Ever wondered what knock looks like...

[dkjones96]

on a dyno chart?



That's a vehicle running WAY too much spark advance or the wrong octane fuel. The choppiness comes from the ECU pulling timing and putting it back.

It's from an SRT4 forum so I'm assuming it was someone running a bad tune on regular fuel.

[DRW]

Are you talking about the constant up/down waviness of the HP/Tq line? Or are you talking about the two downward bumps in the Tq line at 3600 and 4200 rpm?

The constant waviness is due to 'smoothing=5' not being smooth enough. When a dyno infers engine rpm via wheel speed/gear ratio there's a little error involved, and that error shows up just like that wavy HP/Tq line. It all averages out. The dyno operator can choose a higher rate of smoothing to flatten that line.

[bobc455]

I don't know the source of your information, but by looking at the graph I would have guessed that the ignition system is insufficient for the task at hand. I guess if they were logging KR then it would correspond to those peaks & valleys? (I would re-tune his knock retard parameters if possible, if that was the case)

Actually this demonstrates another point about why you should not ever rely on the knock sensor as a band-aid for improper timing or insufficient octane. (If anyone wants me to explain, I will)

-BC

[Jetta90GL]

It looks to me like he needs a higher volume fuel pump, or turn down the boost past 4500 RPM. If he's knocking its from running lean.

[dkjones96]

Infers engine rpm? What dyno does that? All of the dynamometers I've worked around have a tach signal input.

[bobc455]

I suspect that he meant it is inferring overall gear ratio. Engine RPM would have to be measured. (unless this particular dyno works a totally different way by using gear ratio inputs and so on, in which case it could infer engine RPM based on wheel speed, but this would be very very surprising to me)

-Bob C.

[DRW]

Originally Posted by dkjones96

Infers engine rpm? What dyno does that? All of the dynamometers I've worked around have a tach signal input.

The shop I go to has a Mustang dyno, and it doesn't need a tach signal. Sorry I don't know exactly how it's set up, but the dyno is calibrated to the car. Essentially the car is driven on the dyno in 3rd gear at exactly 4k rpm and the dyno operator clicks on some box in the dyno setup program when the engine reaches 4k so it knows which wheelspeed equates to 4k rpm on that car.

My car has an odd misfire issue, so the tach signal is unreliable. We tried using the ignition pickup, but the output on the dyno screen was extremely erratic.

Personally I find it suprising that the ecu on that car would give back timing so quickly that it just knocks again. Most cars I've seen will pull timing at knock and give it back relatively slowly. My car's ecu is set from the factory to return 1/3 degree of timing every .6 seconds until knock is gone. If there's 5 degrees of Knock Retard it would take 9 seconds for the ecu to give back all the timing, assuming there wasn't any other knock events. I have to admit that the factory settings on my car are somewhat outdated, but the theory should still apply.

I still think the dyno sheet shows a lack of smoothing function in the dyno program.

[bobc455]

Originally Posted by DRW

Personally I find it suprising that the ecu on that car would give back timing so quickly that it just knocks again. Most cars I've seen will pull timing at knock and give it back relatively slowly. My car's ecu is set from the factory to return 1/3 degree of timing every .6 seconds until knock is gone. If there's 5 degrees of Knock Retard it would take 9 seconds for the ecu to give back all the timing, assuming there wasn't any other knock events.

That's exactly why I think he had an ignition problem (causing the tach reading to go funny) and not knock. There is no timescale on that chart, but I'd be amazed if any car had such a quick recovery from knock (unless someone deliberately tinkered with the KR settings, which would be dumb because it would be ignoring the thermal aspect of knock events).

-Bob C.

[dkjones96]

That was a Neon so I'm not sure but I've seen Hondas give timing back that fast. If it were a controlled acceleration run it could look like that, I think. I've seen similar from a vehicle running regular when the engine calls for premium but that happened through the entire rev range like shown below. We ran the engine with regular fuel and then premium.

[GasSavers_GasUser]

Originally Posted by dkjones96

That was a Neon so I'm not sure but I've seen Hondas give timing back that fast. If it were a controlled acceleration run it could look like that, I think. I've seen similar from a vehicle running regular when the engine calls for premium but that happened through the entire rev range like shown below. We ran the engine with regular fuel and then premium.

Looks like my last EKG...........what me worry....haha