Knock sensor Alarm for FastEddi

[EagleMark]

Ok so we are trying to get FastEddi to drive his car more then video and watch his laptop while tuning! :laugh:

So his main concern is Spark Knock! I found this little project from a link FSJ left in a thread and was having to much fun with the electronics available there. So here's the noise we need to plumb into his ALDL line or into TunerPro so when Knock is heard he knows.

The project has been done just need to get it set to his knock sensor?
http://www.sparkfun.com/tutorials/328



http://www.youtube.com/watch?v=2pv8CKto7wo&feature=player_embedded

[fasteddi]

:laugh:

Thats awsome!! Well im in the process of taking the old tired tq converter out for a new one that also stalls higher. So we go a few days to think.

Im actually woundering a littler here. Ok now can you explain to me the operation and what makes up the knock sensor? I sorta have some thoughs but I dont wana say anything until I understand it completly. I'll look into it and learn but I know you guys can give me a better explenation then the internet info.

I do ALOT of work on HVAC electronic so I have a decent grasp on solid state/electronic devices and monemtary switches/device and anything else that deals with heating/a.c equiptment.

Also I read up on that link and that is some really intresting stuff.

[EagleMark]

I'm just a GearHead, your going to have to find a uber nerd or geek type to explain Knock Sensors. But bottom line is they pick up a noise in frequency range expected for that engine and pull spark. Some just pull set amount, your's is good that it pulls from a range.

Wouldn't that be cool if we could get that knocker noise he used to come out of TunerPro during knock? :laugh:

I just happened to look around that link and saw some cool stuff then what this guy did to his door and the noise/Knocker just clicked. FastEddi wouldn't have to even watch for knock!

[RobertISaar]

how the knock sensor works? you can explain this in two ways: easier and not easy.

essentially, it's a very small and narrow banded microphone. when it is vibrated by the engine knocking(or what can be suspected as knock), the current limited 5V signal that is fed into it starts moving toward the 5V or 0V range (in a no knock condition, line voltage will be near 2.5 volts) due to the sensor resistance going from roughly 3900 ohms to more or less than 3900, causing the voltage fluctuation.

since the signal basically looks like an AC waveform using a scope, it's one way to think of it.

after the sensor itself is the knock filter, who's job is to further refine the signal coming out of the sensor so that only real knock gets detected as knock. when that happens, a signal is sent to the MCU (68HC11 in your case) and the processor will respond VERY quickly(within a few cycles, and at 2.1 million cycles per second, it is essentially instant) and start pulling timing the next time the timing calculation is done (80 times per second is a normal rate).

basically, the filter takes the roughly AC looking signal and when one of the peaks suddenly moves outside of the range that it's been seeing, that's considered knock and it will react. otherwise, a very slow change doesn't really trigger it.

[fasteddi]

how the knock sensor works? you can explain this in two ways: easier and not easy.

essentially, it's a very small and narrow banded microphone. when it is vibrated by the engine knocking(or what can be suspected as knock), the current limited 5V signal that is fed into it starts moving toward the 5V or 0V range (in a no knock condition, line voltage will be near 2.5 volts) due to the sensor resistance going from roughly 3900 ohms to more or less than 3900, causing the voltage fluctuation.

since the signal basically looks like an AC waveform using a scope, it's one way to think of it.

after the sensor itself is the knock filter, who's job is to further refine the signal coming out of the sensor so that only real knock gets detected as knock. when that happens, a signal is sent to the MCU (68HC11 in your case) and the processor will respond VERY quickly(within a few cycles, and at 2.1 million cycles per second, it is essentially instant) and start pulling timing the next time the timing calculation is done (80 times per second is a normal rate).

basically, the filter takes the roughly AC looking signal and when one of the peaks suddenly moves outside of the range that it's been seeing, that's considered knock and it will react. otherwise, a very slow change doesn't really trigger it.

So after the KR filter, the signal going to the mcu/ and processor is in the form of ac volts that fluctuate from 0-5vts? I was really hoping it was in DC. What I was thinking is how to run a wire off of that in parallel to another device that would sence that and beep when the vts got over a certen pre-determined range.
If 2.5vts is normal and the ohms drop or go up, how does it determine which way is KR? I though that it would only go one way for KR. Not fluctuate higher and lower from what the processor reads as "normal".

[PJG1173]

I wonder if mark at tunerpro could just write this into the next version to play a wav file when knock is detected. I know datamaster will ding at you when it detects errors in the datastream, IKD if it dings when you get knock also.

[EagleMark]

I don't think the question is could he do it.. it's would he do it?

Take into consideration he's a software writer who just happens to be a gearhead and got a Camaro around the time people were hacking GM ECMs and he was a player and built the coolest tuner software available and gives it away free. Asks for $30 donation. Then look at TunerCat and you see a bunch of programs for sale that all do a little piece of what TunerPro does, but never does it in a way truly totally customizable. Even then never does it all! Still need to buy DataMaster on top and a different one for each car...

Get more people to pay the $30 donation and we could see anything is my guess!

Look at windows software you can change any noise it makes to anything you want. Add it in the code, put a check box at each value to turn on sound and a check box to pick sound.

[EagleMark]

Mark at TunerPro responded to request with "Thanks. I’ve this feature on my list for several years. Hope to get to it!"

:jfj:

[PJG1173]

Get more people to pay the $30 donation and we could see anything is my guess!

he already got my $30 and I'll give him another $30 if I have to for the next version! Its money well spent.

[RobertISaar]

So after the KR filter, the signal going to the mcu/ and processor is in the form of ac volts that fluctuate from 0-5vts? I was really hoping it was in DC. What I was thinking is how to run a wire off of that in parallel to another device that would sence that and beep when the vts got over a certen pre-determined range.
If 2.5vts is normal and the ohms drop or go up, how does it determine which way is KR? I though that it would only go one way for KR. Not fluctuate higher and lower from what the processor reads as "normal".

after the filter, it goes through an IC (U13 on a 7727/7730/7749/similar), which seems to perform the function of being an inverting amplifier(turns 0 volts into 5 volts, or 5 volts into 0 volts, stuff like that) on the pins it is connected to. after that, the circuit goes directly to the U2 IC(also known as the P4 timer chip), where it will cause an interrupt to be generated and the MCU will service it.

keep in mind the unfiltered signal is also read as an A/D input and that there are actually two possible knock sensor schemes used on this ECM. there is a pin that goes to the connectors that is tapped into the filtered portion of the signal as well. it is labeled ESCIO# by Ludis Langden and is pin B8 on a 7730/7749. if you monitored that pin with a scope or maybe with a multimeter, you could probably figure out what kind of signal is present during knock and non-knock and maybe make some homebrew hardware to detect when it happens and set off an alarm.

also: that same pin can be used for the earlier seperate knock module type of system (like on a 1227165, 86-89 TPI MAF), so any information applying to it will also be useful here.

http://www.exatorq.com/ludis_obd1/1227749schematic.html

^^ that page is absolutely invaluable for stuff like this when it comes to the 7749/similar ECMs.

[fasteddi]

That is a absolutly amazing website! Thanks for linking it!

Alot of ladder diagrams I can make now to make the schematic more simple. I can read schematics well but ladder diagrams are easier in my mind for sequence of opperations.