CAT 3508 in a Hogzilla

[Tony Wells]

3508 CAT sn BPX00105

I suppose this belongs under Forestry. If not, maybe some Mod could move it for me.

I have a rough running engine in my tub grinder, and have the injectors out. CAT is selling me new o-rings to reinstall them after a mild cleanup. My engine guy advises me that we should try this as a next step, since several of these seals are cracked. Cheap enough to chance it before replacing the injectors. I have a kind of twofold problem. The rough running, possibly, has caused about 3 pages of DTC's on my Diesel Lappie, but I think most are old and just stored in memory. I don't have the printout in front of me atm, so can't provide a list. But my thinking is to clear the codes if I cannot determine which are live/current and which are old, fixed and are just stored for reference and start over with no codes stored at all. Then any that pop up I know would be current. I'd like opinions on that.

The other thing is I am in need of a schematic for this engine control system, based on the overall hodge-podge of codes that are there but don't seem related either to the current rough running or the cracked injector seals. Something tells me there is a problem between the engine and fuel control modules in the harness, so rather than peel it all open and try to find an open or short, or possible bad ground, I would like to have more info. So far, I am told that CAT will furnish a complete engine diagram that includes the ECM and fuel module where I can get the pinouts for all the connectors. I believe I can find a harness issue with that information. We don't do much of our own heavy engine work, and this is about a $100k engine, so I use an outside specialist I've come to trust. Problem is, I really don't want to pay CAT $800 and wait two weeks for their manual only to use (at this point anyway) only the electrical information. But I do need to eliminate any potential problems with the electrical system. If anyone can help me locate a generic 3508 schematic, it might go a long way to help find some ghost problem. I do pretty fair on the electrical/electronics side of things, and my health doesn't let me do the heavy lifting any more so I feel I need to start there. I'm pretty well equipped with instruments of various types to test most anything common in a harness. Plus a have a decent helper to do the climbing around. He's learning.

Naturally, the manufacturer of the machine doesn't offer the same level of information as I believe I need. I want sensor outputs, ranges, etc so I can test them directly after the injectors put back and it is ready to run again. Right now I will not even turn the key. Power is locked out to protect it from dummies anyway.

So I rest it's fate in the HEF Brain Trust to guide me through this, hopefully with some insight and experience. I'll settle for good guesses though. They often lead the right direction. And as always, I appreciate any tips offered.

Some of you fine folks have my email, or maybe it's in my profile, but if you'd rather talk to me off record, feel free to contact me that way. Or PM me and ask for it. I'll happily provide it.

If any additional information will help or is needed, let me know and I'll get it.

Tons of thanks,
Tony Wells

 

[Per Eriksson]

Here's the schematic, RENR4922 for your serial number.
http://www.catsch.info/renr4922.html

Could you post the list of codes you have?

 

[Tony Wells]

Thank you, Per, for that link. I'm sure it will help. Wish my plotter was running, I'd print that out at lease D size so I wouldn't have to use my magnifying glass. I'll just take the laptop along.

Here are 3 screenshots from our laptop. I think there may be a page or two more, but this is all the tech sent me.

It appears that all but one (CODE 91) is just stored in memory. I can't even tell when each of those codes were set, all at once or spread out over years. The TPS code, assuming this engine actually uses one, says it's a valid, current fault. Tomorrow I plan of hunting that down and making sure the harness doesn't have a problem on that circuit, again assuming it actually exists.

I picked up o-rings today and will get the injectors back in and along with my engine guy, get it cranked and see what the laptop tells us fresh. I'll post what we see after we do that. It's possible I guess that it could be fine. This problem came on suddenly after the engine had been running smooth and normal for quite a while, so I really think this is something simple. Sometimes easy to find, sometimes not so much.

Now I also get to track down a main hydraulic pump for an old Galion grader tomorrow. All the brass pistons went to pieces, and are all downline, so flushing it will be a treat. We sold it, but agreed to fix it. It's an early 80's model, so that should be fun. And of course, it's urgent, like everything else. I'll never get caught up.

 

[Per Eriksson]

The only code in the list I think could cause rough running is the speed control code, a bad signal there could make the engine hunt rpm's and that can sound as it is missing or not running on all cylinders.

Table 1
Diagnostic Codes Table
Code and Description Conditions which Generate this Code System Response
91-8 Throttle Position Sensor abnormal frequency, pulse width, or period The Engine's Electronic Control Module (ECM) detects a sensor frequency that is less than 150 Hz or a sensor frequency that is greater than 1000 Hz for two seconds.
or
The ECM detects a duty cycle that is less than 5 percent or a duty cycle that is greater than 95 percent for one second. The code is logged.
The throttle signal defaults to a zero percent duty cycle (low idle).
The speed control is used to provide a throttle signal to the engine control. The output from the speed control for rated speed is a pulse width modulated signal. The signal varies with the position of the speed adjust potentiometer. The ECM calculates the desired engine speed from the signal from the speed control. The rated speed signal is valid when the duty cycle is in the range of 5 to 95 percent. If the ECM detects a rated speed signal that is invalid, the engine will be set to low idle.

Some applications may choose not to use the speed control. Some applications may use a load sharing unit. The unit is configured to provide the engine control with a rated speed signal. Some of these units cannot provide a signal until the engine is running. In these cases, perform troubleshooting while the engine is running.

Test Step 1. Inspect Electrical Connectors and Wiring

1. Turn the switch that provides power to the keyswitch input for the ECM to the OFF position.

1. Turn the breaker for the ECM to the OFF position.

1. Thoroughly inspect the ECM connector J1/P1. Inspect all of the Connectors in the circuit. Refer to Troubleshooting, "Inspecting Electrical Connectors" for details.

1. Check the ECM connector (allen head screw) for the proper torque. Refer to Troubleshooting, "Inspecting Electrical Connectors" for details.

1. Perform a 45 N (10 lb) pull test on each of the wires in the ECM connector that are associated with the circuit.

1. Check the harness and the wiring for abrasion and pinch points.

Expected Result:

The components and wiring are free of the following problems: damage, corrosion, abrasion and incorrect attachment.

Results:

• OK - The wiring and connectors are OK. Proceed to Test Step 2.
  
  
  
• Not OK - The wiring and/or connectors need repair.
  Repair: Repair the connectors or wiring and/or replace the connectors or wiring.
  
  STOP

Test Step 2. Check the Voltage at the Speed Control

1. Connect the red probe of the digital multimeter to the +Battery terminal of the speed control. Connect the black probe of the digital multimeter to the −Battery terminal of the speed control.

Expected Result:

The digital multimeter reads 24 ± 3 VDC.

Results:

• OK - The voltage supply to the speed control is correct. Proceed to Test Step 3.
  
  
  
• Not OK - The digital multimeter did not read 24 ± 3 VDC.
  Repair: There is a problem with the supply voltage to the speed control. Verify that any circuit protection is not tripped. Repair any wiring or connections that are damaged.
  
  STOP

Test Step 3. Check the Throttle Position on Caterpillar Electronic Technician (ET)

1. Connect Cat ET to the service tool connector.

1. Turn the breaker for the ECM to the ON position.

1. Turn the switch that provides power to the keyswitch input for the ECM to the ON position.

1. Observe the position of the throttle on Cat ET.

1. Turn the speed adjust potentiometer to the lowest setting.

1. Start at the low idle position. Slowly turn the speed adjust potentiometer in the other direction. Monitor the position of the throttle on Cat ET.

Expected Result:

The throttle position on Cat ET reads 0 percent at the lowest setting and 0 to 100 percent as the speed adjust potentiometer is adjusted.

Results:

• OK - The speed control is operating correctly.
  Repair: There may be an intermittent problem. If the problem is intermittent, refer to Troubleshooting, "Inspecting Electrical Connectors".
  
  STOP
  
  
  
  
• Not OK - The throttle position on Cat ET does not respond to adjusting the speed adjust potentiometer. Proceed to Test Step 4.

Test Step 4. Check the Speed Control

1. Remove the wire from the S terminal at the speed control.

1. Set the multimeter to "VDC". Press the "Hz" button twice so that the % symbol is displayed.

1. Place one probe of the multimeter on the S terminal of the speed control. Place the other probe on the −Battery terminal of the speed control.

1. Monitor the output of the percent duty cycle from the speed control on the multimeter. Turn the potentiometer from the low setting to the high setting.

Expected Result:

The duty cycle is between 4 and 10 percent at the low setting and between 90 and 95 percent at the high setting.

Results:

• OK - The speed control is supplying the correct signal. Proceed to Test Step 5.
  
  
  
• Not OK - The duty cycle is not correct on the multimeter.
  Repair: Check the speed adjust potentiometer for proper operation. If the speed adjust potentiometer is operating correctly, temporarily install another speed control. Verify that the problem is no longer present.
  
  STOP

Test Step 5. Check the Throttle Signal at the ECM

1. Turn the switch that provides power to the keyswitch input for the ECM to the OFF position.

1. Turn the breaker for the ECM to the OFF position.

1. Remove wire F702-GN from P1-10 at the ECM connector P1.

1. Turn the switch that provides power to the keyswitch input for the ECM to the ON position.

1. Turn the breaker for the ECM to the ON position.

1. Measure the duty cycle between wire F702-GN and P1-5 of ECM connector P1.
   Note: Set the multimeter to "VDC". Press the "Hz" button twice so that the % symbol is displayed.

1. Monitor the output of the percent duty cycle from the speed control on the multimeter. Turn the speed adjust potentiometer from the low setting to the high setting.

1. Turn the switch that provides power to the keyswitch input for the ECM to the OFF position.

1. Turn the breaker for the ECM to the OFF position. Reconnect all wires and connectors to the original configuration.

Expected Result:

The duty cycle is between 4 and 10 percent at the low setting and between 90 and 95 percent at the high setting.

Results:

• OK - The duty cycle is correct. Proceed to Test Step 6.
  
  
  
• Not OK - The duty cycle is not correct.
  Repair: There is a problem in the wiring between P1-10 and the S terminal of the speed control.
  
  STOP

Test Step 6. Check the ECM

1. Turn the breaker for the ECM to the ON position.

1. Turn the switch that provides power to the keyswitch input for the ECM to the ON position.

1. Monitor the percent throttle position on Cat ET. Turn the speed adjust potentiometer from the low setting to the high setting.

Expected Result:

The percent throttle position is between 0 and 100 percent for the above settings of the speed adjust potentiometer.

Results:

• OK - The throttle position on Cat ET is correct. There is an intermittent problem in the wiring and/or connectors between the ECM and the throttle position sensor.
  Repair: If the problem is intermittent, refer to Troubleshooting, "Inspecting Electrical Connectors".
  
  STOP
  
  
  
  
• Throttle Position Not OK - The throttle position is not correct. The ECM is not processing the signal correctly.
  Repair: Replace the ECM. Refer to Troubleshooting, "Replacing the ECM".
  
  STOP

 

[Tony Wells]

Per, thank you for that troubleshooting chart. Once I get the injectors back in I'll go though it. Until then about all I can do is a physical inspection of the harness to make sure there is no obvious damage to it. I really don't think there is, but at the connectors, there always is a chance of a questionable crimp or oxidation...that sort of problem that isn't so visible. I'm sure they could all use a good break and make, with some DeOxit application, anyway. This machine being a tub grinder, with a 2400 lb hammer mill spinning at 1800 rpm, it's not the smoothest environment, and there have been a few quirky wiring issues already. If the harness isn't adequately supported and anchored, the vibration will cause problems, so I envision doing end to end continuity checks while performing the wiggle test. Hopefully this all comes from the injector seals. I removed the old seals today and some had actual gaps of 1/2-3/4" in the o-ring, although I doubt they were in the engine like that, but that tells me there were definitely cracks and that the o-rings are brittle and in need of replacement. Have my fingers crossed.

 

[lantraxco]

Just my $1.98 (two cents with inflation and liability insurance) Make double sure your battery cables and all associated connections are sparkly clean, especially grounds wherever they may attach. I've installed a couple CAT engines in custom machines and CAT is very fussy about how and where everything gets grounded so there's no possible loops or voltage differences to throw sensors off or out of their expected voltage range. That and double check anywhere the engine harness could possibly rub and any conductor short to ground, one short can throw several DTC's if the sensors are all on the same supply circuit. Cheers.

 

[Nige]

Here's a pdf copy of the schematic Per posted earlier. It won't depend on the speed of an internet connection to open if you are out in the sticks.
Also a pdf copy of the CID0091 FMI08 Troublshooting Procedure. It includes some graphics regarding pinouts that may be of assistance.

If you can erase all the Logged Diagnostic/Event Codes it would go a long way to helping you see the wood for the trees. Does your laptop have that capability..?

 

[Tony Wells]

Thank you Nige. I'll see if my old plotter still runs and get a D size print of that schematic. I hate trying to scroll around a large, detailed drawing tracing wires. And I've printed out the test sheet. Very helpful.

I believe I can clear the memory with our laptop, and plan on doing that. The o-rings for the injectors are in hand, and although more urgent matters pulled me away from this, I'll be back on it probably tomorrow. I'll keep posting as I progress.

And thanks, lantraxco. That was part of the plan. I know that grounds can throw a lot of tricks at us, and planned on breaking and cleaning every one, regardless of which element of the circuit I have a problem with. Just an advanced part of PM to me. They are far more important than most give credit for. Weird stuff happens with faulty grounds. Especially on equipment out in the weather. When I am bad ground hunting, I try to do it on a humid day. Seems to make them easier to find. I have a very low current hi-pot tester than generally lets me find them without disturbing anything, which speeds things along.

Thanks all!

 

[Tony Wells]

OK guys, next move. Injectors back in and engine running smoothly. Suspicions of a mechanical problem that is audible. But more on that later. Presently, there is an exhaust leak. Now this is where to me it gets a little weird. I talk to my parts guy, and he tells me there are two basic setups on the 3508. One uses a bellows coupling between the manifold sections, and I don't have that. The other option strikes me as odd because there is no part listing for gaskets between these manifold segments. Now, I have seen metal to metal seals, usually a sort of spherical male/female interface so there can be some misalignment. I'm guessing something like this is used here. I soaked the bolts/nuts down today with penetrating oil with the intent of putting a wrench to them tomorrow, so I'll find out what seals the system then. Supposed to rain, so may not happen.

Can anyone supply me with a R&R procedure for removal of the manifold? I can't hear real well, but everyone who can tells me the leak seems to be on the right hand bank. I'm assuming I can remove only half the system. Even an exploded drawing would help. Or any helpful advice or tips on what NOT to do would be appreciated. After the leak is fixed, the new noise should be more distinct and hopefully easier to localize and identify.

After this, we get to talk about the new mystery noise.

 

[lantraxco]

Your serial number shows a manifold setup with 4 bolt external sleeves. I've never done one like this, but it appears from the drawing you just unbolt the sleeve at each end of the segment you want to remove, maybe loosen everything enough to create some slack and then remove the piece you want? More experienced hands I am sure will join in...

 

[Nige]

......but it appears from the drawing you just unbolt the sleeve at each end of the segment you want to remove, maybe loosen everything enough to create some slack and then remove the piece you want?

Oh that it was that simple ............

To get one or both sides of the exhaust manifold off (I'll come back to the design of it in a minute) the aftercooler and the inlet manifolds have to come off first. It can be done with the inlet manifold left in place but it's a total pain and not worth the trouble trying IMHO because it'll take at least twice as long and be a struggle all the way.

The "joints" between the manfiold sections are 4-bolt and are designed to have a "slip fit" between the sections to allow for expansion as the engine beds in during the first few hours. It's a design that works well on genset engines that run on reasonably constant loads but on any engine subject to any sort of cyclic loading it's a paiin. The problem is that once the sections of the manifold weld themselves together if there is any sort of movement later due to heat cycles it over-stresses the bolts between the manifold and the cylinder heads and they start breaking. I would suspect your exhaust leak is one or more broken bolts. Good luck getting the broken pieces of bolt out - you'll need a very good welder and lots of patience. Sometimes I've seen the head have to come off. I hope you don't have to go that far.

Suggestions:
1. Remove both sides of the exhaust manifold even if only one side is leaking now. With all the preparatory work needed you might as well do both sides while you're in there.
2. When removed it will take a lot of beating with hammers to get the sections apart. At some point you may suspect that some ba$tard welded them together. Once you have them apart assault them with a flap wheel until all the joints are so loose you can literally see daylight between adjacent sections. Yes it will blow a bit when it first starts but the loose fit will give time for the manifolds to find their natural position before they get carboned up and seize solid again.
3. When reassembling a load of anti-seize on the joints between sections is also helpful to let the sections slide through one another as you struggle to get the bolts started on each cylinder head. Be prepared for skinned knuckes and lots of cursing. On an 8-cylinder engine you might not have as much trouble as we do on it's 16-cylinder brother.
4. Use all new exhaust manifold bolts, hard washers, and spacers. The idea of the spacers is to get stretch on the bolts. Don't be tempted to use shorter bolts with washers on them except as called out in certain specific locations.
5. In fact expanding the above I would go with 2.5" long bolts (5P-3961) instead of the 2" listed in the parts manual on all the exhaust manifold-to-head joints. That P/N of bolt would use a 5M-2894 hard washer and a 198-0203 spacer. Basically it's how the 3508 engine manifolds would be equipped in an off-highway truck. See attached parts manual from a 777 truck.

 

[Tony Wells]

Well Nige, you're just a regular ray o' sunshine, ain't ya? Just kidding...this is exactly what I need to know. I suspected it would be fairly involved, but wasn't prepared for all of that. And naturally, this engine is installed in a semi-stationary machine with close fitting safety cages all around. Looks like a painful journey. I don't remember how many hours are on it, but more than enough to get everything settled in and "welded" in place.

Fortunately, I do have a building that I can probably get the machine moved into, and will press for that. It's getting to be winter here, and although we have mild winters compared to many places, it does get plenty miserable working outdoors. We get snow a couple of times a year on average, but 34° and rain/wind makes for unpleasant working conditions. That's about what we get during late Dec through Feb as a typical day, with wild fluctuations. Pretty sunshine and 50's one day, and an ice storm overnight when a front moves through and then it's in the 20's and windy, if not precipitating also. Needless to say, I'm not looking forward to this. I do wish I could hear it clearly myself so I would have a chance of being misled and this turn out to be something else that just sounds like an exhaust leak. That would make me (and the owner) mighty unhappy. I think he and I will have to talk this over.

What's a (almost) funny thing (but sad) is that we have 3 diesel mechanics on the payroll, but he doesn't want them touching this machine. He doesn't trust them as much as he does me, simply because I am very meticulous and view everything as a machinist as much as a mechanic. Plus we've been friends for almost 40 years. I am slower than they are on some things, but I break a whole lot less stuff as I work on it.

Nige, I thank you for the documentation as well as the counsel. It will be taken to heart.

 

[Nige]

Anyone who's been in my line of work for a number of years will have been battling the same way with 3500-series cylinder head-to-exhaust manifold bolt breakage since the mid-1990s.

If you need any more info please ask.

 

[Tony Wells]

Nige, is there a common point of failure in the exhaust system that I should especially look for when trying to decide where the leak is?

As a machinist and mechanic both, I've dealt with a fair number of broken bolts. Some are not too bad, but access is usually key to success in removal. Some are nearly impossible without major disassembly. I've learned several techniques, but none are completely foolproof. I'm preaching to the choir, I'm sure, but hoping not to have to deal with a break in an impossible spot.

 

[Nige]

Everything in my wallet right now is on one or more bolts broken as shown circled in red in the diagram below. That's where 99% of the bolt failures occur. It is uncommon to see bolt breakages in the fixings between the manifold sections, circled in blue. Access to get at the broken piece of bolt that's left in the head is always the problem.

 

[Tony Wells]

That would be natural, since the sleeves are not too difficult to work with. Being a bolt/nut arrangement makes that simpler in itself. It's always the blind holes where the real problems appear.

The high temp spec bolts, are they a stainless alloy or a high tensile carbon steel? If it comes down to welding on them I would need to know.

Guess I'll head out and do some on-site ruminating on how to best approach this. I get notices on my phone when replies turn up on the Forum, so if any more thoughts are to be shared, please continue!

 

[Nige]

All the hardware that connects the sections, as well as the hardware that secures the manifolds to the heads, is manufactured from high-temperature carbon steel not stainless.

Bear in mind that your engine is 15 years old and if no-one has ever been into the top end before I'd suspect that a significant percentage of the manifold bolts will break when removed unless you are extremely lucky.

Another tidbit. I'd use the updated Part Number of exhaust manifold gasket that is now used on machine engines. I see the former 7W-8145 is still listed for your industrial engine application. The 7W-8145 is not cancelled.

 

[Tony Wells]

After consulting with the machine owner, several options were presented. At the very least, the machine will be transported to our main shop where it can occupy a bay to itself as long as needed. More discussion will ensue. This machine has been out of service for some time now, so spending whatever time it takes to do this right will not be a problem. It will be inside a heated building for the duration.

IMO the least desirable option is to simply sell the machine as is, with full disclosure. We have several other grinders in the yard, and our eye on another tub grinder anyway. But we have owned this grinder for some time, and know it well, have good connections with the factory, and have made a few custom changes to suit our work. I'd hate to see it go simply because of a difficult and time consuming repair. I wouldn't protest a brand new one, but this one has grounds hundreds of thousands of tons of material.

Nige, thank you for the service bulletin. I'll make sure I get the upgraded gaskets. And thank you for all the advice. I'm sure we'll be discussing this one for a while, once the job gets underway. I really expect we will be keeping it.

 

[Nige]

Another thought. Generally (but not always) the broken bolts are on the corner cylinders. That would be 1, 2, 7, & 8 in your case. I would say that 80% of the time that’s where you’ll find the broken bolt or bolts.

 

[lantraxco]

I'll stick to Morookas, lol.