Knock sensing ideas and circuit designs

From DIY contraptions to sophisticated FreeEMS-specific designs! Plus general hardware development!
Pilt
TO220 - Visibile
Posts: 9
Joined: Thu Nov 20, 2008 4:37 pm

Re: Knock sensing ideas and circuit designs

Post by Pilt »

OK. I confess

But I think the knock sensor IC's could be the way togo as they seem pretty sorted for the application.

As for DIY'ers getting them they can be obtained for free as samples from TI I believe (or maybe intersil).

I interfaced to one of these for my uni project but have not had it running on a real engine, but would like to.

The dificulty would be in setting these up but from what i have read about knock the bore is the main factor that effects the frequency of the knock signal from the transducer and knock can be simulated by hitting the block with a spanner (so im told). Anyways i imagine anyone with access to a scope could set the chip up for a sensor (with guidence). For those that dont have a scope i imagine it *may* be possible to write a calibration routine that advances the ignition at idle till knock occurs and the IC's settings are altered until the knock is detected correctly. Although this may not be as good at high rpm (due to engine noise) it may help determine the knock freq of the engine. There is a app note for the dev board that may be useful.. http://www.intersil.com/data/an/an9770.pdf

Also looking into the VEMS developent they use this IC so they may be able to help, from what they say they dont use the analog output so maybe it can be read over SPI..

Another thing we do on our older ECU's which i think could be done on the FreeEMS ecu is that a lot of the waveform based sensor inputs are connected to a analogue mux. With code the mux is selected to the appropriate channel and this is routed to an output pin on the ECU. All you have to do then is connect a scope to this channel and you can monitor what the ECU is seeing without having to break into the harness.


P.S Microsoft is a branding hike, we may use there software but they basically do nothing for us!

Cheers

Dan
User avatar
jbelanger
LQFP144 - On Top Of The Game
Posts: 387
Joined: Sat Feb 23, 2008 8:58 pm
Contact:

Re: Knock sensing ideas and circuit designs

Post by jbelanger »

Pilt wrote:As for DIY'ers getting them they can be obtained for free as samples from TI I believe (or maybe intersil).
I wasn't talking about the HIP9011 (which I have samples of) but about the Infineon ICs you were talking about.

Jean
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've got "seeing the waveforms" covered already.

Bed time.
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
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 »

Having re read that, I don't have it covered in quite the same way, but I'm not sure that's necessary on a system like this? Breaking into the box with four screws (or whatever) and attaching a probe to the board or connector should be sufficient if my digital logic analyser style output isn't enough.

I doubt VEMS would want to help, this will be cutting into their profit margins eventually.

I forgot that these were SPI. Still, nothing says someone else can't write SPI code for one and submit it to me to merge in!

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
Movex
TO220 - Visibile
Posts: 6
Joined: Thu Oct 02, 2008 4:54 pm

Re: Knock sensing ideas and circuit designs

Post by Movex »

To make knock detection available on a DIY device the trace data should be accessible by serial port or something similar. Not everyone has got a scope to do this and for saving a trace signal you need a expensive scope.
As proposed by Dan the waveform should be sampled by the MCU itself for this reason.

Attached a example from a PowerPC based ECU where the raw signal and the filtered signal are processed internally and sent to the software running on a notebook.


Image
L98TPI
TO220 - Visibile
Posts: 11
Joined: Thu Aug 14, 2008 4:59 am

Re: Knock sensing ideas and circuit designs

Post by L98TPI »

The TPIC8101 / HIP9011 is on my board, well, will be in the next turn. I saw it as a easy path to basic knock sensing. Delayed now since I got the layoff. Have time but funds are tight. The wife would have a fit if I spent on PCB's now.

Cylinder by cylinder knock sense may be massive over kill. I agree with (insert your name here :) ) that knock sense should not be used as a loop control parameter, more a fail safe, from the diyefi point of view. I am going to get all the other stuff running first, but I want it on the board now.

:shock: Yea I know long time no hear, been busy with actual work stuff.... well till the cutback, gov contracts get cut when the admin changes.

Mark
Any tool can be the right tool...
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 »

L98TPI wrote:The TPIC8101 / HIP9011 is on my board, well, will be in the next turn. I saw it as a easy path to basic knock sensing. Delayed now since I got the layoff. Have time but funds are tight. The wife would have a fit if I spent on PCB's now.

Cylinder by cylinder knock sense may be massive over kill. I agree with (insert your name here :) ) that knock sense should not be used as a loop control parameter, more a fail safe, from the diyefi point of view. I am going to get all the other stuff running first, but I want it on the board now.

:shock: Yea I know long time no hear, been busy with actual work stuff.... well till the cutback, gov contracts get cut when the admin changes.

Mark
Yep i know how you feel fella after being made redundant I'm not able to do much with the car other than basic maintenance and do the never ending and soul destroying job hunts in a recessive economy. I have however had the time to play around with 18F4550 PIC and learn a bit of assembly language.
At the moment i'm making a PIC based DAQ to capture analogue and digital signals and log them in LabView the idea being to piggyback onto the standard ECU and capture the I/O of the oem ecu and form a basis for a base map for FreeEMS and my MS2v3.00. I have had breadboard based circuit sending results back to a VB GUI all I have to do now is write a DLL for labview to call and voila i should have a fairly capable DAQ for the task of sampling the knock signals and monitoring how the OEM reacts to them.
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 »

I've decided to use the TPIC chip because it can be obtained from Newark. Both chips are about the same. The TPIC chip has some items that I should resolve before getting too far along in the schematic capture.

One is the Xin / Xout thing. I plan to copy the application schematic on the data sheet, but I don't understand what this should be connected to. GPIO, DAC, ADC, other? Based on it's functional description, it sounds like it's used to make small adjustments in the internal oscillator. However the block diagram shows it as an inverted op-amp. I don't fully understand this feature.

The other place I'm a bit ignorant is if we should use one of the two available channels, or both channels. This chip has two input channels, but only one can be sampled at a time. The only use for this feature I see is for lowering cost. Saves a buck or two by using one chip for two sensors. I plan to use one chip per knock sensor, so only channel 1, and no channel 2. Does that sound right, or do we perhaps want both knock inputs?

The SPI buss allows multiple devices with the use of the CS line, so we can have multiple chips with little effort from the MCU.
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 »

Err???? looking at the data sheet xin and xout are to do with the timing crystal for the on chip clock. The addition of a crystal and 2 caps to ground off of xin and out will provide an on chip time base.
The internal symbol is that of a NOT gate or inverting buffer.
The Xin probably has a tap off internally to take the clock frequency to the internal workings of the chip.

The Xin can also be coupled directly to the MCU clock so that they are in sync and speakedy the same lingo.

As for the channel select yes you can only sample one channel at a time by sending the select bit to it Via SPI bus but you should only really need to sample 1 at a time unless you have a V12 or something with 4 sensors on 2 banks of 6 then you'll need 2 chips anyway.

Hope that helps.
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 »

My plan is for 4 knock sensors with 4 chips. I think this will cost a pinch more $ in hardware, but will make for easier / more reliable software, while maintaining a more flexible design. I expect typically you'll need 1 or 2 knock sensors, the other two are for rare situations where you may have a deaf block.

I'm planning for one chip per knock because it's easier to keep track of in software. We need to chip select the SPI bus as is, might as well add the two extras in there.

About the clock, I see it now. I confused the other chip (with internal clock option) with this chip. This chip requires it's clock cycles from an external chip, up to 24mhz. I guess we'll have to keep in mind where that signal is going to come from.

About R1 and R2 from the application schematic, I think these will need to be determined on a per sensor basis. Can we make a prediction about what values we should have in there as default, or perhaps what values we should have for a specific sensor?

Can we safely ignore the test line? I believe the chips will come with a proper default config for our needs, so I don't see us wanting to enter this mode. It has it's own internal pull up, and can be jumper-ed to GND if someone wanted to experiment with some exotic features.

See attached draft PDF of the knock schematic.
Attachments
knock.pdf
(20.12 KiB) Downloaded 1202 times
Post Reply