Adding a knock sensor

[nomansland92]

I personly believe in knock sensors for FI. Turbo subarus rely heavily on knock sensors, so much that they use it to determine the fuel octane so they say. There's a base timing table then a second overlap table called a dynamic advanced timing. If the sensor isn't picking up a noise it will add in timing by it self, if it stats to knock like from a lower quality fuel it will pull that timing back out, then it can add it back in gradually when it's not picking up any knock. Granted those cars are very sensitive to knock, but it's crazy how big of a role it plays in the tuning.

[mr2peak]

I have one so I can run nitrous with slightly less fear

[digger]

i dont necessarily agree that a knock sensor isnt beneficial, i agree its rarely needed in most instances, depends on the specs and how good the VE etc.

there are several things that can happen that increase tendency to knock that might not be properly captured in a dyno session.

- high IAT due to heat soaked engine on really hot day with A/C on etc (we had 116F last week). there are mechanisms to cope with this on most ECU but how well are they set up given the conditions might be quite rare and how robust are the capabilities. simlpy pulling a couple degrees timing if the IAT reaches a set value is a pretty big compromise if using a 1D correction table

- bad fuel,

- old engine starts to lose a bit of oil control as mileage goes up and the oil vapour dramatic lowers the knock resistance.

- excessive carbon build up.

- an inertia dyno often wont properly be able to replicate someone trying to accelerate up a steep hill in too high of a gear. a braked dyno does this better

truth is most m20 will have some tolerance to knock so its unlikely to go bang straight away but there was a guy on here (or e30tech) who lost the ring lands on his NA forged 2618 pistons so its not an impossibility

[downforce22]

That's odd about the slugs. We tuned an eta with an 885 head for Time Trials in 2 different configurations so that a simple ECU swap and ballast, it could compete in two classes the same weekend - it got a track record in both classes and the two drivers won a full set of Hoosier's as a result. Anyways, the car made the most power with 27°, and the de-tune was in the low 20's (23° IIRC)

In contrast ITB 2.8 we built was left at 30.6° max, it was losing power with anything over 32.

Either way timing in the 30's *shouldn't* cause knock on an m20 so long as fueling is good. For fun we ran one up to 45 just to see if there was anything audible, but there was none, nor any evidence on the plugs it was just a dog with that much timing. Many times we dyno a stock m20 with stock electronics and they are in the 14 AFR's at WOT (but find mid-13's to be the "sweet spot), they are fairly robust over all, and pretty "detonation resistant" as they were designed.

There is no reason to run a knock sensor on an N/A car once the tune is set up. So this applies to FI applications which is why I'm looking at knock sensors. Interesting note about the stroker engine requiring more timing, I'm at 30 degrees on my 3.1 right now, so nice to know i'm in the ballpark on my street tune. I definitely was getting knock under boost before I retarded my timing.

Any other sensors out there? I think it should be a pretty common component people add to their FI engines and you should know of some commonly used sensors.

[nando]

Doesn't have to be a load bearing dyno. you're just tuning for the minimum ignition advance with the most torque. if the engine isn't knock limited (typically N/A M20s aren't), at some point torque doesn't increase with more advance - so you stop there.

if you hit the knock threshold on the dyno - not a big deal. just back it off. the engine isn't going to explode because of short term knock while tuning (or otherwise).

Common misconception is that you just increase ignition until it knocks, and that's what makes the most power - this just isn't true, unless you're using low-octane fuel or the engine is otherwise knock limited.

There's an optimum timing for the force of combustion to push down on the pistons, which is basically a fixed value - it's basically the same for any engine. The idea with ignition advance is as you increase load and RPM, there is a shorter amount of time for ignition to complete before that optimum timing window passes. That is why you advance ignition as RPM increases - it takes so long for combustion to finish; if you didn't advance ignition the optimum window would be long past by the time combustion ends.

Some engines have slower or faster burn rates. The M20 head is a semi-hemi, which has a longer burn time - modern pentroof designs are faster, so the M20 needs more advance than say, an M54. Old American V8s had huge combustion chambers with slow burn rates, so they needed lots of advance (they also were less efficient because of this). The classic Hemi engines were notorious for this (and that's also why that design isn't used anymore - as far as modern Dodge, it's just marketing).

The faster burn rate and lower required advanced for optimum timing are also reasons why newer engines can run 10:1 or even 11:1 and still burn 87 octane fuel.

[nomansland92]

I have read about this one in a few places and a few people use it. Seems like a smart addition if you are trying to develop a good safe tune...




So then how does one fine tune timing if one has no idea if knock is occurring? By the time you hear it, isn't it already too late?

A load bearing dyno

[tschultz]

This kit here: http://www.viatrack.ca

I have read about this one in a few places and a few people use it. Seems like a smart addition if you are trying to develop a good safe tune...

An NA M20 isnt likely to be knock limited before MBT anyway. I honestly think you are overcomplicating things..

Edit: ForcedFirebird beat me to it and he has even more M20 experience than me. Knock control is a waste of time on an NA M20. My personal engine reached MBT long before the knock threshold.

So then how does one fine tune timing if one has no idea if knock is occurring? By the time you hear it, isn't it already too late?

[nando]

An NA M20 isnt likely to be knock limited before MBT anyway. I honestly think you are overcomplicating things..

Edit: ForcedFirebird beat me to it and he has even more M20 experience than me. Knock control is a waste of time on an NA M20. My personal engine reached MBT long before the knock threshold.

[ForcedFirebird]

by all reports Rama has had great results with flat tops and the 200/731 heads, it would have the same knock suppression capabilities as the 885 with the OEM style slugs. of course this is assuming you set the engine up correctly with suitable static CR , cam timing, piston to head clearances and a proper tune etc

on the other hand ive got couple mates with generic slugs with 885 and they are knock limited quite a bit it seems

That's odd about the slugs. We tuned an eta with an 885 head for Time Trials in 2 different configurations so that a simple ECU swap and ballast, it could compete in two classes the same weekend - it got a track record in both classes and the two drivers won a full set of Hoosier's as a result. Anyways, the car made the most power with 27°, and the de-tune was in the low 20's (23° IIRC)

In contrast ITB 2.8 we built was left at 30.6° max, it was losing power with anything over 32.

Either way timing in the 30's *shouldn't* cause knock on an m20 so long as fueling is good. For fun we ran one up to 45 just to see if there was anything audible, but there was none, nor any evidence on the plugs it was just a dog with that much timing. Many times we dyno a stock m20 with stock electronics and they are in the 14 AFR's at WOT (but find mid-13's to be the "sweet spot), they are fairly robust over all, and pretty "detonation resistant" as they were designed.

[digger]

I have found an n/a m20 stops making power well before it knocks (using the angular squish pistons/head anyways), but the coolant drain hole on the passenger side of the block would be a great place for one. The flat pistons and 200/731 head may be a different story, though.

by all reports Rama has had great results with flat tops and the 200/731 heads, it would have the same knock suppression capabilities as the 885 with the OEM style slugs. of course this is assuming you set the engine up correctly with suitable static CR , cam timing, piston to head clearances and a proper tune etc

on the other hand ive got couple mates with generic slugs with 885 and they are knock limited quite a bit it seems

[Coyote_ar]

what i've read, is that the M20 makes so many noises from its valvetrain, that the knock sensor tends to pick them as knocking when its not really happening. and if you lower the sensitivity of the sensor to avoid false positives, then you wont detect knocking when you can tell its happening (on the dyno).

again, this is not first hand experience.

[ForcedFirebird]

I have found an n/a m20 stops making power well before it knocks (using the angular squish pistons/head anyways), but the coolant drain hole on the passenger side of the block would be a great place for one. The flat pistons and 200/731 head may be a different story, though.

[downforce22]

Anybody get their installed yet?

[berlow94]

says David Redszus on a different forum... cracked me right up.

t

Ya that sounds like somebody stuck using 30+ year old engine managment. Does he know what "knock-limited" means?


Sent from my iPhone using Tapatalk

[TobyB]

knock sensors are more for tow vehicles than race cars

says David Redszus on a different forum... cracked me right up.

t