Best Knock Sensor Position ?
Discussion
V8RX7 said:
They don't.
With a cheap ECU I can get more mpg AND more power than OEM
Because I don't have to worry about fuel standards in India or Emissions tests etc
Similarly OEM count every penny - if it's 10p cheaper to mount it on the block instead of the head - it will be mounted on the block.
They DON'T make the car the best they can - they make the best compromise they can within their budget, time and legal constraints.
You are partially correct. Mazda engineers need to make the car affordable, so in order to reduce the price of the car, they do things like assemble the car on a line rather than hand building everything. The aftermarket provides a way for the end user to trade off affordability for other desired traits; however, your also compromising how well the vehicle was engineered. What I mean by this is, how many different ECUs did you try on your car? Did you take it to a dyno? the track? do any time of endurance stress testing?With a cheap ECU I can get more mpg AND more power than OEM
Because I don't have to worry about fuel standards in India or Emissions tests etc
Similarly OEM count every penny - if it's 10p cheaper to mount it on the block instead of the head - it will be mounted on the block.
They DON'T make the car the best they can - they make the best compromise they can within their budget, time and legal constraints.
Do you think that mazda just did a quick tune on the mx5 and called it good? or do you think they did the absolute best they could do in order to sell more cars? What is more cost effective, spending money to get the mx5 tuned right before manufacturing and selling more cars, or just getting it good enough and selling a lot less units? Many people believe that they can unleash significant power from their car with a quick tune. Aftermarket tunes gain power at the cost of reliability. They usually advance spark and/or adjust KR (both provide more power and MPG at the cost of reliability, and sometimes require higher octane fuel to run properly), another common technique to increase power is to increase boost on forced induction vehicles, which also reduces reliability. When it comes to that cheep ECU producing more "power" and MPG your actually burning cleaner than stock because of your advanced ignition timing, but mazda didn't know that, cause they where too cheap to pay somebody to actually tune the car.
I am not saying mazda engineers can't be out-engineered by the aftermarket, or even an individual, only that it is unlikely. Mazda engineers didn't tune the motor in a day, or even a month and they have way more dyno time with the BP motor than you (or anybody else probly).
dblack1 said:
V8RX7 said:
They don't.
With a cheap ECU I can get more mpg AND more power than OEM
Because I don't have to worry about fuel standards in India or Emissions tests etc
Similarly OEM count every penny - if it's 10p cheaper to mount it on the block instead of the head - it will be mounted on the block.
They DON'T make the car the best they can - they make the best compromise they can within their budget, time and legal constraints.
You are partially correct. Mazda engineers need to make the car affordable, so in order to reduce the price of the car, they do things like assemble the car on a line rather than hand building everything. The aftermarket provides a way for the end user to trade off affordability for other desired traits; however, your also compromising how well the vehicle was engineered. What I mean by this is, how many different ECUs did you try on your car? Did you take it to a dyno? the track? do any time of endurance stress testing?With a cheap ECU I can get more mpg AND more power than OEM
Because I don't have to worry about fuel standards in India or Emissions tests etc
Similarly OEM count every penny - if it's 10p cheaper to mount it on the block instead of the head - it will be mounted on the block.
They DON'T make the car the best they can - they make the best compromise they can within their budget, time and legal constraints.
Do you think that mazda just did a quick tune on the mx5 and called it good? or do you think they did the absolute best they could do in order to sell more cars? What is more cost effective, spending money to get the mx5 tuned right before manufacturing and selling more cars, or just getting it good enough and selling a lot less units? Many people believe that they can unleash significant power from their car with a quick tune. Aftermarket tunes gain power at the cost of reliability. They usually advance spark and/or adjust KR (both provide more power and MPG at the cost of reliability, and sometimes require higher octane fuel to run properly), another common technique to increase power is to increase boost on forced induction vehicles, which also reduces reliability. When it comes to that cheep ECU producing more "power" and MPG your actually burning cleaner than stock because of your advanced ignition timing, but mazda didn't know that, cause they where too cheap to pay somebody to actually tune the car.
I am not saying mazda engineers can't be out-engineered by the aftermarket, or even an individual, only that it is unlikely. Mazda engineers didn't tune the motor in a day, or even a month and they have way more dyno time with the BP motor than you (or anybody else probly).
This Link ECU has DLL software which allows you to log and perform dyno runs on the road and they have proven to be very effective.
The standard car runs too rich and many compromises are made for the various markets etc.
If you wish to keep telling yourselves that std is best then carry on - I meanwhile will carry on tinkering as I have for 20+ years and enjoying better, faster cars.
(Living in Solihull I know several of the test engineers from JLR so I'm aware what testing goes on and how they get their hands tied)
Now there's an argument not to get drawn in to... 
Accurate and timely knock detection is quite a difficult problem. Don't expect to solve it by reading PH, because if anyone has a good solution they'll probably be trying to make money from it and won't be saying much about it!
What's wrong with putting a very simple bandpass filter (need only be 1st order made from resistors and capacitors) between the ECU and the knock sensor as an experiment in signal conditioning? Have the centre frequency at around 6kHz and give it a pass band of say 500Hz.
Accurate and timely knock detection is quite a difficult problem. Don't expect to solve it by reading PH, because if anyone has a good solution they'll probably be trying to make money from it and won't be saying much about it!What's wrong with putting a very simple bandpass filter (need only be 1st order made from resistors and capacitors) between the ECU and the knock sensor as an experiment in signal conditioning? Have the centre frequency at around 6kHz and give it a pass band of say 500Hz.
V8RX7 said:
If you wish to keep telling yourselves that std is best then carry on - I meanwhile will carry on tinkering as I have for 20+ years and enjoying better, faster cars.
You do seem somewhat deluded. With your vast ability to produce such fast, powerful, reliable cars....maybe you should start building them and selling them. They're bound to sell well ! And as you can do it so cheaply, they'll be a bargain.Edited by stevieturbo on Tuesday 9th July 13:01
dnb said:
What's wrong with putting a very simple bandpass filter (need only be 1st order made from resistors and capacitors) between the ECU and the knock sensor as an experiment in signal conditioning? Have the centre frequency at around 6kHz and give it a pass band of say 500Hz.
Because in very simplistic terms:1) The Knocking pressure waves occur in the Crank Angle domain, but the output signal of the sensor is in the Time domain. As such, the average signal to noise ratio is poor if the signal processing is just done within a fixed Time domain.
2) The knock pressure waves are not "pure" they contain harmonics (in fact, a good percentage of the knocking signal energy is held within these higher orders)
3) The crank angle at which knock initiates and then "rings out" is not fixed, but varies with engine speed and load
4) The knock sensor has a high impedance output, a passive pass filter, with it's characteristic low impedance will need to be buffered otherwise a huge signal level reduction will occur.
5) Aftermarket systems make no correlation between knocking pressure (the bit that ACTUALLY matters) and signal level (inc noise)
6) Without windowing, background noise will not be rejected, leading to reduced signal to noise ratio as engine speed / load increases.
stevieturbo said:
V8RX7 said:
If you wish to keep telling yourselves that std is best then carry on - I meanwhile will carry on tinkering as I have for 20+ years and enjoying better, faster cars.
You do seem somewhat deluded. With your vast ability to produce such fast, powerful, reliable cars....maybe you should start building them and selling them. They're bound to sell well ! And as you can do it so cheaply, they'll be a bargain.I said new cars were a compromise - OTHERS said they were made the best they could.
I said they were tied by regulation - If I were to make new cars - I would have to adhere to those regs too !
I also said they were bound by price constraints.
Re read the thread then if you have any balls whatsoever you can apologise.
V8RX7 said:
I bet you got poor results in comprehension at school.
I said new cars were a compromise - OTHERS said they were made the best they could.
I said they were tied by regulation - If I were to make new cars - I would have to adhere to those regs too !
I also said they were bound by price constraints.
Re read the thread then if you have any balls whatsoever you can apologise.
Not in the slightest. You are the one saying you build better and faster cars than the OEM can do. Both are very relative terms.I said new cars were a compromise - OTHERS said they were made the best they could.
I said they were tied by regulation - If I were to make new cars - I would have to adhere to those regs too !
I also said they were bound by price constraints.
Re read the thread then if you have any balls whatsoever you can apologise.
They may be bound by price constraints, but they spend hundreds of millions where you spend a few hundred. There is no comparison.
And of course, they can make their knock sensors work ( most of the time )
Max_Torque said:
acteristic low impedance will need to be buffered otherwise a huge signal level reduction will occur.
5) Aftermarket systems make no correlation between knocking pressure (the bit that ACTUALLY matters) and signal level (inc noise)
6) Without windowing, background noise will not be rejected, leading to reduced signal to noise ratio as engine speed / load increases.
tit for tat aside, a slight segway if that's OK.5) Aftermarket systems make no correlation between knocking pressure (the bit that ACTUALLY matters) and signal level (inc noise)
6) Without windowing, background noise will not be rejected, leading to reduced signal to noise ratio as engine speed / load increases.
My MegaSquirt3 ECU has a 2d table for noise and also does full knock windowing.
Is there a good method for tuning the window without using a dyno (I'm way too cheap/poor/lazy and road tune myself - I'm well aware it's not the most efficient or optimised way of tuning).
The best method is what i would call the "two fuels" method:
Fill up with 95RON fuel.
Map fuelling and spark to produce boarderline knock (how and what you determine this is up to you! When doing things properly i'd normally map to a specific knocking pressure measured with in-cylinder pressure transducers, but this is not an aftermarket solution, so i'd suggest mapping to an approx audible level)
Experiment with moving the window position until the noise integrator value is maximised.
Refill with at least 98RON (>100 better to ensure no knock), rerun tests, and use the new integrator values as your noise floor (higher octain should ensure no knock occurs)
And even better approach is to use a 2 channel high speed logger (a pc sound card can be enough at 44kHz) to log both the raw knock sensor output (through a suitable high bandwidth op-amp buffer) and the Crank position sensor output (through a suitable voltage dividing buffer). Instigate knock through a low grade fuel and analyse data log to determine the knocking angle / window. If you have to artificially advance the ignition angle to initiate heavy enough knocking, then you should take account of this in your window position calcs.
Fill up with 95RON fuel.
Map fuelling and spark to produce boarderline knock (how and what you determine this is up to you! When doing things properly i'd normally map to a specific knocking pressure measured with in-cylinder pressure transducers, but this is not an aftermarket solution, so i'd suggest mapping to an approx audible level)
Experiment with moving the window position until the noise integrator value is maximised.
Refill with at least 98RON (>100 better to ensure no knock), rerun tests, and use the new integrator values as your noise floor (higher octain should ensure no knock occurs)
And even better approach is to use a 2 channel high speed logger (a pc sound card can be enough at 44kHz) to log both the raw knock sensor output (through a suitable high bandwidth op-amp buffer) and the Crank position sensor output (through a suitable voltage dividing buffer). Instigate knock through a low grade fuel and analyse data log to determine the knocking angle / window. If you have to artificially advance the ignition angle to initiate heavy enough knocking, then you should take account of this in your window position calcs.
Gassing Station | Engines & Drivetrain | Top of Page | What's New | My Stuff