Engine Failure - Dry Sump Lubrication

[singh_robbie]

Hello everyone,

I am Rabsimran Gulati from Formula Manipal, a team from India. We recently participated in the Formula Student Germany and Formula Student Czech Republic events. We faced issues with the engine’s lubrication system in both the events.
Before I describe the events, I will explain the system we are using.

We are currently using a Honda CBR 600RR PC37 (2005 - 2006 model) engine and running a dry-sump lubrication system with it. The previous year was the first time we had implemented the dry sump system. Also, we currently use a Motec M400 ECU, which was also implemented in the previous year. We have designed the intake, exhaust and cooling systems with internals being stock.
The lubrication system includes -
Honda CBR 600RR PC 37 engine with a Dailey Engineering scavenge pump , 2 stage pre ’07 model (04-99-2220), without an air-oil separator and 2 scavenge ports, externally mounted reservoir with a running capacity of 3l (Oil being used, Motul fully synthetic 4T 300V 10w40).
Lubrication Plumbing:
AN -10 scavenge lines (from dry sump to scavenge pump inlet). Scavenge ports oriented transversely w.r.t the car, one of either side.
AN -12 scavenge line (from scavenge pump outlet to reservoir inlet)
AN -6 line pressure feed line (from reservoir outlet to engine pressure pick-up port).

Before running in the event, we had the engine running on this setup for almost 50km on a rolling road dynamometer, and about a 150 km of testing. We were always logging data such as the engine coolant temperatures and oil pressures. We made sure that these values never crossed their limits, knowing the desired oil pressures for various RPM ranges.
However, during our pre-competition testing phase in Germany, the engine started making an abnormal ‘tick-tick’ noise that we had never heard before. Soon after the engine stopped cranking, kind of getting locked internally. Assuming this was a block-related internal issue, we switched to our spare block and continued with the testing.
At FS Czech Republic, in the dynamic test area before the Autocross event, the engine started making screeching sound and we went directly to autocross. During the Autocross, the engine failed with the oil cooler flange separating away from the block. The connecting rod of the 2nd cylinder broke and flew out of the block leaving a hole behind. Also, smoke was observed, probably because of the oil burning up by coming in contact with the hot exhaust pipes. Due to this failure we could not compete in the rest of the competition. Upon checking the logged data we observed that the oil pressure values dropped severely below the desired value during failure.
It was suspected that choked flow rate due to the small ID of the -6 line (from reservoir outlet to engine pressure pick-up port) caused loss of oil pressure at high revolutions causing the failure. We then replaced the pressure feed line from a -6 to -12 size and the tested the system. The overall pressure values were higher than that obtained previously, with occasional momentary unexpected pressure drops still occurring.
We managed to fix the issue before the start of the Acceleration event at FSG and competed in the Acceleration and Autocross events. Could not complete the endurance event due to a leak in the fuel tank (not related to the issue being discussed)
Once we were back home we opened up the engine which was making the noise (which was replaced with the spare engine during testing) and we noticed that the journal bearings of cylinder 2&3 were severely damaged. Find the attached pics.
Pic 1: Bearing of 2nd piston. 1086
Pic 2: Bearing of 3rd piston. 1087
Pic 3: Crankshaft journal of 3rd cylinder. 1088
We are confused as to why this sound (tick) and failure was observed at the competition and not when we were in India, during our 3 months of testing with the system. The only variable being, the change of oil brand from Motul to Repsol fully synthetic 10W40 (which has a higher viscosity at 100oC).
We believe that this was due to insufficient supply of oil to the crank journals due to the -6 line. We continued testing with the new system and we ran the engine for about 170 kms. The oil pressure values were within the desired range.
But the engine recently failed with the exact same symptoms. There was a rattling noise heard during testing and the engine was immediately shutdown. Very fine metal powder was found in the sump and at flywheel. Unfortunately, we did not have the oil pressure values this time as the sensor was faulty and did not log any values. We immediately opened the engine and observed a similar failure. The journal bearings of the third cylinder failed in a similar fashion. Find the attached pics.
Pic 4: Metal powder in sump. 1089
Pic 5: Overlapped bearings of 3rd piston. 1090

Since we observed a similar failure, we understood that our engine issues have not been solved by changing the size of the feed line from-6 to -12.
I hope that the experts in this forum can help solve our issue or at least give us a lead on where we might have gone wrong.

Please feel free to ask any questions if I have missed out on any essential information.

[Swiftus]

Sounds like a typical loss of oil pressure due to G's moving the oil away from the pickup.

Perform a G approximation test by tilting your engine assembly and observe if the oil pickup remains covered in oil. If not, well theres is your problem.

[MKakli]

If you're willing Rabsimran, I think seeing what your oil pan looks like may also paint a picture. I would be sure to see if your scavenge pump is okay as well since you had the fine metal powder. Were there any large bearing pieces?

[Z]

Rabsimran,

Your photos show that you have spent much money on a "racing dry-sump system", ... and got the same result as an off-the-shelf wet-sump.

My guess is that your problem is in the design of your oil-tank (ie. what you call "externally mounted reservoir"). Photos or drawings of this tank and its associated plumbing would help pin down the problem.

Z

[singh_robbie]

The designs of reservoir and sump are shown in the images below:

Sump:
1091

Plumbing on sump:
1092

Reservoir side view:
1093

Reservoir rear view:
1094

The scavenge pump was dismantled at the event and was found to be okay. The inline mesh filters at the ports of scavenge pump blocked the large metal chips of journal bearings.

[Z]

Rabsimran,

Firstly, your photos show that it is fairly certain that your failures are due to lack of reliable supply of high-pressure oil to the crank-bearings. For these types of "slipper/shell/plain-bearings" the oil-supply MUST ALWAYS BE AT PRESSURE and have NO AIR-BUBBLES. So no pressure drops, and no "foaming" of the oil is allowed. (In contrast, all-roller-bearing engines do fine with only a fog of oil to keep things slippery.)

So the problem lies somewhere between the crank-bearings and whatever is supposed to be pumping oil to the bearings. Note that the "scavenge section", in itself, is not a problem here. It is there simply to keep the bottom of the crankcase mostly empty (for less power-sapping drag), and to send most of the oil to the external-reservoir, preferably with the minimum addition of air-bubbles in the oil.

But now I need a clarification. In your first post you say,


Lubrication Plumbing:
AN -10 scavenge lines ...
AN -12 scavenge line ...
AN -6 line pressure feed line (from reservoir outlet to engine pressure pick-up port).

This last "pressure feed line" should go (MUST GO!!!) to the INLET of the main-oil-pressure-pump of the engine. This line must have minimal obstructions and supply pure, clean, bubble-free oil from the external-tank to the main-pressure-pump. From this main-pump it then goes under pressure (at ~2++ bar) to the crank-bearings.

So I hope you are not feeding the oil from the external-tank to an oil-gallery AFTER the main-pressure-pump???

To repeat, the pressure-feed-line from external-tank MUST GO to the SUCTION-SIDE of the main-pressure-pump.

Do you have an oil-flow schematic of your system? Such a schematic is usually the first step of the engineering design process.

Z

[singh_robbie]

Z,

The pressure feed line goes directly from oil tank to the pressure pump inlet through a -12 line and not to a gallery after the pressure pump.

The oil-flow schematic of the current design is shown below:
1095

While going through the ECU logs it was found that the oil pressure at gallery just after the pressure pump (where the oil pressure switch sits in the CBR600RR) showed signs of inadequate oil pressure when the car comes out of slalom. Screenshots of rpm, oil pressure and lateral G's are attached below:
1096

1097

1098

1099

We do not have an air-oil separator but the oil tank has baffles installed parallel to the axis of the tank at the peripheries with tangential oil entry to give a swirl action to the incoming oil for air separation.

[NickFavazzo]

The lack of air-oil separator/cyclone is probably the biggest problem here, if air/foam gets into the oil, you will have pretty ordinary oil pressure/flow. That and baffle design allowing starvation under accelerations. The German track could well have had better grip that your home test venue meaning you never saw the same G numbers in testing vs at comp..

[NickFavazzo]

Have you got data from the skidpad event? Or some cornering with relatively steady throttle?

[singh_robbie]

Nick Favazzo,

Thank you for your response. I'll look into our design for deaeration and if needed, will buy an AOS.

I'm afraid we don't have a proper data for the skidpad event. I have data from FSG '16 Autocross, if you'd like to see.

[Z]

Rabsimran,

Your schematic looks OK. However, it is important that all the little details are also done correctly.
~o0o~

OIL-LINES - The most important oil-line to get right is the feed-line from external-tank to main-pressure-pump. Given that the main-pump has to SUCK oil through this line, the line should be as short as possible, and have a large enough diameter to give low restriction. My guess is that your AN-12 line is big enough, but maybe try to make it shorter (see below...).

It is also very important to check that this flexible oil-line is not so flexible that it gets flattened out by the suction, and so blocks-off the flow. This "flattening under suction" is a very common problem on the suction side of water cooling lines. The flattening-under-suction is more likely on the inside of tight bends of the line, so check this.

Also make sure that there is no joint in this line that allows air to get sucked into it. It seems that your main-pump-suction-line attaches to the sump as an AN-hose, and then I guess there is a gallery up through the sump that mates to another gallery in the engine-block, that then leads to the inlet of the main-pump? Make sure that none of these mating joints can allow air to get sucked into the line. So "O-rings" or "gasket-goo" may be needed.

The other lines, from sump to scavenge-pumps, then to external-tank, should be OK. They could be smaller, but no problem as is.
~o0o~

EXTERNAL-TANK (aka "Swirl-Pot") - Your tank is tall and narrow, which is good. (I originally worried that you made a low and wide tank, which would be worse than the original wet-sump.) Standing the tank up vertically, rather than at the slope you have it, would be better (see why below). An even taller tank would be even better, perhaps with the upper section of smaller diameter.

Your infeed to the tank is done correctly, being tangential to the circular tank, which gives the swirl that helps separate air-bubbles from oil. A smaller diameter hose/inlet-nozzle here, and a smaller diameter of tank, would give much increased swirl and so even better air-oil separation. But maybe OK as is.

I think your biggest problem is at the bottom of the tank. The times that the engine-bearings most need a good supply of oil is when the throttle is wide open (WOT) and engine is spinning fast. So typically during hard longitudinal acceleration. Under these circumstances any fluids in any tanks are sloshed to the REAR of the tank. So the fact that your main-pressure-pump is sucking from the FRONTMOST point of the tank is BAD. The rearward leaning tank also makes this worse.

Furthermore, looking in plan-view, your tank-inlet swirls the oil in an anti-clockwise direction. The tank-outlet should take advantage of this, but your outlet is better suited to picking up CW flow.

So one "quick-fix" would be to relocate the tank-outlet (ie. suction-line to main-pump) to the rearmost point of the tank, but still close to bottom of tank. So shift it from right-side to left-side of tank, as seen in your side-view photo. A vertical rear wall to the tank would also help. Also a horizontal mesh-screen/baffle, at about half-height of the tank can also help to keep the air-bubbles up top, and pure-oil down below.
~o0o~

BOTTOM LINE - The "... tick-tick..." sounds you heard from the engine are just as expensive whether the engine has the most exotic dry-sump system fitted, or just a bog-standard, no-cost, wet-sump. In either case the preventative cure for the "tick-tick" is usually very inexpensive in itself, being only a matter of adjusting a few details.

But figuring out those details ain't always easy!

Keep working at it. :)

Z

[welder1]

Cloth braid lines should have a support spring inside if exposed to vacuum...if that's been overlooked. Hardline with tube nuts and sleeves might be even lighter if chasing grammes.

[singh_robbie]

Z,

The hose we are using is a polymer braided bought from BMRS with the innercore having internal glass fiber reinforcement and wall thickness of approximately 3mm. The bend radii are pretty large too, so, I'm sure we are not running into "flattening under suction" condition.


It seems that your main-pump-suction-line attaches to the sump as an AN-hose, and then I guess there is a gallery up through the sump that mates to another gallery in the engine-block, that then leads to the inlet of the main-pump? Make sure that none of these mating joints can allow air to get sucked into the line. So "O-rings" or "gasket-goo" may be needed.
Yes, you are absolutely correct about our design. It's exactly the way you described. We use an o-ring at the mating surface of gallery in the sump and gallery in the block which leads to the oil pump.

A taller and narrower oil tank with conical base has been designed and is ready to be manufactured. The tank has been relocated to allow a vertical and narrower design. It will be tested on the car soon.

The oil pump was opened today and damaged at the tips. The images are attached below:

1100

1101

The oil jet of 3rd cylinder (which sprays oil on the walls of 3rd cylinder) was found to be clogged by some sealant. Though there weren't any scuff marks on the liners, and the piston rings were OK as well.

1102

1103

The clogged oil jet is marked with red.

1104

We are planning to clean all the oil galleries with kerosene and passing compressed air before assembling the engine. Can you suggest a better way of doing that or maybe checking if none of the galleries are blocked?

[Forbes]

And then there is the really obvious: Is there enough oil in the tank to start with?

Forbes

[Z]

Rabsimran,


We are planning to clean all the oil galleries with kerosene and passing compressed air before assembling the engine. Can you suggest a better way of doing that or maybe checking if none of the galleries are blocked?

Yes, it is important to make sure all oil galleries are thoroughly cleaned before building the engine. And sometimes too much gasket-goo is worse than too little. Many an engine has been destroyed because a large lump of sealant blocks an oil line. Worst is if the sealant blocks the entry to the main-oil-pump.

The standard way of modifying an OEM motor for "more reliable oiling" is to drill out the main galleries to a larger diameter (typically out to 1/2" or 13mm +), then fit screw-in end plugs rather than the original press-fit plugs. This way, after the major engine parts are bolted together, or even inbetween races, the end-plugs can be unscrewed and the galleries visually inspected for debris, lumps of sealant, etc. Also much easier to clean a large diameter hole than a very narrow one.

But it may not be possible to increase the diameter of the galleries on your CBR? Probably not enough wall thickness? Nevertheless, you should be able to remove some of the press-fit plugs and replace them with screw-in plugs (<- just drill-and-tap, then fit a short bolt + loctite). This makes it much easier to clean/inspect all the galleries.

Main thing is to be careful when bolting everything together. Clean everything thoroughly first. Use just enough sealant, but not too much. Use compressed air to blow out the galleries before, during, and after assembly. Blow in and out every orifice you can find, with motor turned every which way up. Sometimes little things get stuck in the funniest places.

The above careful build-process, together with an external-tank that is guaranteed to always feed pure oil with no air-bubbles to the main-pump, should cure your "...tick-tick..." problems!

Again, my current guess for most likely reason for your problems is the main-pump sucking from the wrong place on external-tank (ie. too far forward), and occasionally gulping in some air.

Z

[singh_robbie]

Forbes,


Is there enough oil in the tank to start with?
The recommended volume of oil by the manufacturer and the volume of oil in plumbings along with a buffer volume was calculated and used.

[singh_robbie]

Z,

The conical bottom reservoir with spiral baffles in ready to be put up on the car. The plumbings are now way shorter, with minimum bends. Also, main pressure feed line has been replaced by a -16 line just to be on safer side as we have the Indian event coming very soon.

Two failed engines were cleaned thoroughly and assembled after changing the bearing, bolts and gasket. The engine was run on a jig for about an hour without any tick-tick sound.

We are currently waiting for our pressure sensor which should arrive within 2 days. The new dry sump setup will be tested on the car after that.

[Steve Slowboy]

I suspect you're not running with enough oil. You mention 3 litres? I would expect you to need over double that volume to ensure there's enough air-free oil in the bottom of the tank at all times especially at high RPM and WOT.