×

Sale ends todayGet 30% off any course (excluding packages)

Ends in --- --- ---

# Calculated vs observed Knock Frequency ..sanity check

### Tech Articles

Discussion and questions related to the course Practical Standalone Tuning

Hi Andre,

tuning a 3L 4 cyl Porsche 944 turbo - bore is 104 mm using Haltech NSP on an Elite 1500. The calculated theoretical knock frequency is 5.5 Khz. When i do an intentional knock at 2000 RPM 60% throttle with 30+ degrees of additional advance over base map. I can clearly hear the "ping" of knock in the Knock ears and my interpretation of the spectogram is that knock frequency is about 7K hz. I understand that theoretical is subject to a lot of "it depends" but is this a reasonable variation from theoretical?

Peter

Peter,

Have you tried 11KHz, the second harmonic based on bore size?

There's certainly room to use data gathered to make alternate decisions, but I'd try HPAs usual suggestion and see what you find.

Mike,

I don't remember seeing any spectogram activity at 11Khz, there was some at 5.5 Khz but the 7Khz was significantly stronger "by eye". My understanding is that knock voltage levels are reported on a logarithmic scale in dB and 6dB = a factor of 2, so the difference, for example, between 60 dB and 66 dB is 200% not 10%... a 3dB difference is 40%. i did the forced knock test twice on two seperate days at the same operating point and got the same result.

To clarify, my original question was how much varaince is there between the theoretical knock frequency and the "real world". I am concerned that the 7Khz is an artifact of the particular engine operating point where i forced the knock and not real knock. In other words - if you have an ancient pushrod cast iron straight four 2V with a bore of 104mm and a high tech V12 4V variable DOHC , aluminum block with forged and titanium bits with a bore of 104mm, how much difference could be expected in knock frequency? 100's of Hz? or Khz?

an old timer here said you could distinctly here the difference in ping between a '71 Ford V8 and a '72 Ford V8 - identical engines (i think Cleveland 351) in every respect except for the pistons and compression ratios (11 vs 8.5) .

all that being said, i will review the ECU logs from the forced knock and see what's at 11Khz and also take a harder look at the 5.5 and 7 Khz. I may also try forcing knock at a couple of other low power operating points and see if the pattern is consistent.

Peter

so googling around and i came across an article "Combustion Knock Sensing: Sensor Selection and Application Issues" in the SAE journal from 1990.

Despite being a technical journal article it's actually mostly understandable https://www.jstor.org/stable/44548143

TLDR: generally speaking fundamental knock frequency IS a function of bore diameter and combustion temperature. combustion temperature can cause a +/- 400 Hz deviation. There can be secondary modes (related to the axial direction, perpendicular to piston top ... i beleive) - fundamental frequency multiplied by 1.65, 2.07 and 2.26. Secondary phenomena can mimick knock.

Swirl (the rotational motion of the incoming air charge about the cylinder's axis in a TWO valve engine ) combustion chambers lead to very different knock frequencies 1.25-1.3 calculated Freq (..no theory explains it..as of 1990 :) Turbochargers may also impact the frequency, not clear what the impact is from this article. so since i have a turbo, i'm not sure where this leaves me...for the rest of you , seems like the formula using bore diameter and possibly one of the secondary mode multipliers is a resonable bet.. as has been the published advice, except now we know why a factor of 2 (2.07) makes sense.

there is also this more recent article that talks about classic engines from the 30's vs current: designs. i have not read it yet, but it doesn't appear to cover the topic of turbo charged engines

"Distribution of Knock Frequencies in Modern Engines Compared to Historical Data" 2018-01-1666

https://www.sae.org/publications/technical-papers/content/2018-01-1666/

likely there are a bunch more articles out there

this would make an intersting topic for a webinar as the concepts are not all that hard to grasp and it would take at least some of the"magic" out of choosing a knock frequency.

Great info! Thanks for sharing the source...

To me this all tracks with what Andre often says regarding verifying knock with audio detection. Theory and reality often don't perfectly align; I feel this is the case here.

I've understood that factory OEM sensors that are 'narrowband' or tuned frequency don't work well once you modify the engine to an extent, and now hearing combustion temp and cylinder pressure affect the knock frequency, that makes sense too. I know air density effects waves in the intake (Helmholtz resonance) as sound travels slower through denser air.

I've also always wondered if block metal/alloy and sleeves make a difference in the frequency as well...

i attached the SAE article from 1990.

All this theory is great, but it would be nice to hear from the real world. I understand that lots of people are using 2X fundamental, but has any one observed a fundamental knock frequency that was significantly different than computed ...say more than +/- 750 Hz ?

So kudos to @Mike McGinnis - changed out gas to 87, re ran intentional knock at 2000 RPM 20Kpa/45% TPS and 40 degrees of total advance. 11Khz lit up like a beacon, could clearly hear pinging in knock ears and knock signal was a good 10-15db over background noise. The 7 Khz was definitely a "secondary phenomena mimicking knock"