Part Throttle Knock Events: VVT | Cam Overlap | Ignition timing

Hi..

Currently working on the part throttle areas of my mapping.. 1.6 litre 4 pot turbo with cams.

Low speed cruise in Vac - perfect

Full bore 1.5bar boost - perfect

>50% throttle - perfect..

But 30% throttle openings, 300mbar of boost cylinders 2/3 go noisy at 10deg advance..

Now this is not dangerous territory as load is low and knock events never exceed knock threshold. They also do not look like knock. but kind of engine noise.

Things I have tried.. Pulling some exhaust retard out of the VVT map - did nothing but hurt spool a bit.

Next thoughts are add some more IGN to these areas, thoughts being that with the overlap. Potentially I am getting some revision and the charge is unstable.

If this doesn't work. Then I will look at Inlet VVT map, as we have the inlet cam advance cranked in hard to aid with spool.

I am open to all ideas, I would really like dive into the detail around the theory for mapping with bigger cams. I have even bigger ones to go in next. :-)

Logs and maps attached.

I would try adding fuel in that area (of just those cylinders if possible) to see if I could aid in cooling the combustion temperature. If the problem persists, the I would try reducing the intake VVT just in that area. I bet the spooling when at WOT wouldn't be affected.

Are you controlling boost pressure (i.e. pre throttle), or manifold pressure (post throttle)? If only measuring manifold pressure, then you might be creating more boost / heat than you really want at the 30% throttle condition.

Hi David,

Thanks for that advice.

I am controlling boost utilising MAP. Which should be ok. I monitor TIP/EMAP as well.

I 100% have the ability to do individual cylinder fuel trims and I have noticed from my EGT logs that cylinder 2/3 do run the hottest.

This is what I'm thinking. add in 2/3% more fuel as a starter..

These are the EGT numbers including the compressor outlet temps. But we ambient into the engine.

It's unlikely adding ignition advance will reduce knock, so I'd look in other directions like reducing VVT, testing the impact of enrichment.

Mike,

I agree. But I don't think it is knock per say. I've got a feeling is a slightly bad DI injector that is struggling to fire at low Pulse width.

So far I have done the following.

1. less ign: -2 degs no change

2. more ign - +1 degs no change

3. Added 5% more fuel to cylinders 2/3 no change. Did see a small change in EGT (but that was on cylinder 2)

4. took out 10degs of VVT advance - no change.

Now I spoke the ECU manufacturer and they sent new firmware. Which I need to try today and see what it does. :-) I have also lowered the HPFP target as a lower rail pressure will give me more pulse width.

If that doesn't do it. I will be replacing the injector and testing again.

To add to this thread.

The final test I made was to add 20 litres of MS109 Reg.. 109RON race fuel. Set the Stochiometric for the fuel and gave it a general run.

There was absolutely no difference in the knock trace. So this is leading to me to think it's another noise source. potentially gearbox bearings loading up.

The hunt as they, continues. :D

Richard,

Good work narrowing things down.

It sounds like the knock frequency being monitored perhaps isn't ideal. What's your bore size and what knock frequency are you listening on?

Mike,

The knock frequency is set correctly. 7300hz based on bore size of 79mm.

Now I have found one thing as started looking at engine mounts. I was previously running a poly bushed rear engine mount. I found this to have excessive wear in metal insert. It audibly clonked when slacked off. This has now been replaced with an uprated rubber mount to match the engine mounts.

I will test again and report back with the new mount.

HPA suggested frequency would be 14,600 Hz.

As the second harmonic?

Based upon this thread: https://www.hpacademy.com/forum/webinar-questions/show/083-understanding-knock-1/

Ok. I will have a look at that.

What you describe sounds very much like low speed pre-ignition to me. Small boosted GDI engines are very prone to this, especially if the engine has some miles on it and hasnt been apart. Im no expert on LSPI and there are various contributors to it, but it is generally believed to be due to larger fuel droplets mixing with the small amount of engine oil that leaks past the rings and forms a localised pocket of low octane/low flashpoint "fuel" that self ignites during compression. Obviously spark timing has no effect when it is pre-ignition.

The main LSPI mitigation factors are generally related to fuel and oil droplet size as well as the normal pre-ignition factors such as carbon deposits in the chamber. Ensure you are using an SP rated oil (SP has a specific LSPI test requirement), use the lowest oil viscosity possible (believe it or not a thinner oil leaks past the rings less), run the engine and oil temps relatively hot (no low temp thermostat, keep the oil/coolant heat exchanger in place etc). Fuel droplet size has a big influence on the evaporation at low speed so dirty injectors or injectors modified for increased flow would be another factor. Experimenting with rail pressure and injector timing in the problem area may help.

Adam,

The engine is built. Mahle pistons, bridge way rods and crank, piper cams. We have fitted 30% larger DI injectors and stroked the pump an additional 1mm to get the fuel in we need.

The turbo charger is significantly larger than stock to support almost 400hp at 1.6 bar out of 1598cc.

We use Motul 300v 5w40 which has an LSPI package in the oil.

Temps for oil and water run at around 83degs C which sufficient but not massively hot.

Now, we know the 30% injectors so not have the OEM spray pattern. I may need try some of the Nostrum 50% over as they have the OEM spray pattern.

I have tried the second harmonic and this has cleaned up the signal a little.

I have a plex v3, so this is getting wired in as well.

Update for everyone that has contributed.

Went back to basics and looked engine at health. Compression test has shown that No3 pot. The noisy one. Is 10% down on compression.

Motor has been pulled and will be stripped and verified. Think we have caught it early enough that it shouldn't have caused too much damage.

The mad thing is the motor is showing no signs of ill health, starts perfect, idles perfect and made stout power on the dyno.

Richard,

Yes to the 2nd harmonic. That will hopefully give you better meaningful signal to noise ratio.

Regarding LSPI, it certainly causes damage, but wouldn't have shown up as knock since it doesn't occur in the knock window or have the same ringing behavior in the bore.

Good oil choice.

Great work on diagnosis. I'm glad you caught things while it will be a less expensive repair.

At only 10% down on compression, usually an engine will only loose 10% of what that cylinder made in power, so 2.5% total engine power on a 4 cylinder. This doesn't necessarily remain consistent as compression gets real bad, and somewhat depends on where the damage is i.e. valves, piston, but if you think about it you have 3 cylinder doing what they usually do, and one that's making a little less cylinder pressure than usual, but still operational.