Electronic 'detonation cans'
#1
Elite Member
Thread Starter
iTrader: (15)
Join Date: Dec 2007
Location: San Antonio, Texas
Posts: 4,847
Total Cats: 27
Electronic 'detonation cans'
So I finally installed the Knocksense unit in my car and hooked it up to my MS. As many have noted it is tough to tune the threshold of the Knocksense device. I followed the instructions on the initial setup: idle the car, turn the adjuster pot until the knock LED comes on, then back off 15-20 degrees.
But, how do you REALLY know the sensitivity is set correctly? Too high, and it triggers when there is no knock. Too low, and it may not trigger when there is knock. If one could listen for knock, preferably in a manner other than just listening to the engine (bad knock at that point) then one would have a tool for setting the knocksense sensitivity.
One such solution is detonation cans, or 'det cans'. Info on this can be found on the net, but basically you are bolting a stethoscope to the engine. It works, it is simple, good for the dyno maybe, but not that great for driving around.
Another approach is to do this electronically. I have seen some around the net either use a knock sensor (existing or installed) as a microphone, and using a set of computer speakers in the cab to amplify the signal. Again, it works, but who wants computer speakers in their small-assed Miata cab?
So, I started thinking... hmmm... I have an unused line input on my car stereo. What if I use that to amplify the signal? So I went for it, and tried out the idea. I used the existing knocksense connections to tee in a RCA line signal cable. I just cut off the end to expose the conductors, added a wire to extend the shield, and hooked it up to the MS. The other end plugged into the line input on my stereo.
It works!
I started the car and turned the stereo all the way up and viola, I hear the sound of the engine ticking away. I then took it for a drive to see how well it worked. I took my spark map, and added some timing across the board to see if I could induce some knock. I gave it partial throttle at low to mid RPMs (gently). Sure enough, I could hear the knock through the stereo. I have heard it described as the sound of BBs rattling in a metal can which is a fair description. It is clearly audible above the rest of the engine noise, and has a random sound, not the regular ticks of the engine. Normal timing, no knocks. Advance the timing 5-6 degrees, audible knocks. Return the timing to normal, no knocks. Repeatable.
And, better yet, when I turned down the stereo, I could not hear the knock coming from the engine under the same conditions that produced clearly audible knocking sounds on the stereo. This means you can hear the knock in the stereo before you are able to hear it otherwise. Thus, it is a more sensitive tool for hearing knock then just listening to the motor from the cab.
There is some ignition noise in the sound, like what you hear when you have a grounding problem with the stereo. I am guessing it is just coming from inductive pickup from the ignition system, or some ground loop thing. But, it is not loud enough to obscure the engine sounds, and the knock is clearly audible over the ignition noise.
So, I still have some knocksense tuning to do. But now I have an easy, effective tool to do it with. And, even if you do not have a knocksense, one could use this to check the aggressiveness of your timing map without blowing up your ****. Sweetness. Knocksense going off? Thought you heard a knock? Filled up with some low grade shitty gas? Switch the stereo over to the aux input and give it a listen. Easy.
But, how do you REALLY know the sensitivity is set correctly? Too high, and it triggers when there is no knock. Too low, and it may not trigger when there is knock. If one could listen for knock, preferably in a manner other than just listening to the engine (bad knock at that point) then one would have a tool for setting the knocksense sensitivity.
One such solution is detonation cans, or 'det cans'. Info on this can be found on the net, but basically you are bolting a stethoscope to the engine. It works, it is simple, good for the dyno maybe, but not that great for driving around.
Another approach is to do this electronically. I have seen some around the net either use a knock sensor (existing or installed) as a microphone, and using a set of computer speakers in the cab to amplify the signal. Again, it works, but who wants computer speakers in their small-assed Miata cab?
So, I started thinking... hmmm... I have an unused line input on my car stereo. What if I use that to amplify the signal? So I went for it, and tried out the idea. I used the existing knocksense connections to tee in a RCA line signal cable. I just cut off the end to expose the conductors, added a wire to extend the shield, and hooked it up to the MS. The other end plugged into the line input on my stereo.
It works!
I started the car and turned the stereo all the way up and viola, I hear the sound of the engine ticking away. I then took it for a drive to see how well it worked. I took my spark map, and added some timing across the board to see if I could induce some knock. I gave it partial throttle at low to mid RPMs (gently). Sure enough, I could hear the knock through the stereo. I have heard it described as the sound of BBs rattling in a metal can which is a fair description. It is clearly audible above the rest of the engine noise, and has a random sound, not the regular ticks of the engine. Normal timing, no knocks. Advance the timing 5-6 degrees, audible knocks. Return the timing to normal, no knocks. Repeatable.
And, better yet, when I turned down the stereo, I could not hear the knock coming from the engine under the same conditions that produced clearly audible knocking sounds on the stereo. This means you can hear the knock in the stereo before you are able to hear it otherwise. Thus, it is a more sensitive tool for hearing knock then just listening to the motor from the cab.
There is some ignition noise in the sound, like what you hear when you have a grounding problem with the stereo. I am guessing it is just coming from inductive pickup from the ignition system, or some ground loop thing. But, it is not loud enough to obscure the engine sounds, and the knock is clearly audible over the ignition noise.
So, I still have some knocksense tuning to do. But now I have an easy, effective tool to do it with. And, even if you do not have a knocksense, one could use this to check the aggressiveness of your timing map without blowing up your ****. Sweetness. Knocksense going off? Thought you heard a knock? Filled up with some low grade shitty gas? Switch the stereo over to the aux input and give it a listen. Easy.
Last edited by ZX-Tex; 06-12-2008 at 10:36 PM.
#3
Elite Member
Thread Starter
iTrader: (15)
Join Date: Dec 2007
Location: San Antonio, Texas
Posts: 4,847
Total Cats: 27
Yes I agree. Though I mentioned using the Knocksense connections, really all I am doing is connecting to the signal coming from the existing knock sensor. So, if one does not have a knocksense, find the knock sensor wire going to the ECU, tee into that, and connect it to an RCA cable, with a ground connection nearby for the cable shield.
EDIT: Let me qualify that. The stereo has high impedance inputs so I think there is little risk of damage here. But I have no idea if tapping into the stock sensor line is OK to do with a stock ECU. I am guessing it is also a high impedance input but don't know that for sure, so try it at your own risk.
EDIT: Let me qualify that. The stereo has high impedance inputs so I think there is little risk of damage here. But I have no idea if tapping into the stock sensor line is OK to do with a stock ECU. I am guessing it is also a high impedance input but don't know that for sure, so try it at your own risk.
Last edited by ZX-Tex; 06-13-2008 at 12:31 AM.
#5
Former Vendor
iTrader: (31)
Join Date: Nov 2006
Location: Sunnyvale, CA
Posts: 15,442
Total Cats: 2,100
For those of us without stock knock sensors, I set my KS up by filling up with 2 gallons of 100 octane on an empty tank. I then did 3-4-5 pulls with my friend adjusting the knock sensor so it just BARELY wasn't going off. I then filled up with 91, and voila, it picks up every single slight hint of knock and nothing that isn't.
#8
Elite Member
Thread Starter
iTrader: (15)
Join Date: Dec 2007
Location: San Antonio, Texas
Posts: 4,847
Total Cats: 27
The previously mentioned electrical anomalies were unrelated so I am cleaning up this thread accordingly to remove the unrelated side issue.
Last edited by ZX-Tex; 06-16-2008 at 09:34 AM.
#12
Elite Member
Thread Starter
iTrader: (15)
Join Date: Dec 2007
Location: San Antonio, Texas
Posts: 4,847
Total Cats: 27
I'm thinking an audio isolation transformer like this might do the trick? These can also be bought pre-packaged from Radio Shock. I have one at home, somewhere...
http://www.circuitsarchive.org/index...Audio_Isolator
Thoughts?
http://www.circuitsarchive.org/index...Audio_Isolator
Thoughts?
#13
Elite Member
Thread Starter
iTrader: (15)
Join Date: Dec 2007
Location: San Antonio, Texas
Posts: 4,847
Total Cats: 27
Here is a simple op-amp based amplifier for knock sensor audio amplification. I would think the op-amp would provide very good isolation since their input impedance is typically several megohms IIRC. If the gain is set very low with the pot I am guessing it would be OK for a line input jack?
Scroll to the bottom of the page:
http://home.netcom.com/~bsundahl/kno...n.htm#Recorder
Thoughts?
Scroll to the bottom of the page:
http://home.netcom.com/~bsundahl/kno...n.htm#Recorder
Thoughts?
#15
Go to the shack and get one of these:
277-1008
http://www.radioshack.com/sm-buy-the...i-2062620.html
Its a stand alone batt power amp, should isolate it from the car.
277-1008
http://www.radioshack.com/sm-buy-the...i-2062620.html
Its a stand alone batt power amp, should isolate it from the car.
#16
Elite Member
iTrader: (22)
Join Date: Dec 2006
Location: Sunny Spanish speaking Non US Caribbean
Posts: 3,224
Total Cats: 3
ZX, I've got no idea about electronics but I really want you to succeed with this one.
BTW, I see that you're running your MS in parallel while I on the other hand am running the AEM standalone; will that make any difference?
Thanks for trying to fix this issue
BTW, I see that you're running your MS in parallel while I on the other hand am running the AEM standalone; will that make any difference?
Thanks for trying to fix this issue
#17
Elite Member
Thread Starter
iTrader: (15)
Join Date: Dec 2007
Location: San Antonio, Texas
Posts: 4,847
Total Cats: 27
After running around this morning looking at various iPod type MP3 amplifier gadgets I decided to go with the radio shack device that Mike mentioned above. Though I did want to integrate this into the stereo, the little amplifier unit is small, has a volume control, and fits nicely in the glove box. The battery power isolates it well from the vehicle as Mike pointed out, so the only thing the knock sensor and the knocksense sees is the 5kohm load for the input. I made a simple twisted-pair suicide cable terminated with a 1/8" mono jack, connected it up to the knocksense same as before, plugged the other end into the amp, and mounted the amp to the back of the glove box with some velcro. When I idled the car and turned on the amp, it was as audible as the stereo setup was (on one channel). The ignition noise was a little quieter and the engine noise (lifters, etc.) was clearly audible. And, no electrical anomalies.
I have not road tested this yet but I expect it to work well. I'll update this thread after I have tried it out some more. So basically the idea is to leave it connected in the glove box, turned off. It is small so it does not take up much room. To use it, just reach over, open up the glovebox and turn on/up the volume control. Easy.
It has an earphone jack so that could probably be used to connect it to the stereo again. But as long as it is not in the way and the battery lasts awhile (it should) then I'll just leave it as-is.
I have not road tested this yet but I expect it to work well. I'll update this thread after I have tried it out some more. So basically the idea is to leave it connected in the glove box, turned off. It is small so it does not take up much room. To use it, just reach over, open up the glovebox and turn on/up the volume control. Easy.
It has an earphone jack so that could probably be used to connect it to the stereo again. But as long as it is not in the way and the battery lasts awhile (it should) then I'll just leave it as-is.
#18
Elite Member
Thread Starter
iTrader: (15)
Join Date: Dec 2007
Location: San Antonio, Texas
Posts: 4,847
Total Cats: 27
I would not think so. The idea is to tap into the knock sensor signal with a high impedance connection so that it does not alter the existing circuitry; if the impedance is high enough it is effectively like not having anything connected at all. I would try it with something like the radio shack amp above, battery powered (not connected to the car for power). BUT, YMMV, go forth at your own risk, blah blah.
#19
Elite Member
Thread Starter
iTrader: (15)
Join Date: Dec 2007
Location: San Antonio, Texas
Posts: 4,847
Total Cats: 27
OK well the Radio Shack amplifier has a problem. It definitely impacts the sensitivity of the knocksense. I had a chance to do some more testing today and it was clear that the knocksense was much more sensitive with the amplifier unplugged. I tried to compensate by raising the sensitivity of the knocksense while plugged into the amplifier but from what I could tell the performance was not consistent.
I measured the impedance across the knocksense input to knocksense ground and it is on the order of 250Kohms. Thus it is quite possible that the amp does change the behavior of the knocksense as I observed since the amp has a much lower impedance.
On a positive note I tried connecting the earphone output from the amplifier to the aux input on the stereo using a 1/8" mono to dual line output adapter. That worked very well since it preamplified the input into the stereo. In fact, ignoring the knocksense impedance problem, this still makes a good tool for listening for knock. I bumped the timing up and down in 2 deg increments to induce knock (very carefully) and it was readily detectable over the stereo. In fact, for my normal map, I am reducing the timing on the famous Ben's timing map by 2 degrees since there was some faint audible knocking at those advance levels.
I measured the impedance across the knocksense input to knocksense ground and it is on the order of 250Kohms. Thus it is quite possible that the amp does change the behavior of the knocksense as I observed since the amp has a much lower impedance.
On a positive note I tried connecting the earphone output from the amplifier to the aux input on the stereo using a 1/8" mono to dual line output adapter. That worked very well since it preamplified the input into the stereo. In fact, ignoring the knocksense impedance problem, this still makes a good tool for listening for knock. I bumped the timing up and down in 2 deg increments to induce knock (very carefully) and it was readily detectable over the stereo. In fact, for my normal map, I am reducing the timing on the famous Ben's timing map by 2 degrees since there was some faint audible knocking at those advance levels.
#20
Elite Member
Thread Starter
iTrader: (15)
Join Date: Dec 2007
Location: San Antonio, Texas
Posts: 4,847
Total Cats: 27
Well even though this seems to be a conversation of one at this point, I'm still looking for some help on creating a high-impedance pre-amp for this. As mentioned the Radio Shack amp only has a 5K ohm impedance. Last night I measured the input impedance on my stereo and it was only 10K ohm. I would like to get the input impedance up much higher than the 250K ohm impedance of the knocksense so that this setup does not alter the knocksense gain.
Any analog circuit designers out there got any ideas?
Any analog circuit designers out there got any ideas?