EFI misconceptions and rumors

[robains]

Thanks for the clarification Kris.

Rob.

[dmh]

...a few more questions regarding application...

>>>The EFI ecu controls all the gauges (mileage, tach, speedo, etc) and warning lights (oil pressure, check engine, etc), correct?

>>>What is the EFI's knock stratagy?

[Kris from DRS]

Originally Posted by dmh

...a few more questions regarding application...

>>>The EFI ecu controls all the gauges (mileage, tach, speedo, etc) and warning lights (oil pressure, check engine, etc), correct?

>>>What is the EFI's knock stratagy?

There are 2 configurations for instrumentation.
1. With the full harness system it requires a dash display.
2. With the splice in version the stock ECU controls the instrument panel.

The current system supports a knock input via a stand-alone "knock box". As long as whatever module is used has a 0-5V output the EFI can use it.

The new version of the ECU (release date pending) will have built in CAN for the stock instruments if someone wants to use that option. This ECU will also have internal knock.

My concern is that people will rely too much on knock control, so in the past I have been reluctant to bring it into the system.

Kris

[dmh]

Originally Posted by Kris from DRS

There are 2 configurations for instrumentation.
1. With the full harness system it requires a dash display.
2. With the splice in version the stock ECU controls the instrument panel.

The current system supports a knock input via a stand-alone "knock box". As long as whatever module is used has a 0-5V output the EFI can use it.

The new version of the ECU (release date pending) will have built in CAN for the stock instruments if someone wants to use that option. This ECU will also have internal knock.

My concern is that people will rely too much on knock control, so in the past I have been reluctant to bring it into the system.

Kris

...but the slice in does indeed control all the warning lights so I do not have to install add-on warning lights, correct?

[robains]

Speaking of knock sensors, any reason why Toyota only use one? I can't see how it would be at all useful (located center between 2 and 3) -- their are 4 cylinders so I doubt one sensor could adequately determine detonation on all 4 cylinders.

I recall a problem I had with my old STi where the single knock sensor was falsely reporting detonation and cutting timing and then sometimes it would not report detonation causing short bursts of pinging/detonation til it finally realized it was detonation. It was a very inaccurate system, but the dealer said "this is normal operation". Several dealer reflashes and the problem never went away.

So just how accurate are these knock sensors and why do most manufacturers use just one?

Rob.

[Kris from DRS]

Originally Posted by dmh

...but the slice in does indeed control all the warning lights so I do not have to install add-on warning lights, correct?

all the warning lights remain functional regardless of the EFI, except for the CEL.

thanks
Kris

[Kris from DRS]

Originally Posted by robains

Speaking of knock sensors, any reason why Toyota only use one? I can't see how it would be at all useful (located center between 2 and 3) -- their are 4 cylinders so I doubt one sensor could adequately determine detonation on all 4 cylinders.

I recall a problem I had with my old STi where the single knock sensor was falsely reporting detonation and cutting timing and then sometimes it would not report detonation causing short bursts of pinging/detonation til it finally realized it was detonation. It was a very inaccurate system, but the dealer said "this is normal operation". Several dealer reflashes and the problem never went away.

So just how accurate are these knock sensors and why do most manufacturers use just one?

Rob.

Rob,
In most cases OEM knock control does not really work all that well. If it did the engine would never ping. Some systems will measure false knock and that results in a needless performance loss. I have literally seen a race clutch that has a loose fit on the mainshaft splines produce a flase knock and the ECU will retard the timing or put the ECU in limp mode.

More advanced knock systems will have multiple knosk sensors evenly postioned on banks or between the cylinders. I.e. cars equiped with Bosch ECUs.

Kris

[turbophil]

Computers are only -so good- at truly hearing knock, IMO. The best kind are knock sensors are your ears (w/ headphones) and a knock mic. You can make these little contraptions with relative ease and very low expense. I have not used one on my car yet, but I used to use one quite a bit. I'm not saying knock sensors are junk, but they can be a pain unless calibrated very well...

I personally wouldn't mind having one, but it wouldn't make me tune any more aggressively than what I currently do. Afterall, it's just one more layer of safety... Again, the best tool for tuning off knock is your ears with a mic (a knock sensor is essentially a mic)...The idea for track tuning, IMO, is to build in adequate saftey and not tune to the ragged edge of blowing the car up...

Best,

Phil

[2-Y-CHROMOSOMES]

Originally Posted by Kris from DRS

Rob,
In most cases OEM knock control does not really work all that well. If it did the engine would never ping. Some systems will measure false knock and that results in a needless performance loss. I have literally seen a race clutch that has a loose fit on the mainshaft splines produce a flase knock and the ECU will retard the timing or put the ECU in limp mode.

More advanced knock systems will have multiple knosk sensors evenly postioned on banks or between the cylinders. I.e. cars equiped with Bosch ECUs.

Kris

I have heard multiple times that the knock sensor can mistakenly retard timing on supercharged cars. Not sure why that would be.
ps: Kris did you get my pm?

[darkSol]

Kris, if you were to just place a J&S Safeguard inline with the coils then I would think that would be sufficient insurance against bad gas, etc... They have their own knock sensor and sensitivity adjustment.

I used one "back in the day" on my Hondata ECU'd delSol and it saved me from bad gas more than once. The display lit up like a Christmas tree!

[robains]

Wonder just how much an aftermarket exhaust might cause false readings from the knock sensor, or worse yet never permit a correct reading even when it is detonating...

I suppose the same goes for headers as they would have a tendancy to generate interferance and prevent the knock sensor from being accurate.

Probably better to not use the stock Knock sensor at all if you have any mods that might influence exhaust sound levels and/or air intake levels.

Knock sensors in each coil pak, now that is pretty cool!! Are these available for the 2ZZ?

[turbophil]

I've had long talks with J&S in recent months... He's probably still a couple months out from making any more knock boxes... His truly are amazing units. Much more to them than one might think...

Part of the reason m62 is so hard to tune the knock sensor is because of all the clattering noise the blower makes... IN fact, if you look at the blower outlet, you'll find machine slits next to the main outlet. Those are there to help quiet the clattering sound down and were an after thought from eaton due to all the complaints that there was something wrong with the blower when there wasn't... All that noise shows up in the knock sensor... You can tune the ECU to not listen for those noises, but then you may find the ECU bypassing real knock... THey're very nice safeguards, but not great for tuning... again, your ears (amplified) are the best for tuning--- if you're so inclined to push the tuning efforts that far...

Best,

Phil

[AustinP]

In my experience (on a former car), a knock sensor can be pretty useful for tuning. When you look at a datalog of the knock signal, actual knock (versus valvetrain noise, or whatever) has a characteristic spikey look to it. If you see something in the datalog that you're suspicious of, and assuming the AFR is OK (adjust your fuel if not), you can try pulling some timing from the suspect map cell(s). Or try adding octane (i.e. add a gal of toluene to your tank). Then take another datalog, and revisit the suspect cell(s). If the spikes are still there, it's probably noise; if the spikes go away, you have your answer. I also used to have a knock display (led's), which saved me a couple times when something went wrong.

[turbophil]

No doubt- I like them too, but too many people put too much stock in them. They're just one more tool for the tuning bag. Many folks wouldn't have gone through the same analysis that you did and just flood the car with timing and let the knock sensor do the work. I think that's what Kris is getting at...

Cheers,

Phil

[robains]

Ed,

Agree that knock sensors are good to have, but I think the current position of the single knock sensor on the 2ZZ is not much help (especially if you go outside of any stock parameters). As Phil pointed out, it's right next to the blower.

Keep us posted on J&S Phil.

Rob.

[AustinP]

Originally Posted by turbophil

No doubt- I like them too, but too many people put too much stock in them. They're just one more tool for the tuning bag. Many folks wouldn't have gone through the same analysis that you did and just flood the car with timing and let the knock sensor do the work. I think that's what Kris is getting at...

Cheers,

Phil

Understood. And in line with your comments, I was taught "it's a lot easier (and better) to avoid knock, than to kill knock". Just another way to say the same thing.

Regards,

[dstevens]

One problem we have is that the obvious knock control algorithms have been patented, and thus it is difficult to implement any sort of decent knock control without jumping through flaming hoops. Another problem is that almost any change to a stock engine generates false knock. The next problem is that nothing negates the need to correctly tune the ignition timing for the fuel. We do notice a spike in knock related problems in our help lines around November each year along with the change from Summer to Winter fuel in CA.

[robains]

I'm curious, how can you patent a sound - didn't Harley try this and get shot down in court? Assuming knock algorithms look at a select frequency -- it is basically just a microphone right? Toss in some duration and viola, your knock trigger. Someone was actually able to Patent this? Geez, what's next patent how one drinks a beer.

Sounds (pun intended) like Lawyers out of control.

Interesting info though.

Back to relating this to EFI 1.2 (Lotus) - what's the scoop on that?

Rob.

[stinkyonion]

Can anyone give me an idea of how many hours it took them to install the EFI, how many hours you spent tuning before car ran smoothly, and hour many hours tuning before you felt reasonably optimized ? Also curious as to total cost with EFI, guages, etc.

Just trying to get the big picture. Everyone talks about the flexibility of EFI, but im curious as to the total typical costs in terms of time and money.

[turbophil]

Well, things have a come quite a ways even since I jumped on board with it earlier this year as we have more SC'er maps floating around now then we did then...

Here's the big picture from my perspective.

Extra gauge/sensor:
Buy a WB 02 (although you should really have one whether you use EFI or not): $350

That's all you need extra, IMO. You can get away with out the above, but the I wouldn't recommend it...

Time:
Install: 3-6 hours. I know that's a big difference. I had the clam up and was done in 3 hours. Without that luxury (or similar just having the clam off), I think you can add another few hours for the PITA factor

Tuning: For the SC'er apps, we have some really good base maps floating around... They load up and will run a lot of cars reasonably well, but will require some tuning to dial in the cars street and track characteristics. That's the unknown time issue. Chances are that the base map will work just fine on the dyno, so you may not need much if any dyno time. However, you may find that you need to drive around with a buddy behind the wheel and you on the laptop with the EFI software doing some tweaking of the fuel map to clean up any odd fueling issues, such as rich or lean cruising, hesitation upon accell, etc... It can take some time, but it's usually good bonding time with a buddy I did all my tuning by myself on country roads and then added some input from Dave's map and have the car pretty much perfect with more hp and tq than ever. I've posted up several dyno plots and you can see that you can make more torque sooner and hold the torque longer with the EFI than the available off the shelf flashes.

Now that doesn't mean you can't fly Charlie out to do the same thing on the stock ECU, as he could obviously custom tune your car to achieve similar power--- so I'm not making case for either solution.

You have to like to fiddle with stuff and have to like to gather an appreciation for how stuff works to really enjoy any standalone... If you choose the standalone, you'll likely find yourself modding the car more since you can tweak your tune for each mod and that requires a dyno. So in the long run, the EFI is much more expensive, because most won't be able to just plug in the SC and walk away. You'll want an intake, exhaust, then a smaller pulley, and/or water/meth injection, etc... Then you may switch gears to a different induction system all together just because you can...

On a final note, you will need some over the phone hand holding. I try to make myself available as needed. Dave knows the software real well now as well. Kris and EBturbo know are resources as well... Casey at FF is also available, along with some others than can pipe up if they like. If you're looking for plug and play, this is not your ticket... get a flash if that's your gig. Looking for an extra 20+ hp and a flatter TQ curve, and want to figure the tuning thing out, then the EFI can be really fun and quite rewarding...

Best,

Phil