Jump to content
SAU Community

Recommended Posts

Sorry if im commandeering the thread

i have been reading around and getting some mixed views on this topic, what id like to know is, what is the raw output of a stock knock sensor, and what is the ecu looking for after the knock sensor is amplified?

so just in regard to the Knock amp/filter, they are bloody expensive, has anyone rigged up one from say...a jaycar kit?

any advice is much appreciated :D

Zurlox

mine is running off a jaycar kit but they are hardly compact.

The ecu is looking for a 0-5v input 0v being no knock and 5v being your engine about to explode with a signal somewhere in the region of 5-9khz.

Sorry if im commandeering the thread

i have been reading around and getting some mixed views on this topic, what id like to know is, what is the raw output of a stock knock sensor, and what is the ecu looking for after the knock sensor is amplified?

so just in regard to the Knock amp/filter, they are bloody expensive, has anyone rigged up one from say...a jaycar kit?

any advice is much appreciated ;)

Zurlox

I looked into this a few months back. The knock sensor is basically a microphone. The amplifier just amplifies this so that it can be detected by the ECU. The problem is that the engine makes noise at all frequencies. We only need to detect the knock which occurs at specific frequencies. The frequency of knock depends on a few things. Bore size was one of them I believe. I found a formula somewhere that calculated the frequency for you.

So the external amplifier needs to amplify to right level and filter out any unwanted noise outside of the knock frequency. There is an IC out there that is designed specifically for this (I can look it up if you want). It is surface mount and not commonly available from shops like Jaycar.

Also, knock only occurs at a certain point in the engine cycle. Therefore, a smart amplifier/ECU will only switch on the knock sensor input at those points in the cycle where knock is expected. I'm not sure if the Vipec does this or not.

After reading all this I was thinking I'd go for the amplifier sold by Vipec/Bob, but then you need to induce pinging in your engine to set the gain correctly, etc. That kinda scares me, so I've stuck with no knock amplifier and trusted by tuner's tuning :D

After reading all this I was thinking I'd go for the amplifier sold by Vipec/Bob, but then you need to induce pinging in your engine to set the gain correctly, etc. That kinda scares me, so I've stuck with no knock amplifier and trusted by tuner's tuning :laugh:

if your smart you can do it in a non detrimental way, obviuosly you dont want to induce knock at full load, full boost and so forth but it is simple enough to do it safely at the right time.

Edited by URAS

Zurloc PM'd me for further details on the knock amp, but I thought I'd reply here for the benefit of everyone else.

The knock amp IC is the TPIC8101 http://focus.ti.com/lit/ds/symlink/tpic8101.pdf

Autospeed article on DIY knock detection

http://autospeed.com/cms/title_DIY-Detonat...48/article.html

According to that article the frequency of knock can be approximated as 900/(3.14*Cyl Radius). I read somewhere else the frequency also depends on cylinder temperature and the location of the knock sensor, so I dunno how accurate that formula would be.

I was trying to find a Subaru site that I read a few months back. It had lots of info including sound recordings etc. Anyway, if you do a Google search you will find a mountain of information.

I wonder how different the knock amp settings will be from one engine to the next. Perhaps it would be good to collate everyone's settings in a database or maybe the Vipec forum. It would be awesome if we could say RB26 = frequency...., gain....

The stock ECU and the power FC must do the amplifying and filtering on board, so I guess engines must be fairly similar.

Zurloc: Keep us posted on how you go :down:

Edited by BoostedBarge
  • 1 month later...

Update:

i got the knock kit from jaycar, but the frequency band is wrong, it detects noise within 4.8Khz to 6.4Khz, which is too low, i am looking for a band of about 6.5Khz to 9Khz, so in order to achieve this i need to change a couple resistors. the one you change are the 2.7K resistor (inbetween the two 5.6K resistors top right of board) and i changed it to a 1.6K resistor to beef up the low pass filter, the other one you change is the 10K resistor (inbetween two 6.8nF caps top middle of board) and i changed it to a 12K resistor.

Now i havnt confirmed these changes but they should be pretty close :ermm:

Right from the horses mouth:

"Changing to the higher frequency range is easily done by changing the capacitor values in the filter.

The 6.8nF capacitors in the high pass filter become 5.6nF and the 3.3nF to 2.2nF and the 12nF to 10nF for the low pass filter."

This has come from the designers of the Jaycar Kit KC5444 to achieve a frequency band of 6-6.5Khz high pass to 8-8.5Khz low pass.

consult this diagram to get a better picture or things - Knock Schematic

Zurlox

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now


  • Similar Content

  • Latest Posts

    • Could be. Could also be that they sit around broken more. To be fair, you almost never see one driving around. I see more R chassis GTRs than the Renault ones.
    • Yeah. Nah. This is why I said My bold for my double emphasis. We're not talking about cars tuned to the edge of det here. We're talking about normal cars. Flame propagation speed and the amount of energy required to ignite the fuel are not significant factors when running at 1500-4000 rpm, and medium to light loads, like nearly every car on the road (except twin cab utes which are driven at 6k and 100% load all the time). There is no shortage of ignition energy available in any petrol engine. If there was, we'd all be in deep shit. The calorific value, on a volume basis, is significantly different, between 98 and 91, and that turns up immediately in consumption numbers. You can see the signal easily if you control for the other variables well enough, and/or collect enough stats. As to not seeing any benefit - we had a couple of EF and EL Falcons in the company fleet back in the late 90s and early 2000s. The EEC IV ECU in those things was particularly good at adding in timing as soon as knock headroom improved, which typically came from putting in some 95 or 98. The responsiveness and power improved noticeably, and the fuel consumption dropped considerably, just from going to 95. Less delta from there to 98 - almost not noticeable, compared to the big differences seen between 91 and 95. Way back in the day, when supermarkets first started selling fuel from their own stations, I did thousands of km in FNQ in a small Toyota. I can't remember if it was a Starlet or an early Yaris. Anyway - the supermarket servos were bringing in cheap fuel from Indonesia, and the other servos were still using locally refined gear. The fuel consumption was typically at least 5%, often as much as 8% worse on the Indo shit, presumably because they had a lot more oxygenated component in the brew, and were probably barely meeting the octane spec. Around the same time or maybe a bit later (like 25 years ago), I could tell the difference between Shell 98 and BP 98, and typically preferred to only use Shell then because the Skyline ran so much better on it. Years later I found the realtionship between them had swapped, as a consequence of yet more refinery closures. So I've only used BP 98 since. Although, I must say that I could not fault the odd tank of United 98 that I've run. It's probably the same stuff. It is also very important to remember that these findings are often dependent on region. With most of the refineries in Oz now dead, there's less variability in local stuff, and he majority of our fuels are not even refined here any more anyway. It probably depends more on which SE Asian refinery is currently cheapest to operate.
    • You don't have an R34 service manual for the body do you? Have found plenty for the engine and drivetrain but nothing else
    • If they can dyno them, get them dyno'd, make sure they're not leaking, and if they look okay on the dyno and are performing relatively well, put them in the car.   If they're leaking oil etc, and you feel so inclined, open them up yourself and see what you can do to fix it. The main thing you're trying to do is replace the parts that perish, like seals. You're not attempting to change the valving. You might even be able to find somewhere that has the Tein parts/rebuild kit if you dig hard.
    • Can you also make sure the invoices on the box (And none exist in the boxes) are below our import duty limits... I jest, there's nothing I need to actually purchase and order in. (Unless you can find me a rear diff carrier, brand new, for stupidly cheap, that is for a Toyota Landcruiser, HZJ105R GXL, 2000 year model...)  
×
×
  • Create New...