I don't think such a glow plug exists. Good idea though!

I have never been comfortable with any explanation of bending rods that does not involve things in the cylinder that are not designed for....Oil, water, carbon or solid objects.

Tom W

If there is a leak in the seal surrounding the cylinder, there will be evidence on the metal seal and on the mating aluminum or steel surface. Pulling a head and seeing perfect shiny mating surfaces and no carbon or corrosion would suggest the seal is good.

Further, even at 3bars and the intake stroke creating a slight vacuum, it would need to be a very large leak to allow even a few mL of oil past in .00025 seconds (the duration of the intake stroke at ~1000rpm). The pressure in the oil system drops quicly as the engine is shut off, I have some difficulty believing that pressurized oil is getting into the cylinder through the head gasket fast in any quantity. Gas goes through the gap faster, would probably push the oil out of the leak/channel at 400psi-1800psi much faster than it could seep into the leak/channel at ~45psi. JMO.

But if I understand his thinking correctly the oil pressure system would be pressurized beyond its normal pressure. I suppose the pressure release valve would release it though, but at considerably higher than 45 psi, I believe.

I am still pondering how a head gasket leak might result in more powerful or somehow more jolting combustion pressure in the cylinder. Extra oil would not do so would it?

If some water or oil leaked into the cylinder enough to increase the compression how will that affect the injection of the fuel when it occurs?

Folks laughed at my thought that vaporizing water might cause a very strong pulse, is there anybody with actual knowledge of what would occur with significant amounts of water in the cylinder? (not like running into standing water though).

Is 1000psi the actual expected combustion pressure?

Tom W

Full open throttle should be about 1,000PSI no more than 1200PSI.

The 3.5 liter 603 cylinder is 9.64 square inches, less 5% larger than the 3.0 liter 603 cylinder.

The piston forces on the 3.5 piston could normally approach 10,000 pounds, but then it is still less than 5% more than the 3.0 liter engine.

With the right amount of trapped oil, the forces could be far higher especially if you have just the right amount so that the rod is nearly fully extended when the piston stops runs out of room to move.

I hope to start testing my engine tomorrow in a day or two.

In the end, knowing why probably wont help me much. but it could help many to prevent failures. Something MB should have done, maybe would have done before in went out of fashion to be honorable in business.

why
 

why is so hard to believe that mb designed a weak piston rod,they redesigned it stronger to solve the problem.
larry perkins

The seemingly random nature of the failures.........no definitive pattern. With a weak rod, there should be a considerable number of failures by 100K and nearly complete failure by 150K..........as an example.

Some engines go 300K on the original rods...........so, that's not a basic design problem.

We don't know Larry.

If we knew more things, such as the distribution of the failing engines: If the failures were more common on engines produced in Q3 & Q4 of '91, or that the failed engines all had head-gasket leaks, or that all of the failed engines had the EGR still enabled, etc.

The expansion rate of water to steam is 1600:1 Tom, don't know if that helps. Yes, oil in the cylinder is more fuel so more heat, regardless of whether it enters through the induction system, injector, or a leak in the seal.

It's hard for me to believe that the oil supply system could become super-pressurized and force oil back into the cylinder, the oil would become gasous and blow a lot of oil out of its way into the oil journals though and likely this process continuing would create an oil-starvation where the gas was bathing journals instead of oil.

I'm sticking with my pet theory, where oil is ingested into the front cylinders from a build-up in the induction system, an excursion into high-airflow / high-rpm allows at least one gulp of oil into a cylinder (or cylinders) large enough to do damage, can happen more than once, in early heads it would crack the heads in later-stronger heads with the 3.5L rod geometry it would bend the rods. Works for me, my induction system is clean and EGR disabled.

Now I can happily listen to others' theories and wait for a better explanation.

<edit> I mean wait for an explanation I like better.

It seems to me to be a well thought out theory.

Tom W

i agree. Secret to keeping the ghost in the machine... Might be a resivoir right below some of the cylinders?

I don't follow anything except the 'I agree' part.

Tom W

A couple of problems with the oil theory.

The early heads suffer from cracking due to excessive temperatures. We're all fairly sure about this. It's easy to get a 603.961 to go 300K on the original head if it's never overheated. Therefore, the argument must exclude the 603.961 due to head cracking. Furthermore, the crack in the head is very slight when it begins..........nowhere near sufficient to prevent a bent rod if the oil scenario occurs. The compression is usually decent with a cracked head and many people drive them a long distance in that condition.

So, the question again begs..........how does the 603.961 live with rarely a bent rod (not impossible) but the 603.970 suffers from the malady on a much more frequent basis.

Each of your theories..........water or oil.........would be viable if some explanation could be provided that covers the older engine. To date, nobody has satisfactorily done it.

The older engine does not have the block flex that the 3.5 has.

I agree that the cracked heads on the 3.0 did not come from hydrolock.

Tom W

Its a gusher!!!!
 

Today we tested engine #235. Removed #1 injector, removed the air pipe and the blow-by tube. It was down 5 quarts of oil so we filled it. We fired it up and let it idle. Every time the piston stroked up, compression and exhaust both, oil sprayed out of the pre-chamber throughthe injector bore. We let it idle for about 5 minutes. There was oil all over the shop and the shop owner. We then shu tit down and took it apart.

The intake manifold was oiled everywhere, wet but not soaked. The inlet tube going from the air filter housing to the turbo was wet from the cam cover vent tube inlet the turbo compressor inlet.

The exhaust manifold was wet with oil inside.

We removed the head and found the gasket blown out at the oil channel adjacent to #1 cylinder. (now, AKA the aorta on 603 engines) The seal ring was broken and the ring was bend outward on both sides of the break. the gasket material was burned or melted or otherwise alblated all the way acros the oil channel and through to the chain tower. So the gasket was breached all the way across the oil channel and into the chain tower.

Clearly, oil was pumping at high volume into #1 from the breach. Far more than most of us would imagine. After running less than five minutes, we drained four quarts from the engine. So at least 3 quarts came out through the open combustion chamber. With the injector in, the dealer estimated it lost only 3 quarts in 45 minutes. I would assume that oil was then leaking into the cylinder and then being pumped out through the breach by the piston when the injector was in.


We will measure the piston clearances to look for any signs of bending in the rods. If there is any it will require a dial indicator to tell. Visually the pistons are all flat and even against the cylinder wall edges. The cylinder walls look nice with no signs of unusual wear. Everything looks really good in the head as well.

There was excessive blow-past noted when the engine was run with the injector in place. Ufortunately, we did not check for blow-past today. We will do that when the head is back on. I expect virtually none.

I do believe that engine oil was being sprayed and atomzed at high pressure through the chain tower/oil channel breach and then being sucked into the turbo via the cam cover vent tube.

We will be putting the head back on tomorrow or the next day.

Here is a link to some photos: http://www.misterdiesel.com/Albums/92SD/index

There is how you go through 2 quarts in 45 minutes of idling.

Now when you measure you will measure from the deck to the top of the piston, right?

Tom W

New Rod Bender Theory - Oil Mist Pre-Ignition
 

After examining the failure of the 3.5 liter turbo diesel engine # 603917 -12-000235 mentioned in the "Rod Bender” thread http://www.peachparts.com/shopforum/diesel-discussion/217243-rod-bender-true-stories-how-broke-why.html, I am proposing a new failure theory that, if correct, would explain virtually all bent rod failures in these engines.

Oil Mist Pre-Ignition Theory

Summary:

Oil mist suspended in the gasses found inside the engine crank case, cam cover, and chain tower are evacuated by the turbo charger via the cam cover vent tube, then mixed with compressed air and forced into the intake manifold. This is a normally occurring process in all turbo charged 617 and 603 engines. Normally the amount of oil is small. Normally, the amount of new gas formed chemically by burning during the compression stroke is small. Combustion pressures are nominal.

On some 3.5 liter 603 engines (maybe 1%) the area of the #1 cylinder seal adjacent to the oil channel in the head, and next to the chain tower leaks. The leak almost always begins as combustion gas leak at high (near full throttle) pressure. The leak develops over hundreds of thousands of cycles. At some point combustion gases escaping past the loosely held seal ring burn through the gasket material to the chain tower, creating an oil atomization site. As this continues, there comes a critical point at which the intake gas charges contain enough atomized oil, and ignition occurs early enough in the compression stroke, that the amount of gas(mass) in the cylinder is large enough to overload the cylinder(a bit over 2,000PSi should do it). This pre-ignition occurs sporadically, only when conditions are right. Probably higher rpms, fully warmed-up engine, colder days (nice thick intake air).

Please challenge this theory. If it is weak, it should be destroyed. If it survives attack, and it is correct, it means rod bending can be predicted by monitoring oil in the vent tube, and prevented by re-routing it.

Fire at will!