Crankwalk Theory

/ / Magnus Motorsports Blog

Crankwalk Theory

February 2007 – This will be my last update to this. I have left this up on my site as a testament to how naïve and stupid I was when I was younger. I should have listened to my father who had been a mechanic for 40 years. He said, “Change that clutch that’s in there and your problem will go away.”

In the end, I changed the clutch to an OEM clutch, and my problem went away.

Remember the days when your Mitsubishi had a stock clutch and it shifted so quickly? How come it can’t do that with your performance clutch?

All that is listed below means absolutely nothing unless your clutch is properly adjusted and is disengaging properly, have it inspected by a professional. Read below to see how I wasted my time and money, if not for anything I enjoyed the cut up blocks and good times.  Well not all was lost, This research and more like it that followed, allowed us to know exactly how to make this engine perform the way it does now.   Years ago Magnus overcame all these barriers that others are still trying to catch up to now.   How?

Sound engineering + skilled  problem solving = Knowledge.  The most reliable, most powerful Mitsubishi 4cylinders on the planet from years of knowledge.

Crankwalk Theory – The year 2000:

Seeing as how the 2g cars are starting to get up in the years in age and coming off of warranty our shop is beginning to see more of these failures. With this situation becoming more commonplace, and cars getting heavily modified, we will undoubtedly be seeing more and more cars affected with this situation in the future. Unfortunately its starting to get to the point where the amount of cars that I’m seeing is becoming quite alarming, I really only have maybe 50 or so 2g customers at this point, the number is always growing. But I am beginning to see and hear of at least 2 cases a month in the Toronto area alone, through contacts and calls from other shops and dealerships. That’s quite a few affected cars. What adds more fuel to the fire is the fact that 2g owners are basically helpless, Mitsubishi, and Chrysler will not own up to the problem, and no one can come up with a solution to the problem, nor even an answer as to why?

Case in point, we recently installed a fully built bottom end in a customers car (I should point out it had no thrust problems prior) the crankshaft endplay had started to change within 1000kms. It was greater by .002″. That didn’t have us worried yet. Had the customer not cared to come in on time for his first scheduled oil change it may not have been detected in time. We really couldn’t believe it could have been a thrust bearing failure so soon, considering, my machinist (in my books one of the best) who takes great care in assembly had checked the crank after it was ground, and then checked it again after he had assembled the short block. Being a stickler, and I hate leaving things to chance knowing the finicky nature of the 2g, I also checked the endplay. It was within spec. The crankshaft endplay had been checked on three separate occasions, yet the thrust bearing still failed. Why? The car had not been abused, the owner was very careful with the car obeying the break in rules. He had even requested I go with him the first time he put the car thru it paces, after the break in. At 1500km’s we went for that pass. Something didn’t sound right, we came back to the shop and heard what sounded like the early signs of crankwalk, the trigger plate hitting the cranksensor.

Measuring endplay with a dial indicator, we were now out of spec. We were screwed. Who is at fault for all this and who is gonna pay for it? I can personally tell you from experience that I do not like to do things twice, especially an expensive shortblock like this. All I can do is pray for no damage and swallow it hard. Up to now I had been lucky, almost all of the 2g rebuilds that we had done had turned up shiny. So I did some research, I came up with no answers. Basically people told me to keep doing what I was doing and pray they hold together. Forget it, I needed more definite answers than that. Everyone I spoke to in the last 2 weeks, I had quizzed them on the topic, be it parts counter guy, old school domestic car racer, engineers, whatever. Everyone had either no answer or a theory. I had my own. But still no answers.

Something Jeff Hill at Turbotrix told me, was he had heard about a misdrilled oil passage. Danny, my machinist mentioned something about core shifting of castings happening on old chev smallblocks possibly obstructing flow. My good friends Jimmy + Ryan at J+R Trading and I spoke about it over the phone in great detail, we mentioned “block seasoning”; a process some european manufacturers partake in. They cast a block, then leave it outside well oiled to “season” making sure there is no nasty core shift. Nowadays they are “pasteurizing” them or “pickling” them to speed up the process, whatever you wanna call it, manufacturers everywhere are taking shortcuts to keep the cost of automobiles down and their profits up. Besides, who has time to keep oiling a block in the backyard for a few years, by that time most engines are out of production.

This theory intrigued me, I wanted to cut some blocks apart and check these oil passages. I just happened to have a few scrap blocks kicking around. A 95 that had lost a rod bolt and grenaded, then a 93 that had cracked an engine mount hole. We went to Jamaican Racing Woodbridge HQ (J+R) armed with these and a digital camera, with the hopes of slicing them to pieces on a free band saw. If we slice right thru the galleries we can see what the hell is going on and make a more educated guess as to what is is the difference between these engines, why does one suffer from the affliction while the other doesn’t? Jimmy and Ryan were definitely up to the job, everybody loves a good dissection.

So there we were, Jimmy, Ryan, myself and The Wandering Mercenary for Hire (Peter), late night, an empty machine shop, pack of Dumaurier’s, bucket of chicken and a large band saw. Here are the photos and our theory.

Crankwalk Theory Part 2 (click images for larger view)

1995 Talon block destroyed from a rod nut coming loose, a prime candidate for the slicing.

This is the 93 block getting chopped in half thru the main gallery, notice the one hole in the mains casting that feeds each main bearing. The 95 has an additional hole in the first 4 mains, for the different oil squirters.

The 93 block getting cut down the center main journal, through the oil gallery.

Here’s a comparison pic after the carnage, the cut on the 2g block was not exactly thru the center of the gallery, measurement with a bore gauge showed that they were in fact the same diameter, But you can just barely make out the notch in the 95 main journal (arrow) where our little friend the hidden 2g oil squirter popped out of when we were cutting.

The block on top is the 93, you can tell by the oil squirter passages just below the cylinder bores, they take their oil from the main oil gallery where there is plenty of oil flow and volume, right after the filter. The block below, the 95, has none because the squirters are located in the main journals pointing upwards toward the piston. They are sharing oil from the main bearings oil supply. This is the only major difference between the two blocks that we could find. It turns out to have more significance than previously thought read on.On a side note, the 95 has a larger volume in the main gallery, whereas the 93 gallery volume tapers like a cone the further you go down the gallery (away from the oil pump) probably just the manufacturer taking a shortcut to save time in this case.The core shifting theory flew out the window at this point, both gallery’s were unmolested, and could carry more than adequate flow. We were speechless this is when the bucket of chicken and cigarettes came into play.

This is the 95 block cut about 6″ down. You are looking at the block the way it sits in the car, except the top 6 inches have been removed.In the foreground (A) you can see where the oil squirters reside. They are pressed into this hole. They share their oil supply with that supplied to the main bearings.In the background (B), you can see the size of the main oil gallery. It has plenty of volume, but once again more than the 93 (hmm… this gives credence to the clutch switch theory. The larger volume to fill will take longer on start-up to build pressure in the system. Possible but pretty far fetched)At the bottom right hand corner of the picture, you see the main oil gallery and a main bearing journal gallery (small black hole). Oil flows from the pump to the main oil gallery, then through the main bearing journal gallery, into the journals, then feeds the main bearings, and crankshaft then connecting rods.

In the case of the 95 after it gets to the main bearing journal, It is also split up to feed the oil squirters (which cool the underside of the piston in case you don’t know at this point)

The 93 blocks do not have this, they are squirting oil through a restrictor in a different location (as seen in the previous photo), from the main oil gallery where there is plenty of oil flow and volume.

Many arguments followed, and our focus switched to the little understood 2g oil squirter.

So here It was, the oil jet. didn’t look like a restrictor, such as is featured on the 1g block, but looking down one end it looked like a check valve, blowing air through it proved it was. Taking it apart answered a few questions.

It is a check valve. A damn small one, and it was filled with crud. After more argument we realized that these things could clog real easy. The point being, not that they get blocked, but that they could very easily get hung open.

The check valve squirter was probably a good idea in some office at Mitsubishi HQ, but it didn’t look so hot at J+R HQ. With the squirter relocated to the mains you would not want to be squirting constantly at lower oil pressures, like when the car is at idle. Nor would the bottom of the piston need cooling at idle. But in the same regard the fact that it is so small, it could easily fall victim to contamination and allow the ball not to seat properly. When disassembling engines it is commonplace to find some scratches right at the orifice of the oil feed on the bearing surface, this comes from particles that the oil filter has not caught. Remember if an oil filter gets clogged, some have bypass valves which allow the engine to continue to get oil. You get unfiltered oil, but you still have a motor.

The orifice in the check valve appears to be between .030″ and .035″, without a needle gauge set we won’t know for sure. But is that enough to bleed off enough pressure to cause massive thrust failure? Possibly, but what other major differences are there between the two blocks? We found no other major differences between the 93 and 95 block (90-91 have larger main journals). This could explain why there is no actual way to predict thrust bearing failure, some engines could be dirtier than others.

The 1g blocks use an oil squirter from the main oil gallery where there is always adequate volume and flow. The 2g squirters are taking oil from the main bearing journals which share oil with the mains and connecting rods.

Picture this scenario: A particle has wedged itself between the check ball and seating surface of the squirter, the engine is at idle the clutch is depressed, we are now receiving less oil flow and pressure than the main should be getting. Now keep in mind the only way the thrust bearing is fed oil, is from oil that bleeds past the main bearing looking for an escape, it then flows to the thrust. If that check valve is contaminated and stays open, then you are losing some pressure and flow. Today, a day after I posted these findings, I received an email explaining that, they look very similar to ABS unit check valves, which should be kept clean at all times. Those manufacturers put filters in front of them to avoid this sort of thing. (Thanks for the info Joe Enamait) more on that as I get it.

If I had a greater command of mathematics I could calculate the oil flow and pressure lost using Bernouli’s theorem, but I don’t so I will leave that to some other skilled individual. These are our findings and I hope sharing this information can help us find a solution to this problem.

In the meantime, over here at Magnus we will no longer be assembling 2g blocks with these oil squirters in place. Like the balance shafts we find them too risky, any time we can eliminate a source of failure on a motor, we’ll do it. I’m not asking anyone to pull their motor right now to do this, but any new block that is going together around here will be machined to accept the 1g oil squirters, and the 2g squirter passages will be blocked. Another thing we are considering is shimming the oil pressure relief valve to provide a bit more pressure, but haven’t looked into it yet, nor have we compared pumps. Knowing what we know we see the 2g oil squirter not entirely as a risk, because some motors have lived long lives with them, but we see them as maybe a way to end our premature engine failures, end the stressfully hair pulling, and generally calling each other nasty names around the shop.

If anyone has any further information they would like to add to this, any other info or discussion would be much appreciated.

Crankwalk Theory Part 3 (click images for larger view)

Here are is a pic of the 1g oil squirter mounted to the cut block without the eyebolt. The main oil gallery is on the other side of the squirter. 4th pic further down on this page there is a pic of the other side.

Here is a close up of the same thing.

Ohhh, the top view of the 1g squirter, be still my beating heart.

The alleged culprit, the 2g squirter, looking down the block from the top.

The 1g main journals.

The 2g main journals, hmm notice where the squirters reside? Not quite as much volume in the journal. If that squirter hung open, it could definitely starve the thrust of pressure. SOme have brought up the point, “Why aren’t the mains and rods failing” Remember a motor spinning freely at idle really requires very little lubrication, on the other hand though, push on that heavy clutch pedal with little oil pressure and the thrust is now getting quite workout.

Comparison shot, this is where the 1g squirters take their oil from. If there is 10 psi of pressure there at idle, like were seeing in some cars, the volume would still be adequate.

The Actual thrust bearing, from the 95 block.

Front side of that bearing, still looks sharp.

Back side of that bearing, after about 100,000kms starting to wear out pretty good, huh? This car ran pretty good 12.7, no crankshaft doing the harlem shuffle either, but it was getting there. Your probably wondering why there is still some more bearing material at the edges. That is just the way they are manufactured. It is not caused by uneven wear, we can assure you, what was left of the surface was flat.

I hope these clearer pics will help clarify some of the questions some people had, stay tuned as we do some actual testing.

Some thing to keep in mind, were looking into oil pressure’s from car to car and checking it with a mechanical gauge. The chrysler manual says “oil pressure at curb idle should be between 80kPa or 11.4 psi”

More Tests
Let me just explain something first, the response we have received regarding these findings has been overwhelming. I try to respond to everyone’s email concerning the subject and I think I got back to all of you, If I didn’t try me again. When looking at the facts and the way we presented them, most people said “hey that makes sense”. But a few people were still not convinced. Sure why should they be? After all, It was just a theory.

Basically, some told me that our theory was weak, and that it holds no water. Well for a little while we started to question ourselves and wonder, could this be the only culprit? Were we barking up the wrong tree? What did we have to do next to prove that maybe it is the squirters, or is it something else? We had to prove it to the jury.

Oil pumps were dissected and examined, cars were tested for oil pressures, no conclusive evidence was found not even something to raise an eyebrow, but we did replaced a few oil pumps, just for cheap insurance.

After much deliberation, and a dissection of another engine that had fallen victim to crank walk, we decided to do some pressure tests on the squirters. Here’s what we did:

Same people were all at Magnus HQ the other night (J+R Turbo, the Mercenary), lotsa coffee, lotsa cigarettes and a serious chicken shortage. Collecting all the 2g blocks we could find, we pounded out the squirters, they come out pretty easy, with a long brass punch hit the little bastards from the top side and the come right out the bottom. In total we had 9 good oil oil squirters. First thing we did was to check if any of them were clogged like we had suspected. We simply blew through them with our lips; made us look and feel stupid, considering none of them seemed to be clogged except one which was questionable. “There goes the theory”, we thought. Still not the answer we were looking for.

We then cleaned them in the parts washer tank with some varsol, and blew them dry with an air hose. Using an air regulator with a gauge., we hooked them up one by one to see what we could find. Here’s the results.

  • #1. opens @ 18 psi
  • #2. opens @ 14psi
  • #3. opens @ 22 psi
  • #4. opens @ 25 psi
  • #5. opens @ 20 psi
  • #6. opens @ 25psi
  • #7. opens @ 16 psi
  • #8. opens @ 22 psi
  • #9. opens @ 25 psi

Something of interest to note was that the squirters locked up closed at anything past 40 psi. Just in case your doing this at home don’t be alarmed. You’re probably wondering, “are they closing up in my engine”. The only reason we saw this was because we were using air, not oil to pressurize them. With over 40 psi of pressure the check ball would close, sealing up against the outlet inside of the squirter, closing it. Why are we not worried? Lets do this with oil. When the check ball lifts off the seat, oil begins to pass around it through the spring and then out the orifice. This constant flow, which cannot be compressed, keeps the check ball in suspension, in the halfway position, the open position.

What does this mean to you? well the first 4 were from one engine, 5-8 were from another engine, and the last one was from the cut block. It looks like a healthy one should be opening @ 25 psi, logically you would want them to open when the engine oil pump has more speed; therefore pushing more volume. So let us say that they are supposed to open @ 25 psi. Most healthy car are between 15-22 psi @ idle at operating temperature, the manual states that 11.4psi is when the technician should call the oil pump anyway. If we have some check valves that are opening too early because of a relaxed spring, like say #2 in the test, we now have that pressure leak we were talking about. This is happening only at idle, when when most thrust load is taking place (sitting in traffic? foot on the clutch?)

Is this evidence conclusive enough? For some the theory may still be weak. When we went over it again, and looked to find any major differences between the 2 blocks this is the only basic difference between the two that we could find, remember, they are in such a critical oiling area. Denying that this could be a problem and ignoring it, is in my opinion foolish. Like I said before I will not put an engine back together with these squirters in place, way too risky for me, and when you take risks with other peoples cars, they tend not to like it.

Currently we have purchased a milling machine so we can modify the blocks in house. After careful examination of the cut blocks, there really is a lot of casting flash in some areas, they will all receive a cleanup, to help flow. The 1g oil squirter mod is taking shape and we will be offering the service as well as a how to instructions and kit. We may also be trying a new thrust bearing coating, or a lot further down the road, we are trying to find someone who will produce an entire bearing for us out of a few exotic bearing materials we have found.

Stay tuned as promised there are still more tests we want to run.

Mar 17th
A lot has happened in the last few months of building and examining motors. We have not yet given up on the 95+ motors, our conclusion as to what may be happening is pointing more and more to our theory being correct.

All the motors which we have disassembled and were had experienced thrust bearing failure, have had questionable pumps. When I say questionable I mean upon inspection and measurement, they are no good. Mind you some 1g cars that we have found pumps at or beyond the service limit have still run fine and after an oil pump replacement have been found to operate flawlessly.

Then there was that 93 1g that came in with a low oil pressure complaint. After checking it with a mechanical gauge and finding it to be too low, we took down the pan and began to check out the engine. The oil pump was found to be bad, the rod bearings seemed questionable, and the crankshaft endplay was .010″. “Ouch”, I thought, after explaining to the customer his problem, he decided to replace just the Oil pump and take his chances from there, he really could not afford a rebuild and decided to buy some more time till he could build a killer motor. This will be an interesting car to watch as it is completely stock and we’ve caught it in the earlier stages, Yes I know it is not a 2g, but the blocks are very similar and they do share the same crankshaft. We can now watch and see a car that is on its last legs, and see how far its going to get along, or if the oil pump change can save it. After the pump replacement oil pressure was back to normal.

A built motor that had walked was also disassembled and a full postmortem was done. The crankshaft had dug itself through the thrust bearing and into the block. All that survived was the Pauter rods, the rest was junk. Opening up the oil pump, it was the worst oil pump I had ever seen. The block reeked of fuel contamination. The pistons were scuffed. A number of thing killed this motor.

Firstly the pistons (which shall remain nameless) had insufficient clearance, this not being the fault of the builder, but of the piston manufacturer, they claim .0035″ is all the clearance their pistons need. I find them to be a bunch of morons and will not use their pistons again. Customer service was horrible, they of course stand by their story. My question is this, A small block chevy that make 250 hp on 8cyl and 5L can probably run .0035 wall clearance on a forged piston, but a 250hp 4cyl 2L, that experiences double the cylinder pressure should not run these tight clearances, It must run at least .005″, the heat generated will expand the piston a lot more. From now on we will only use ROSS pistons.

Secondly, an aftermarket fuel pump and no engine management had this car running extremely rich, thick carbon on the piston tops, and tonnes of fuel in the oil, combined with the scuffing pistons, had ruined the oil pump. This motor was doomed.

So all we were left with was a nice set of rods, In the meantime I had to build find 2 more blocks and cranks to build motors from, because of catastrophic failures on some other engines. As any of you who have looked would know finding these items is very difficult, we managed to locate 3 used blocks and 5 used cranks. The blocks were checked for straightness, magnafluxed and the such, we ended up with out of all those parts 2 good blocks and 3 good cranks, 1 block was cracked and the 2 cranks were at their thrust limit of .007″ and one was slightly bent.

The good cranks were knifedged and balanced, the bearings were specially coated, the oil pumps were new from Mitsubishi, the squirters were blocked, yes blocked and no squirters were put in place. Why the no squirter? When we machined a block to accept the 1g squirter we soon enough realized that no other machine shop out there was going to be wiling to do this kind of work to the block. Us selling kits wouldn’t matter if people had nowhere to do the install. Instead taking from what we had done before we eliminated the squirters’ altogether. We had been doing this for some time now in race motors to eliminate the possibility of windage, small loss of power due to oil being squirt onto the rods and pistons. Many V-tec honda motors have them welded up to take advantage of this effect. Now before you all jump out and say “we need the squirter”, take this into account, I have heard of many cases of the squirters’ on 1g’s breaking off and falling into the oil pan because of harsh vibration (out of phase balance shafts), we have even seen them in person, doing oil changes, and dropping oil pans.

In my 1g 6bolt motor they were gone a long time ago, I eliminated them in my build-up. That particular engine ran many trouble free miles and is still running today. No ill effects from this mod. I am not going to discuss the virtues of squirter or no squirter because I can go on forever, I am just going to tell you if you so desire we can retrofit the 1g squirters to your block just ship it to us.

So far 2 engines built are running flawlessly and we are happy to report no failures thus far, only time will tell. As extra precautions we have removed the clutch switches to help on startup.