Rob Edwards

Thanks for the list Tuomo- I have been perusing Speedtalk for a while but don't know the players, most of the conversation is wayyyy above my pay grade.

I was struck by the following quote in one of the longer threads on plenums (plenii?) and airboxes volume calculations and their design.




The past couple of weeks have been fascinating to me (it's easy to learn a lot when you start with no knowledge ) and I think that Greg's gotten some good insight from the datalogging and dyno runs. Can't wait to see the next iteration!

ptuomov

PorKen --

I prefer to think about it slightly differently. Air flow is proportional to displacement, rpm, and volumetric efficiency. My starting point is that a 5 liter engine will make the same power at 6400 rpm as a 6.4 liter engine at 5000 rpm, if the volumetric efficiency is constant. Volumetric efficiency isn't constant, but it's a useful starting point, especially if the intake valve area in the two engines is about the same.

In other words, one could take the power curves and shift the rpm axis of the 5.0 L engine by dividing it by the 6.4/5.0, then compare the curves. If they look different, then it's things like runner length etc. changing the volumetric efficiency.

ptuomov

Rob --

I don't think that simple spreadsheet formulas or simulators like EAP are going to get the Helmholtz resonator volumes exactly right. However, some notes: First, they probably do get one to ball park as a starting point. Second, once you have one experiment done on dyno, the formulas appear fairly accurate in predicting the peak rpm changes in response to plenum volume changes.

Third, the manifold we're talking about in this thread is a single plenum manifold with all eight cylinders being fed from the same plenum. The usual Helmholtz equations are probably not a productive way to think about this manifold, because the pulses cancel inside that manifold. Instead, the equations that are used for individual isolated runners seem like a better way to think about it.

In contrast, if there were two plenums and the pulses would be separated correctly (think S3 manifold, but just bent and tucked in the valley), then some Helmholtz equations might be useful. I think I posted some months ago the formulas that I used to try to figure out how much the lower torque peak rpm drops in the S4 manifold if a spacer is added.

(Digression: I think that those individual isolated runner equations are themselves also derived from the Helmholtz equations, it's just the resonating cavity is the cylinder itself and the volume is taken as an approximation the volume at half stroke. The plenum manifold Helmholtz formulas are in fact derived from a two-chamber resonator. I think this might be the original reference from 1953, it's Engelman's dissertation: http://minds.wisconsin.edu/bitstream.../Engel1953.pdf)

Rob Edwards

Tuomo- Just for purposes of discussion, if you wanted to reality test the equations on a plenum design that only feeds four cylinders (say, the Threshie CF intake), and for which there are lots of dyno runs and info on the engine configuration, what measurements would you need to have? I can't do math, but I can measure stuff.

PorKen

Is it a single plenum? To the extent that Greg has been moaning about development time, I would say the interior is more involved.

I would imagine that the box has two levels, with the same cylinders feeding a layer as they would a S3/4 plenum.

Packaging aside, it would be sweet if the two levels could have a flappy connection to restore the first torque peak.


FWIW, when I had my S4's intake off I thought that more common plenum space could be created by putting a spacer between the upper and lower parts of the S4 manifold. (Would need to find a smaller TPS.)

simos

Gentlemen, don't get me wrong, I'm not any specialist just tried to explain the reason for those lows and highs of torque curve.
As Greg pointed out, those lows and highs are probaply undetectably in driving, whereas the stock S4 intake effects are.

Playing with intake port CSA and port shape is different thing to tune flow and velocity. My heads are made and designed by this guy:
https://www.facebook.com/pages/anorr...38511279525280

I'll stop here

danglerb

Plenum, full of holes, weird or no Helmholtz.
Cylinder, less holes, Helmholtz effects maybe more likely.

The way I think about it is two fold; most of the flow is filling the cylinder as the piston drops, and what happens in the immediate time vicinity of BDC and intake valve closing. The first part is like HVAC ducts, so much pressure differential, length, cross section, gives X amount of cfm. The second part is more about velocity, mass of air, speed of sound and distances within the cylinder and within the intake system.

ptuomov

I listed some of the measurements that I think are relevant in the stock intake manifold thread.

For Threshie's CF intake, the optimal plenum volume is as large as possible. This is because the pulses are combined into two plenums with unequal spacing. The plenum resonance will never be right, so it needs to be minimized. The two ways to minimize is to add cross-over pipe(s) between the plenums and to make those plenums as large as possible.

For GB's custom intake manifold, the plenum volume is close to irrelevant. This is because all eight pulses are spaced so close to each other that the plenum pressure about averages out. All that the plenum needs to do is to be large enough to slow down the air enough to give equal cylinder filling with steady state flow.

For a dual-plane manifold, such as the stock S4 manifold with the flappy closed or the S3 manifold, the plenum volume matters. Increasing plenum volume will move the peaks to a lower rpm and reducing it will move them higher up. Once you know the volumes and the current peak, the formulas are pretty accurate on how much the peak moves for a given change in the plenum volume.

I am pretty sure that it is a single plenum manifold. The torque curve is such. If it were a dual plenum manifold, we'd see significant peaks, not one hill. It's also the case that the runner arrangement on the top, as displayed in the photos, works for a single plenum manifold, but would make the bottom part unnecessarily complicated for a dual plane manifold. For these reasons, I'll take an even odds bet that it's a single plenum manifold.

Just speculating, the fabrication difficulties are likely related to the entry angles of the runners to the plenum. The fabricator probably wanted to have the runners enter normal to the cover, and each runner has a slightly different entry angle. Otherwise, the stories about small aluminum pieces wouldn't make much sense.

An alternative construction method would be to measure the entry angles from this manifold, then cut a flat plate with holes that correctly project the hole shapes when a round runner intersects the flat plane at that angle. Then, one could just push the bellmouth pieces thru from below and weld them in place. This is what I would try if I were reconstructing this manifold.

Another difficulty I see in this manifold is that there's about 90-degree turn for the air flow near the throttle plate. For this reason, one should probably use a throttle plate on the bigger side and also large diameter pipes on both sides of the throttle plate. If I were reconstructing this manifold, I'd try to fabricate the piece between throttle and the plenum from bent pipe section. If it won't fit, then it's more involved project of welding together an angled box (like a piece of a periscope) to feed the plenum.

PorKen

I suppose you are right. Wishful thinking on my part. It probably doesn't have variable length trumpets, either.

The straight sections for prototyping hose connections did not help matters. I doubt they will be there in the final version.

dr bob

Just so we're clear on choked-flow parameters and performance: Decreasing the cylinder pressure drops the density, moving the sonic velocity higher. The choked flow doesn't increase on increases in upstream pressure, but is still sensitive to changes in downstream pressures. One 'challenge' in modeling dynamic flows like this is how any reversion will affect sonic velocity upstream of the choke point. In a straight-walled tube, the choke point moves with that reversion pressure 'wave'.

I often use a Slinky to illustrate the effect, with the downstream end oscillating in sync with flow changes as the valves open and close. The upstream end is linked to the other Slinkies sharing the plenum. How they are coupled depends on the relative positions of the other intake runners and the flow/firing order of those coupled runners.

The Good News, at least for me, is that my little process modeling package is used for no more than four sources with one destination, and is looking at only one on-off flow condition at a time. No wildcards like fuel spray expansion to convolute the calculations.

Back to our regularly-scheduled car stuff...

ptuomov

bob --

That's not how I've understood the physics to work.

http://en.wikipedia.org/wiki/Choked_flow

However, I don't know what I am talking about. I have no idea whether the intake port flow ever goes sonic or not. Maybe it's just turbulence starting to choke off the flow. I don't know. The point was that at some air speeds the frictions start growing fast with increases in air speed, and there's effectively a speed limit on how fast air can flow in the intake tract.

Vilhuer

How much $$$$ you spend on it? We have four heads waiting for Niklas and Teemu. Guy we planned on using isn't doing too well so we need to find another place for at least two heads going to Niklases engine. Valve seats are inserted for 39mm intakes but porting needs to be done.

simos

The price depends on what else needs to be done for the heads( valve work, seats etc.) But roughly less or little more than 200/cylinder is what Andre is charging. If you are interested, I can send you more details.

Better to send them asap. as those race quys usually are priority #1 customers and they starts their projects some time after Christmas.

Andre also welded some hooks to bottom of the port as an extra safty to keep epoxy in place next thenths of years. That was my extra requirement.

Tested heads at last summer with my partially flow enhaced and balanced intake and the difference was amazing what came to torque and power increase.
Didin't dyno as intake was so badly balanced, but was very happy since cylinder #5 was knocking together with #2 and #6. I did hell lot's of work to make #5 to get better statc flow

I have also contacts to other guy who's doing plasam welding and re-grounding to cams, the Viiala guy doesn't wnat to do it anymore

You need to give some details of your project and he always want's to measure the cams to make heads to work perfectly with them.

EDIT:
Here is series of fluid mechanics lectures from 70's, wery informative.
http://www.youtube.com/playlist?list=PL0EC6527BE871ABA3

Vilhuer

Everything else is already done except modifying ports. Valves and seats are done so it should be cheaper to finish the job. We aren't looking anything radical like you. I can get his email from web page so its not a problem. Good to know he already has 928 head experience.

We don't have cam problems right now since Niklas has Colin's cams and Teemu needs to keep hp under 360 due to racing club power to weight limits. Normal S4 cams with headers and X-over will do just fine so there is more low end torque. Teemu will use stock S4 heads instead of ported ones in racer but it makes sense to have his half done pair finished at same time.

Too bad he's finally retaired. I quess usual S3 cam to S4 head modification would be doable by Andre. Don't have any sets waiting for it at the moment but I'm sure there will be need in future.

Sending cams to him is not a problem but we probably have them modified first in Lahti. As you might remember we ordered them in S3 format and modification isn't done yet. Should be easy job for anyone with lathe. We are taking few cranks etc there for balancing next week anyway.

dr bob

Two glasses of wine after a very long workday. Not a good combination with the keyboard.

The ratio of upstream and downstream pressures are a good yardstick way to determine at what point the flow is 'choked'. As downstream pressure changes, the sonic velocity changes with the density of the gas. Once the flow is sonic/choked, changes in downstream pressure change the mass flow only as it affects the density of the gas. Changes in upstream pressure directly affect the volumetric flow. Discounting the change in temperature as the gas expands beyond the choke point, flow is almost perfectly related to changes in upstream pressure.