Detonation detection

[2.2=8]

A question for those of you with engine dynos, How do you detect the early signs of detonation? Do you fit the engine with Piezio sensors?, read spark plugs under magnification? or are there other tricks of the trade? I'm dialing in a car I built for my son, its a turbo charged 4 cyl Daytona. using Electromotive tec II operating system, so making changes to the timing map is easy but I dont wish to destroy an engine trying to find the optimum max advance. So far we have run 10.9 @ 125mph on C16 and a little alky (@35 psi boost). The TecII has a detonation sensor but is all but useless at higher rpm because of ambient engine noise. So I was wondering what methode you dyno operators use for detonation detection. Warren

[2.2=8]

Stupid question?, or secret answer?

[la360]

My friends shop that has a Superflow dyno started off using a MSD knock sensor unit, before upgrading to a Motec Unit. He also runs this with a Lamda sensor, so he will generally see if the engine will knock before it even gets to that.
AL...

[hotrod]

You may also want to investigate the knock link.

http://link-electro.co.nz/knocklink.html

This unit is very popular with the turbocharge subaru WRX tuners and seems to give reliable knock detection if the sensitivity is properly adjusted.

You can buy one from a local US vendor, as several list it on their product listings. They sell for about $150.00 USD. If you want first person experience with these, contact King Pin performance in AZ and speak to Clark or Mike. They do chassis dyno tuning on high power turbocharged cars like the WRX, and have lots of experience with this unit.

Larry

[ChrisU]

It can be detected on the plugs, but that's not exactly 'early' detection.

[maxracesoftware]

A question for those of you with engine dynos, How do you detect the early signs of detonation? Do you fit the engine with Piezio sensors?, read spark plugs under magnification? or are there other tricks of the trade?

i don't know if you could classify this method as a trick of the Trade,
but we sneak up on ignition timing usually 2 degrees at a time from starting what we consider safe starting point.

if we get to point where we just increased ignition timing another
2 degrees and saw a Peak HP or Peak TQ gain or even a gain in the average amount of HP or TQ, but the starting RPM numbers began to decrease, its usually a sure sign not to try anymore ignition advance and to try 1 or 2 degrees less.

>>>>read spark plugs under magnification?<<<<
at that point you usually observe a few black specks or too many black specks fused to white porcelain..what you don't want to see are very tiny aluminum specks/balls fused to porcelain/base ring/ground strap or beginning eroding/melting of porcelain ditch around center electrode.

a few instances we will leave ignition timing that made best average Peak/TQ , but started to loose bottom RPM HP/TQ numbers, based on experience how long or if the Engine ever is run at those lower RPMs. most all other times, we retard ignition timing 1 or 2 degrees to make the Engine safer at those lower RPMs if the Engine happens to run there and at higher than normal temperatures we expect.

it depends on some Car/Engine Combos as to how long in a certain RPM range/load the Engine will stay in , as to how prone it might become to chance of detonation.

the lower the operating RPM , the more time for detonation to occur.

[2.2=8]

Thank you all for your replies and helpfulness, But I'm going to have to find the optimum timing for this engine the old fashion way at the track as I have no Dyno available. could I simply keep adding a degree or two untill either the MPH starts to drop or the plugs start to show signs and then back off a couple, or is there a better way (without a dyno)?
Also, speaking of black specs on the plugs after a clean cut, are the specks caused by carbon being blasted off piston and chamber, or is it from boiling porcelain? So far the plugs do show signs of black specs. Though not visable without high magnification. And the porcelain "ditch" look fine...............Warren

[Grocerius Maximus]

You'll find most sensitive modern factory knock sensors essentially useless above 7000 RPM or so.

Do you have a way of datalogging how much, when and which cylinder is knocking?

[2.2=8]

Yes, I can datalog sixteen channels, one being knock, but only one knock sensor for all four cyls........I get false knock activity after about 4500 rpm. I know that it is junk signal because going from a slider cam to a roller, knock activity disapeared with no other changes. engine noise diminished considerably. valvetrain noise must be close to same frequency as knock sensor..9000hz?..........Warren .............or is that 900-1000hz?

[Grocerius Maximus]

Sounds like an ECU/software issue more than anything else, 3000-6000 is where a properly set up knock control will save your butt.

Can you change the knock sensor calibration and control in the ECU at all?

How much ignition timing at what boost on C16?

[2.2=8]

23 degrees total @ 35psi boost, engine runs great and MPH is good for the ET. I think I'm bloody close on the timing but I just don't want to leave anything on the table........Going to try a different knock sensor. warren