Kubota L4740 Engine "knock" Help?

[pmsmechanic]

It would be worth it to loosen the belt and see if the pulley is loose.

 

[Alien]

Yes it sounds a bit like a conrod big end but if so then fine metal should be visible in the oil.

 

[jshwhite]

That is my fear but can't say for sure. As I said earlier, after hearing your video just at the end it sounded like a rod to me. It is very hard to diagnose noise from a video, for me at least.

Yeah I know it really doesn't come through on the video really well. Wish I knew a pro diesel mechanic who could actually come and have a listen.

Boy that video is really a stumper,I don't think there is enough pulley movement there to make that much noise.It's seeming more & more like a rod but the only way to know for sure is to pull the pan and see. Not an easy job as in a lot of cases you have to pull the engine to get the pan off. Depends on how you are set up, shop wise, if you can diy.

I have a good shop set up, no problem on the diy. It's a split pan type with the driveshaft running through the middle. Not sure if that shaft can be removed and drop the pan with the engine in place or if I have to remove the engine. I'll have to have a look at it tomorrow.

It would be worth it to loosen the belt and see if the pulley is loose.

Good idea. I'll do that.

 

[jshwhite]

Checked a couple other small things this morning.

Ran it without the exhaust/muffler just to be sure it wasn't a bad muffler. Still knocking.

Ran it for a few seconds without the belt to eliminate the water pump and alternator. Still knocking.

The crank pulley is definitely tight, as is the shaft. I grabbed it and shook it around pretty hard and couldn't get any movement out of it.

I wanted to run a compression test, but looks like I don't have the right adapter for the gp threads. Not sure if that would tell me anything though.

Pretty well boiled down to something in the low end, most likely rod knock, right? Or am I missing something else?

I downloaded a service manual, and it looks to me like the propeller shaft can be removed and the oil pan dropped without any other disassembly, so I guess I'll start with that. Any other thoughts or ideas welcome.

 

[rScotty]

Based on everything that's been eliminated at this point, rod knock is about the onlything left, am I correct?

Too bad that the valves are fine. You know, I assume that those are solid lifters, but I don't know. Best check it out with Messicks. if they are not, we still have the possiblity of a collapsed hydraulic lifter knock.

I'm not buying the rod knock. The problem I have with it being a rod knock is that I believe you said you loosened each injector nut until that cylinder didn't run any more. At that point, I would expect a rod knock to make a very obvious change in the sound. Same for small end or piston slap. When it isn't firing shouldn't it be different? Run the exhaust out the door as we discussed so you can hear it.

What if it is a worn piston/ broken rings and the accompanying piston skirt slap? That's a very hard one to diagnose, but it brings to mind another idea. If you have an inspection scope - and here is an excuse to get one - then are not expensive anymore - then you could pull injectors and if it is direct injection you can stick the scope down the hole and look at the cylinder wall for scoring. Sometimes you cannot get a scope in if it is indirect injection. But maybe through glow plug hole. The service manual may have cut-away view of combustion chamber.

I'm not sure what to say about the front crank pulley. I'd sure have a look though. Is that a two part pulley?? Could it be as simple as that? After all, a two part pulley with slop would also wear out that oil seal.

First thing I'd do is pull the fan belt off and run it sans belt to see if that makes any difference in the knock. It's simple and might tell us something Be nice to have a high speed camera and mic.

One other thing you could do is compare the frequency of the knock with the RPM of the motor....not quite sure how to do that, but it shouldn't be too hard to get an idea of the ratio, and there is a ton of information to be had there.
luck, rScotty

 

[jshwhite]

One other thing you could do is compare the frequency of the knock with the RPM of the motor....not quite sure how to do that, but it shouldn't be too hard to get an idea of the ratio, and there is a ton of information to be had there.
luck, rScotty

They are solid lifters, yes. And yes, I ran it without the fan belt on - no difference.

Funny you mention the frequency of the knock. I had the thought too. And actually I do some audio recording/engineering for a hobby, so I have the capability I did two takes - one at 1500 RPM and one at 1800 (measured with a tach on the crank pulley). At 1800 rpm, it occurs 15 times per second. At 1500, it's 12.5 times per second. Do the math and that's exactly 1 knock every other revolution. (1800 rpm = 30 rev per second. Divide by 2 =15. Same for 1500.)

Also, when looking at the audio on the spectral frequency graph, the knock occurs every 4th "bang" or "pop" or whatever you want to call the cylinder firing. So the rhythm is "POP, pop, pop, pop, POP, pop, pop, pop"

Not sure what that tells me, but it's food for thought...

 

[crazyal]

How many hours does it have on the oil? I think I would change it and send a sample to a lab to have it analyzed. If it is a main or rod bearing that'll tell you.

 

[metalbender]

Pull the oil filter, cut it open, check for metal particles in the pleats of the element.

 

[K5lwq]

If you load the motor and the knock gets lighter, it most likely will be a rod.

 

[rScotty]

They are solid lifters, yes. And yes, I ran it without the fan belt on - no difference.

Funny you mention the frequency of the knock. I had the thought too. And actually I do some audio recording/engineering for a hobby, so I have the capability I did two takes - one at 1500 RPM and one at 1800 (measured with a tach on the crank pulley). At 1800 rpm, it occurs 15 times per second. At 1500, it's 12.5 times per second. Do the math and that's exactly 1 knock every other revolution. (1800 rpm = 30 rev per second. Divide by 2 =15. Same for 1500.)

Also, when looking at the audio on the spectral frequency graph, the knock occurs every 4th "bang" or "pop" or whatever you want to call the cylinder firing. So the rhythm is "POP, pop, pop, pop, POP, pop, pop, pop"

Not sure what that tells me, but it's food for thought...

I want to think on that... What I'm trying to do is picture how that fits with no overt change in the knock when you disabled each injector.
And so far I can't reconcile the no change and the no injection. I was just so sure that would at least point to the correct cylinder even if it didn't tell us the cause.

A good way to trouble shoot the difficult ones is to divide up the problem. One problem is we don't know which cylinder, the other problem is we don't know why the knock is happening in that cylinder. We ought to be able to at least get it down to the proper cylinder.

Er, I'm assuming that your spectrum analysis is pointing towards one cylinder. Do you agree with that?

Could it even be something crazy? ....like a dangling glow plug end?

 

[dodge man]

I wish I could help more but my reaction is you are delaying the inevitable. Just pull the engine, crack it open and figure it out.

 

[tbk5]

To the OP, I am sorry for your troubles. But, tbis is one of the most informative threads I have read in a long time. Kudos to all responding and TBN

 

[sixdogs]

To the OP, I am sorry for your troubles. But, tbis is one of the most informative threads I have read in a long time. Kudos to all responding and TBN

Very true and what a great thread. I sure hope this has an easy solution.

 

[Tx Jim]

jshwhite
Someone may have already suggested this but have you loosened the inj lines 1 at a time with engine running to determine if knock can be isolated to a certain cylinder? If knock can be isolated to a cylinder I strongly suggest to have that injector or even all injectors pop-off pressure tested.

 

[jshwhite]

If you load the motor and the knock gets lighter, it most likely will be a rod.

Thanks for that info. I did try that and I think it's slightly lighter under load, but not a big difference. Not sure how much difference it should be.

I want to think on that... What I'm trying to do is picture how that fits with no overt change in the knock when you disabled each injector.
And so far I can't reconcile the no change and the no injection. I was just so sure that would at least point to the correct cylinder even if it didn't tell us the cause.

A good way to trouble shoot the difficult ones is to divide up the problem. One problem is we don't know which cylinder, the other problem is we don't know why the knock is happening in that cylinder. We ought to be able to at least get it down to the proper cylinder.

Er, I'm assuming that your spectrum analysis is pointing towards one cylinder. Do you agree with that?

Could it even be something crazy? ....like a dangling glow plug end?

I would have thought disabling the injector would identify it also. No such luck though. But yes the audio certainly is pointing to one cylinder. But I think we knew that already.

I wish I could help more but my reaction is you are delaying the inevitable. Just pull the engine, crack it open and figure it out.

Yeah, I know. I was actually thinking the same. Just hate to do that and find out I didn't have to.

To the OP, I am sorry for your troubles. But, tbis is one of the most informative threads I have read in a long time. Kudos to all responding and TBN

That's what I love about TBN. Lots of helpful and experienced folk on here. I've helped a few people, but I think I've received a lot more help than I've given.

jshwhite
Someone may have already suggested this but have you loosened the inj lines 1 at a time with engine running to determine if knock can be isolated to a certain cylinder? If knock can be isolated to a cylinder I strongly suggest to have that injector or even all injectors pop-off pressure tested.

Yep, been there done that. All it does is drop the RPM, but the knock stays the same.

How many hours does it have on the oil? I think I would change it and send a sample to a lab to have it analyzed. If it is a main or rod bearing that'll tell you.

Yep, that's what I'm doing. It's got a bit under 100 hrs on the oil. Probably not enough to tell in an analysis, but I'm sending a sample off this morning and will see what I get back.

Thanks again everyone. Going to put this on hold for the moment while I wait on the oil analysis results.

 

[jshwhite]

Update: I got the oil sample results back. High aluminum levels, indicating a bearing. They said of course aluminum could be piston, but chrome (rings) and iron (cylinder liners) levels are normal, so most likely a bearing. Rod bearing of course being most likely, although could be a main crank bearing I suppose.

Busy this week, but I'll tear into the engine next week and see what I find.

 

[K5lwq]

Thanks for the update. I would definitely fix this one. Still a very nice tractor. Good luck!

 

[TMGT]

Well that's not what any of us were hoping to hear.
Makes you wonder what the PO did to cause such an early failure.

 

[goeduck]

This will be very interesting. I hope it is relatively easy to correct, considering. Looking forward to your data/pictures.

 

[jshwhite]

Re: Kubota L4740 Engine "knock" Help?

Ok, finally got some time to start on this project. I dropped the oil pan today and had a look up in it. Do you think I should be able to feel some play by hand on the offending part? I can't seem to get any movement other than side-to-side for each rod end.

There were several metal shards stuck in the pickup screen. And a lot more in the bottom of the pan. See picture. Seems that should give a clue?

Those of you with more experience in this, what do you think? Is there some way I should be able to identify the issue from the oil pan, or should I just bite the bullet and tear into the engine?