Messing with timing...
You should set the range for your timing a lot smaller so you can actually see when timing is being pulled by the ECU. Make the upper and lower Y-axis like 0-25. Any time you have a spike in knock sensor voltage look at the timing map there. If there is any timing at all pulled, it is knock.
I've made many threads about this a while back and I have PM'd Tom about it, too, but knock voltage is proportional to RPM and you can't just use a set value and call any number under that safe and any number above that knock. It is hard to tell by just looking at the knock sensor voltage, but any rapid increases (spikes) are most likely knock and if you look at the corresponding timing, you can tell if the ECU thinks so, too.
The stock ECU pulls about 1 degree of timing for every 3 'counts' of knock, at least in the DSM ECU. This doesn't really correlate with any preset voltage...it's any spike above a dynamic knock threshold, above which, and depending on how much, the ECU will pull the appropriate amount of timing. I'm pretty sure the EVO ECU does the same, but not positive. With DSMLink, you can log 'knock retard', which is the actualy amount of timing the ECU is pulling due to knock. The stock ECU does a fantastic job of monitoring knock, but you have to know when this is happening.
Using ECU+, the only way that you can really do this, is to open up your timing window nice and big and look at the knock sensor voltage spikes and see if there is any dip in timing. You don't wan't any dips in timing after peak torque. If there are some, that is most likely knock.
I'm just trying to save some of you guys from evetually messing up a motor. Knock is the most important thing to monitor on a modified engine. Sure, you can make more HP/TQ by increasing timing and pulling fuel, even while knocking, but if you knock for too long of a time or knock high enough one time, you could lose your motor.
Just be smart about tuning and always tune for a knock-free tune, not the most HP/TQ.
Eric
I've made many threads about this a while back and I have PM'd Tom about it, too, but knock voltage is proportional to RPM and you can't just use a set value and call any number under that safe and any number above that knock. It is hard to tell by just looking at the knock sensor voltage, but any rapid increases (spikes) are most likely knock and if you look at the corresponding timing, you can tell if the ECU thinks so, too.
The stock ECU pulls about 1 degree of timing for every 3 'counts' of knock, at least in the DSM ECU. This doesn't really correlate with any preset voltage...it's any spike above a dynamic knock threshold, above which, and depending on how much, the ECU will pull the appropriate amount of timing. I'm pretty sure the EVO ECU does the same, but not positive. With DSMLink, you can log 'knock retard', which is the actualy amount of timing the ECU is pulling due to knock. The stock ECU does a fantastic job of monitoring knock, but you have to know when this is happening.
Using ECU+, the only way that you can really do this, is to open up your timing window nice and big and look at the knock sensor voltage spikes and see if there is any dip in timing. You don't wan't any dips in timing after peak torque. If there are some, that is most likely knock.
I'm just trying to save some of you guys from evetually messing up a motor. Knock is the most important thing to monitor on a modified engine. Sure, you can make more HP/TQ by increasing timing and pulling fuel, even while knocking, but if you knock for too long of a time or knock high enough one time, you could lose your motor.
Just be smart about tuning and always tune for a knock-free tune, not the most HP/TQ.
Eric
Thread Starter
Joined: Aug 2003
Posts: 14,092
Likes: 1,090
From: Mid-Hudson, NY
Thanks Eric, you basically are saying the same thing as me but with a bit more background info.
Any knock spike that pulls timing is a bad thing, and even knock spikes in general over the norm should be looked at.
Any knock spike that pulls timing is a bad thing, and even knock spikes in general over the norm should be looked at.
I also get a lot of noise (don't remember which forum I mentioned this on) with the K&N at low RPMs. I haven't synched up my laptop logs with my desktop logs but I'll do so within the next few days and see if there are any nice examples.
Here's one for now (2/24/06), but my wife was in the car during the run. I'm tuning on top of a DynoFlash which had 2-3* less timing in a lot of places. The AFR is actually 1 point lower than what you see (I'm using the PLX WBO2 tweak for the LM-1 as suggested by Tom to get more precise readings).
Here's one for now (2/24/06), but my wife was in the car during the run. I'm tuning on top of a DynoFlash which had 2-3* less timing in a lot of places. The AFR is actually 1 point lower than what you see (I'm using the PLX WBO2 tweak for the LM-1 as suggested by Tom to get more precise readings).
Thread Starter
Joined: Aug 2003
Posts: 14,092
Likes: 1,090
From: Mid-Hudson, NY
Originally Posted by KevOVIII
I also get a lot of noise (don't remember which forum I mentioned this on) with the K&N at low RPMs.
How much timing are you running 3500-4750?
Thread Starter
Joined: Aug 2003
Posts: 14,092
Likes: 1,090
From: Mid-Hudson, NY
Originally Posted by forum04pl
just a quick question, what is the max timing i should be seeing on 93?
Every car is different, monitor your knock / afr / timing.
Here is an example of lots of knock, with no timing retard. My car does this to varying degrees in the same area every pull (5700-7000). I'm guessing it's the cams doing it but I'm not sure. This pull dyno'd at 396/393, and peak torque is right at 4000 rpm, with virtually zero knock. Should I even worry about the knock at 6k? Torque is falling at that point obviously, but it does not show any dips and timing still rises to 21-22 degrees at redline. Boost was around 22 (before I got my MAP sensor) and I run only 93 octane. Mods are TBE, MBC, GSC 272/264, 680cc injectors, Walbro 255, ECU+, Injen UICP/Intake, Coolingmist alky kit.
Sorry if this is the wrong thread for this...
Sorry if this is the wrong thread for this...
Thread Starter
Joined: Aug 2003
Posts: 14,092
Likes: 1,090
From: Mid-Hudson, NY
Originally Posted by 992gnt
Here is an example of lots of knock, with no timing retard. My car does this to varying degrees in the same area every pull (5700-7000). I'm guessing it's the cams doing it but I'm not sure. This pull dyno'd at 396/393, and peak torque is right at 4000 rpm, with virtually zero knock. Should I even worry about the knock at 6k? Torque is falling at that point obviously, but it does not show any dips and timing still rises to 21-22 degrees at redline. Boost was around 22 (before I got my MAP sensor) and I run only 93 octane. Mods are TBE, MBC, GSC 272/264, 680cc injectors, Walbro 255, ECU+, Injen UICP/Intake, Coolingmist alky kit.
Sorry if this is the wrong thread for this...
Sorry if this is the wrong thread for this...
Wow, 3.5V of knock and no timing pull. That's surpising or a glitch in something. Your timing curve looks pretty damn good to me, same with afr. That alky seems to really smooth everything out.
Do you run the alky on the open track as well?
I'll be running for the first time this friday at Mid Ohio. Everything on that pull looks perfect to me - except the knock around 6k. As far as it being a glitch, I can show you a bunch of logs that look that same in that respect, that's why I think it's a harmonic or something and not real knock. I do have a dip in torque shortly after the peak that I want to tune out yet, but I'm pretty happy with things the way they are now.
3.5v is about as high as I've ever seen it, btw. It's usually around 2.5-2.75v.
3.5v is about as high as I've ever seen it, btw. It's usually around 2.5-2.75v.
Last edited by 992gnt; Apr 16, 2006 at 06:09 PM.
992gnt,
Can you post that same log, but with the timing window at a lot smaller range, like say between 5* at the bottom and 20* at the top. It looks like you are getting some timing retard in there, but with the range set at -50* to 50*, you're just seeing the overall shape of the timing curve.
Eric
Can you post that same log, but with the timing window at a lot smaller range, like say between 5* at the bottom and 20* at the top. It looks like you are getting some timing retard in there, but with the range set at -50* to 50*, you're just seeing the overall shape of the timing curve.
Eric
With the higher resolution, you can see the timing plateau for a (split) sec, but it really doesn't get pulled by any appreciable amount. I would expect the timing "curve" to be stair steps at that resolution- advance, check for knock, advance, check for knock, etc. Would that be a correct assumption?
Thread Starter
Joined: Aug 2003
Posts: 14,092
Likes: 1,090
From: Mid-Hudson, NY
Originally Posted by l2r99gst
992gnt,
Can you post that same log, but with the timing window at a lot smaller range, like say between 5* at the bottom and 20* at the top. It looks like you are getting some timing retard in there, but with the range set at -50* to 50*, you're just seeing the overall shape of the timing curve.
Eric
Can you post that same log, but with the timing window at a lot smaller range, like say between 5* at the bottom and 20* at the top. It looks like you are getting some timing retard in there, but with the range set at -50* to 50*, you're just seeing the overall shape of the timing curve.
Eric
I posted two below with timing window set like you said. I have no idea why I didn't think of this before, it does make it much easier to see timing pulls. One thing I did notice is there are some points on the curve where timing drops by 1 degree but no knock spike was present. Is that because the ECU+ sample rate just didn't pick up a knock spike there or just the ecu doing its thing?
The first screenshot is my last log I posted above with my timing tweaks, the 2nd one is before I started messing with timing much, notice the 2nd one has much less knock voltage noise.

Thread Starter
Joined: Aug 2003
Posts: 14,092
Likes: 1,090
From: Mid-Hudson, NY
Originally Posted by 992gnt
I'll be running for the first time this friday at Mid Ohio.






