Made myself a pair of electronic det-cans today. Hopefully I can once and for all determine if my weird part throttle knock is false or not. As well as some other areas.

They seem to work well but I was curious who else is using det-cans for tuning and where do you usually place the clip or bolt to your engine? Some say block some say right above block on the head. My particular setup has the mic on a large aligator clip. Is there a good spot that anyone has found for this type of mount?

I read somewhere that somebody thought the mic on these cheap devices couldnt pick up the higher frequencies that knock makes. It sounds pretty sensitive and has a broad range of higher frequencies to me. Anybody have any experience with these?

Thanks - Ethan

 

Interested to hear this as well. It looks like you tried to upload an image, but it's not displaying correctly for me.

On a similar note, have you logged the raw knock voltage? I believe the channel is called "knock_adc_processed". You can log this, along with RPM, "knock_base" and "knock sum" and NOTHING else, so it logs at a very high sample rate. This may give you an idea if the knocksum you're experiencing is real. Knock_base is the floor, or threshold, that the ECU creates as a result of your knock multiplier settings, and its dynamic adjustment to the threshold based on continuous measurement of noise from the knock sensor. If your knock_adc_processed exceeds knock_base, it is considered to be knock by the ECU and will be reported in knocksum.

Typically, actual knock will be significant spikes above the nominal amount of knock "noise" in the sensor. Continuous "noise" that builds with RPM and crosses the knock_base value may just be that, noise, and nothing else.

 

as I've been told evoscan's logging speed is slow for that

https://www.evolutionm.net/forums/ec...-knockadc.html

 

The image appears for me. Ill try adding it as just an attachment.

I have tried logging just the knock voltage but after a handful of good logs with knock examples and posting a few here i couldn't get any real results. I was basically look for the spikys and what not. Im not really comfortable just say hmmm theres no real spikes so i guess its not real knock...However I was only looking at knock voltage so let me reread your post and the link that was posted to see if I can get a better idea.

Here is a portion of a log I took this morning. It so happens that I had knock voltage checked...Its weird how the knock just appears. It love that 3000-4000 rpm area. It makes the car inconsistent as it doesn't always knock like this. Plus the timing is quite low in these areas. 3°? And the ECU retards it to -1 and dumps fuel down to 10.4 AFR? Seems werid to me. Its been like this for a while though. Id like to nail it.

Ive also tried recording my det-cans while logging but for some reason my computer is recording noise from the car even when the det cans arent plugged in. So I need to sort that out first.

LogEntrySeconds	TPS	RPM	TimingAdv	KnockSum	KnockVoltage	Load1B	JDMMAP	WideBandAF
118.70293	13.7254902	2218.75	23	0	0.468	15.6	-9.66325	15.3615
118.81242	13.7254902	2093.75	22	0	0.429	16.8	-9.85672	14.7588
118.9176	13.7254902	2125	22	0	0.6825	16.8	-10.05019	14.7588
119.02728	13.7254902	2093.75	22	0	0.585	16.8	-9.66325	14.3766
119.14244	16.8627451	2062.5	22	0	0.4875	18	-9.27631	13.9062
119.24938	19.21568627	2062.5	23	0	0.702	22.8	-8.6959	13.9062
119.35958	19.60784314	2093.75	26	0	0.5265	25.2	-8.50243	15.1851
119.46936	20.78431373	2062.5	30	0	0.429	28.8	-8.30896	20.3007
119.57651	21.56862745	2031.25	31	0	0.507	31.2	-7.34161	20.3007
119.68635	23.52941176	2062.5	30	0	0.9165	38.4	-6.56773	31.5756
119.79627	23.92156863	2093.75	30	0	0.5655	40.8	-5.60038	30.2085
119.90328	29.41176471	2093.75	29	0	0.6825	57.6	-3.08527	30.2085
120.01338	31.76470588	2156.25	25	0	0.819	68.4	-2.11792	17.5518
120.12334	31.76470588	2125	24	0	0.7995	69.6	-1.92445	14.2884
120.23028	33.33333333	2187.5	24	0	0.741	73.2	-1.34404	14.2884
120.34065	36.07843137	2250	20	0	0.702	79.2	-0.76363	15.1557
120.45034	36.47058824	2281.25	18	1	0.702	80.4	-0.76363	14.7294
120.56059	36.8627451	2343.75	15	0	1.6185	84	-0.37669	14.7294
120.66928	38.82352941	2375	14	0	0.8385	88.8	-0.18322	14.5971
120.77935	41.56862745	2437.5	10	0	0.4485	96	0.39719	14.4648
120.88939	41.96078431	2468.75	8	0	1.3065	99.6	0.78413	14.4648
120.99924	47.45098039	2531.25	6	0	0.702	106.8	1.55801	14.553
121.10927	47.45098039	2593.75	4	0	0.7215	110.4	2.13842	14.5824
121.21635	47.45098039	2656.25	5	0	0.4875	115.2	2.71883	14.5824
121.3263	47.05882353	2687.5	4	0	1.053	120	3.10577	14.0679
121.43635	47.05882353	2812.5	4	0	1.482	126	3.87965	12.7302
121.54357	46.2745098	2843.75	3	0	0.8775	132	4.26659	12.7302
121.65336	46.2745098	2968.75	3	0	1.053	139.2	5.23394	12.4215
121.7636	46.2745098	3000	3	0	0.9945	146.4	6.00782	11.9217
121.87041	46.2745098	3125	3	0	1.482	153.6	6.97517	11.9217
121.98024	46.2745098	3187.5	3	0	1.3065	163.2	8.13599	11.7012
122.09024	46.66666667	3312.5	3	0	0.9165	171.6	9.29681	11.5101
122.19727	46.66666667	3375	3	0	1.053	183.6	10.6511	11.5101
122.30755	47.45098039	3562.5	3	4	2.1645	194.4	12.19886	11.319
122.41735	47.84313725	3625	0	7	1.833	213.6	15.10091	11.0544
122.52429	54.11764706	3812.5	0	7	2.028	247.2	18.3899	11.0544
122.64226	56.47058824	3968.75	-1	7	1.911	271.2	21.09848	10.584
122.75826	58.43137255	4093.75	-1	6	0.8385	267.6	22.64624	10.4076
122.86425	72.94117647	4343.75	0	6	0.663	274.8	23.61359	10.4076
122.97439	76.07843137	4531.25	1	5	1.0725	268.8	24.194	10.4811
123.0815	76.07843137	4687.5	1	5	0.819	265.2	22.64624	10.4811
123.19137	57.25490196	4875	1	4	1.287	236.4	18.97031	10.9809
123.30124	34.90196078	4937.5	5	3	0.468	100.8	0.59066	11.0397
123.40857	14.11764706	4718.75	37	0	0.741	26.4	-8.50243	11.0397
123.51827	14.11764706	4531.25	38	0	1.287	21.6	-10.05019	10.0401
123.62855	14.11764706	4312.5	38	0	1.2285	18	-10.43713	11.1132
123.73537	14.11764706	4093.75	38	0	0.546	14.4	-10.6306	11.1132

 

Does it not respond to timing advance changes?

I'd actually try increaseing the timing at the low RPM/load areas that get pased through while the turbo is spooling up.

I could be wrong, but my theory is the ECU is only looking at the knock sensor for a small window of time. If you start retarding the timing too far, you start getting more of the ignition noise in that window and it can lead to false knock detection. I was running a decent amount more timing at that load and RPM range and if I dropped it any lower, I got all kinds of "knock."

 

I havent really tried giving it more timing but was pondering that too. The car does feel quite sluggish down low and gets worse when it actually get "knock" and retards even more.

I am looking into the DMA live map and it might make making changes here and there easier on the fly. The only thing that sucks is they done have my ROM ID patchable. So I am looking into switching to 95630706 ROM ID.

Thanks for your input