3.2 intake plenum flow testing... interesting results.

[gliding_serpent]

This originated in this thread, but figured I would give it it's own publicity.

I got my plenums back from my local machine/performance engine shop. I was curious, so had it flow tested. Why? Because some say extrude honing can add 5 or so hp on the top end, or at least balance air delivery and AFR's to each cylinder. I have come across opinions from respected members of the 911 tuning/engine building world that have differing opinions on extrude hone benefits, ranging from never having seen power benefits, to ~5-7hp gains in max rpm ranges.

The owner of the shop did the flow testing, and was polite, but not impressed with the results (they make highly tuned and blueprinted vintage muscle-car and dragster engines).

I was initially impressed that things were pretty close, until i saw cylinder #6.


#1 261 CFM (cubic feet per minute)
#2 251
#3 252
#4 245.5
#5 248
#6 238

Lets run some numbers...

- The largest spread is between #1 and #6: 23 CFM. That is a near 8.8% difference.
- The mean CFM is 249.25.
- Thus deviations from the mean are as follows...
#1 +4.7%
#2 +0.7%
#3 +1.1%
#4 - 1.5%
#5 - 0.5%
#6 - 4.7%

Now, my engine will be punched out to a 3.4L (actually 3367cc, or 205.47 cubic inches). Lets calculate the total CFM demands of my engine...
- max rpm (6760) x cubic inches (205.47) = 1,388,977.2
- 1,388,977.2/3456 = 401.9
- 401.9 x 0.85 (presumed volumetric efficiency) = 341.6CFM
- Thus, an ideal intake should provide about 342CFM to meet my 3.4L engine's intake needs. In other words, 56.9CFM per plenum!!!
- My intake (at the plenums) provides 249.25CFM on average per plenum. The worst is 238CFM (#6). That is still over 4 times the flow needs of each cylinder at max rpm.

But what does this mean?

Such info in my mind suggests that extrude honing of the plenums makes little difference if any, as this simply is not an area of flow limitation at these power levels... not by a long shot. I get the matching AFR benefits, but as each plenum is so far from being flow-limited, I can not see that you could get measurable benefits from extrude hone unless you have an engine with maybe 100% volumetric efficiency and 3+ times the displacement.

This would support an opinion shared to me from a well known tuner that they had never seen dyno proof of gains from extrude hone. More obvious intake bottle necks to rectify would be the low hanging fruit at the airbox (mine will be opened cup-style, Steve Wong data suggests a 5hp gain in a chipped 3.2), the throttle body (mine will be bored out from 63mm to 66mm, felt by SW to add maybe 1-2hp, not low hanging fruit), or the AFM (not low hanging fruit).

I welcome thoughts, counter-points. I suspect this is not this easy. I will also look to see if there is a more accurate way to estimate the total CFM needs of an engine. In other words, this is work in progress, and not the be-all end-all on this matter.

[Elombard]

Interesting. I wonder if smaller diameter runners with faster flow would actually improve things? I am sure its more complicated then that.

[gliding_serpent]

Originally Posted by LJ851

That is not how it works, only one cylinder pulls at a time.

I pulled this comment from the original thread.

My take on things was that any one of the six cylinders pulls air at any one time. As such, if the total theoretical CFM (cubic feet air per minute) of the 3.4 engine were 342CFM, than 342/6 would be the CFM demands of any one cylinder. If it were a 4-banger it would be 3.42/4 and so on.

[LJ851]

Flowed Carrera 3.2 Intake

[gliding_serpent]

Originally Posted by LJ851

Flowed Carrera 3.2 Intake

Awesome! I was just about to link that very post to this thread. You beat me to it. That was 2002 mind you. Last month Steve told me he has not personally seen dyno data to date showing gains from extrude honing. Man sees a lot of dynos. I have not found any either, but I am only googling. Mind you, maybe he has never seen a car dyno pre and post honed intake swap on the same day/same dyno. Not an easy thing to do same day, unless you have two (one honed, one not).

I am trying to figure out if this extrude honed thing is just dogma. I like to see data (dyno curves).

[LJ851]

Trying to find the cfm rating of a 3.2 intake port, i guarantee it is WAY higher than 56 cfm. The intake manifold (ideally) should not be a restriction.

[Tippy]

I really appreciate this data.

[gliding_serpent]

Originally Posted by LJ851

Trying to find the cfm rating of a 3.2 intake port, i guarantee it is WAY higher than 56 cfm. The intake manifold (ideally) should not be a restriction.

I look forward to this info and thanks for helping to figure this all out. Of interest, the intake ports of the 3.2 engine are said to be overkill for the engine (better than the 930)... good to 320-350hp or so based on my readings around here (assuming my memory is up to task).

I found one thread that had it in the ~300cfm per port range (EDIT: those numbers may have been post extrude honing of the heads)... had it earlier in the evening, will see if I can find it again. The folks in that thread figured those numbers would be good for an engine making 600-1000hp.

[LJ851]

Originally Posted by gliding_serpent

I am trying to figure out if this extrude honed thing is just dogma. I like to see data (dyno curves).

I completely agree with you.

[LJ851]

The stock 3.2 engine is (from what i understand) is on the threshold of utilizing 46mm carbs/ITBs. As in it will make more power with 46mm vs 40mm. A single 46mm TB will flow around 150cfm so each cylinder of a 3.2 must be using close to that amount.

[gliding_serpent]

Here is Steve Werner weighing in on the issue...

3.2 Intake Manifold Runners-FLOW

Originally Posted by Steve@Rennsport

Andy:

I don't sell the process of honing these intake manifolds and have no vested interest in anyone doing such things, but we have done extensive flow testing of these since 1984 and there certainly are variations in airflow from branch to branch in many of these things.

We've seen up to a 30% difference in runner airflow that does affect cylinder power balance. Assuming for the moment that each cylinder receives the same amount of fuel, balancing airflow in each runner does improve HP and lessen crankshaft stresses.

Speaking only for myself, I would tell anyone that its worth having the intake manifolds flowed to see what kind of benefit one might gain from such a process.

JMHO, as I have no "dog in this hunt",... I simply have learned what works and what doesn't.

Logic would tell me that they know far more than I do... but I have yet to see the dynos. Also not sure if 8.8% difference from best to worst plenum, or 4.7% difference from average flow is worth the trouble/cost in my case. The 15-30% folks, sure. That 30% case sounds like a rodent nest to me...

Edit: Thinking about what Steve said some more, in theory this kind of proves that extrude hone works (with one caveat). The idea being that reduced flow into each cylinder with a fixed fuel volume will produce a rich mixture, which may bring the AFR below optimal power. Increase that flow, and the AFR increases, giving more power in that cylinder (assuming your DME is programmed to produce an AFR that gives max power). The trick here is that this concept looks good on paper, but if each plenum produces flows 2-4x the CFM that each cylinder needs, than a plenum with 30% reduced flow would still provide more flow than required, and thus not producing much if any reduction in AFR/power. Or this is the logic I am running by.

My suspicion is that the reduction in plenum flow is not proportional (1:1) to the reduction in AFR/power given that each plenum can flow far more air than the cylinder needs. So a plenum with 30% less flow would not have 30% less air, and thus 30% less power from the associated cylinder. If the plenum provided JUST ENOUGH flow to meet the needs of the cylinder, this would make sense, but again, if each plenum can provide 2-4x the needed max flow of the cylinder (and remember engine CFM numbers above were for 3.4L, not 3.2), I don't see this as being a significant bottle neck. Thus I don't think you will see significant gains from extrude hone. I will re-think my logic if data shows that my CFM needs per cylinder are off (by a factor of 4).

Mind you, 5 hp is 2.2% on a 230hp engine (typical healthy chipped 3.2). So it might make sense that one plenum that is 30% weaker for flow might turn into a very non-linear 2% power gain when you correct that reduced flow. So with my 8.8% max flow difference (best to worst plenum), or two with 4.7% (from the average plenum flow) flow difference, I could only expect maybe 1/3 to 1/4th the gains compared to the 30% scenario. Lets assume 15% is the variance in flow needed (or 7.5% from the mean for two plenums) for the extrude hone to make an engine 6hp extra. I am about half of that... so I could expect half the gains. So, best case scenario, assuming extrude honing works, for 600$ + shipping I could get maybe 3hp.


Will extrude hone prevent a cylinder from running too lean due to imbalanced plenum flows?

Time for some more armchair logic. Of interest, I have read many folks mentioning that preventing one cylinder from running too lean is one reason to extrude hone. I think this is not the case. Let me explain.

When the intake valve opens (only one is open at a time), there is a vacuum process. Air gets sucked in through a straw into the combustion chamber of the piston. The volume and rate of that air is the CFM number. For my 3.4L engine, lets call it 100CFM, for simplicity sake. Now, say that 5 of 6 plenums on our car can provide 100CFM to their respective cylinders. Lets say the last, cylinder #6 can only provide 70 CFM due to poor casting. Now, lest assume a 13:1 AFR gives max power. Thus, our injectors are giving 7.7units fuel per 100CFM (13:1). But, in cylinder # 6, you only have 70CFM to the 7.7 units fuel. this AFR is closer to 9.1:1. Power is lost. Increase the flow, increase the power. Cool.

But wait, how does one get a lean condition? You get a lean condition if your injectors are not producing the expected 7.7 units of fuel (clogged, out of spec, or you just have the DME inject less fuel). In theory, you could also go lean if you could ram more air in the cylinder than intended. This is not a turbo, so this can not happen... air enters the cylinder due to vacuum, not push. (NOTE: not fully true as wave pulses of air going up and down the plenums can in effect ram air in the valve... thus why intake length is key. I am not sure how much of an effect this has on things). So, if plenum #6 is now 130CFM, will it force in more air? No. Why? Because the vacuum is only 100CFM. The extra flow capacity of that plenum is just not needed. AFR is still 13:1.

Now, where I can see you getting into trouble is with a split header system with separate banks of 3 cylinders (i.e. 1-3 and 4-6 seperate) and the AFR data is taken from one side. Go back to my first post and note that my plenum flows are less on the 4-6 side of the engine. My system has SSI's. Lets say I will be getting AFR data from the side of the engine that is less intake limited (1-3). As such, if I have steve wong tune my ECU based on this AFR data, I will be safe as this will be the "leaner side." However, the 4-6 side of the engine will run a bit more rich (power loss) assuming the flow differences are significant (and I am arguing they are not).

If I had my AFR sensor on the 4-6 side (less intake plenum flow), and then calculated my afr, and optimized the AFR... the 1-3 side might run slightly lean (due to less fuel being programmed to go in). Take home here would be to check the AFR of both sides at dyno time, and tune your AFR to the leaner side.

So, if extrude hone was able to balance all plenum flows, you could avoid the above issues (assuming your injectors were to spec...). My argument is still that if each plenum can flow 2-4x the air needed by any one cylinder, small changes in plenum flow from cylinder to cylinder will not be significant. Basically, the engine is not intake limited at this point, so a few percentage points here or there do not matter.

I know this is an oversimplification, but this is how I am currently processing this.

[gliding_serpent]

Now, the next test of interest would be to run the same flow testing, but with the stock throttle body, afr, and airbox in place (starting from the snorkel). You could then see how things like an enlarged throttle body helped out, and how much the upstream intake components contributed to a end flow reduction. I suspect that at the end of the day, you might find flow getting closer to the 56.9cfm need of each cylinder.

Not sure i will bother testing this (would need to pay to setup the flow testing again), but i am curious.

As i see it, when it comes to flow testing, it is less about the absolute numbers, and more about the variation seen in flow from plenum to plenum. If your differences are not that much, you will probably not benefit from extrude hone.