Notices
944 Turbo and Turbo-S Forum 1982-1991
Sponsored by:
Sponsored by: Clore Automotive

Knock counting, What's good

Thread Tools
 
Search this Thread
 
Old 08-14-2007, 11:14 PM
  #106  
samluke
Burning Brakes
Thread Starter
 
samluke's Avatar
 
Join Date: May 2004
Location: St Augustine FL
Posts: 1,077
Received 2 Likes on 2 Posts
Default

Jim/Oddjob,

I had a blast, but my car pss's me off. I had overtaken JR in the Endurance race when I blew a Turbo hose and finished dead last. In the sprint I was about to pass JR in 1/2 when my motor went bad. I couldn't find any problems, and In the Loonacy race I finished dead last again.

My car runs great on boost, I have learned to lift and re-apply on the straights and can get pretty much seamless full boost. I would love to do Brainerd again, turns 1 and 2 are the scariest I have encountered, even scarier than the kink at Road America.

I have another two sets of computers that I agoing to try with different chips, but I doubt it will change anything. I was running a 1/5th mix of 112 octane with 91 and getting approx 3,000 knocks per day.
Old 08-15-2007, 10:28 AM
  #107  
Oddjob
Rennlist Member
 
Oddjob's Avatar
 
Join Date: Aug 2003
Location: Midwest - US
Posts: 4,661
Received 73 Likes on 56 Posts
Default

Originally Posted by Pauerman
Wow, 12,000 knock counts! You'd think with that much knock activity that the KLR must be pulling timing to protect the engine. Maybe that's why you feel like your car is down on power compared to others on the track.

I sure would like to know how many knock counts it takes for the KLR to start to pull timing. In my datalogs, I see 1-1.5 volt knock spikes at 12-13 psi but at 14-15 the knock spikes increase to upwards of 3 volts. Funny thing is is that none of the knock activity is audible and none of my knock spikes are back to back - there're usually separated by few hundred RPM. I've recorded some of the erratic and consistant knock activity and this is usually the knock you can hear. I wish I knew that the knock I'm seeing is "normal".

So you see a difference in amplitude of the knock signal, not frequency? Interesting, I never thought of that. I figured the knock signal was a "yes" or "no" occurrance signal, not a level/intensity signal.

Keep in mind that the 12000 knock counts were during track use - significantly different level of abuse than street use. I only get a few knock ticks during full throttle runs on the street. Also, EGTs run about 150 to 200F hotter, during sustained track driving, than WOT runs on the street or dyno. And the 12k counts were during a previous event, not the latest race that I was questioning the power output (only got 1500 at this last race).
Old 08-15-2007, 10:42 AM
  #108  
Oddjob
Rennlist Member
 
Oddjob's Avatar
 
Join Date: Aug 2003
Location: Midwest - US
Posts: 4,661
Received 73 Likes on 56 Posts
Default

That would be about 95 (91 + 112 mix) octane. I would think that should be enough with standard/stock timing and boost. You could always use the FQS in the DME to retard the timing by 3 degrees, and see if that reduces the knock count frequency. Or try the 100 unleaded or 112 leaded straight; if youre still getting high knock counts then, either the counter is in error, or there is something wrong with the motor/controls.

The 3 seconds of boost, then drop-off is an odd symptom. I know you tried about everything to fix that, and I cant think of anything else offhand to recommend.

BIR is my home track, been driving there since '92, but 1 and 2 can still raise the hair on the back of my neck on occasion. Learning to go fast in those corners does desensitize most of the Nord Stern guys to other tracks. I do pay very close attention to the RA Kink, but it doesnt put the fear of god in me, like being on the outside of turn 1 @ BIR, going two wide - or having 996 Cup overtake you at the turn-in of 2.

Originally Posted by samluke
Jim/Oddjob,

I had a blast, but my car pss's me off. I had overtaken JR in the Endurance race when I blew a Turbo hose and finished dead last. In the sprint I was about to pass JR in 1/2 when my motor went bad. I couldn't find any problems, and In the Loonacy race I finished dead last again.

My car runs great on boost, I have learned to lift and re-apply on the straights and can get pretty much seamless full boost. I would love to do Brainerd again, turns 1 and 2 are the scariest I have encountered, even scarier than the kink at Road America.

I have another two sets of computers that I agoing to try with different chips, but I doubt it will change anything. I was running a 1/5th mix of 112 octane with 91 and getting approx 3,000 knocks per day.
Old 08-15-2007, 05:14 PM
  #109  
Pauerman
Addict
Rennlist Member
 
Pauerman's Avatar
 
Join Date: Apr 2002
Location: Valley of the Sun
Posts: 863
Received 0 Likes on 0 Posts
Default

Originally Posted by Oddjob
So you see a difference in amplitude of the knock signal, not frequency? Interesting, I never thought of that. I figured the knock signal was a "yes" or "no" occurrance signal, not a level/intensity signal.

Keep in mind that the 12000 knock counts were during track use - significantly different level of abuse than street use. I only get a few knock ticks during full throttle runs on the street. Also, EGTs run about 150 to 200F hotter, during sustained track driving, than WOT runs on the street or dyno. And the 12k counts were during a previous event, not the latest race that I was questioning the power output (only got 1500 at this last race).

Yeah I do see a differences in amplitude when I record knock. I'm just wondering if the knock amplitude has to exceed a certain voltage value before the KLR starts to pull timing.
Old 08-15-2007, 05:50 PM
  #110  
Oddjob
Rennlist Member
 
Oddjob's Avatar
 
Join Date: Aug 2003
Location: Midwest - US
Posts: 4,661
Received 73 Likes on 56 Posts
Default

Originally Posted by Pauerman
Yeah I do see a differences in amplitude when I record knock. I'm just wondering if the knock amplitude has to exceed a certain voltage value before the KLR starts to pull timing.
Youre logging the KLR knock output signal to the DME, correct? Not the raw sensor signal?

So there are two variables in the signal from the KLR that the DME is evaluating, voltage (amplitude) and the occurance rate (frequency), before it decides to respond.

I read from a factory source, that timing is pulled on the particular cylinder where knock is detected, not on all 4. So something in that signal pulse would need to be indicating in which cylinder the KLR detected knock.

I would have assumed that the amplitude would be the indicator of the intensity of the measured knock pulse, and the frequency is how often the knock is being detected. But it might be possible that the voltage range of the signal identifies a particular cylinder (just totally guessing), and not intensity. Maybe it doesnt care about the intensity, just if any level of knock is detected, and how often it is occuring.

Not sure of a way of knowing w/o figuring out a way to log ignition timing with no/low knock counts, then induce higher knock counts (if you dare) and see at what frequency/amplitude the DME starts to retard timing.

With the stock CV system, the KLR is also supposed to drop boost in 40mbar increments when knock is detected. Same questions, at what occurance rate and/or amplitude of the knock signal, will the KLR start doing this? And its unclear if the KLR drops boost at the same time that the DME retards timing, or if it does so, only after the DME retards timing yet knock is still occuring.
Old 08-16-2007, 01:00 PM
  #111  
nize
Banned
 
nize's Avatar
 
Join Date: Aug 2004
Location: seattle, washington - usa
Posts: 1,822
Likes: 0
Received 0 Likes on 0 Posts
Default

guys, keep in mind that the stock klr can only pull timing 3 degrees. if knock is still detected it's supposed to then pull boost and put the car into 1.2bar cripple mode to save the motor.

if your dme/klr is chipped, you've eliminated the stage2 boost-pull knock protection. in that case if your motor still knocks with negative 3 degree timing retard (and it will on hot/track days), you're out of luck and will be damaging your motor.

this is why i've been stressing the importance of maintaining the stock 2-stage knock protection with boost-pull but nobody listens.
Old 08-16-2007, 01:10 PM
  #112  
Oddjob
Rennlist Member
 
Oddjob's Avatar
 
Join Date: Aug 2003
Location: Midwest - US
Posts: 4,661
Received 73 Likes on 56 Posts
Default

The limp home mode is triggered if a fault is found in the knock sensor itself by the KLR, not if it detects knock.

Timing can be retarded up to 6 degrees total in response to knock, if there is still knock after the initial 3 degree retard. As mentioned, boost will be reduced in 40mbar (0.6 psi) increments - the KLR doesnt just dump the boost.

Knock protection is not eliminated on most aftermarket chips. Just the DME overboost protection/fuel cutoff. In fact some of the aftermarket chips are DME only chips, and you still retain the factory KLR chip. Yes, however, if you remove the CV from the boost control system and run a mechanical boost controller, then the KLR has no way of reducing boost during knock or other problem.
Old 08-16-2007, 01:24 PM
  #113  
nize
Banned
 
nize's Avatar
 
Join Date: Aug 2004
Location: seattle, washington - usa
Posts: 1,822
Likes: 0
Received 0 Likes on 0 Posts
Default

Originally Posted by Oddjob
Knock protection is not eliminated on most aftermarket chips. Just the DME overboost protection/fuel cutoff. In fact some of the aftermarket chips are DME only chips, and you still retain the factory KLR chip.
i believe this is false, and that all aftermarket chips remove the car's ability to pull boost regardless if the cv is in place or not.

i remember recently confirming this with russell's 951max chips, at least.

to clarify, the porsche factory has a 2-stage knock protection;
1) retard timing. if knock is still detected,
2) pull boost

aftermarket chips retain stage1, but eliminate stage2.
Old 08-16-2007, 03:20 PM
  #114  
Oddjob
Rennlist Member
 
Oddjob's Avatar
 
Join Date: Aug 2003
Location: Midwest - US
Posts: 4,661
Received 73 Likes on 56 Posts
Default

So if using a factory KLR chip with a Vitesse or MAX DME chip (Russell is now recommending to use the factory KLR chip with his DME chip, Vitesse always has), exactly how are the factory KLR knock functions changed?
Old 08-16-2007, 03:30 PM
  #115  
nize
Banned
 
nize's Avatar
 
Join Date: Aug 2004
Location: seattle, washington - usa
Posts: 1,822
Likes: 0
Received 0 Likes on 0 Posts
Default

Originally Posted by Oddjob
So if using a factory KLR chip with a Vitesse or MAX DME chip (Russell is now recommending to use the factory KLR chip with his DME chip, Vitesse always has), exactly how are the factory KLR knock functions changed?
i don't make the chips so i don't know. why not ask them?

also, keep in mind there are two completely separate but related 'knock functions'. the timing retard is always there, it's the boost pull that i'm referring to.

every chip maker for the 951 that i've asked: can your chip pull boost when knock is detected?

has answered: no.
Old 08-16-2007, 04:48 PM
  #116  
Oddjob
Rennlist Member
 
Oddjob's Avatar
 
Join Date: Aug 2003
Location: Midwest - US
Posts: 4,661
Received 73 Likes on 56 Posts
Default

Originally Posted by nize
i don't make the chips so i don't know. why not ask them?

It was a rhetorical question...
Old 08-16-2007, 04:48 PM
  #117  
Rogue_Ant
Addict
Rennlist Member

Rennlist
Small Business Partner

 
Rogue_Ant's Avatar
 
Join Date: May 2006
Location: Denver
Posts: 5,252
Likes: 0
Received 5 Likes on 5 Posts
Default

The KLRs ability to pull boost is only effected when the CV is taken out. An aftermarket DME chip does not change the KLRs funtion.


Rogue
Old 08-16-2007, 05:20 PM
  #118  
nize
Banned
 
nize's Avatar
 
Join Date: Aug 2004
Location: seattle, washington - usa
Posts: 1,822
Likes: 0
Received 0 Likes on 0 Posts
Default

from russell himself;
http://forums.maxhpkit.com/viewtopic.php?t=15&start=15

Originally Posted by russell
Originally Posted by nize
Originally Posted by russell
Originally Posted by nize
i think that knock protection is critical for reliability, and is the main reason i have been considering moving to the wolf3d system. there are a lot of 'tuning' solutions out there, but it seems almost all of those that allow you to increase boost will remove the knock protection.

the wolf3d system is the only one i have been able to confirm so far that still provides knock protection in a turnkey out-of-the-box solution.

megasquirt is another, but that's way more work than i (and most others) are willing to invest.
Agreed, a lot of aftermarket chips map out the knock protection for some reason, I suppose because valve chatter at high boost can cause false detection. The only difference between the stock KLR and mine is that mine doesn't complain about overboost conditions. My DME chip would work fine with a stock KLR, but if you ever did blink codes it would say there had been an overboost condition and error for it. But if knock is detected the standard timing retard is in place.

Regards,
Russell
what about boost reduction? the porsche factory knock protection operates in two stages as follows:

1) knock is detected and timing is immediately retarded.
a) if the knock goes away, timing is advanced again automatically.
b) if knock is still detected with retarded timing, boost is immediately pulled and the car goes into 'limp mode' running crippled .2bar boost.


because phase b) knock protection is difficult to deliver (you have to be able to control the wastegate via the klr/cycle valve for it to work), most chips and tuning solutions provide only phase a) knock protection via timing retard.

very few chips and almost no tuning solutions provide phase b) knock protection via boost elimination, which is the more critical protection.

i'm wondering if your solution provides phase b) knock protection.
No, only the timing retard. In fact I advocate bypassing the cycling valve when using a manual boost control or electronic because the cycling valve does not function properly under high boost, and the MAP sensor in the KLR cannot register high boost properly. No limp home mode.

With properly tuned software these things become non-issues. I only leave the timing retard in as a safeguard, as I try to incorporate as much protection as possible to retain reliability along with performance.

Regards,
Russell
Old 08-16-2007, 05:44 PM
  #119  
Oddjob
Rennlist Member
 
Oddjob's Avatar
 
Join Date: Aug 2003
Location: Midwest - US
Posts: 4,661
Received 73 Likes on 56 Posts
Default

The premise of your discussion with Russell is incorrect, in that you are asking if the KLR will drop boost and go to limp home mode when knock is detected. It doesnt do that even with the factory chips, so of course it wont do it with the aftermarket chip(s). Go read the FSM, there is no blink-code fault for a knock condition, only for knock sensor failure.

Russell states that his KLR chips do not have the diagnostic fault "blink-codes" and the limp home mode boost reduction, so they do not dump boost if a fault is detected. But that is a different function of the KLR, and not related to knock detection and control.
Old 08-16-2007, 06:17 PM
  #120  
nize
Banned
 
nize's Avatar
 
Join Date: Aug 2004
Location: seattle, washington - usa
Posts: 1,822
Likes: 0
Received 0 Likes on 0 Posts
Default

Originally Posted by Oddjob
The premise of your discussion with Russell is incorrect, in that you are asking if the KLR will drop boost and go to limp home mode when knock is detected. It doesnt do that even with the factory chips, so of course it wont do it with the aftermarket chip(s). Go read the FSM, there is no blink-code fault for a knock condition, only for knock sensor failure.

Russell states that his KLR chips do not have the diagnostic fault "blink-codes" and the limp home mode boost reduction, so they do not dump boost if a fault is detected. But that is a different function of the KLR, and not related to knock detection and control.
again, i believe you're mistaking limp home mode with stage2 boost-pull knock protection. those are two completely different features, and neither are present in aftermarket chips while both are present in the stock factory configuration.


Quick Reply: Knock counting, What's good



All times are GMT -3. The time now is 03:09 AM.