I'm curious... (about per cylinder knock, not my sexuallity)

[HolyCrapItsFast]

I was thinking about per cylinder knock today and I was wondering about the logic behind it. I see you can log per cylinder knock on the 08+ and I would like to start adjusting based on per cylinder knock.

This is my take on it... Per cylinder knock sum has little to do with where the sensor is located or even the intensity of the signal. The knock sensor sees knock from all cylinders and as long as a threshold is reached it is registered. The ECU has further filtering logic that isolates the knock event and correlates it to crank and/or cam position. From this it can very accurately determine what cylinder a knock event occurred on and then reports it to the log.

Of course this is an assumption and before I start messing with things based on per cylinder knock i would like to get a take from the more astute of you.

If my assumptions are correct then logging these parameters is in my best interest for making safe tunes for other people. I can optimize my tuning methods and include per cylinder timing and fuel optimization. Ideally, tuning these parameters requires the use of per cylinder EGT and AFR gauges, but that is very impractical. The per cylinder knock parameter in conjunction with global EGT and AFR will suffice

Any thoughts?

 

[Vermont]

I would say that per cylinder knock does give you a good case for doing per cylinder timing improvements. The only problem I do see is that if you have to retard timing for the 1 or 2 of the cylinders due to knock only being there. Then you run the risk of any small problems causing massive amounts of knock across the board with all cylinders since you are running the other cylinders closer to the thresh hold. I think it is more use full at showing you that you have problems. Because if you are only getting knock on one cylinder then maybe just maybe you might have a problem with that cylinder alone.

Also I fail to see the uses of any of the per cylinder adjustments expect as methods to which control knock. Do you think that any real power goals can be achieved by advancing timing for only 1 or 2 of the cylinders? Or even if it will be worth it to do that? Also with out the ability to see what conditions each cylinder is seeing as far as EGT's and AFR's could potentially be very bad. After all if you are only seeing knock on cylinder 2 or example then reduce timing for just that cylinder, well you run the risk of decreasing it to much and heating up just that cylinders temps and causing it to burn out. The increase in temp over all would be so marginal and diluted from the others that it could very well be missed and by that point damage could be done.... Maybe I am just acting like the boy who cried wolf, but to me messing with any adjustments for individual cylinders with out seeing the conditions and data for each individual cylinder is just asking for trouble.....

 

[HolyCrapItsFast]

I would say that per cylinder knock does give you a good case for doing per cylinder timing improvements. The only problem I do see is that if you have to retard timing for the 1 or 2 of the cylinders due to knock only being there. Then you run the risk of any small problems causing massive amounts of knock across the board with all cylinders since you are running the other cylinders closer to the thresh hold. I think it is more use full at showing you that you have problems. Because if you are only getting knock on one cylinder then maybe just maybe you might have a problem with that cylinder alone.

Also I fail to see the uses of any of the per cylinder adjustments expect as methods to which control knock. Do you think that any real power goals can be achieved by advancing timing for only 1 or 2 of the cylinders? Or even if it will be worth it to do that? Also with out the ability to see what conditions each cylinder is seeing as far as EGT's and AFR's could potentially be very bad. After all if you are only seeing knock on cylinder 2 or example then reduce timing for just that cylinder, well you run the risk of decreasing it to much and heating up just that cylinders temps and causing it to burn out. The increase in temp over all would be so marginal and diluted from the others that it could very well be missed and by that point damage could be done.... Maybe I am just acting like the boy who cried wolf, but to me messing with any adjustments for individual cylinders with out seeing the conditions and data for each individual cylinder is just asking for trouble.....

Oh absolutely not :tup: What I'm suggesting is using these tables to curb knock in that cylinder rather than remove timing globally. I believe this is a useful tool in making an engine safe. Each cylinder has a specific profile and thereby different requirements. This could be due to several things like port runner length in the heads or pressure differences due to exhaust manifold runner lengths or excessive heat from being so close to the turbo or differences in cooling due to differences in the water jacket configurations. There are a multitude of reasons why it would be prudent to touch these tables when diagnosing knock. You can do it either by adjusting fuel and or timing per cylinder.

Regardless I think EGT is a big key but as I said, measuring this per cylinder is not often feasible. So I suggest using the per cylinder knock parameters in conjunction with global EGT's and AFR to achieve ideal timing and fuel per cylinder.

The other thing that crossed my mind is... Are there parameters for monitoring per cylinder fuel trims? I have not looked into that.

 

[HolyCrapItsFast]

We already know you swing both ways....


OH! about cars....
Faak per cylinder... you make that one work and I'll stop tuning my own car.

:lol:

 

[Td_d]

Interesting thought George - I think if you were to play with these tables, it would be a defensive strategy to curb knock, rather than optimise timing for power, as you noted above. I think the knowledge that the different cylinders (stock at least) are experiencing uneven heat due to the manifold and fueling setup (and the prevalence of failure in certain pistons over others) would justify this approach - but you would need an EGT probe per cylinder to really get value out of it. I recall a thread somewhere where someone did do this exactly (and also confirmed that tables A, B, C, D correlated to cylinders 1 - 4).

Per cylinder fueling is more interesting for me, given the way the tables are set up - if you have the patience, can give you a great amount of control, as the axes are "previous" pulse width referenced.

 

[HolyCrapItsFast]

Thank you TD... I was unsure of the correlation between the table designation and the cylinder itself but you have now confirmed that for me...

With regard to EGT I am beginning to think it is not a huge deal... If we are talking a few degrees of timing then the impact to EGT is minimal at best and wouldn't be far off from the global reading. Not to mention we are not talking about removing timing through out the entire RPM/Load range for a particular cylinder, only the area that needs it. Those transitions will only last mere fractions of a second in most cases. For safety sake I would use these tables to reduce knock in a particular cylinder but I would also inherently increase fuel slightly for the same cylinder and in the same area of the map. :tup:

I don't think there is to much thinking to be done here now that I have given it some thought. :lol:

 

[Td_d]

Now this makes sense... I would also adjust the injector size per cylinder also using the flow test sheets that come with them. Usually there is only a small percent variance, but with the per cylinder injector compensation tables, you can even out the AFRs better. Plus you can make the dog leg run a little richer if you want to.

I guess you could do that by basically multiplying the entire table for each individual cylinder by the required factor in terms of the flow scaling - although it would be a pretty small margin.

I'll see if I can find that thread, interesting reading.

 

[Td_d]

I don't think there is to much thinking to be done here now that I have given it some thought. :lol:

Chuckle... because we don't over think shit at all!

That's half the fun!

 

[HolyCrapItsFast]

I'll leave mine the way it is until I can get some EGT sensors in each runner. That's when I'll play with it.

At the very least this will help to gather data and/or prove some stuff... This is something I have wanted to do for the longest time... That and the AEM Multi AFR Thingy.

 

[Td_d]

It was NSFW who was going to do this - should've guessed!

http://forums.nasioc.com/forums/showthread.php?t=1911733

 

[HolyCrapItsFast]

Nice... a little pricey but nice...

here is the AEM... http://www.aemelectronics.com/wideband-air-fuel-systems-15/4-channel-wideband-uego-controller-60/

@ TD_D, I read that thread over at nasioc and I think NSFW has a good method there... Run up fuel in on cylinder till a measurable difference in EGT is seen. Do that for each cylinder and you will better know the effects per cylinder fueling has on overall EGT.

Kinda wish he was here though.

 

[Td_d]

This is BadNoodle's take on these tables - which makes the most sense - firing order:

After doing some header research I had a though about this.
Rather than 1-4 -> A-D maybe they're lined up by firing order. Meaning the firing order is 1-3-2-4 so
A=1
B=3
C=2
D=4

Now looking at the manifold design, cylinders 3 and 4 scavenge of 1 and 2. Therefore you would have to shorting the pulse wide in 1 and 2 and making in longer in 3 and 4 would make sense.

In terms of A,B,C,D that would mean B and D would have have positive values While A and C have negative values.

Looking at the 08 sti values, thats exactly what's going on.


I would love to see what happens when someone running EL hearders zeros out the comps.

 

[Td_d]

Kinda wish he was here though.

Me too, he has a wealth of knowledge.

 

[Td_d]

Who is he and where is he at?

He's one of the coding gurus and mods on the Romraider Forums - he's the one who wrote the LC/FFS code for the open source SD patch, for example.

 

[smitty5788]

Ok... This wishlist is getting expensive...

$632 for 2 gauges that both read two EGT sensors... guess I'll keep looking... 4x http://www.siliconeintakes.com/prod...d=980&osCsid=3e75361026bb3a8e005c8ccc969fb6fa would be about the same price....

Wish I could find a 4 channel gauge

I am a bit late to this party but don't you remember the one I posted a while back on IWSTI Ed?

Here it is ...

http://www.chiefaircraft.com/aircraft/engine-instruments/ei-instruments/ei-egt.html

Consiter what other options there are vs the price of this, no reason you could data log it as well.