Hoping VegasInvasion will comment on this, but currently im trying to diagnosis why my knock sensor is seeing occasional detection of pinging when letting off the throttle between shifts....here's my setup

Stock B16a/B16 LSD tranny (JDM)
T04E 57trim .63A/R
38mm Tial WG
Spoolin Performance log
3" Ex
440cc Injectors w/255 GPH Walbro
Greddy BOV
KnocksenseMS w Siemens knock sensor
Crome tuned tuned at 7psi (255/175 WHP/WTQ)


I'm in the process of getting my KnockSenseMS fine tuned however I think I'm right where I should be at exactly half-pot since I've seen cause and effect changes by reducing timing and adding fuel which has nearly eliminated any blinking of the LED light from this unit with the exception of when I make quick throttle changes while letting off throttle under boost during shifts.

One theory is that perhaps the quick rush of air from my BOV during throttle release is just enough pitch to trigger a false knock detection however it doesnt happen every time I release the throttle under boost that subsequently make noise by the BOV

My concern is fuel and Ign changes from tip-in or tip-out that results in a momentary ping that the Knocksense is picking up that other wise would never be heard to the naked ear.

In Gravity Imports manual to Crome tuning, In the Fuel Multipler advanced section it's noted (pg 13) "For general purposes, O2 and tip-in should be kept at 1.00, unless working with special fuels (E85, race fuel, nitro). Cranking and Post-Startup should be slightly higher than the Final Multiplier."

My question is, even in the case where I'm running larger than stock injectors (440cc) and the fuel multiplier value is now "0.545", do I adjust the tip-in at "1.000" or should the tip-in value in the fuel multiplier match the value of the adjusted value being "0.545" as the injectors are set to?

 

DC, sorry to jack your thread but I see you are using the Megasquirt KS unit instead of the regular board and since you mention the Crome manual I'm assuming that's you engine management.

I'm looking into this option and would like to know why you selected this unit. I apologize if this has been covered as I did not search before posting here but was intrigued by your post and quickly hit reply.

 

No worries....I made the thread vs. PM'ing VI to see what others had to say even if not directly related to the topic.

The KnockSenseMS was suggested by others here on HT however once I saw that it was originally designed by and used by the MegaSquirt crowd it had my interest especially at the cost of roughly $85 shipped to my door, sensor and all. Very easy installation and the sensitivity is tailored to my bore size.

I've read where many feel this unit still triggers false knock however my own experience with it so far seems to be pretty accurate because of the cause and effect changes I've made since employing it and noticing the difference in the frequency I've seen the trigger light/ LED go off.

And yes, I'm using Crome for now until I move onward to a Hondata. I purchased my car from a guy who tuned it on Crome at a conservative boost level so I didnt see any reason not to try to work off this since the tune it self was 85% where it needed to be however with the right tools I felt I could get a bit more out of the lowcam power which I have since using the KnockSense unit and keeping a close eye on the condition of my plugs.

 

I went looking back through S Managers features and noticed something that might remedy my problem (once I get the S300).....I suspect the issue is slight detonation under throttle cut in which a momentary lean condition occurs as the injectors are cutting off before they should.


Fuel Overrun Cutoff by TPS or MAP


Granted Crome does this too however its not as diverse as Hondata's feature. With Crome I can try to play around with the MAP pressure level for which the injector cut-off point by mBar is however the cutoff pressure that I'm running is already at 93 mBar which is well below the first column on the lowcam tables to begin with.



I should have mentioned too that I use an OBD1 throttle body which has chambers for air bypass to the intake when the TB is closed so in theory I'm wondering if the MAP is being jolted by the quick Vacuum levels of closing when between shifts however the pressure quickly increases in the manifold through the TB bypass chamber enough to cause a momentary lean spike while the injectors are cut

 

I'm subscribing.

Been looking for an affordable knock sensor.

OP, where did you purchase from?

Is it DIY / assemble and solder yourself?

You logging the output of the knock sensor in crome pro?

 

http://www.viatrack.ca/

Read through it, email this guy with bore specs and the type of unit you want. You can get a unit that is universal vs. one thats tailored for MegaSquirt...

The unit it self has both an LED indicator for visual confirmation of knock and a 5v output for Auxilory logging with Neptune or Hondata....and no I'm not using CromePro but only because the vehicle was already WOT tuned on a dyno by Crome so street tuning is what im doing currently for everything at PT.

I could invest in all the cost to get CromePro and the Ostrich but might as well spend a bit more and get whats better.

No soldering needed however I generally solder wire connectors and use shrinkwrap for durability on wiring anyhow, the unit has tabs for wire inserts and is super simple with a tuning **** for sensitivity.

when I get an S300 I'll be logging the knock events on a non-logic system to where basically its on me to let off the gas if it pingings but so far as I've said at WOT and anywhere else except momentarily letting off throttle do I get an indication of ping since using this device so I love it already...

 

Another item you may want to grab is an inexpensive stud that Toyota makes......It was recommended by others on the board here on HT and it has the same thread pattern to certain block cavities.

I ended up placing the sensor mid-block to get away from noise from the cylinderhead on Cyl#3 to a spot where a bolt to the previous intake brace use to reside....

In the pic of the stud I put a threaded cap on the big end so I could tighten the stud more securely. The smaller end is what threads into the block...

 

Thanks.

I sent an email, so hopefully I can get it ordered asap.

I'm tuning my autox car on jrsc + b18c5 + crome pro, so I think this will help for the tip in also.

I have aem ems and s200 as well, but thinking of going with s300 soon.

Anyway, I'm interested to see how this turns out for you.

 

Thanks.....I always keep my threads updated right to the root of the issue so if someone else finds themselves in the same situation they don't get a disconcluded thread wondering WTF happened in this situation....

You'll like this device, beats trying to mockup your own det cans which aren't bad at all but to try in decipher detonation by ear everyday vs. a visual que + logging is an advantage since this device allows a greater range of flexibility over a stock knock system which makes it sort of the wideband of knock sensors

 

Did some more driving today with a longer round trip running this place and that place and noticed more of where my problem areas are in the low cam timing.

I've found that the issue happens most often when under light load (1st & 2nd gear) while under positive pressure at vacuum (while spooling) between 10"-20" hg from glancing at my boost gauge when I would get a blip on the knock LED.

I need to move the LED closer to the boost gauge because where I currently have the LED is on the other side of the steering wheel from the boost gauge to my left.

As I've said this most typically generates a knock detection as I'm building boost but under vaccum and then when I let off the throttle between shift's such as 1st to 2nd it happens. I noticed today though that it happens the worst in 2nd gear and even under acceleration before transition from vacuum to boost.

I've targeted mostly the table cells from 3000 to 6000 rpm from as low as 114 mBar to right before positive pressure at 981 mBar as the potential area of concern so I'll try trimming back these values by .75-1 deg to see if this reduces the frequency of detection or eliminates this all together.

 

I'm running the MS unit, with the Toyota stud, as well. The difference between the standard and the MS is that the MS version has an output so that you can log knock. I tapped mine into B6 on my ecu, but I might use it with a transistor to ground one of the switched inputs for a proper true/false input (that I can use to retard timing or switch maps in the future).

Seems to work well so far. I also used the extra cable supplied along with a headphone jack from Radioshack and a couple of .01muF capacitors (for noise isolation) to wire in a port to use with a headphone amp or into the aux in on my stereo for det-can use.

If the led starts to blink on me I just tab the stereo over to aux-in and check for knock. I'm kind of shocked by how much (very light) knock I've been living with at part throttle, and I've been working through my maps to get rid of it.

It took a couple of days to get it really dialed in, I just left it in my glove box with the back cover off and kept a screwdriver handy while I drove around listening to my motor through the car stereo. I haven't picked up any false knock (except from shifting into reverse), but I am on a very stock NA build so things are pretty quiet.

 

As a KnockSense user I can tell you that the KS is very sensitive and will detect noise as knock.
When my down pipe was rubbing on my oil pan under high rpm decel, the KS would light.
When my pistons were slightly contacting the oil squirters, the KS would light.

What I think you are regarding as knock is more likely mechanical noise. It could be caused by engine movement or even VTEC engagement.
A simple test would be to remove a known safe amount of timing in that region or the entire table, say 3* and retest. If you still get "knock" then its not real knock and disregard it.

 

It might very well be but as I've seen, every cause and effect change I've made to ignition timing has reduced the frequency of knock indication so I believe the unit is doing its job with little false indications.

Today I didn't see it go off not once (honest truth) and this is after I reduced timing slightly as I mentioned in my last post.

As for sensitivity to other mechanical noise I can see this happening based on where its placed but where I have it mounted is by use of the Toyota stud just next to the oil filter (lower right) where it should hear knock from the 3rd cylinder mostly instead of valvetrain noise which I'm fairly certain is not an issue as far as I've seen.

 

Yeah I have to thank you and Vegas Invasion for recommending this device....works great for me so far! I think one reason i might not be seeing so many false indications is due in part to the motor being all stock as far as internals, I'm curious to see how well it performs when I rebuild for higher boost levels and forged pistons go in which may or may not cause false triggers.

 

You might be right Muckman.......granted I have noticed a correlation between removal of timing and the frequency of knock indications from the LED however as I've been driving the car now on a regular day basis (after getting the registration squared away) its still happening in light of me pulling timing here and there.

I haven't taken a whole lot of timing out, just trimming it in regions where the MAP x RPM is when the knock occurs, however it hasn't stopped completely.

I even put one gallon of Sunoco 110 in over 6.5 gallons which would have brought the octane level up to 95 roughly due to the octane calculations on Sunoco's website and I still witnessed this last night on my way home over the stretch of 15-20 miles.

What I have noticed is that it occurs 99% after the motor is fully warmed up so I wonder if the IAT and ECT corrections need a bit more attention.

First before screwing with those though I'll remove a good chunk of timing and see if it happens as a test.

I'm also considering switching to a newer 2nd Gen B16 manifold I have which doesn't have the same bypass chamber into the manifold as my 1st gen does.

Generally the way to get the LED to go off happens best in 2nd gear after the motors warmed up by letting the rpms climb into VTEC transition or slightly before VTEC engagement (it doesn't always happen at VTEC engagement but almost always after shifting when a large amount of spool is generated) then shift into 3rd which then subsequently the spool collides with the TB when closed, the BOV releases the excess spool however some of this makes its way through the bypass in the TB which I believe is raising the pressure in the manifold right after low vacuum just enough to created a problem with fuel and ignition because it happens so fast that the MAP doesn't even pick this up in order to react to the pressure spike.

Admittedly this may be the result of how my waste gate is configured too. I'm not running a boost controller, just the 7 pound spring..... The car was setup this way from the previous owner whom did not route a line from a nipple on the compressor outlet to the bottom of the wastegate, instead the pressure line to the wastegate is coming from my intake which puts the pressure source after the TB so when the TB closes the WG see's vacuum and stays shuts leaving the BOV to bleed off the excess of spool.......If this is the source of my issue I'm really going to feel like a dumbass

 

Yeah, you should also be using some sort of headphone setup while you're tuning. It's hard to say for sure without the audio to back up the knock sensor especially if you've got it set to catch the light stuff.

 

I missed this thread, but I have been using the KnockSense MS and Toyota stud for a couple different setups now. I have installed them on a few different setups as well. I personally think they qualify as barely an aid to tuning but definitely not a replacement for a DET Can. They are OK for tuning completely stock engines with low miles, but not higher mileage engines or modified engines. As Muckman said, they are VERY sensitive. We spent a LOT of time and money on plugs chasing phantom knock in both stock and modified setups.

Using it for knock Input with a factory knock board doesn't work unfortunately with a P30 or P72 (tried both). Some engines have a lot of normal noise that can't be tuned out with the KnockSense dial or by adding a dashpot / potentiometer. In the end we have learned to not dive into the rabbit hole very far while paying attention to the knock light. If you are getting flickering then reduce timing and if your AFRs are still good just learn to ignore it. Plugs will still tell you the whole story at the end of the day. We went has far as using E98 to be sure it was just false engine noise knock plaguing us on different setups and it always was (tuned for 93 pump gas cars).

As far as knock on deceleration / gear change is concerned, things like piston slap (load to no load), timing belt deflection (load to no load), valve train slap, and other similar things will cause this to happen. For instance I have an engine right now that develops false knock over 8k RPM from piston slap. How do I know that is what it is? My DET Can showed me the answer. Don't underestimate the power of using your ears for tuning, but unfortunately at the end of the day nothing beats the Dyno for the bulk of the tune with street for map clean up.

 

Or non-vtec b-series. Those rockers are NOISY above 5k!

 

Yeah, I agree this is a nice device but not the end of means solution for ignition tuning. I'll probably mock up some Det cans next and do more comparative testing.

Yeah I recall seeing the amp unit your showed in another thread that you were using in conjunction with the KS unit. How did you end up wiring this together if you dont mind my asking? I cant recall if it was you or another person that wanted to rig an amp+headphones up to their OEM knock sensor so they could fine tune the sensitivity of their KS unit ....I would guess you could either amp the 5V output coming out of the KS however I dont see where this would be accurate since that should be strictly a 5V output each time the unit thinks its heard a knock so the only other way is to tap the signal coming from the knock sensor before the signal wire is feed to the KS. My only concern is wouldn't this impede on the clarity of the signal to the KS once feed through the amp input?

 

Yeah there's nothing analog to amplify on the Knocksense output. It's just a line that's held at 5V until there is a knock and then dropped low (or vice-versa depending on the logic).

All I did to use the headphone amp was take the extra shielded wire that came with the unit (he includes quite a bit) and add a headphone jack to one end with a couple of .01muF capacitors to isolate the circuit. Take the clipped end, strip and solder it so it looks like the end that was already prepped on the KS wire, and hook it up to the same tabs on the unit. You now have a listening port in parallel to the Knocksense. As long as you use shielded wire and capacitors it shouldn't impede the function of the unit.

Even without the Knocksense unit you could wire directly to a Bosch (or any, really) knock sensor and have strictly a listening device.

Read this article:

https://uprev.com/documentation/Knoc...n%20Device.pdf

 

Awesome!! Many thanks, common sense told me it had to be setup in this fashion but just wanted to be certain+ the capacitors were something I was unaware of. I'll try incorporating this in to the mix this weekend.

One last question.....the guide says to use .01uf capacitors. RadioShack has these but theirs different voltage ranges, most common are 50V thru 1000-2500V ranges. Obviously this wont see more than 12-15V max if that so will 50V caps work or what range did you end up going with?

Also, if your interested, I have an old copy of Sony Soundforge 7 I might use to record samples and ill PM a DL link with a keygen if your interested.

 

Yes you'll be fine. There's barely any voltage generated by the piezo element, hence the need to amplify it to hear anything.

Just for fun though, here's an article on how capacitors work and why you should know:

https://www.sparkfun.com/news/1271

Thanks for the software offer, I still have a few things on my computer from my music days. I use Labview for all my signal analysis these days.

 

I ordered my knocksenseMS the other day, thanks to you DC_Legacy.

Since I'm on crome pro right now also, I'm assuming you're using the tip-in throttle compensation as well?

Did it help your problem, or are you just pulling timing?

 

Right on......Thanks for all your input!

Awesome! Seems this forum now has a fair share of KS users with you and me as the newest probably.

I'm using the tip-in compensation however the issue isn't so much on the tip-in as it is the "out".

As I've mentioned it has its tendencies but isn't 100% predictable because it can happen under certain conditions some times and sometimes not.

What usually brings on a knock reading is either 1st or 2nd gear shifts (after motors full warmed up, never happens cold after startup) where when I let off the throttle as the turbo is spooling that triggers the knock indicator LED to blink. This tends to happen generally between 4-6k rpm when between 10-25 HG of vacuum (when spool is built up enough to trigger the BOV, but doesn't happen very much during loud 3,4,and 5th gear shifts).

Some have said I'm chasing mechanical noise which might be the case and if it is I'll have to get this thing dialed in better by use of audio input now from the knocksensor as spAdam pointed out.

I've pulled a fair amount of timing and have tried high octane blends tested at 95 octane and still had blinks but I'm not at all discouraging use of this unit as I'm certain it has picked up knock that I cant hear. Since using it and making timing adjustments the frequency of knock indication has slowed so at this point what I might be seeing is mechanical triggers and I'll soon find out since I'll be incorporating an isolated audio tap into the KS feed so I can listen for knock on top of having the device do its thing to help me get the unit dialed in even better or help me tune my P/T timing better

 

Just ran a test by reducing timing to 1.00 Deg per psi (prior settings were .875 per psi) and this seemed to eliminate nearly all any knock indication between shifts (except a minor low intensity blink after letting the car sit an idle over 5min then resuming driving which is understandable since IAT's increased as it sat) .

I definitely felt a loss of power/response that came with it so I'll try maybe reducing by .9375 per psi and see if the frequency of indication increases any.

It's funny because it seemed that the area of timing concern was below positive pressure but I suppose this will be a mile stone to a work in progress.