Knock Headphone Setups?

[Mitco39]

Hey guys,

I am looking to find what you are all using to listen to knock on the engines you tune. I have made a knock listening device for the Nissan cars I have tuned and it works well but I am looking for other ideas on the least intrusive way to monitor these engines for knock. I want to be able to listen along side of what the knock sensors are saying to rule out any false knock (or vice versa). I would prefer not to have to hack any harness's or cause any irreplaceable damage to the vehicles being tuned.

I am currently tuning a LS1B controller however I would like to have a solution for all the EFILive supported controllers.

If anyone cares to see the device I made here is a link to the instructions to build it, while it would work I would imagine on these domestic vehicles I have to splice into the harness (or use a patch harness) to listen in.

https://uprev.com/documentation/Knock%20Listen%20Device.pdf


Thanks in advance.

Mitch

[Mitco39]

I guess knock detection is not a big deal for guys on here? Incase anyone ever finds this via search I decided to go with this setup.

http://thetuningschool.com/product/plex-knock-monitor/

[joecar]

It's a big deal to me, I use a modified Chassis Ear, but it seems to pick up too much noise.

I'll see what the engine shop uses.

[pinstripebob]

So if you make or buy knock phones, do you watch a log and listen for knock at the same time and make a note of false/real knock?

Sent from my Nexus 5X using Tapatalk

[Mitco39]

I like to listen to the engine to determine if the engine is detecting false knock and go from there. Without headphones you are at the mercy of what the ecm thinks is knock never knowing for sure if its safe to desensitize the sensors to stop what you think might be false knock.


I was joking about not finding it important, lol. Everything I found on this plex setup is good so I am curious to see how it goes. Shitty part is you have to swap out the factory sensors but it does get around needing to worry about wiring that way.

[joecar]

So if you make or buy knock phones, do you watch a log and listen for knock at the same time and make a note of false/real knock?

Sent from my Nexus 5X using TapatalkYes, on the dyno, as you hold a cell in the VE or MAF table, you listen.

[pinstripebob]

That makes sense. Would knock phones be usable on the street?

I've attached a photo of some knock retard I'm seeing periodically, as well as the log and tune it came from for reference. In the picture, I had about 5.5 degrees of knock retard spike and decay over 5 seconds.

20456

20457
20458

[pinstripebob]

Thinking about this some more, would it be possible to pull the signal from the factory knock sensors similar to Mitco39's setup in post #1 and split that off to his amplifier? Then from the amplifier (or some "black box"), split to Mitco39's headphones as well as a 0-5V analog signal into the Flashscan V2? Then you could chart what you hear, correlate it to the data the ECU has, and determine action taken from there?

[Mitco39]

Yes you can do that, thats how I have my Nissan wired up to listen to it while its running. The issue with this is if you are doing customers vehicles you dont really want to be cutting or splicing into their wiring harness. But if it is your personal vehicle then this is what I would exactly do.

When I took the car down the Vegas last year I would put the knock headphones on with every tank of fuel and set the boost accordingly lol. If I heard knock id turn down the boost till the knock stopped for that tank of fuel. Overkill but I like to go as fast as possible lol.

[joecar]

That makes sense. Would knock phones be usable on the street?

I've attached a photo of some knock retard I'm seeing periodically, as well as the log and tune it came from for reference. In the picture, I had about 5.5 degrees of knock retard spike and decay over 5 seconds.

20456

20457
20458Increase the decay rate (so it decays quicker), then when it comes back, observe if it is a single spike, or a series of a few or more spikes.

[joecar]

If you do splice in, I would use make sure the signal to the PCM is not degraded... i.e. whatever you splice in should have high impedance (this is what the 2 capacitors do in Mitco's circuit) so that the original signal is not pulled down... if you know electronics you might use an op-amp or some low voltage analog buffer.

[pinstripebob]

If you do splice in, I would use make sure the signal to the PCM is not degraded... i.e. whatever you splice in should have high impedance (this is what the 2 capacitors do in Mitco's circuit) so that the original signal is not pulled down... if you know electronics you might use an op-amp or some low voltage analog buffer.

Yeah I would be a little worried about not getting the PCM the signal it needs. I work with a pretty good EE, I'll go through this with him and see what he comes up with.

[pinstripebob]

Alright I got a good log. The screenshot below shows the one of the "better" knock retard spikes. It still seems to have a pretty slow decay rate. The peak shown is 4.1 degrees. I multiplied B6207 by 5 in an attempt to make the knock retard decay more quickly. Should I go ahead and make B6207 more like 0.50 or something? It was 0.01 and now around 0.07.

20501

20502
20503

[joecar]

Try that and take a log, you will see if that is enough or if you need faster decay rate.

[pinstripebob]

I changed B6207 to 0.50 across the board, that definitely made a difference. Any knock retard events are fairly narrow spikes and decay very quickly. The screenshot below shows one instance where there were two events back to back.

20514

20515
20516

[joecar]

Good job.

If you see rapid successive spikes (sawtooth pattern) then that is real knock.

[pinstripebob]

Ahh okay. I haven't had a lot of successive events, but good to know that I'm not seeing false knock. Is it safe to keep the 0.50 decay rate, or are there other steps I should take (I assume drop the spark table in that area of I consistently see knock).

[joecar]

Try 0.25, and post another screenshot.

[pinstripebob]

Got it, set B6207 to 0.25 across the board. I had three KR events on the drive in to work, one of them gave me 3.3 degrees and had a couple successive spikes.

20526

20527
20528

[joecar]

Ok, that looks good to me, when there's real knock then the decay rate is just slow enough to take care of the knock, and it's just fast enough to recover quickly.

[Mitco39]

Thanks for the info again Joe! I will do the same thing next time I am on the dyno. My knock setup showed up so I am curious to see how those results compare with the factory ecm.

[joecar]

Mitch,

Which knock phone did you get...?

I'm thinking that the knock phone output might be recordable via the FSV2 AD inputs (if has range 0-5V)...

[Mitco39]

I got the Plex V2

http://www.plex-tuning.com/products/plex-knock-monitor/


What I like about other softwares (and this may be a limitation on the OEM not eflive) but with ECUtek for example you can see the readings from the knock sensors as a raw signal, so instead of just seeing the measured knock as per the ECM you get the raw signal as well as the amount or retard the ecm thinks is a valid amount. Looking at all the reviews on this it seems like this will be the best way to see and hear that raw signal as well.

This one does have the ability to output knock readings.

I am excited to play with it and see how it works.

[joecar]

On some LS1B's the following pids provide raw knock counts:
- GM.ESC1DNL
- GM.ESC2DNL

I've seen those pids work on some PCM's and not others.

[joecar]

That is a nice knock phone, has good features.

[dian]

so you spent $1000 on that devise. what about just a microfone and phones? you could add an adjustable filter/amplifier. knock is around 3000hz with our bore, right?

[PLohr]

I would like to toss in a few more names to get feedback on some other chassis ears / knock headphone setups.

1. J&S Safeguard (http://www.jandssafeguard.com/) (not the main purpose but it does have a headphone output)
2. Phormula (https://efi-parts.co.uk/product/phormula-ks-3-knock-detonation-detection-kit/) (is it still around?)
3. The Knock Box (https://theknockbox.com.au/)

Any thoughts on any of the above?

To make the list more complete and put everything in one place, most of these have also been discussed:
Mitco39's DIY setup (https://uprev.com/documentation/Knock%20Listen%20Device.pdf)
Plex Knock Monitor v2 (https://www.plex-tuning.com/products/plex-knock-monitor-2/) (discontinued)
Plex Knock Monitor v3 (https://www.plex-tuning.com/products/plex-knock-monitor-v3/) (not yet available)


Thanks for any help.

[joecar]

Thanks for posting.

The V3 is expensive, but has many features including CAN bus connectivity, programmable knock threshold, and advanced filtering.

[Doc]

I have been on the Plex V3 pre-order list since March. Hopefully should arrive sometime soon. Really interested in testing it out. Are there those on the forum that have already received theirs?