Engine Damage - Help me find the issue

Hi all!

A few weeks ago we had my RB26/30 on the engine dyno and at 6000rpm, 1.7bar of boost and around 630hp the spark plug gave up causing a hole in the cylinder head and therefore breaking the engine.

Now, the interesting part is actually, why the spark plug gave up so I brought my engine to the engine builder with a very good reputation and we started dismantling it.

Besides the obvious hole in the cylinder head, the first thing my engine builder saw were the cylinder walls. They had surface marks on them. These marks cannot be felt by finger or by fingernail.

When he looked at the pistons he said: you fucked it by knocking, took some pictures and videos, sent them to some local tuners, they all aggreed, to high temperature on the surface of the cylinder and knocking.

Further more, after removing of the pistons, we noticed that on all but one piston (no. 3) the top rings were stuck at exactly the point where he saw small pittings on the crown (pittings which led him to believe it was knocking in the first place). These stuck rings were probably the cause for the cylinder surface marks. What also stood out was that the pittings were on the intake side rather then the exhaust side.

So, now comes the part where I lost faith in everyone and why I am writing this here.

During the tuning sessions we were three people on the dyno. The tuner who is working for a huge company as an gasoline engine developer and who tuned the most OEMs gasoline cars driving on roads worldwide for the last 25 years, the dyno guy who has a tremendeous experience in engine tuning (30+ years) and me, who was an application engineer for 7 years in the same huge company my tuner, where my job was to educate OEM engineers and engine developers on what knocking is, how to detect it and how to properly set up the engine dyno for ignition retarding for specific knock events on this companies specific software. So in other words, 3 people who definitely know what knocking is and how it sounds.

Now, we had all the sensors in place: 2 knock sensors, 8 EGT sensors, 2 temp sensors before charge cooler, 1 temp sensor in intake manifold, pressure sensors before and after charge coolers, oil pressure and temp, engine temperature, water pressure and temperature, 4 lambda sensors and so on. Only thing that was missing is the EMAP sensor which we do not need as the exhaust goes from 2x2.75" to 1x3" and is not restricted after that.

We started tuning knocking thresholds at around 3500rpm. We had a pressure sensor in cylinder 1 for this purpose. Relatively fast we had our thresholds, moved to 4000 and 5500 to see how the knock levels move and behave and set everything up accordingly and with the help of the experienced tuner, extrapolated the thresholds to higher rpm and to have enough safety margin.

For speeds above 4500 for full load tuning, we removed the in cylinder pressure measurements due to experience showing a catastrophic end of an engine if the quartz part falls out - which they regularly do with higher boost applications.

Also, to verify if the engine is knocking and if the ecu is doing the right thing, at later point we disconnected the rear knock sensor and connected it to headphones so that we could hear knocking in the booth if there was any.

During all of the operation from 4500 to 5500 everything was smooth, we had to slightly readjust the knock thresholds as there was no visible or hearable knocking at all.

Only thing noticeable on the headphones was that we might have heard misfire during WOT at 4500 and 5500rpm. However we could not find the reason for that because a) the lambda was stable the whole time, there was no indication of lambda change (which would have happened if one cylinder is misfiring), b) all the EGTs were within 10-20°C within each other and were plausible and c) there was no rpm or engine jitter at all, hence why we have dismissed the idea of misfiring. We even removed the spark plugs and they were exactly how You would expect the plugs to look like with optimal combustion.

As for the tuning procedure:

- the ignition was set up with the ecu knock threshold in mind. so even if we didn't hear any knocking, we didn't want to exaggerate it and put the timing as where we though the engine will be safe. tuner even mentioned that from his experience, we are still to conservative and are leavin a lot of torque behind.

- the injection was set up according to the EGTs. our max threshold is 920°C before turbine, which is exactly what we aimed for. in our opinion this should be perfectly safe. We are running Lambda 0.83 at 5500 and were running 0.8 at 6000 (target at 6000 was 0.83)

So, now for the final part:

From all people we asked, involved or not, there were several opinions and most of them do not match.

What I can tell, just from the hearing aspect, we did not have any knocking, temperatures were fine everything was smooth.

What we have gathered and what seems to be the most common opinion after looking at all parts and logs, is that the knocking that occured is something that very rarely happens as it is not knocking on top of the piston but between the piston crown and the cylinder, in this narrow gap. We also heard some suggestions on how to overcome this issue, but nobody was 100% sure as noone knows how this type of knocking even occurs.

Also, in the meantime, we tried to rehone C5 and C6 and found out that for unknown reasons, both of the cylinders are wider than before from top to like 3cm down.

Questions:

Why did this damage happen?

Are we naive in thinking that these old engines are best setup according to EGT?

Do we need NOT to rely on EGTs and just go to Lambda 0.75/0.76 to be sure and cool the cylinders and pistons and avoid side knocking?

Do we need to coat the piston crown to avoid heat getting into it?

I am open to any suggestion and for a nice discussion as new parts are coming and we need to know how to avoid this happening in the future before putting everything back on the dyno.

Engine data:

Block: RB30

Head: Tomei RB26 Cylinder Head (1mm oversize valves, and so on)

ARP Studs top (11mm) and Bottom

Pistons: Nitto 86.5

CR measured: 9.0:1

Fuel injectors: Bosch 980cc original

CAM: 260in/ex Tomei poncam B with adjustable CAM gear, set to slight overlap

Turbos: 2x Garrett GT2860 with smaller compressor housing.

OEM port matched exhaust manifold

Mines super outlet pro II turbo outlet pipes

Nismo intake manifold

OEM throttle bodies completely restored and sealed but changed to DBW.

Max boost: 1.7bar -> 270kPa absolute Pressure

Fuel: 95ROZ (includes 10% bio ethanol standard)

ECU: Emtron KV8

Spark plugs: BKR8EIX

Engine was honed with torque plate

Main bearing gallery was honed

Piston cooling nozzles installed

Oil return galleries opened to 10mm

Crank was balanced

Lambda demand

Ignition timing table

Thanks in advance to everyone!

Some further information:

Oil temperature never went above 95°C

Water temp was 80-90°C

Room temp was between 35-45°C

Intake manifold temperature during boosted operation was usually 30-35°C

Temperature before intercooler was around 120°C

Oil pressure is at 5bar from 5000rpm, around 3bar at 3000rpm

I am not seeking astronomical numbers, hence why I chose these small turbos. My wish is to have 600hp at the wheels, around 700 at the engine and to maximize the torque curve, which with these turbos it is quite achievable. Peak torque we have seen were 780Nm at 5500rpm and 1.7 bar boost pressure.

Some questions I forgot are:

Do these engines need more fuel at higher boost anyway?

Why did the rings get stuck? How to avoid this in future?

Thanks!