www.gt4dc.co.uk
Maintain, Modify and DRIVE your GT-Four


It is currently Fri Mar 29, 2024 1:23 am




Post new topic Reply to topic  [ 19 posts ]  Go to page Previous  1, 2
Author Message
PostPosted: Mon Jan 06, 2014 2:43 pm 
Offline
Club Staff
User avatar

Joined: Wed Jul 13, 2005 6:54 pm
Posts: 3484
Location: On Top
Car Model: None
sorry - I was in the middle of editing my post!

brain fart part deux...


3) If knock can occur at 0-70 deg crank angle on the power stroke then assuming all 4 cylinders are being monitored then we have a knock detection window of a total of 280 degrees out of 720 degrees engine cycle - this is a large amount of time and I wonder if it will allow sufficient monitoring of regular engine noise. What I did find interesting however is that the cylinder nearest the knock sensor (PRESUMABLY) has considerably higher knock sensor o/p than the others, as such the "threshold" is different for each cylinder. This then raises the question about individual cylinder knock monitoring/thresholds or single cylinder analysis. using a single cylinder for analysis may however mean that my initial bright idea of using the differential between regular engine noise and knock event noise is not possible as an attenuated knock event from the other cylinders would be picked up.

Other engine noise events are a major concern especially for any gravel rallying etc! The natural frequency of a steel plate (sump or sump guard?) is 12700Hz/mm so a 1.8mm steel plate has a similar natural frequency as the knock sensor! (12900Hz/mm for aluminium so it is still very close.

_________________
JP
GT4DC Treasurer

Grey St185 RC
Silver BMW E46 M3
Multicolour yawn Honda VFR400
Silver Honda Civic Type-R


Top
 Profile  
 
PostPosted: Mon Jan 06, 2014 6:12 pm 
Offline
Group B
User avatar

Joined: Sat Sep 02, 2006 3:13 pm
Posts: 3679
Location: Bournemouth
Car Model: None
The knock chip integrates over the 'knock window' period, so all time periods within this have equal sensitivity, so it's not possible to compare 'knock window' vs 'non knock window'.

In my opinion, a peak detect would be far more effective as knock is, theoretically, a very short burst of high output from the sensor. Therefore 10mS at 0.1V output (engine noise) would be equivalent to 1mS at 1V (knock) leaving a very poor signal to noise ratio (1:1) when compared to peak detect (10:1)

My reasoning for proposing this chip is that firstly 'this is how it's normally done, surely they can't all be wrong' and secondly the sensor is so resonant that a knock event will make the output large for most of the knock window thus going some way towards making it a peak detect anyway.

Thinking about it a bit more, I may look into doing my own DSP function using a dsPIC and going for peak detect during knock window only, with extra weighting in the area most likely to have knock.

My original knock unit was peak detect only without knock window, and seems to work fairly successfully in terms of lighting up like a christmas tree whenever I can hear det. (window open, next to armco). I haven't done any precise testing on a dyno as yet hence I haven't marketed it actively.

Quote:
The TPIC8101 is designed for knock sensor signal conditioning in automotive applications. The device is an
analog interface between the engine acoustical sensors or accelerometers and the fuel management systems
of a gasoline engine. The two wide-band amplifiers process signals from the piezoelectric sensors. Outputs of
the amplifiers feed a channel select mux switch and then a 3rd order antialiasing filter. This signal is converted
using an analog-to-digital conversion (10 bits with a sampling frequency of 200 kHz) prior to the gain stage.
The gain stage is adjustable via the SPI to compensate for the knock energies. The gain setting is selectable
up to 64 values ranging from 0.111 to 2.0.
The output of the gain stage feeds a band-pass filter circuit to process the particular frequency component
associated with the engine and transducer.
The band-pass filter has a gain of two and a center frequency range between 1.22 kHz and 19.98 kHz (64-bit
selection). The output from this stage is internally clamped.
The output from the band-pass filter is full-wave rectified with its output clamped below VDD.
The full-wave rectified signals are integrated using an integrator time constant set by the SPI and integration
time window set by the pulse width of INT/HOLD. At the start of each knock window, the integrator output is reset.
The output of the integrator is internally clamped and the digital output may be directly interfaced to the
microprocessor.
The integrated signal is converted to an analog format by a 10-bit DAC. The microprocessor may interface to
this signal, reads this data, and adjusts the spark ignition timing to optimize fuel efficiency related to load versus
engine RPM.

_________________
If at first you don't suck seed, try drier grain.

Image


Top
 Profile  
 
PostPosted: Mon Jan 06, 2014 10:50 pm 
Offline
Club Staff
User avatar

Joined: Wed Jul 13, 2005 6:54 pm
Posts: 3484
Location: On Top
Car Model: None
I see your reasoning and on the face of it would agree with your thoughts on using the peak value.

My only possible concern is the sampling frequency used for the signal analysis. the 200 kHz sample freq used in the off-the-shelf chip initially (and without running through the maths) seems to offer low levels of resolution for peak detect? At 8K RPM you have the crank rotating at 48000 degrees per second that will give you a resolution of around 1/2 degree? considering the highly resonant nature of the Toyota knock sensor it is assumed that it also has a very low damping ratio and therefore will continue to "ring" after the knock event and hence basic knock detection of a circa 6-7kHz event will not be an issue and well above the Nyquist requirements. Is 1/2 degree crank rotation resolution enough for engine health monitoring? I guess running through it now after my waffle - probably!

_________________
JP
GT4DC Treasurer

Grey St185 RC
Silver BMW E46 M3
Multicolour yawn Honda VFR400
Silver Honda Civic Type-R


Top
 Profile  
 
PostPosted: Sun Feb 16, 2014 9:07 am 
Offline
Group N

Joined: Wed Oct 20, 2010 8:37 pm
Posts: 285
Location: uk southampton
Car Model: ST205
i was looking for the hks module (dis/cont) but could only find this http://www.ebay.co.uk/itm/Kit-Cars-Knoc ... 2c7b3ac33f


Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 19 posts ]  Go to page Previous  1, 2


Who is online

Users browsing this forum: No registered users and 23 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum

Search for:
Jump to:  
cron
Powered by phpBB® Forum Software © phpBB Group