Intake runner cover 5

[Louie928]

I'm designing an individual throttle body per cylinder intake system. I will need an air filter assembly with a closed top over the air horns. The question is how close to the bell mouth opening can I have the air filter top, i.e., the clearance space between the bell mouth opening and the filter top. Engine Analyzer Pro says the peak flow velocity will be 500 ft/sec in the runner.

Thanks

Louie

 

[Retracnic]

Louie928,

I don't have a concrete number for you. I can say, as a general rule, you would want to locate the filter element as far away from the throttle-body opening as possible. The reasoning behind this is a simple one. No matter how efficient the filter media you chose, there will be a pressure drop in the incoming air charge as it passes through the filter. The further away we can locate the filter, the more time we can give the charge to equalize pressure and adopt a more laminar flow pattern.

Good Luck
Bryan Carter

 

[Louie928]

Thanks Bryan,
The filter element itself will have enough space between it and the runner entrance. What I can't find is the acceptable distance from the runner bellmouth to the lid, or cover, (sheet aluminum) that is above it. Fluid dynamics books I have don't address this issue and I can't find it on the SAE site.

Thanks,

Louis Ott

 

[patprimmer]

I don't know, but instinct tells me to be at least the O/D of the bellmouth clear of the bellmouth, but then when you think of valve lift vs size, I think about 30% of dia allows maximum flow, provided there is no shrouding.

1/4 the dia clearance gives the same cross sectional area between the valve seat and the head as the area of the face of the valve, so theroreticaly it need only be 1/4 the dia of the bellmouth.

I think I just confused myself.

I think if it is a good design with no obstriction to flow from any side of the bellmouth, then 1/4 dia should just scrape in

Regards
pat

 

[Retracnic]

Louie,

Sorry, I slightly misinterpreted your original question. I'd have to say once again that you'd want your lid as far away from the bell mouths as space allows. Unfortunately I don't have any hard data at hand, but my reasoning goes something like this:

We know from basic aero/fluid dynamics, that air does not like turning corners. Typically, an air stream that is forced to turn more than 8-10 degrees becomes turbulent. If our filter/lid assembly is too close, then all of our incoming charge will be forced to make a sharp (nearly 90 degree) turn to enter the throttle bells. This is not good. If we made use of an assembly that provided more clearance, then a larger percentage of our charge would be able to make a straighter shot.

Pat's theory has merit, but seams a little on the short side to me. If I understand it correctly, that would imply that an 80mm throttle bell would require only 20mm of clearance between the opening and the lid. Appears rather small to me. I would say as a guess, that you need to be looking at 50% of the bell diameter. But I'll be the first to admit, that it's a guess.

In any case, it's a problem that can be solved with some cardboard, duct tape, and a few hours on the dyno.

Good Luck,

Bryan Carter

 

[patprimmer]

Bryan

Mines only a guess to, based on airflow in a port vs valve lift vs valve dia.

I would prefer to see 50% or even 100% if there is room.

Regards
pat

 

[evelrod]

Louie, the old man speaks again. Your set up sounds very similar to the Lotus twincam engines I raced for years. My setup was two 45mm DCOE 9 Weber carbs. Over the years several intake systems were used and it was found that the air horns shape was more critical to airflow than length and , to answer your question, the gap between the airbox and the air horns needed to be no less than 1 1/4 inch. Below that we experienced a reduction in power above ~8000 rpm. OEM setups on Lotus Elans, Cortinas, etc. used gaps of less than this but also never operated much over 6500 rpm in stock form. Our end setup was closer to 2 inches and my current air box on my racing mini (45 Weber, very short horns) is about 4 inches, but we tried 1 inch and it functioned well up to my 7200 rev limit (I just like the bigger air filter).
It's more than just the gap between the air horns and the cover as airflow patterns within the air filter/airbox assembly can play havoc on carb setups. I use good filters to slow and defuse the air a bit, "ramair" is tough to jet for, not consistant over the broad speed range.

Rod

 

[Louie928]

Rod,
Thanks for your personal experience info. A friend did some computer flow analysis and came up with a 1.75 inch spacing as being good, and that correlates with your experience as well as some other's thoughts. The engine is a Porsche 928 32 valve, 6.5L. I am using 4 dual throat FI throttle bodies, similar in appearance to the webers. They are 50mm ID, and the bellmouths have a radius of .25 x the diameter of the runner ID, i.e., 1/2" radius. Fluid mechanics formulas indicate that this inlet radius is good. I'm thinking of casting some others with more of a taper to the inlet rather than the simple radiused opening.

Thanks,

Louie

 

[Flowczar]

Louie,

The inlet bells make a large difference in airflow at high rpm, especially when combined with a taper through portions of the inlet tract. A good general rule for taper is between 1 and 1.5 degrees per foot. Experimentation proves valuable when tuning and optimizing for higher prm ranges. There are times when a flat plate, positioned above a correctly radiused intake bell, will reflect an inlet pulse down the inlet tract and negatively effect the tuning of the runner. Steady state testing will not show it, but you will notice it in a running engine or on the test sheets. An experiment to help locate these tuning "dips" would involve measuring the air into the complete inlet system and comparing it to a system without the plate (or filter cover) in place.

Have a great time, sounds fun!

 

[PFM]

Flowczar,

Funny you would mention the plate above the bell and reflecting the "wave" or pulse back down the inlet. I have seen this plate "tuned" on the dyno to improve output in a given range as well.

Again you cannot see this on a flow bench.

Regards

PFM

 

[evelrod]

PFM, you can see this "stand off" while the engine is running at moderate to high rpm on the dyno quite easily. In the good old "bugeye Sprite" days it is one of the things we looked at to determine what length trumpet, what shape and how close to the end to make the airbox cover for those little SU's---besides looking at what the automotive gurus had written and 'spying' on the competition a bit. ;-)
I was a LOT younger then.

Rod

 

[Marquis]

Louis
very interesting, I kind of kept quiet in this thread until I saw the numbers 928!
I own a Porsche 928 32 valve S4.
I have also helped a friend make a 6.5 litre version of this engine using Porsche 968 VCT units! It wasn’t easy and the cams needed to be reoptimised but on a dyno jet it produced a lot of power. I used my professional engine performance development experience and this engine produced about 85 BHp/litre and lots of torque with a wide band. I say “about” 85 Bhp/litre” because the test work was carried out on an aftermarket dynojet 248 which although is very consistent it may not provide a valid comparison to the OEM manufacturer Schenk dynos I’m used to calibrated to DIN 70020.

What is your application?

85 Bhp/litre may not sound a lot but remember how big the cylinders are and how much air must be fed to them each!

Now, next, lets tackle your initial question: the distance from the bell mouth to the flat plate/wall.
I tested a high horse power per litre (90 bhp/litre) 3 litreV6 on an engine dyno a while back for a production application. On this engine one bank had a plenum wall about 40mm next to the 3 runner bell mouths the other had lots of room. On test there was no dicernable difference in bank to bank AFR –therefore the air flow per bank was not particularly different.
Analysing it, it’s not difficult to see why. Assuming you have a good peripheral area around the bell mouths you can get quite close.

The engine in question had 64 mm diameter runners at the plenum end. As Flowczar correctly said, this isn’t a steady state problem but a transient one. You want to make sure that at peak gas flow demmand this area around the bell mouth-close to the wall doesn’t represent a restriction.
A initial funademntal way –that simply makes sure that this periphery area doesn’t represent a restriction is to equalise the entry area of the intake runner to the periphery “cylinder” area that the runner will draw from:
In my engine example for instance we had
Runner entry area needs to be greater then or equal to the periphery cylindrical area
or
Pi* (R)^2 = 2*PI * (R) * X
Where R is the radius of the entry to the runner
Pi is 3.14159
X is the distance to the plenum wall

In my example we fine
R was 32 mm and X was 40 mm
And 2*PI*( R) * X was far greater then pi* (R ) ^2 hence there was no retriction there. In fact if we cancel the above formula we arrive at a wall distance X of R/2!!

This reflects Patprimmers emipirical recommendation! So as long as you stay above one quarter of the runner entry Diameter you should present no restriction here.

In terms of effecting the WAVE pulse refelction by the use of this wall it is possible and this could be beared out not only with test but using simulation tools too. However this is likely to effect the peak torque “tuned” region and could quite easily adversely effect peak power region If not done correctly.

When considering an intake system for any cruciform crank V8 it’s worth thinking about Charge robbing.
The problem with cruciform crank V8s is that although they are even firing as a whole engine they are uneven firing when you consider one bank and some cylinders fire VERY close together.
The 928s firing order is 1-3-7-2-6-5-4-8. You’ll notice that 6 and 5 fire only 90 degrees apart on the same bank and cylinders 3 –7 and 4-8 fire consequitives across banks. This could lead to charge robbing if you package the runners badly. In my opinion Porsche went “over board” with the original 32 valve S4 design grouping up equal phase cylinder in separate plenums to make a twin plenum design (grouping cyls 5,2,3 ,8 and 1,6,7 and 4!. Although this gave a degree of resonance charging tuning, it also lead to tortuous runner paths. I prefer the solution on the BMW M5 V8 which all runners into a Large plenum but with each runner subtley directed in different directions in sympathy with it’s firing order!

Going off topic a bit but staying with Porsche 928 tuning.
The 928 pent roof chamber has 37 mm inlet valves which is plenty for 450 Bhp plus (400 BHP BMW M5 has 35 mm valves), but you might start to struggle if you want masses of horse power per litre with such huge cylinders! A good point is that if you’ve ever sectioned a cylinder head you’ll see that the valves are NOT bore shrouded and there’s room to go to at least 38 mm valves each I’d say. I’d aim for as straight a run as I could when considering the runner and port section as a unit. Unforunately the ports aren’t the straightest and steepest but a lot can be achieved by a bit of cleaning up and porting work! Remember the 944S achieved quite respectiable BHP/litre using similar heads!
In terms of exhausting I aimed for minimum interactions- grouping up cylinders that were 270 firing degrees apart (360 isn’t possible unless you cross over under the sump)-this was a 4 into 2 system. I kept the secondaries very long spanning benneath the car to get the low speed torque up.

The 928 has a really robust bottom end, with a forged and heat treated crank- alhough you’ve probably gone for a different crankshaft. Beware if you perpetually use high revs you’ll need oiling modifications on two of the cylinders bearings ( I don’t recall which right now). The early S4s had oil cooling jets for the pistons, but this was deleted later- this is worth having on high output applications along with the sump from the GTS with it’s better windage tray for track use.

In retrospect, I would have liked to have gone for an even bigger capacity. The engine has bore centres of 122 mm which means you can grow the bore out a lot if you get custom made pistons. ( I went for the 104.5 mm overbore 968 units).

A well optimised modifed 928 engine will easily embarrass the worthy and good but seriously over hyped “pushrod fraternity” at a suprinsigley good price by comparison even with a smaller capacity.

 

[Louie928]

Marquis,
Thanks for your informative reply. I also thought the space of runner diameter/2 might work, but had no experience in the matter to know for sure. One concern I had was that with a gap that tight the air would have to accelerate up to speed within the fairly short distance of the radius of the bell mouth lip. In my case, that's the runner diameter/4 (runner dia. = 50mm , opening radius = 12.5mm). Some flow bench experimentation did show that a gap of about 18mm had no flow difference as compared to no top on the plenum. I wasn't sure if that test was valid since the steady state runner velocity was not as great as the peak velocity would be on the engine. I did note that the plenum top has to be very stiff since the high velocity air moving past the lip of the bell mouth creates quite low pressure in that region. I also thought a conical projection (with concave sides) fastened to the plenum top and projecting down into the bell mouth opening might help the air turn into the runner and prevent a vortex from forming. Looks like there will be a lot of dyno testing before this is over.

Regarding the 928...
This engine will go into my '90GT which suffered the 2/6 rod bearing failure you mentioned. Too many high G, high rpm, left turns. I persued the reason for the failure and discovered some interesting information through video taping the oil flow pattern inside the engine while on the track. The rod bearing failures are not wholly due to the crankshaft oil passage drilling deficiency. I had the GTS sump baffle and it is not effective. I have a better one now as well as oil scrapers.


The "new" 928 engine has 968 (104mm) pistons, 968 (39mm) intake valves with some porting and a different (non-968) valve seat for better flow at partial valve opening. The crank has a stroke of 95.25 mm and has improved oil passage routing. I'll use a cam set with more lift and duration than the GT cams. This intake system development came about with the experience of others with this type engine. Their dyno charts showed that the S4 intake would not permit the engine to develop sufficient torque at the higher RPMs. After the torque peak, the torque fell off at an alarming rate. Also, the low & mid torque is enough to break the transmissions without much provocation. The goal is to move the torque peak up in the RPMs and at the same time limit the low & mid range torque to a safe level that the transmission can handle. I have the throttle bodies canted inward at the top to provide a straighter path into the port and intake valve. Each bank is angled at 16 degrees from vertical.

There's more, but I don't want to stray too far afield here.

Thanks again,

Louie