Knock Retard processing speed Stock ECU vs JS box and keeping it alive

norbs

Classic fast, XFI, SPortsman & MS3 programming
Joined
May 25, 2001
There has been a lot of talk on this forum about knock retard and the mysteries associated with it. 1st of all I would like to thank Dennis Leek for contributing the stock ecu knock information that I am about to discuss. To sum up the stock ecu infomation we have to think mentally what is going on at 6000 rpm for an example. At 6000 rpm we use a 3 tooth reluctor, so we actually have 18000 spark events per minute at this speed and if we divide this by 60, we get 300 sparks per second or .3 sparks per ms of time. The stock ecu needs 5ms to recognize you have knock present and another 12.5ms to actually process the retard between knocks, so in theory 17.5ms of time has gone by since the 1st knock event, before any retard has taken place. This is 17.5ms x.3ms=5 spark events or 5 cylinder firings with "knock". Next is the recovery time is every 200ms, which means you have 200 x .3ms=60 spark events with reduced timing before the timing is put back to normal after not seeing knock.

What the JS box does, is that it can see the knock instantly it does not have to wait 5 firings to retard the timing, it is done on the next firing, and the recover rate is done on the next firing also. So what this means to me and you is that the timing is pulled 5 times faster and recovers 60 times faster at 6000 rpm with a JS box. If some numbers don;t make sense please correct me. If we could get John from J&S to chime in here to verify this would be even better.
 
I read somewhere that after the first knock event. That you have to wait a while to let the cylinder cool off, before you can try to restore the timing back.
Otherwise knock will occur sooner.


x188: Knock Retard. In degrees, the amount of retard to use
during normal engine operation. Based on map VAC.(LOAD)

0x18d: Retard attack rate. Amount of degrees per millisecond
that any required retard will occur at, based on rpm.

--- When knock is detected, a count is available to the
ecm code. This is read from the ecm hardware. This
count is the number of milliseconds that knock has
been occurring. This is the 'knock count' displayed
on a scanner.

The total amount of retard is calculated by:

1). Multiplying the attack rate by the 'count'.

2) Lookup on either PE retard table if WOT, if not in
PE, then it uses the normal table.

The amount of added retard will be the lessor of the
above two values; the attack rate, or the table value.
It is then divided by two. Yes, it appears that the code
divides the retard value in half before subtracting it
out from the total spark advance.

Of course, a maximum retard check is then done using the
value in location 0x01b. The ecm will not retard the
total timing any more than this value.

0x196: Retard recovery rate. Percent per millisecond that any
retard recovers back to normal at, based on rpm.

The decay routine runs 10 times a second. The decay starts
immediately upon the addition of retard.

Again, required for a smooth transition?

/* ------------------------------------- */
Juan :)
 
Pulling timing was never meant to save an engine, its only designed to give you an idea if your going to far, or if something abnormal is going on somewhere.

If you see knock the damage is already done.
 
Pulling timing was never meant to save an engine, its only designed to give you an idea if your going to far, or if something abnormal is going on somewhere.

If you see knock the damage is already done.


That would depend on the severity of the knock I really don't agree with that statement.....we are trying to reduce the knock before it gets to that point...
 
That would depend on the severity of the knock I really don't agree with that statement.....we are trying to reduce the knock before it gets to that point...

A stock-ish build can knock for a while before she lets go, on a race high horsepower application not so much


Posted from the TurboBuick.Com mobile app
 
I read somewhere that after the first knock event. That you have to wait a while to let the cylinder cool off, before you can try to restore the timing back.
Otherwise knock will occur sooner.


x188: Knock Retard. In degrees, the amount of retard to use
during normal engine operation. Based on map VAC.(LOAD)

0x18d: Retard attack rate. Amount of degrees per millisecond
that any required retard will occur at, based on rpm.

--- When knock is detected, a count is available to the
ecm code. This is read from the ecm hardware. This
count is the number of milliseconds that knock has
been occurring. This is the 'knock count' displayed
on a scanner.

The total amount of retard is calculated by:

1). Multiplying the attack rate by the 'count'.

2) Lookup on either PE retard table if WOT, if not in
PE, then it uses the normal table.

The amount of added retard will be the lessor of the
above two values; the attack rate, or the table value.
It is then divided by two. Yes, it appears that the code
divides the retard value in half before subtracting it
out from the total spark advance.

Of course, a maximum retard check is then done using the
value in location 0x01b. The ecm will not retard the
total timing any more than this value.

0x196: Retard recovery rate. Percent per millisecond that any
retard recovers back to normal at, based on rpm.

The decay routine runs 10 times a second. The decay starts
immediately upon the addition of retard.

Again, required for a smooth transition?

/* ------------------------------------- */
Juan :)


This info is based off the stock gn ECU or some other ECU? The JS box pulls 2 degrees max per knock, so it can be a smooth transition......
 
I read somewhere that after the first knock event. That you have to wait a while to let the cylinder cool off, before you can try to restore the timing back.
Otherwise knock will occur sooner.


x188: Knock Retard. In degrees, the amount of retard to use
during normal engine operation. Based on map VAC.(LOAD)

0x18d: Retard attack rate. Amount of degrees per millisecond
that any required retard will occur at, based on rpm.

--- When knock is detected, a count is available to the
ecm code. This is read from the ecm hardware. This
count is the number of milliseconds that knock has
been occurring. This is the 'knock count' displayed
on a scanner.

The total amount of retard is calculated by:

1). Multiplying the attack rate by the 'count'.

2) Lookup on either PE retard table if WOT, if not in
PE, then it uses the normal table.

The amount of added retard will be the lessor of the
above two values; the attack rate, or the table value.
It is then divided by two. Yes, it appears that the code
divides the retard value in half before subtracting it
out from the total spark advance.

Of course, a maximum retard check is then done using the
value in location 0x01b. The ecm will not retard the
total timing any more than this value.

0x196: Retard recovery rate. Percent per millisecond that any
retard recovers back to normal at, based on rpm.

The decay routine runs 10 times a second. The decay starts
immediately upon the addition of retard.

Again, required for a smooth transition?

/* ------------------------------------- */
Juan :)


This info is based off the stock gn ECU or some other ECU? The JS box pulls 2 degrees max per knock, so it can be a smooth transition......
 
This is very interesting. Can this work with a fast system, can it be incorporated into the system if feasible?
 
No and yes, the js is a complete knock system for about $700 it will work with the stock ecu and coil packs or with a fast. Im actually using the js system along with the factory esc module so either fast or js can retard the timing.
 
No and yes, the js is a complete knock system for about $700 it will work with the stock ecu and coil packs or with a fast. Im actually using the js system along with the factory esc module so either fast or js can retard the timing.

Norbs,

Here is a good question and maybe on for Bob Bailey. As you know I am running the TR6 with XFI. Since the TR6 is a stock replacement/factory coil and esc module, I wonder how this will work/install? Wonder if it will mess with the rev limiter and 2step?


Clint

Posted from the TurboBuick.Com mobile app
 
Norbs,

Here is a good question and maybe on for Bob Bailey. As you know I am running the TR6 with XFI. Since the TR6 is a stock replacement/factory coil and esc module, I wonder how this will work/install? Wonder if it will mess with the rev limiter and 2step?


Clint

Posted from the TurboBuick.Com mobile app


That is a good question, the TR6 is to new to me, so I don;'t know much about it. You would have to work with BOB and John@JS to see if it would be compatible.
 
That is a good question, the TR6 is to new to me, so I don;'t know much about it. You would have to work with BOB and John@JS to see if it would be compatible.

It is a wastespark system....


Clint

Posted from the TurboBuick.Com mobile app
 
Norbs: Your unit can retard one, two, or three steps per knock event, up to a maximum of ten steps. If mode switch #1 is down, each step is one degree, giving a max of ten degrees knock retard. Double that if the switch is up. The advance rate is one step every ten engine cycles (twenty revolutions).

In 2008 I was beta testing the Vampire on a Ford Lightning. The owner wanted to test on 89 octane, while tuned for 91, but on the street, no boost. Pulling away from a light it would rattle for a couple seconds before it retarded enough to kill the audible knock. I had him connect his Predator tuner to see what the timing was doing in his tune. It would go from about 32° at light load to 13° at zero psi. The tune was backing off the timing, but there was a minor lean spot at what he called "tip in" that he didn't feel the need to correct.

We went back to his place to talk about it, and I told him "It's not designed to do what you are trying to do." I felt defeated, like I was letting him down, so I returned to World Headquarters and thought about it, then said, ok, stop crying about it, it's simple. Change the code so it does what he wants.

I worked on the code so it retarded UP TO seven steps with one knock event. I went to see him the next week and Yee Ha! Stomp on it and it pulled up to 14° instantly on one cylinder, no audible knock, and just trace knock in the headphones (it sounded like birds chirping). After about fifteen minutes testing on the street he headed for the freeway onramp. Going up the onramp he said you can't even feel it pull timing, yet you look at the display and it's like Christmas lights.

Video of the display as we do three 70-90 freeway pulls:

The truck had pulleys to make 19psi at the time, but he was holding it to 8psi. That was the end of a thirty day test, when I turned the unit over to him. Since then, he changed pulleys to make 21psi, and it still has that lean spot, but he doesn't care, the unit covers it for him.

Four years later he's still running the stock block, and the heads have never been off. "Stew" is bgstew6, a moderator on lightningrodder.com. At the time he was a Marine officer at Camp Pendleton.
 
this is quite interesting i turn knock control off after 4000rpm due to engine noise triggering it above that.my factory ecu (not a buick turbo ecu but still uses the maths above) cuts it off at 3600rpm stock.i only run 8 degrees of possible retard and it comes back in at 29 degrees per second ive never felt it come on or off.i also run tip in spark retard that can pull timing on throttle opening to prevent knock even occuring if needed.factory its pulling timing fast enough that you cant even hear knock on the knock ears.its only got the knock sensing still connected purely for safety when boosting while extremly hot or should the fuel filter block or an injector start failing.i can boil the car on the burnout car and it survives no worries
 
The Interceptor version will work with the TR6. Order the four channel version if you are running waste spark, or the six channel version if you want full sequential. The six channel version will work in three channel waste spark mode.

I have a attached a partial schematic of the I/O.
O.png
 
The Interceptor version will work with the TR6. Order the four channel version if you are running waste spark, or the six channel version if you want full sequential. The six channel version will work in three channel waste spark mode.

I have a attached a partial schematic of the I/O.View attachment 196766

Now that is what I'm talking about!!
Thanks!!


Clint

Posted from the TurboBuick.Com mobile app
 
John, I;m very pleased you have chimed in here on this thread, and clarified how your unit processes, knock. Seems like all the Buick guys think the stock system is fast enough to get by?
 
Pulling timing was never meant to save an engine, its only designed to give you an idea if your going to far, or if something abnormal is going on somewhere.

If you see knock the damage is already done.


John, I;m very pleased you have chimed in here on this thread, and clarified how your unit processes, knock. Seems like all the Buick guys think the stock system is fast enough to get by?


The stock Buick computer isn't fast enough but all the late model cars are plenty fast to in fact pull timing fast enough to prevent damage. That's why the tunes always have hp left in them. They make even Corvettes run fine on 87, just with less power, but the ecm in that car is PLENTY fast to pull timing and save the engine and this is exactly what the knock system was designed to do; just not enough processor to accomplish this back in 1987 but the thoughts were the same.
 
Norbs, if cylinder #6 knocks, do you want to retard the very next cylinder (#5)? Or do you want wait until #6 comes around again?
Eric
 
Norbs, if cylinder #6 knocks, do you want to retard the very next cylinder (#5)? Or do you want wait until #6 comes around again?
Eric
Eric:

The SafeGuard waits until the knocking cylinder is about to fire, then dials in the calculated amount of retard for THAT cylinder.
 
Top