Cam Design (Or lack thereof) for the MZR DISI

Thanks for plugging away at those number for me.

I plan to run +2mm exhaust valves with port work to accommodate them to get more exhaust flow.
My split duration of 8deg is less aggressive than stock (16deg) to be more balanced flow of exhaust vs intake (your grind has no split with equal intake and exhaust durations of 224/224).
My logic behind more exhaust duration and flow is to spread out the time it takes to get the exhaust through the turbine housing instead of trying to force it through in a quick bursts. Now I know exhaust velocity helps spool the turbo and lengthening the exhaust duration might make it more mellow, but on the flip side opening up the exhaust earlier (hopefully not too early) might be letting out some residual combustion similarly to anit-lag, but much less aggressive, so maybe spool wouldn't be infringed? My other logic is the intake side has help getting air in, so increasing the exhaust flow will help with getting all that extra air back out.

 

I'm sure you're aware of the reasoning behind size differences in intake and exhaust valving, and that past a certain point there's severe diminishing returns in cylinder pressure vs crank angle. Thus, most HP gains from hotside work are going to come from cam duration (giving more time to evacuate exhaust gasses).

When comparing to my grind, keep in mind my engine is reduced in displacement and optimized for higher RPM running. At high RPMs, there's a lot less time to fill and drain the cylinder. Below is what my curve looks like compared to the others, factoring in my engine setup:



The most important part of engine design is matching everything so there are no conflicts in performance; headwork, valve sizing, cams, displacement, turbo size and aspect ratios, and even transmission gearing and tire size all are major factors. That's what I'm attempting to do with my build; match everything.

Sure, changing valve sizes might net a benefit, but it's also likely to cause valve shrouding (the cylinder head and wall impeding airflow past the valve due to proximity). There are a lot of variables, and it takes a lot of time to figure out what works.

 

I didn't figure increasing the exhaust valve size would cause a shrouding issue when it would still be 3mm smaller than the intake valve.

 

I can run it in the sim if I get some correct valve sizes. So far what I've found is 35mm intake and 30mm exhaust, is this correct? I've got 28mm exhaust in the sim for whatever reason I can't recall.

Edit:
Redid the numbers assuming 35/30 and compared to +1 exhaust valve size.

30mm:
399 @ 7000
337 @ 5500

31mm:
412 @ 7500
339 @ 5500

So +13 peak CHP and a 500 RPM shift of powerband (with your grind).

 

35/30 is correct for stock.
I was wanting to do a full rebuild on my head anyways, so I am strongly considering bumping the exhaust valves to 32mm. The additional cost of the valves themselves (32 vs 30) isn't bad at about $80 more for the set, but the machine work required to make them fit in the head plus clear the pistons is where the real cost falls.
But, I had already planned on getting bowl work done on both intake and exhaust, so maybe the additional cost to redo the exhaust valve seats while they're in there wouldn't be too astronomical by comparison. I'll just have the exhaust port cleaned up, not really opened up.
And by looking at the EVC and IVO events of my grind vs other aftermarket cams that work with currently available pistons, I was thinking maybe only the diameter of the valve relieve would need to be enlarged and not so much the depth of the valve relief. Of course, I would still measure the clearance to be sure.

 

Is that really worth it for an estimated 13 crank horsepower (roughly 11.5 at the wheel)? Inconel valves are heavy as shit and require a lot more spring to control them, limiting revvability. This is why I left my valves stock and upgraded to the Supertech springs (which they suggest will just handle 9k rpm, but not too much more). If you're looking to spin to 8 or higher, that might be an issue.

 

Arent our stock valves inconel?

 

Indeed they are, but my point is bigger valves are heavier and even with upgraded springs controlling them at higher rpms is a challenge

 

I will inquire with supertech what their valves weigh, both intake (nitrided 35mm and 37.5mm) and exhaust (inconel 30mm and 32mm).
Perhaps we can compare to the stock valves.
If the smaller inconel valves weigh the same or less than the larger nitrided intake valves, then there won't be an issue considering the valve springs are the same for intake and exhaust and the intake cam is typically more aggressive than the exhaust cam.

But that doesn't really address the simulated small gains with a +1mm valve. Is seems a 3.33% increase in valve diameter (+1mm) yielded about a 3.26% gain in CHP (+13hp). So a 32mm valve is a 6.66% increase which napkin math says may result in a 6.52% increase in CHP (+26hp). BTW, a $600 set of Corksport cams only adds 20 something hp. So if we're looking at $/hp, increasing the valve size isn't all that much of a waste if the machine work can be done for a similar cost.
All of these comparisons/simulations have been done at 15psi. What is stock vs cam/valves at 30psi? Perhaps the gains will be more pronounced at higher boost levels? I've been running my motor at 28psi.

Just to be clear, if it wasn't obvious already, I don't know everything.
I'm hoping the conversation will bring one of us or all of us to a realization of what can and can't be done or what should and shouldn't be done. Worst case, I waste money and take one for the team. Not that there would be an adequate apples to apples test to see if the larger valves did anything at all since I would be upgrading cams at the same time.

 

I actually did the weight math a long time ago and that's why I originally chose .300 lift intake and .320 lift exhaust with the stock valves; unfortunately, that's not what I got in the regrind, but that doesn't matter much. I wanted more reliability at higher RPM than just raw airflow, as we can always turn the boost up. Also, I did the sims and from talking with racers (like my dad and his drag racing buddies) one thing was clear: lift is great but duration is king.

The thing about the CS cams power gain is we don't actually know if anything other than the cams were changed; I'd wager if they were tuned for, and the ECU wasn't pulling load (like it should with a drastic change in VE increasing airflow), then power gains should be much higher, all else the same. The sim seems to back me up on this; swapping from a 277 hp 283 ft-lbs dyno (so, stock, or as close to it as I can reasonably approximate) cam and HP setup to the CS cams nets 366 hp and 350 ft-lbs...A fairly hefty increase. The stock cams appear (to my lightly educated eye, anyways) to be centered around safe fueling and emissions for the most part, and to provide enough torque to make the car feel peppy. The overly small turbo helps with this. Meanwhile, going +2mm on the valves (both intake and exhaust) only nets 22 hp and 16 ft-lbs, while the headwork (using stock valve sizes and stock cams) nets 27 hp and 18 torque (and acts as a VE multiplier when cams are added later).

Grant it, my valve sizes for this testing may be off but you get the idea; valve size helps, but having your parts match your design goals for the powerband is far, far more important.

On the subject of more boost, keep in mind the sim software isn't that great for this as it may or may not account for stuff like backpressure and reversion; for reference, though, it says my build (with headwork, custom cams, etc) goes from 530 hp @ 9k and 386 tq @ 5500 with 15 PSI to 608 hp @ 6500-7k and 561 hp @ 5500 with 30 PSI; about 200 HP per atmosphere (that's with stock valve sizes).

Basically, I think the way CS did their test might be flawed; hard to tell because there's no info (that I saw with a quick search) on what they did in the tune to test this.

As for not knowing anything: Don't worry, I don't either; I will wager that I've spent more time researching this kind of stuff than you have, though, but other than that I don't think there's as much knowledge gap between us as you might think.

If anyone's taking one for the team here, it's probably me; I've already written my engine off as a loss, and I don't even have a mile on it yet. At any rate, several people already have CS cams and @VashEXE (I believe) even has the same headwork as me. He's currently turbo/flow limited and is likely fighting both reversion from backpressure due to FullRace exhaust setup as well as running a top mount intercooler. Hasn't stopped him from making over 500 though (roughly 580-600 crank HP, also with stock valve sizes, if memory serves).

Suffice it to say that unless you're going for every last bit of efficiency/power you can get, shelf cams and bowl work will get you 90% of what you need to get to your goals (so long as they are reasonable).

If you want to move forward, I'm not stopping you or even trying to convince you otherwise; just saying that bigger isn't always better and changing some parts can yield greater rewards for less money than others.

 

What is making you consider your motor a loss already?

 

Multiple issues with the (brand new) block, issue with the head build, etc.

 

So not the concept just the execution?

 

No I've got the research down pat. I also have absolute shit luck; check the DISI-MZResponse thread for more info.

 

I'm leaning towards the CS cam dyno numbers were based on a boost tune, not load tune.
Because you'll notice in the attached shot of SP63 tests show similar "weak" gains with all the cams tested and they show boost being the same for all pulls. It seems quite strange to add that much duration and show such a relatively small gain.

 

TLDR:
Attack of the $800 regrinds...LMAO

Aiight I'll be up front: I don't like SP63 so take the following with lots of salt:

These are just stock cams with more lift and a slightly adjusted center, which means the buildup on them is going to be off center...Take that for what you will.
On a stock engine they are simming to 364 HP @ 5500 and 352 TQ @ 5k; nice narrow power curve.
With modified heads (same as the other tests), they up to 385 HP @ 5500 and 368 TQ @ 5k.

Here's what they look like; first, the grind:


Then, the curve (no headwork):


Oh and here's your cam grind in case you forgot:


I find it amusing that if I didn't know which grind was which in a blind test, I'd still bet on the car running yours to win based on this curve alone.


In my opinion, the only thing stage 2 about these SP63 cams is the price; that said, I expect some nuthugging hatemail in the morning saying I don't know what I'm talking about, etc.

Again, lmao


EDIT:
Oh and to add to your actual question about why there's so little gains on a boost tune with no other changes, well, if there are any load limits kicking in it could be pulling timing, etc. There's actually any number of things to consider for that.

As a quick side note, here's what my cam grind looks like on the same engine setup (stock, cams only and no headwork):


Note that I'm only running .353 (less than stock) lift intake and .321 (basically stock) lift exhaust...While also making 17 more HP and 14 more TQ while also spreading the powerband by an additional 1000 RPM.

In a nutshell, lift is important, duration is better, but all the numbers associated with the cams *DO* actually matter; just throwing a shit load of lift at it and getting more duration as a result (because it does work that way for regrinds) is a laughable way to address cams in general.

Oh and my cams were $200 shipped. Lol x2

 

Comparing our grinds, it seems you are running a lot more overlap than i am (22deg more in fact). Although you did say your grind has some VVT built in...as evidence by your cam being 10 degrees more advanced than mine (or is mine retarded...no pun intended). I was trying to keep overlap to a minimum at 0 VVT advance for high rpm WOT. But with that said, the sim of my grind with VVT makes more power across the entire curve vs no VVT. Seems the opposite should be true for top end power. I expected more low-mid range with VVT advance but also expected it to fall off hard on the top end.

My uneducated guess:
-Short version, our stock intake ports can only flow so much, so holding the intake valve open too long ABDC may not yield much additional cylinder filling which would mean the last degrees of intake duration may be useless.
-The long version, I've read that most of the cylinder filling happens between 45 and 135 degrees ATDC, which makes sense as that's when the cylinder volume is increasing the fastest. It would seem a lobe centerline closer to 90 degrees would be "ideal" for maximum cylinder filling, but that obviously can't happen with high duration cams because that would put the IVO event deep into the end of the exhaust stroke. Now, I'm thinking the reason intake lobe centerline is often larger than 90 degrees is due to continued cylinder filling after BDC due to the inertia of air in the intake ports (which is why some NA motors can exceed 100% VE). Larger ports will create higher inertia at higher RPM and the IVC event can happen later. This skews the "ideal" lobe centerline higher than 90 degrees. Add boost to the mix and lobe centerlines can be pushed even further. BUT, with the stock ports being so small, there isn't a lot of inertia there and even with mild boost, there probably isn't much flow either. WIth my lobe centerline being 124 degrees and IVC event happening at 56 deg ABDC, I'm likely pushing it too far...with stock ports/stock valves and mild boost. It shouldn't be as bad with higher boost. Your grind IVC is 46 deg ABDC, very similar to the stock cams (42 deg ABDC).

Additional data:
-According to the CS Cams white paper, the stock intake port flow starts to flatten out around .300-.350 lift which is about 80-90 degrees after TDC for my grind @ no VVT advance and remains in the "max port flow" zone until 10deg BBDC. Looking at it that way, it seems silly to still be in the max flow zone when the cylinder is no longer expanding and only a few degrees from starting the compression stroke.

-Lift aside, our exhaust grinds are pretty similar. I open up the exhaust 4 degrees sooner and close it 12 degrees sooner...yielding the 8 degree shorter duration in my grind.

-Comparing your grind to Piper Stage 2 (again, lift aside), your IVO and EVO are arguable identical. Piper holds the intake open way longer but closes the exhaust a few degrees sooner.

 

Actually I never posted my grind, nor will I until I confirm they work well. My overlap is minimal, but not zero.

Your grind is making more with VVT because it's in a weird state where the benefit of increased dynamic compression is outweighing the detriment of overlap. You're dead on about the intakes closing after BDC; there's a bit of resonance and that's based on runner length; you can see how important this effect can be by reading up on the 787b's variable length intake runners (not piston engine, but still).

I'm not sure which grind profile you think is mine, but the one in the last post was for SP63 and it's garbage. Even then, I'll remind you my cams are designed for a 2.0 running 9k RPM, but they look like they'd work OK on a 2.3 as well.

 

That's my bad. You posted a grind that was "close to your current grind but with a lot more lift". So I shouldn't have been referencing values as "your grind" when they aren't.

 

Correct; that other grind is close, but it's still tailored to a 2.3 and ~7k rpm IIRC...All the numbers are different on my actual grind though.