Myth Busters - Does a Larger Throttle Body Add Power to a 3.2?

[Peter M]

I've been pondering this for years having owned an '85 and currently a '88 3.2.

Sure the larger butterfly gives more part throttle power for each degree of opening due to the larger area and therefore flow but does it add power at WOT or is the 3.2's airflow ultimately constrained by the intake manifold or the air filter?

Searches have bought up very little and mostly circumstantial opinion. Probably the best being:

http://forums.pelicanparts.com/porsche-911-technical-forum/126005-3-2-intake-mods-pics-info-good-read.html

Where it was stated an enlarged TB and extrude hone of the inlet manifold would give 7 Dynojet HP increase.

http://forums.pelicanparts.com/porsche-911-technical-forum/1015032-3-2-throttle-body-question.html

Where it was strongly inferred that an enlarged TB would provide measurable benefits but fell short of actual data.

However when a friend mentioned he had a throttle body being enlarged I thought this was a golden opportunity to find out definitively by back to back testing.

Unfortunately not having a dyno at hand (or a standard 3.2 for that matter!), I though I could do some ramp runs using my modified 3.2 to measure the change in acceleration times between two specific rpms, say 2,500 to 6,500 rpm using the data logging function in my Motec M130 ecu. (I don't have a wheel speed sensor configured to my ecu but can log engine rpm and the time it takes between these chosen rpm points).

I also measured the net difference between atmospheric air pressure and MAP at WOT at redline for both throttle bodies based on the premise that torque is directly related to intake manifold pressure when all else is kept constant. Whilst change in acceleration times is a real world measure, I can't easily adjust for climatic conditions unlike the MAP so there will have to be that caveat. Both measures however give some solid evidence of the value of this modification.

Depending on what I find above, I may have to consider and test if the air cleaner/air box is becoming a more significant obstruction. I found from earlier testing that I found by placing a couple of large ports in the air filter lid, the airbox now flows as much as the not having a filter element and standard lid at all. (ie the MAP for both configurations were near enough the same.) However keen Mezger students will recall that whilst Porsche continued to use the same AFM from the 3.2 to the 964, they enlarged the airbox along with the the intake manifold which suggests the 3.2 one's were at their limits.



How much can a 3.2 throttle body be enlarged?

SWChips suggests the bore can be enlarged from 63mm to 66mm or about a 10% increase in area. Maxbore (maxbore.com/porsche.html) indicate an increase from 63 to 67.5mm. I measured Mike's enlarged throttle body at 66.5mm.

In terms of flow, based on laminar flow and Poiseuille's Law, a bore increase of this order is equivalent to around 27% increase in flow! However, I'm sure the TB as installed would not satisfy the laminar flow assumption and the potential increase in flow would be much less. However let's not sweat the modelling numbers and just say the change is "significant"



Comparison between the standard 3.2 throttle body and an enlarged version.

I'm also interested in assessing if the larger body makes the car unpleasantly sensitive to throttle, particularly small openings typically used around town. It's pretty sensitive/responsive even now since I've converted to Motec with MAP load sensing.

[Peter M]

Initial Impressions

I installed the larger body without dropping too many washers and clamps down into the dark recesses of the engine bay. The first drive was noteworthy for:

I didn't notice any change in behaviour at part throttle. I couldn't detect the promised "like punching the sport button on a modern car" feel promised by some adverts. It just felt it's normal smooth and creamy responsive self!

At WOT above 6,000rpm it felt great! Willing and revvy, I couldn't get enough of it. On the way home I was constructing my "Want To Buy" advert to hopefully tease out an unwanted enlarged throttle body from an ex-3.2 owner's box of bits.


Unfortunately most go faster threads end about here but I wanted some solid data to base my decision on whether to buy one or not for my own car.

[Peter M]

Test Method

I took a couple goes at this to come up with something that was achievable, reasonable safe, repeatable and minimised induced errors. I adopted the following:

1) Same road, same direction, no traffic, slight uphill rise to increase loading to better exaggerate any differences in engine performance
2) Tried to keep reasonably consistent weather conditions
3) Measure the acceleration time between 2,500rpm and 6,500rpm at WOT in 3rd gear. ie 61 to 158kph. I don't have a speed sensor but can datalog rpm easily
4) At a steady throttle at 2,000rpm in 3rd, approach a fixed point on the road, in my case a shallow causeway, at the fixed point go immediately to WOT and hold it flat until rpm reaches around 6,600 or 6,700rpm. I found starting the acceleration early and ending late further reduced driver induced errors.

I was going to average a number of runs but found using the above method gave very consistent results. However I noted other data that could be used to interpret the results such as ambient temperature and pressure, inlet manifold pressure, cylinder head temperature and Lambda.



My M130 has data logging capability that can then be displayed in a graphics package like below:



Whilst this is an old screenshot of previously testing, I used similar screens for these ramp runs. The screen can be stretched to allow you to pick up extremely detailed data. The time increment shown on the screen above for example is 0.2 seconds. Click on the Values screen and then it gives times to the millisecond along with data from all the other sensors.

[Peter M]

Test Vehicle

Whilst a standard 3.2 wasn't available, I thought my 3.4 would be representative of many other 3.2 owners who have also been thinking about the value of larger TB's for their cars.

A summary of my modifications are:

98mm Mahle pistons, flycut for additional valve clearance giving a measured 10.4:1 compression ratio

993SS cams

Motec M130GPA ECU

Twin Plug direct fire Nippon Denso CoP's

Full sequential Bosch 4 hole fuel injectors 0280 155 868's

AFM delete. Load modelling based on manifold absolute pressure, air temperature and throttle position

Twin knock sensors mounted on 964/993 knock bridges attached to the underside of heads

Twin wideband O2 sensors to allow each cylinder to have individual fuel and ignition trims

Cam position sensor to correctly time the direct fire ignition and injection

BBE generic 911 1 5/8" headers and Dansk Motorsport 2in/1out muffler

Modified airbox lid to improve flow. I've done previous testing that indicates the standard lid is a restriction and that the modified lid flows similar to a no lid, no filter element airbox:



It is a nice car:

[Peter M]

Results

[Superman]

People have been trying to out-think automotive design engineers for as long as cars have been manufactured, with dubious results. I suppose they are imagining the design engineers erred in selecting a too-small throttle body, unnecessarily leaving some performance (torque and horsepower) on the table. Good luck with that.

Consider a K&N air filter, which is said to increase power by 10%. LOL

[Black 993]

Interesting that even with all those mods the bigger TB made no real difference. I'd guess they reduce power on a stock motor by degrading VE.

[mepstein]

I think the idea is that it won’t increase hp on its own but it won’t be a choke point if you use more aggressive cams, ported heads, etc and need more airflow.

[ADDvanced]

Can you resize your pics? This forum doesn't do that automatically so it makes them hard to see without zooming in and out.

[Mr. Merk]

Originally Posted by Black 993

Interesting that even with all those mods the bigger TB made no real difference. I'd guess they reduce power on a stock motor by degrading VE.

It made a huge difference. .12sec quicker to 6500rpm. I'm sure most of that was up at higher RPM. I'd like to see a the same tests but with just the RPMs between up shifts. (4300-6500?)

I'm thinking back to the front straight at Indy where even a .2 margin at 120+mph would be a 2.5 car length gap. If you could do that on just front straight alone, think about the entire track, over multiple laps. Just from a throttle body change compared to the other guys.

[Peter M]

Originally Posted by mepstein

I think the idea is that it won’t increase hp on its own but it won’t be a choke point if you use more aggressive cams, ported heads, etc and need more airflow.

Mark,
I understand that view but looking at most 3.2 "builds", my 3.4 engine is very typical of these in having the standard 3.2 inlet manifold paired with 993SS cams because the original intention was to run the Motronic with a Steve Wong chip. The common view is that 993SS cams are the most aggressive cams that can be run with Motronic and still be streetable.

Doing the enlarged throttle body ramp runs with this enhanced "3.2" shows that the throttle body is not a "choke point" for 3.2's and our typical 3.4 build. My suspicion is that significantly more airflow would only occur with a change to a later 964/993 intake manifold or ITB's and different cams.

[Peter M]

Originally Posted by Mr. Merk

It made a huge difference. .12sec quicker to 6500rpm. I'm sure most of that was up at higher RPM. I'd like to see a the same tests but with just the RPMs between up shifts. (4300-6500?)

I'm thinking back to the front straight at Indy where even a .2 margin at 120+mph would be a 2.5 car length gap. If you could do that on just front straight alone, think about the entire track, over multiple laps. Just from a throttle body change compared to the other guys.

Mr Merk,
If you look at all 3 runs you will see that the enlarged TB did not always provide a reduction.

The ramp runs were specifically designed to eliminate errors that would be induced by gear changes so only 3rd gear was used between 2000rpm and approximately 6,600rpm. Only the time between 2,500rpm and 6,500rpm measured to further reduce driver induced errors.

The variation in measured acceleration time between the 3 runs was around 1%.

Consequently I am confident to say that a change in throttle body will not change my track position at Indy.

[Peter M]

Originally Posted by Black 993

I'd guess they reduce power on a stock motor by degrading VE.

Black993,

I suspect it would not make any difference on a stock 3.2 as I suspect the airflow is predominately controlled by a combination of the standard airbox lid, inlet manifold and cams.

I deliberately tracked the Lambda, air temperature and pressure, cylinder head temperature and MAP on my runs to ensure they were reasonably consistent. I thought that any significant variation found would effect my VE and invalidate my data. The consistency of this data I think validates my acceleration data and by inference, engine power output.

[Peter M]

Originally Posted by Superman

People have been trying to out-think automotive design engineers for as long as cars have been manufactured, with dubious results. I suppose they are imagining the design engineers erred in selecting a too-small throttle body, unnecessarily leaving some performance (torque and horsepower) on the table. Good luck with that.

Consider a K&N air filter, which is said to increase power by 10%. LOL

Superman,

I think you have got to the nub of it all, we are simply too trusting and too willing to spend money on no other justification but blind faith.

As you can see I started this testing with the expectation of a power gain, even bought a spare TB that I could send away to get enlarged. The first drive suckered me in as my "butt dyno" said it was a good thing!

I think it all comes back to: "Without data, it is just an opinion".

[Peter M]

Verdict:

[Classic11]

Hello Peter M, I really appreciate the time and effort you put in this. Thank you.

[T77911S]

i thought it would only make a difference at WOT.

i know some of the autocross rules do NOT allow for bigger TB.

i put the TB from an SC on my 77s as it was bigger.
i couldnt tell but maybe a dyno could.

[onboost]

Ahhh, so I know I'm late to this party.
Not that it doesn't make sense in theory. However, this has always been one of the mods that one begs to question.

If we're boring or using a larger throttle body, then unless you’ve also enlarged the hole in the intake itself, where the throttle body is bolted, where is your benefit? The throttle body measures right about 63mm, and the opening in the intake is the same. If you enlarge the throttle body, one should also remove and machine the intake opening i would think as you then create uniformity in flow. If not, then you now have slight ledge or lip which the increased air tumbles across as opposed to flows through... as it did with the stock configuration. Or is the intent to created that turbulence? I would think not but maybe I'm wrong.

[Peter M]

Originally Posted by onboost

Ahhh, so I know I'm late to this party.
Not that it doesn't make sense in theory. However, this has always been one of the mods that one begs to question.

If we're boring or using a larger throttle body, then unless you’ve also enlarged the hole in the intake itself, where the throttle body is bolted, where is your benefit? The throttle body measures right about 63mm, and the opening in the intake is the same. If you enlarge the throttle body, one should also remove and machine the intake opening i would think as you then create uniformity in flow. If not, then you now have slight ledge or lip which the increased air tumbles across as opposed to flows through... as it did with the stock configuration. Or is the intent to created that turbulence? I would think not but maybe I'm wrong.

Better late than never Onboost!

I didn't actually measure the diameter of the inlet manifold as none of the suppliers of enlarged throttle bodies have raised the potential mismatch as an issue. However I certainly get your point that the throttle body bore and the manifold should match.

However, I don't think the mismatch is significant for the following reasons and this is also why the suppliers of enlarged throttle bodies probably don't raise it as an issue either:

1) The majority of the boring occurs to the "stepped" venturi section of the throttle body with it being enlarged from 63mm to 66.5mm. The actual diameter of the standard throttle body at the manifold interface is 65mm. Due to the o-ring seal groove on the throttle body, this limits the bore being enlarged much beyond 66.5mm anyway. Bore too large and the o-ring could be sucked into the engine! The attached photo probably explains it better than my words:




2) Even if there is a little mismatch between the throttle body and the inlet manifold, I can imagine the airflow to be very turbulent and inefficient as it enters the plenum cross piece as it is essentially a sharp right angled tee with very little flow smoothing.



So from what I can see, even if I corrected the possible small mismatch between the throttle body and inlet manifold, I would still not realise a measurable performance gain.

[Peter M]

Someone suggested that my results are invalid because I didn't retune my engine to suit the bigger throttle body.

This would be certainly true of any modification that results in the engine now performing outside of the existing "tuned" cells of the VE/Load table or if my ECU's load modeling was based on "alpha N" throttle angle (like ITB's are predominately modelled on) rather than Manifold Absolute Pressure. My results are valid for the following reasons:

1. I observed no measurable change in the VE of the engine as the intake flow appears to be restricted by the inlet manifold and other components and not the throttle body.
   
2. The engine was tuned at near sea level and the testing of the 3 configurations detailed above was completed at 1000m above sea level. The resulting reduction of air pressure of around 5% well and truly ensured the testing was completed in "tuned" cells and not outlying "extrapolated" cells.
   
3. Further I observed consistent AFR's, Air Inlet Temperatures and Manifold Absolute Pressures across the 3 test configurations confirming no change in VE.