3500 vs 3900

[SappySE107]

Beware flow numbers you find on the internet. I have read some pretty crazy shit for iron heads, 3400s, etc. I am highly suspect of 3900 flow improvements based on my own research.

[NateD4]

Just my two cents.....

I'm working on a 3900 VVT build with the LZG (DOD and E85 version of the LZ9). I have complete stock cam specs if anyone wants them.

Recently I drove a friend's 2007 G6 GTP (or was it GXP) with the 3900 LZ9 engine. I honestly don't think the head flows and cam are a limiting factor on it. The engine revved past 6500 with no noticeable drop in torque. Based on the throttle body size of the LZG and the size of the ports I'm not sure much gain is to be had unless you bump the redline up to 7500 RPM.

Perhaps someone from WOT can give agree/disagree with me. But I think the 240 HP from the stock engine is close to its capacity. Headers and exhaust would probably help. The stock cam is already pretty good.

IF someone can get me some engine log data (I don't have a running 3900 ATM) from a car with one in it I can probably figure out whether my above claims are accurate.

As someone else mentioned the 3.6 DOHC in the Camaro/CTS does make 300 HP stock with I believe a 500 or 1000 higher red line. I also know there is a place in Australia that sells a custom intake for the 3.6L. With cams, headers and intake the 3.6L made 422 HP at nearly 9K RPM. (www.performancemanifolds.com.au)

With that said I think there might be some room for the 3900 to flow more and make more power. However to get there is not going to be easy. Even if you flow more I suspect that you'll need to crank the red line up to get more HP from it. I would feel pretty comfortable using the 3.6L as a baseline HP/Liter/RPM... so 300Hp/3.6L = 83.3 HP/liter divide by 7000 RPM and that's 0.0119 HP/(liter*RPM). Recompute that for the 3.9L and you'd have 0.0119 * (3.9*7500) = @348 HP. If you carry that to 9K RPM you'd have 417 HP. Based on that I'd say the mechanicals of the 3.9 are a large hold up to power than the flow or cams are. Sure cam and flow is going to add some power. But if you aren't careful you'll lose on the bottom end of the RPM band. The cleaner way to get there is to wind the heck out of a 3.9L which will probably take new internals (I'll be able to verify in a week or two one the dynamic modeling of the crank and valve train is done since I'm trying to hit 8500 RPM with out going to full race lifters).


If I were doing a swap I'd just go with a 3.6L DOHC engine. Buy the cams and manifold from the guy in Australia and run it to 9K RPM to make 422 HP. Otherwise turbo a 3.9 will get you the same result with a lot less hassle..

Just my 2 cents...

[NateD4]

Also forgot to mention the 3.6L in Australia is not direct injected (I think). So it may be possible to get close to 500 HP from one N/A with that manifold and cam combo.

As to the 3900.... if you are interested I'm doing a pretty serious build program on it. New pistons, H beam rods, custom cam, custom lifters, pushrods, possibly even custom rockers. I also am looking into a billet low inertia crank and possibly even a dry sump oil pan. Again if you are interested I maybe able to have some spares made....

[SappySE107]

The 3900 is held back by the drive by wire and entire computer program. 240 is sorry. 222 like my parents' 2008 G6 GTP Convertible is laughable. Its a heavy car and I was able to get it to work stupid amazing until the computer limits the throttle to 8% because the controller cannot do an absolute = input to output ratio. for example, 50% throttle is good for about 20% I think. Not looking at it, but its pathetic. 75% is maybe 35%. Its probably worse than that by a lot, but you get the point. 90% = 80%, 100 = 100. That is a wicked curve, and explains the horrible throttle response stock. The cam timing is also pretty restrictive, giving 9 degrees advance over 113 stock 3400 cam at start up. it isn't bad, but retuning did improve the throttle response and perceived acceleration.

[NateD4]

I'm not 100% sure I can believe that. While GM may limit the throttle curve there may be good reason to do so in the transient.

I venture to say there is some benefit from having the throttle open progressively compared to all at once. I wouldn't be surprised if GM found some benefit to metering the air with the RPM as opposed to dumping it.

Perhaps you are right and it simply limits the engine. However, given the other cars GM has I can't imagine the throttle body tuning is that big a hinderance. The 3.6L, the Northstar and LS engines all use similar throttle mapping strategies. I'd be will in to be GM simply gave the engine some manors and tuned it for drivability.

I will be looking at this once I get my engine on the dyno. I'm hoping to get some engine dyno time prior to installing on the car.

The cam timing issues I'm not sure what to make of. What kind of results did you get when you re-mapped the engine? Did you get the VVT to update with the timing improvement you suggested?

Why not re-map the throttle body response? Can't be too hard its only a stepper controller.

[NateD4]

Also the throttle settings you discussed. Were they at idle or on the dyno?

With modern controllers I imagine GM has a neutral limiting throttle position so the car doesn't over rev in neutral. Perhaps in gear and at speed you are able to see full throttle opening??

[Fastmax32168]

Beware flow numbers you find on the internet. I have read some pretty crazy shit for iron heads, 3400s, etc. I am highly suspect of 3900 flow improvements based on my own research.

Ben,
I believe all the flow numbers I used, both stock and ported are yours.
I have been playing with different camshaft combos and everything I try tells me that the 3900 really needs exhaust help. The intakes flow great but the exhausts are not even as good as 3500 exhausts and just cant keep up. Adding additional exhaust duration really cuts it loose in the simulations. I think I am going to steal a good LS grind and have you make it for me.
Roy

[Joseph Upson]

Ben,
I believe all the flow numbers I used, both stock and ported are yours.
I have been playing with different camshaft combos and everything I try tells me that the 3900 really needs exhaust help. The intakes flow great but the exhausts are not even as good as 3500 exhausts and just cant keep up. Adding additional exhaust duration really cuts it loose in the simulations. I think I am going to steal a good LS grind and have you make it for me.
Roy

I'd expect the VVT to compensate for some of what you're seeing with the exh flow because it will certainly change with the cam variation relative to the crank position. When I manually activate the camshaft in my engine there is an audible change in the exhaust note to a much smoother tone accompanied by a ~.5 drop in AFR.

After discovering that my cam specs were too conservative at 216/213 for the stroked high compression 3.9 I sent a cam recommendation request to Crower and the following is what they sent for a turbocharged street application:

@.050
230 IN
221 EX
Lift
.617 IN
.600 EX
Lobe Center 114

I used Sappy's numbers for the head flow although I accidently set the exh valve and seat diameter on the spec sheet to the same 1.52". I don't want lift above .525 though.

[SappySE107]

I have never posted ported flow numbers, or pics, or cc measurements.

The 3900 has the same exact exhaust port as a 3500, with the valve seat opened up for the larger valve, and the throat modified to blend it back to the stock 3500 size. If you compare flow with manifolds, its not that biased at all.

As for the throttle body, believe what you want. You can bring up other motors and your opinions to doubt me, but I have tuned a 3900 with and without the drive by wire. Yes, throttle maps and all. If you go too far with the DBW maps, it goes into safe mode. It may be a dumb stepper motor but it has physical limitations for how fast it can react. The software is setup to recognize this and in turn, cut you off from the fun and make 100% input = 8% output. The cam timing helped improve response and output, cutting spark timing to get rid of KR helped, and changing the trans settings helped a lot as well. Nothing compared to make that throttle table a 1:1 ratio across the board. That included laying into it. HUGE DIFFERENCE.

I can't believe you asked me if I tested this shit at idle.

[NateD4]

Can you post some times or dyno numbers without the DBW throttle body?

I have a hard time believing that GM put such a large TB on it and won't let it open fully.

What is the TB position at full pedal and 60 MPH?

I'll be able to confirm with a bench test of the ECM I have.

[NateD4]

Well it would make sense to have a different control strategy in neutral then while driving.

[SappySE107]

All street tuning. This isn't about WOT only testing. Its about trying to change lanes and not get hit by the truck in the next lane because the damn DBW delayed the throttle so the trans could shift. Its all integrated to be a warranty saving, detached feeling driving experience.

I don't need a dyno to tell me its an insanely huge difference. But I am wasting my time. Go do your thing, and share it with us.

[NateD4]

I never noticed that on the G6 GXP I drove. It was responsive and as snappy as a GM auto could be.

Did you alter the OS on the DBW microcontroller by chance?

[SappySE107]

I don't know how you could alter the OS on the DBW controller.

[NateD4]

From what I know it's not easy, but not impossible either.

What has me puzzled is that I never hear of GM cars with this complaint.... Making me wonder if something happens when you modify the settings that messes with the phase lag. Or is it a swap related issue?

The G6 I drove was Auto which means when you stomp on it it downshifts first. That causes a slight lag. But not something that has ever seemed out of the ordinary to me.

Do you by chance have an Oscilloscope?