better lower that waaay down before trying it after recoding.
)
12-30-2014, 02:05 PM
#65
Inventor
Rennlist Member
Rennlist Member
Thread Starter
I wouldn't discount the possibility that there's some issue somewhere.
Another way would be to change your cylinder grouping a bit to 1/6, 2/8, and 3/4/5/7. The first group has the same bank blowdown from another cylinder after 90 degrees, the second group has the same bank blowdown after 270 degrees, and the last group has the same bank blowdown after 180 degrees.
Another way would be to change your cylinder grouping a bit to 1/6, 2/8, and 3/4/5/7. The first group has the same bank blowdown from another cylinder after 90 degrees, the second group has the same bank blowdown after 270 degrees, and the last group has the same bank blowdown after 180 degrees.
IIRC the S4 cams have a lot of negative overlap. Would the blowdown still be a factor? (Cool stuff, BTW.) At some point I will make the cylinder grouping more easily selectable so I could test the ideal pairing.
12-30-2014, 02:26 PM
#66
Nordschleife Master
(Not that I know anything about this, but it seems to me that a little more overlap and long-tube headers would work nicely in improving the gas exchange. All theory and no practice on N/A headers...)
Even if there were no overlap, the exhaust gas pressure in the combustion chamber would be different between #6 and #8 at their respective exhaust valve closings because of the blowdown interference effect. Consequently, #6 would have more hot exhaust gas mass in it after the exhaust valve has closed.
------ Valve Flow & Cam Calculations
Overlap Area, deg*sq-in 0.0
Vlv Area, deg*sq-in in 237.7 ex 193.3
Total Exh/Int % 81.3
Total Avg Flow Coef in 0.309 ex 0.335
Lobe Separation, deg 106.5
Lobe Area, inch*deg in 24.05 ex 20.40
Overlap, deg 7
Duration @ .003, deg in 230 ex 217
Overlap @ .050, deg -21
Opening Events, deg in 0 ex 30
Closing Events, deg in 49 ex 7
Last edited by ptuomov; 12-30-2014 at 02:52 PM.