The thread on CKP trigger wheels got me thinking about how much error it might get when the crank is accelerating or decelerating quickly.

Out of curiousity I examined the characteristics of driveline windup and ringing in my datalogs. Wondering how much acceleration it had. See below. I floored the throttle at ~2000 RPM to have a look.

The RPM trace shows the characteristic ringing which sounds and feels like a sort of "boioioing".

Here's what I found. The ringing is at 6 Hz, and the effect on RPM is around 240 Hz peak-to peak. IOW if you start at 2000 RPM and floor it, for the first few cycles it will peak at 2120 RPM, dip to 1880, repeat, until it dies out. The peak rate of change appears to be close to 1000 RPM/sec, on top of the 300 or so from accelerating the car (3rd gear, ~100 kPa).

I ran some math assuming sinusoidal frequency modulation of the RPM trace.
If you modulate RPM at 6 Hz, and the driveline lash is 3° is either direction, I get very close to 240 RPM peak to peak, and 700 RPM/sec peak acceleration.

FWIW decelerating in neutral with my 11 lb flywheel is around 1500 RPM/sec.

So the engine appears to rock 3° in each direction.
More later on the resulting error in crank angle...

Upper trace is RPM, lower is TPS.
Horizontal scale is 1 sec per division, I rolled the throttle in in about 0.5 sec.

 

It's cool to know that while I'm at home, alone, in the cold, pretending the cat understands me while I drink beer and play a video game with a steering-wheel clamped to a coffee table, there is a team of engineers on Miataturbo working around the clock to make my car faster, more reliable, and writing it at a 3rd grade level so I can interpret it. It would have been nice if you guys could have done all this four years ago so I don't have to waste my precious time doing things twice.

 

I told you you shoulda bought the Mazdaspeed motor mounts!

--Ian

 

Good fuing lord that sucks


anyways

How do you like the AEM? Obviously your logging resolution is MUCH higher than my hydra.

 

My DD has failed mounts, it's halarious on low rpms with the amount of rumble.

 

that wiggliness is partly related to throttle enrichments. when they are really bad, I get visibly larger (and larger feeling) wigglies.

 

Nah those wigglies are driveline windup. I don't have any tip in misfire there.

LOL@ hustler's post.

 

Best post of 2012.

 

Sampling period is 50 ms (20 Hz) if you use the laptop to log.
It's 200 Hz for the internal logger. Great for doing torque/power extraction runs, troubleshooting CKP and CMP sync'ing, and tip-in enrichment.

 

What makes this especially hilarious is that, as I am home sick today, I have been experimenting all morning with different circuit topologies for trigger input decoders. Picked up a handful of CD40106es earlier in the week, and I'm doing variants on the JasonC schmitt-trigger circuit at this very moment.

(I am finding, for some reason, that the hysteresis of the circuit is not meeting published specs. Not quite sure what's up with that yet...)

 

Jason, what's the effect with, say, a 4000RPM tip-in? I know that I avoid lugging like the plague. However, the engine rock would still be there at about the same frequency at high RPM -- and because that's a higher torque area, I'd expect the rock to have higher amplitude. To lazy to do the math myself.

Plus, I'm still using a CAS, so now I have to factor in the timing belt. Darn it, you're going to keep me up nights!

 

CD40106's and their ilk have loosey goosey input hysteresis specs. But they do work well.

 

What do you mean by trigger input decoders? Are you trying to design a hardware solution to decode different trigger wheel designs. i.e. 12-1, 12+1 ...

I am going to use JasonC's schmitt-trigger circuit to feed my MS3 running my vvt engine.

-Raj

 

You know, this has always been my test as to whether someone is legit or a poser. Every profession has it's lingo, and oftentimes you encounter people that use nothing but "lingo." Having worked in multiple professions, I've found that nearly every human concept is capable of clear expression at a 3rd grade level. However, only those that REALLY UNDERSTAND the concept can reduce it to that level.

For some reason, we have a lot of these people on MT.net. It's why I'm addicted to it.

Of course, there is also newb buggery.

 

It'd be interesting to see how much engine mounts and diff mounts increase the windup resonant frequency!

 

I've thought about this and my conclusion is, according the "principle of superposition", that the ringing frequency, 6 Hz, and the delta RPM, (240 RPM peak to peak), will be the same. Yes torque at 4000 RPM is greater, but the torque rate of rise would have to be faster than 1/12 sec in order to excite the 6 Hz resonant frequency - 1/2 of a period of 6 Hz is 1/12 second). It is much easier to do that at low RPM, because the motor will develop close to full torque at something like 5% throttle. Whereas, at higher RPM, it might take 50% throttle. As your foot rolls in the throttle, it will reach 5% throttle 10x quicker than 50%. So at low RPM you can hit full torque in less than 1/12 second. This is why the boioioing driveline windup is much more noticeable if you tip in at low RPM than at hi RPM.

As for timing errors, it will always be proportionally worse at low RPM. E.g. say you have 2° error doing this at 1000 RPM, it will be 0.5° at 4000 RPM. This is also why if you do the throttle-blip-in-neutral timing test where you set the ECU to do constant timing, you will see the most error initially at low RPM, and it diminishes as RPMs rise.

 

The 74C14 (same functionality) appears to have better specs for our purposes.

 

In fairness, I have a handful of the 40106s sitting here and I've only tested one gate of one chip so far. I'm going to go through the rest of them and see what kind of spread I find.

 

FYI:

 

Is the external OS the thing you bang into?