Knock sensing ideas and circuit designs

From DIY contraptions to sophisticated FreeEMS-specific designs! Plus general hardware development!
davebmw
LQFP144 - On Top Of The Game
Posts: 331
Joined: Sun Jul 13, 2008 2:58 pm
Location: South Wales, UK

Re: Knock sensing ideas and circuit designs

Post by davebmw »

No biasing is required whatsoever, the piezo element generates electricity when it is shocked or moved in the manner to which its tuned.
its used just like a microphone a tuned one at that.
Biasing is for Fault logging of the sensor element and adjoining cables only.
93'BMW 325is M50B25TU, Rebuilt 06/06, JE10.5:1, polish&port. Scorpion BB, K&N CAI, TEJ21 WBO2, '07 M3 Evo 18" 225F, 255R, EBC Kevlar, Bilstien Sprint, Polyflex. Head rebuild Oct'08, OEM+FSE FPR, MS2v3.0_DJB Custom, Extra 2.0.1
User avatar
jharvey
1N4001 - Signed up
Posts: 1607
Joined: Tue Jun 10, 2008 5:17 pm

Re: Knock sensing ideas and circuit designs

Post by jharvey »

The more I read about the TPIC8101, the more I like it. Fairly low cost (couple bucks), allows lots of flexibility, multi vendors, backed by a large company. If we ever had trouble obtaining it, we could do the same thing with a AVR, so supply should never be a problem here.

It offers the physical connection to two knock sensors, allowing for digital or analog data back to the MCU.

So we can use it local, than connect the AN out line to Dave's windowing circuit. Which means every install will require the installer to be good at the black art of noise reduction. However it's much easier to develop now and get a working version up and running.

Or we can put it remote, run the SPI lines back and do the window / detect in software. This will take much longer before software is stable and usable, however, it removes almost all of the black magic, making for easier and less problematic installs.

I think we really want to use this chip as our plan for both PCB installs, and remote installs. Both setups will require SPI interfacing, and the remote install is little more that keeping space for a header for SPI and perhaps power. I think this chip makes it easy to do both local and remote.
User avatar
jharvey
1N4001 - Signed up
Posts: 1607
Joined: Tue Jun 10, 2008 5:17 pm

Re: Knock sensing ideas and circuit designs

Post by jharvey »

About biasing and not needing power on a microphone, I'm assuming you refer to non-phantom power Mic's. Most Mic's these day are phantom powered requiring bias.

I kind of thought the Mic wouldn't need it, but does it produce a better signal with a bias? By measuring the noise floor, that should tell you if the sensor is working, but doesn't tell you why it failed, if it fails. I would tend to think we can use a meter to figure out why it failed, after the brain knows it has failed. So perhaps biasing isn't important to think about. I still wonder if a better quality signal could be obtained with a bias.
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: Knock sensing ideas and circuit designs

Post by Fred »

SPI is a short distance interface only.

Fred.
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
User avatar
jharvey
1N4001 - Signed up
Posts: 1607
Joined: Tue Jun 10, 2008 5:17 pm

Re: Knock sensing ideas and circuit designs

Post by jharvey »

The distance is dependant on the data rate. I've seen estimates that claim 100K should allow for around 28ft, or about 9m. However, I haven't found something that shows actual data rates vs distances obtained. Could also use this device to extend it if required.

http://www.avrcard.com/products/can_adapter.htm

Hmmm if only I could find that CAN knock sensor chip.
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: Knock sensing ideas and circuit designs

Post by Fred »

I'll tell you this right now, it will be a long time before CAN or SPI or I2C are supported in a clean releasable way. Focus on direct interfaces, we can expand later...
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
davebmw
LQFP144 - On Top Of The Game
Posts: 331
Joined: Sun Jul 13, 2008 2:58 pm
Location: South Wales, UK

Re: Knock sensing ideas and circuit designs

Post by davebmw »

Phantom powered? I assume you mean what we call electret microphones that require a supply of 1.5 to 3volts to power the internal amplifier.

Piezo mics and knock sensors generate their own voltage through the piezoelectric effect. in the knock sensor the transducer is mechanically trimmed and tuned to respond only to the frequency band required in this case 5KHz to 22KHz +-3dB with the largest knock component between 6 and 7KHz depending on the size of the combustion chamber.

I see what you are saying by biasing the element to offset the noise floor you could strip away a lot of unwanted crap, but in all honesty I don't think that there is that much there as the sensor is tuned to reject most of the noise, mostly produced by the alternator and ignition/injection system.
As long as care is taken to correctly screen the cables at the ecu end only we should be in a good position to detect what we want.
I am adamant that the noise i was seeing on my Hameg CRO is as a direct result of me rushing to measure the knock by clipping the scope probes onto a spare bosch connector with an unscreened cable.

In the interests of automotive science, this afternoon when my wife gets back from work i will have a go at plugging the MS2 in and connect the knock sensors directly to the line in on my laptop. then by introducing some particularly weak mixture and a little too much advance I should be able to log the Knock in stereo as a wave file.
With a stethoscope i should be able to hear mild knock and note the time index of the wave file giving us a wealth of data to post up and sift through. (one proviso though the weather looks like its improving if it starts raining forget it!)

Also found this article too: http://www.siliconchip.com.au/cms/A_108910/article.html
you have to buy the article but the PCB overlay is free to download on the site, no schematic or BOM though. what i can make out from the picture its a dual op amp arranged as a low pass filter and a high pass filter no windowing whatsoever nothing clever, just an output that goes high when knock is detected.

I have been looking at the clever IC's too the TPIC8101 looks very similar if not the same as the HIP9011 from intersil. trouble with these is we would need to interface to the XDP that is a little down the road yet but would be the golden solution when we get to that point. it may be worth noting that neither of these IC's have a bias on the peizo element either.

I am wondering how easy it would be to have a pulse coming out of the XDP for every 6 degrees of crank rotation, this way we could much more accurately place the detect window based on crank angle. is this what i have heard referred to as angle clock??
93'BMW 325is M50B25TU, Rebuilt 06/06, JE10.5:1, polish&port. Scorpion BB, K&N CAI, TEJ21 WBO2, '07 M3 Evo 18" 225F, 255R, EBC Kevlar, Bilstien Sprint, Polyflex. Head rebuild Oct'08, OEM+FSE FPR, MS2v3.0_DJB Custom, Extra 2.0.1
davebmw
LQFP144 - On Top Of The Game
Posts: 331
Joined: Sun Jul 13, 2008 2:58 pm
Location: South Wales, UK

Re: Knock sensing ideas and circuit designs

Post by davebmw »

jharvey wrote:The distance is dependant on the data rate. I've seen estimates that claim 100K should allow for around 28ft, or about 9m. However, I haven't found something that shows actual data rates vs distances obtained. Could also use this device to extend it if required.

http://www.avrcard.com/products/can_adapter.htm

Hmmm if only I could find that CAN knock sensor chip.
That kind of data rate would be very difficult under the bonnet/hood as there is a lot of EMI floating around, that is why CAN was invented.
93'BMW 325is M50B25TU, Rebuilt 06/06, JE10.5:1, polish&port. Scorpion BB, K&N CAI, TEJ21 WBO2, '07 M3 Evo 18" 225F, 255R, EBC Kevlar, Bilstien Sprint, Polyflex. Head rebuild Oct'08, OEM+FSE FPR, MS2v3.0_DJB Custom, Extra 2.0.1
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: Knock sensing ideas and circuit designs

Post by Fred »

Dave, if you haven't seen this, you should read it :

http://www.diyefi.org/forum/viewtopic.php?f=35&t=193
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
davebmw
LQFP144 - On Top Of The Game
Posts: 331
Joined: Sun Jul 13, 2008 2:58 pm
Location: South Wales, UK

Re: Knock sensing ideas and circuit designs

Post by davebmw »

Here is the program I'm using to capture the signal it's very useful not just a scope but a sig gen and spectrum analyser too!

http://zeitnitz.de/Christian/Scope/Scope_en.html
LVScope.jpg
LVXYScope.jpg
LVSpecAn.jpg
93'BMW 325is M50B25TU, Rebuilt 06/06, JE10.5:1, polish&port. Scorpion BB, K&N CAI, TEJ21 WBO2, '07 M3 Evo 18" 225F, 255R, EBC Kevlar, Bilstien Sprint, Polyflex. Head rebuild Oct'08, OEM+FSE FPR, MS2v3.0_DJB Custom, Extra 2.0.1
Post Reply