fed up with knock control
Yeah I've heard about cars stalling when harleys pass them. I'll be honest, I haven't had my evo in many loud situations. The DSM though, I can bang on the block with a hammer and all is fine, actually I can do that on my Evo as well and it won't stall. Go figure.........
Evolved Member
Joined: Apr 2008
Posts: 604
Likes: 0
From: Sydney, Australia
I am definitely no newbie, but I found this interesting. I still find it hard to believe the knock sum is interpretted first, and then from that the timing is pulled. From my experience the knock sum is just a reported amount of knock for the tuner or tech to see, and it is based off the amount of timing that is pulled.
The point of my post was to think about what knock sum is actually reporting. If it is directly showing us the amount of knock, then why is there a decay function? The process is easy to understand... engine knocks, knock sum reports it and timing is pulled, knock stops , retarded timing decays back to normal. If it is meant to show the amount of knock present, why is knock sum decayed when the knock has already stopped? I understand completely the timing needing to be decayed, but I dont understand why reported knock sum reflects this unless it is based off the timing values.
On another side note, has anyone ever found if the ecu will use fractions of a degree of timing? i.e. will 1 knock sum pull 1/3 of a degree but we just dont see it in the logging? Or is there an interpolation between two map points that are 1 degree apart? I doubt it does but I am interested if anyone has seen anything to the contrary.
The point of my post was to think about what knock sum is actually reporting. If it is directly showing us the amount of knock, then why is there a decay function? The process is easy to understand... engine knocks, knock sum reports it and timing is pulled, knock stops , retarded timing decays back to normal. If it is meant to show the amount of knock present, why is knock sum decayed when the knock has already stopped? I understand completely the timing needing to be decayed, but I dont understand why reported knock sum reflects this unless it is based off the timing values.
On another side note, has anyone ever found if the ecu will use fractions of a degree of timing? i.e. will 1 knock sum pull 1/3 of a degree but we just dont see it in the logging? Or is there an interpolation between two map points that are 1 degree apart? I doubt it does but I am interested if anyone has seen anything to the contrary.
I am with you mattjin. The knock sum is really a timing retard value, and it decays from there.
The DSMLink guys have been logging knock retard from the beginning and that is where I got my 3 counts per 1* of timing rule. It's actually a fraction of 255, but I forgot the exact number. I could look through the DSMLink forums, as it was years ago since I read that post.
I do recall in that post them talking about fractions of a degree able to be pulled as well, but don't quote me on that. I'm just going off of my memory.
EDIT: Here are some responses to a post I found on the DSMLink forums when someone asked about knock counts and timing retard and how it differed from other loggers, since most use knock sum and the DSMLink guys always knew it was knock retard. The replies were from twdorris and dmertz, the DSMLink creators:
twdorris:
dmertz:
and further down in same thread, posted by twdorris again:
Eric
The DSMLink guys have been logging knock retard from the beginning and that is where I got my 3 counts per 1* of timing rule. It's actually a fraction of 255, but I forgot the exact number. I could look through the DSMLink forums, as it was years ago since I read that post.
I do recall in that post them talking about fractions of a degree able to be pulled as well, but don't quote me on that. I'm just going off of my memory.
EDIT: Here are some responses to a post I found on the DSMLink forums when someone asked about knock counts and timing retard and how it differed from other loggers, since most use knock sum and the DSMLink guys always knew it was knock retard. The replies were from twdorris and dmertz, the DSMLink creators:
twdorris:
There is nothing different in DSMLink related to how much timing retard is associated with the raw "knock count" variable. You are misinterpretting something somewhere. TMO called the knock retard value a "count" presumably because at the time they were not aware of its true usage. It *is* a timing retard value. As soon as you see any knock "count" you're getting timing retard. You're just not getting very much at 0-3 counts. Probably not enough to even distinguish in a TMO type datalog. DSMLink just displays that "count" as what it really is...some fractional number of degrees of retard applied to ignition advance. Multiply "counts" by about 0.35 to get the number of degrees of retard you're really seeing.
This loggable value translates directly into degrees of timing retard, as previously stated, whenever the engine is operating above the thresholds configurable in DSMLink. Any statements that say that the ignition timing is not directly affected by the TMO "knock count"/DSMLink KnockRetard value are incorrect. The "knock count" value is directly and unconditionally applied to the ignition timing; the code for this that I am looking at is quite simple. If a log shows a non-zero knock count, the ignition timing in degrees has been retarded by that value divided by 3.2.
Knock "counts" as Todd Day called them back in the day IS ignition retard. Multiply knock "counts" by 90 then divide by 256 and you have the number of degrees pulled from ignition.
Eric
Last edited by l2r99gst; Nov 11, 2008 at 08:31 PM.
I am with you mattjin. The knock sum is really a timing retard value, and it decays from there.
The DSMLink guys have been logging knock retard from the beginning and that is where I got my 3 counts per 1* of timing rule. It's actually a fraction of 255, but I forgot the exact number. I could look through the DSMLink forums, as it was years ago since I read that post.
I do recall in that post them talking about fractions of a degree able to be pulled as well, but don't quote me on that. I'm just going off of my memory.
EDIT: Here are some responses to a post I found on the DSMLink forums when someone asked about knock counts and timing retard and how it differed from other loggers, since most use knock sum and the DSMLink guys always knew it was knock retard. The replies were from twdorris and dmertz, the DSMLink creators:
twdorris:
dmertz:
and further down in same thread, posted by twdorris again:
Eric
The DSMLink guys have been logging knock retard from the beginning and that is where I got my 3 counts per 1* of timing rule. It's actually a fraction of 255, but I forgot the exact number. I could look through the DSMLink forums, as it was years ago since I read that post.
I do recall in that post them talking about fractions of a degree able to be pulled as well, but don't quote me on that. I'm just going off of my memory.
EDIT: Here are some responses to a post I found on the DSMLink forums when someone asked about knock counts and timing retard and how it differed from other loggers, since most use knock sum and the DSMLink guys always knew it was knock retard. The replies were from twdorris and dmertz, the DSMLink creators:
twdorris:
dmertz:
and further down in same thread, posted by twdorris again:
Eric
As for the stalling at idle issue, I'll look into that. I'll grab my MAS and shake it at idle sometime... I know I've seen knock when my car shut down though. Unfortunately I can't find a log amongst my millions.
I am definitely no newbie, but I found this interesting. I still find it hard to believe the knock sum is interpretted first, and then from that the timing is pulled. From my experience the knock sum is just a reported amount of knock for the tuner or tech to see, and it is based off the amount of timing that is pulled.
The point of my post was to think about what knock sum is actually reporting. If it is directly showing us the amount of knock, then why is there a decay function? The process is easy to understand... engine knocks, knock sum reports it and timing is pulled, knock stops , retarded timing decays back to normal. If it is meant to show the amount of knock present, why is knock sum decayed when the knock has already stopped? I understand completely the timing needing to be decayed, but I dont understand why reported knock sum reflects this unless it is based off the timing values.
On another side note, has anyone ever found if the ecu will use fractions of a degree of timing? i.e. will 1 knock sum pull 1/3 of a degree but we just dont see it in the logging? Or is there an interpolation between two map points that are 1 degree apart? I doubt it does but I am interested if anyone has seen anything to the contrary.
The point of my post was to think about what knock sum is actually reporting. If it is directly showing us the amount of knock, then why is there a decay function? The process is easy to understand... engine knocks, knock sum reports it and timing is pulled, knock stops , retarded timing decays back to normal. If it is meant to show the amount of knock present, why is knock sum decayed when the knock has already stopped? I understand completely the timing needing to be decayed, but I dont understand why reported knock sum reflects this unless it is based off the timing values.
On another side note, has anyone ever found if the ecu will use fractions of a degree of timing? i.e. will 1 knock sum pull 1/3 of a degree but we just dont see it in the logging? Or is there an interpolation between two map points that are 1 degree apart? I doubt it does but I am interested if anyone has seen anything to the contrary.
My car has been particularly sensitive to this since I got it. Some people I talked to at AMS said they hadn't heard of this issue, so I was flipping trying to think of what it could be.
A week or so ago I had a bike ride by me while I was waiting for a light... the guy was literally 1 lane, 2 turn lanes, and a wide (2.5+lane) median away from me and I got the idle dip!
That is probably the worst case, but at the track my car has noticeably done it since it was stock, so I didn't think it was modification related. I don't even want to run near the big crazy slicked up V8s anymore because I imagine they are pwning me at launch. I do get the dip with almost anybody revving next to me though.. evos, hondas, even had a grand prix with a smallish exhaust leak drive by with a lane between us the other day and got it also.
Evolved Member
Joined: Apr 2008
Posts: 604
Likes: 0
From: Sydney, Australia
One thing I just noticed is that the Max Knock Sum Routine adjustment is in the same subroutine as the ignition coil charge time calculations. So this would tend to mean, without digging deep enough to send me insane, that the knock retard is applied at the very very end of the calculations. I also wonder if charge time is reduced during knock to help lower the chance for full combustion.
I am interested in how I can get a ahold of the DSM Link software and a solid definition?
I am interested in how I can get a ahold of the DSM Link software and a solid definition?