Continuing ignition timing debate from the 416 thread.
Your not telling me why. You just keep saying you can't. Only louder each time. lol. You see the difference right?
You don’t grasp the why? I guess I’ll try it this way one time and then I just don’t give a ****.
If you have ever read the SuperFlow dyno manual you’d know this. If you don’t have a copy, get one. Then read it because I’m not going into that kind of detail here.
Everyone with a water brake dyno regardless of what data acquisition you use should read it. There is valuable information that I’m not going into here.
So take this for what it’s worth because I doubt anything I say to you will change your thinking.
Superflow says it takes 250 cycles to affect a change. That means a change in A/F ratio, timing, whatever. I’ve talked to people who say that’s a bit generous.
I do most of my testing at 300ish rpm/second. I know guys who never test at less than 600 rpm/second. I believe those power numbers are more accurate than the slower speeds, you can’t get enough data in a pull that fast.
That’s because it takes at least (according to SuperFlow and two of my mentors, one of which as I said has developed more testing protocols and has done and analyzed more dyno data from every conceivable dyno out there) 250 cycles to affect a change.
I believe it to be more like 300 cycles. Even if your software is sampling (or whatever they call it now) at 100 times a second you are still missing too many data points.
As a side note when I first worked with data loggers they were 5 samples per second. That’s horrible and you’d make tuning changes that went sideways or backwards because you were always at least late with the data.
The next up was 20 times a second and for a low 9 second car it was acceptable, but for a faster car you were in the same boat.
I have no idea what the sampling rate is for Pro Stock or fuel burners but I’m betting it’s over 100 times a second. For a 6.5 second pass you’d get about 700 data points. That’s pretty slow IMO.
It’s the same on a dyno. Because it takes 250 cycles to affect a change you can not, I say again CAN NOT use a sweep test to determine a timing curve.
You are not getting enough data to do it.
That’s why on sweep tests you can get too much if timing at peak torque and not see it.
If you steady state test at say 3000 rpm and even if you just do it every 1k after that to whatever peak rpm is, you’ll see it wants a curve. Every time.
And when you get the curve right the curve is almost always right. By that I mean in the car it may take more timing so if it wants X timing at peak power on the dyno, in the car it may take 2-4 degrees or more timing at peak for the best time slip, because the engine isn’t loaded the same in every gear, nor is it loaded like it is on the dyno.
But the shape of the curve will be the same.
I have a 395 inch small block I built that made 505 on my dyno, with OE iron heads and 2.02 valves. It’s a measured 11.xx:1 compression.
To get it happy I did a steady state test from 2500, then every 500 rpm after that to peak torque and after peak torque every 1k rpm.
The curve is 25 initial at a 1k idle at peak torque it’s at 30 and that’s at 5k. It tops out at 34 at 6500.
That is almost exactly 1.5-2.0 degrees per 1000 rpm. Bill Jenkins published that in 1975 and as crazy as it may seem, it’s just about spot on with all my testing.
I forget exactly how much power locked out timing cost this engine but it was at least 20/20. And it picked up from the bottom of the pull (2500) to peak rpm (6900).
It is grossly under valved and it’s severely under area for ports. Yet it made over 500 hp on bowl ported iron heads most guys would throw away.
So the upshot is you can’t sweep test for timing because it takes too many cycles to cause a change.
You go so fast that the engine always wants more timing because you don’t have enough time to let the engine tell you what it wants.
So you steady state the engine and move the timing around and watch the numbers. The engine will tell you exactly what it wants for timing and it’s a curve and it’s usually less total than locked timing.