My friend and I have been on a quest to see how much power we can squeeze out of his car. He made 248hp/253ftlbtq @ 7300rpm with 5lbs of boost on a dyno the other week. He has now bumped the boost up to 12lbs and we can hear knock.

In the cells it's at 32*. I know there is a general rule to follow but everything i have been able to find is general theory and each document changes it's formula for calculating spark advance per 1 pound of boost. Plus since the engine is full bridgeport the base timing is around 15* already.

I have used the knock sensor with the ms unit and has proved to be useless unless you dial it in good, i have yet to get it right myself.

Can someone here help me understand how to setup timing under boost for a rotary? Thank you, Russ.

 

my recommendation would be to check out the spark advance maps that BDC posted up here a while back... i think there was something in there for significant porting. Also, what split are you running?

 

Spark RETARD under boost! Retard 2 degrees per PSI. If you are at 32 degrees under boost, no wonder you are knocking. At 12 PSI, depending on the turbo of course, you should not be running more then 18 degrees of timing. 15 is a hell of a lot safer. I rarely go above 15 degrees above 12 PSI.

 

2* retard per 1psi increase in manifold pressure is a bit heavy assuming it's linear across the board but I do agree with the latter advance figures. You can do about 17*BTDC at 12psi of boost, taper down to 13-14*BTDC by 15psi of boost. Of course this is also contingent upon the fuel being used.

B

 

The split is for the most part at 15* accept under 10lbs and above, i run it at around 20* split.

So less main timing is safer?

 

Sounds about right. Split should decrease as RPMs rise and manifold pressure increaes, then increase again under boost.

Yes, you MUST run less "main" timing advance under boost! This is fundamental. Keep it conservative under boost.

 

We used these guidelines http://www.turborx7.com/tuning.htm
and got worse results. For now we just lowered boost back down to 5lbs.

 

Is it possible the crank angle sensor isn't zero'd out properly to the MS?

B

 

i have a general question for you guys:
when you say to "run x* advance at y psi boost" at what rpm are you refering to?
i do know that you need less timing with higher charge density (aka higher boost/closer to peak torque) and i know that one needs more advance as the rps go up (faster receading metal surface=start ignition sooner to match peak pressure with peak mechanical advantage of rotor onto e-shaft)...but again, at what rpm are you guys refering to?
Thanks,
Max

 

Therein lies the question of tuning.

Generally I will set timing advane under boost as if I was thinking of manifold pressure only. That is, after about 2500 RPM all areas of the map that say 10 PSI (as an example) will show 16 degrees. Under 2500 RPM I pull a few degrees out. Low RPM higher boost is the danger zone where you will blow engines with too much timing and not enough fuel.

Generally speaking, as RPM increases timing also increases. You will run most of your timing at peak torque, then taper off. Timing increases to atmospheric then starts dropping under boost. High vacuum conditions should get a lot of timing to make the engine feel crisper and combat popping. Timing under cruise should be rather high (30+ degrees) as long as EGT doesn't syrocket for the most fuel economy.

 

ok, but i was under the impression that advance should go lower closer to peak torque.
basically, if you just consider the NA case, "peak torque" is where the filling efficiency is highest. therefore you have the highest charge densities at peak torque, lets say it's at 5500 rpm. so at WOT your advance will gradually increase all the way from idle to redline, except that it is reverese bias to the torque curve. by that i mean that the greater the torque (the greater the VE), the less timing you want to give it becasue here is where more molecules of air and fuel will be in the charge, and therefore you will get a faster flame front (so you need less advance). this is pretty much the same as with forced induction, except here your effective "natural" VE curve (or torque curve) is getting multiplied by the elevated pressure ratio you are running. disregarding such thigs as compressors heating the air, a pressure ratio of 2 (14.7psi boost) should double the number of air molecules in a given induction stroke (in reality it will be less than double becaue the air will be hotter than it was NA). anyway, assuming no heat addition working to lower the density, this will effectively double the entire torque curve, except for the part in which the turbo has not met it's boost threshold yet (low rpms). so in this way, the same rules apply as stated above in the NA case. except now, there is so much more molecules per charge, the timing must be further retarded at peak VE.

the point of all this is, weather you are NA or turbo, the point at which your advance is at a minimum (disregarding the rpm advance) should be at WOT at peak torque (say 5500rpm), right? at WOT at higher rpms than peak torque it should start to be a little more advance, say 1 degree more per 1500rpm or so?

so maybe something like this:

......................2000......4000.....5500....6 000.....8000
150kPa(a).......2*..........16*.......15*.......17 *.......19*
100kPa(a).......8*..........19*.......17*.......22 *.......30*
50kPa(a)........10*.........20*.......22*.......30 *.......45*

(i hope that text format doesn't look wierd)

 

I am having hell setting my timing too, I've got a motor that I sear NOONE else has... and i don't understand it either.

 

swear^

 

What is your air system set-up...i'm mainly asking about your:
port config (do you have aux ports?)
boost level
boost threshold (rpm you hit your boost level target)
do you have VDI?

of course you could answer all of thoes questions in one shot if you tell me what series motor you have and if it's turbo or not (unless you are doing something wierd like mixing and matching rotors for different compression ratios/side housings for different port configs/forced induction on an NA block)

 

In general you're right about timing curves... you have less advance right at peak torque, but you're forgetting one important thing. At low RPM, the air and fuel is in the chamber for much longer and has more of an opportunity to absorb heat from the walls, bringing it closer to its auto-ignition point. If you fire the spark too early, the flame-front will heat up the charge that hasnt burned yet and will cause it to auto-ignite, resulting in a very early CA50 and probably knock and a loss of power. At higher RPMs this is less of a problem because the fuel has a finite resistance to autoignition, and these events just happen too quickly for the fuel to autoignite. The fuel's resistance to autoignition is the octane rating. Therefore at low RPMs and high pressures (atmospheric and above), you want to run minimum spark advance.

 

topless:
ok that's a very good point, and that makes snece to me. so you are saying that not only should one factor in the slower rotor speed at low rpm when giving it less advnace, but factors like oil and coolant temp come into play more heavily here too.

for a while i was trying to program the ECU we are building to use a single equation to calculate spark advance, instead of using tables. i'm not too sure this will be possible, but the variables in the equation would consist of:
rpm, charge air density, AFR, fuel octane #, meth%, water% (for obvious reasons)
air temp (colder intake air, more advance)
oil temp, coolant temp: (the hotter these are, the less timing - but as you said, they are more critical at low rpm, so there will be a weighting function for these baased on rpm. *these are used to determine rotor temp and rotor hsg temp)

i have been scouring text books, but i am unable to find a good equation that describes the flame front velocity for "near-homogenious turbulent flamelets-in-eddies" flow regime (same regime for all non-direct-inject otto cycles), but have had no luck. i think i am going to be stuck with a basic 3D ignition map...

i think i'll just tune advance with transducers in the trailing plugs and shoot for MBT on a dyno.

 

We played with the timing using the megaquirts knock sensor (witch we got working right) and we found the sweet spot with every rpm and boost level. Tuning timing is so touchy that 1 degree makes a huge difference. we found that for his cat 20 degrees was a safe starting point and worked from there. We tried 15 degrees and 30 degrees but they were both one way to extreme.

 

So you got the ms knock function working right? Are you using the knocksense kit?
Hope you dont mind me asking?

 

I would do two things:

1) double check total hardware latency: to do this set a fixed timing that still lets you rev from idle to 3000-4000 rpm. Then while watching timing with a light, rev the engine slowly to 3000-4000. If the timing retards while you rev, increase the hardware latency until it doesn't.
2) reset to using the table for timing, and retune the timing. As revs climb you will have to back off timing, but at least the timing in the table will be correct.

Ken

 

Also, although it would be possible to set up an equation to calculate what the advance should be and have it work reasonably well, that wouldnt take into account manifold dynamics at specific speeds/loads and the extra or lack of air mass that would not be accounted for with simple volumetric efficiency calculations, not to mention that the spark advance req'ed for MBT generally does not follow linear trends, will change with differences in AFR that will be counter to the affect that the VE table has on the AFR.