308 V12 conversion begins

I let the software go on the cam timing to flatten the torque curve but this time with tighter constraints then I tried yesterday and I used a wider rpm range for the optimization. I lost some hp for yesterday but man did it flatten the torque curve...for an engine with ITBs at least.

The cam timing is a little odd so this mgith be pointing to what Steve was talking about. It was the LSA angle at 108.3 which is a good number but then the intake is 101.5 and the exhasut is 115. Kind of odd so it might mean nothing....and I'd like to get the hp back up over 800.
Image Unavailable, Please Login

 

It's turning into a give and take game now. It gave me 804 hp but took about 15 ft-lbs of torque down low.
Image Unavailable, Please Login

 

A few more weeks I'm afraid. I have an EFI conversion to get done. I can play computer while I watch the little guy so thats the only reason there are even graphs to look at.

Software guy says he found the problem and a new release to fix the bug will be out tomorrow.

 

That's a lot of timing looking at a computer screen! ...this has been going on too long now.

 

i'm going to play the optimization game again tonight with a plenum type inke manifold and see if there is a net gain. To get a decent looking torque curve with ITBs I need pretty short duration cams which kill the top end power. A plenum intake might let me run enough more duration to get more power up top without killing the bottom, we'll see.

Steve is going to give me some actual real working cam profiles to try as well instead of the simulated profiles I'm messing with, that will be a big help.

 

I personally like thinking and talking about the design part of the project better than the actual doing part. The stuff just takes so darn long to make that I get bored doing it so doesn't really bother me not actually being in the shop working on it….but there is a lot more satisfaction in seeing a finished part than in seeing a finished design.

The cam/header/intake/head flow combination design has to be done and is probably the most important part of the project. This is what determines how the engine is going to feel to drive, what quirks it's going to have, how much power it's going to make, how long it's going to live, ect. This is going to take a several weeks of work iterating and refining to get it to where I understand what is possible, where compromises need to be made and I'm comfortable with the plan.

Up to now it's been a concept really, so I didn't see much point in spending a lot of time on details. But now most all of the major technical hurdles have been cleared and I'm pretty sure I will get thing working. It's time to sort out the details and I think it's clear in the graphs I posted that a lot more goes into making a good engine then just the hp and torque peaks. 2 engines can make exactly the same peak hp and peak torque at exactly the same rpm and be nothing alike with one being evil and ill tempered while the other would be as acceptable for Sunday drives with grandma as it would at the drag strip. Time spent matching, optimizing, and compromising will make all the difference in the final project result.

One thing I learned the meaning of the hard way is the shape of the torque curve. When torque is increasing with rpm it causes an unstable situation with the car. Lefts say you are rolling though the corner at the limit of traction and start your acceleration so rpm is increasing and so it the torque, but you were at the limit of traction so the increasing torque mean you need be rolling out of the throttle as you accelerate to prevent the tires from breaking loose. On the track the classic answer is you set the shift points so that you never go below the torque peak and don't have to think about it. On the street you end up using the whole rpm band and often don't know where the limit of traction is, particularly in wet conditions. A spike on the torque curve can put you off the road in a big hurry.

My lesson came on a bone stock 1990 Kawasaki ZX6 that had about a 5% dip in the torque curve between 9000 and 9500 and rapidly recovered from 9500-9800….very nearly causing me to high side the bike the first time I came into a corner in the wrong gear with the tach at 9k instead of 10k. A flat torque curve is best, but a curve with relatively smooth changes is nearly as good. A curve that looks like a mountain range is no good on the street and I'm happy to give away hp to avoid it.

A lot of the work over the next couple weeks is going to focus on the shape of the curve more than the specific numbers. On the last graph I posted, the green line is a pretty good-looking curve, nice and smooth but less low end then I would like to see. The red line has a good solid bottom end, but the dip in the 4k-6k range is a little sketchy. The blue line flattened out most of the dip while keeping most of the bottom end power and also closed about 1/2 the peak power gap between the red and green lines. I like the blue line the best I think, but I might try to move it a touch closer to the green line shape.

The torque peak is coming a bit later than I'd like which means the intake ports are too big….maybe a slightly smaller port size and more cam duration to make up for the reduced flow would make a better package.

It's all give and take.

 

About 1/2 the people here call me Mike, no worries. I thought it was a cool car name type car forum name...but mostly it's just mike

 

So the software has 3 modeling methods. The filling and emptying is the same as in my old software. The wave model is the most coplicated and then there is a hybrid that is supposed to use the strengths of both. You can see thay all produce differenct numbers....and I have no idea which is the best for what applications. Clearly the hybrid method takes most of it's information from the F&E model whatever that means. My plan is to make sure I'm happy with all the results and then live with whatever I get.
Image Unavailable, Please Login

 

I basically agree with you, but this is a sports car and it should actually pull to redline. My plan is not to go after a peak number, but the flattest torque curve I can get and still get the hp peak right around 8500. Small stock ports and cams would do better down low but that is not what I'm after here.....and this engine will make a lot of power anyway. Part of the 24 injector plan has to do with understanding just how hard it will be to control idle on what is basically a 9k race engine. I'll get it tamed without haveing to give up on the nice round 800 number.

 

It looks like the premade engines loaded in the software have much better model to model agreement then I do.......more investigation is required.

 

You surmise exactly correctly. In the wave model, the wave forms are graphed. This lets you see the intake and exhaust coming into or going out of tune and the hp/tq graph shows the effect of it. The F&E model basically assumes all the components are optimized, so it's a "best case" type analysis that tells you what's possible.

Steve said the low end numbers never happen and that seems right with ther amount the wave model exceeds the F&E model. The top end number in the F&E model always seemed high to me in my old software which got me inthe habit of telling lies to the software to pull the top number down a bit.

The hybrid analysis in the software looks to be about 100% the F&E model down low and about 80% the F&E model up top. The effect is very similar to what I was doing manually in the past to knock the top down. I think I'm going to proceed basically the way I have been which is to tune in the wave model then believe the hybrid model.

 

You're thinking right, the problem always is in collecting all the information needed fro existing engines. That is why steve mentioned he was happy with his 10 year old software version and my buddy who does heads is the same way. They have 100s or dyno runs and simulations to compare. I don't. The models I looked at in the software aren't a lot of help because they clearly were not real engines.

There is a TR in the library so I thought GREAT!.....but it’s got the wrong bore and stroke, the head is a generic selection from the menu with much bigger valve and way more flow than even I have so impossible then its got 14 inch header pipes and a few other bizarre things……it’s make believe and no help at all….and makes a lot less hp than mine on any of the 3 models anyway.

After I get something I like I’ll have my buddy run it in his software and if the 2 answers are both acceptable I’ll build it. That is really all I can do.

 

Paul,
I just measured. The 308 CV is 108mm OD which I think is the 934 part so I think I'm good to go. I've never heard of one breaking and there are a cfew hot turbo cars out there making the same torque number I'm looking at, but without change in drop gear ratio I did that drops the torque to the trans by 24% vs stock.

 

More computer stuff. Steve Send me a few complete cam profiles vs the estimates in the software's cam liberary. The answers cam out very similar but now I now the cams I using can be made and are safe so it was a big help. I was thinging the sweet spot was 235-245 duration at .050" lift for the intake and about 5 degrees less on the exhaust. With the real cams the best overall performance was 240 intake, 235 exhaust.

Here's some graphs. The patch for the software isn't out yet so I still working with the 3 cylinder model. I was getting tied of moving everything to excel to multiply by 4 and make new graphs so I didn't bother this time. The real answer is times 4 what you see.

I ran everything a few different ways. The Wave model only runs with open headers as the exhaust, but the real car will clearly have a muffler.

The first graph is the wavwe vs FE model with open exhaust an dthe best I think the heads will flow when finished. I think it's safe to say this represents the highest and lowest the output might be.

The second graph is the highest and lowest flow I expact (the low number is what I have right now with unfinished seats) using the wave model and the 3rd is the samething in the FE model. This shows that the engine will perform well with all the specified components acroos the possible flow range without casuing any weird things to happen.

The 4th graph is what I think is the most probable output for the finished engine. The higher and current flowing head lay almost right on top of eachother which says I have enough air flow right now without having at it with the welder.....although with a little better port I could go to smaller intake valves and maybe pull the bottom up a bit more, but over all I'm happy.

The 5th grapg is what the most probable, moved to excel and fixed for 12 cylinder. I through the ignition timing on the graph too (reads on the RIGHT scale) because I was impressed how close is was to factory timing. A QV is 32 degrees, I think Russ found 33 on his 328 with it's little bigger bore and here the computer says 33.5 with my bigger yet bore.

Here's what I've got:
Bore - 86.5mm
Stroke - 78mm
CR - 10.5:1
Intake - ITBs with 10.4" intake track (valve to top of stack) with a 4.8 degree taper top to bottom.
Intake cam - NET is .435" lift with 240 duration at .050 lift, 277 degrees at .006" lift (SAE seat to seat)
Exhaust cam - NET is .395" lift with 235 duration at .050" linft, 272 degrees at .006" lift
Headers - Step design, 22" long primaries with 1.625ID at engine, 1.75" ID at collector, collector is 4" ID x 15.5" long
Image Unavailable, Please Login
Image Unavailable, Please Login
Image Unavailable, Please Login
Image Unavailable, Please Login
Image Unavailable, Please Login