| Author |
Message |
Sean Welsh (Sean)
Rating: N/A
Votes: 0 (Vote!) | | Posted on Monday, May 23, 2005 - 8:23 pm: | |
Folks,
As many of you know, we are at PEDCO in southern California, where we've just had an in-frame rebuild of our 8V92-TA.
The rebuild was completed successfully, and we've double-checked that, mechanically, the engine is in good shape (cylinders, heads, turbo, blower, injectors, etc.), however, we have been troubleshooting a lack-of-power complaint.
This is a 475 horse motor, and, even though the coach is heavy at 45,000lbs, we can't get decent power on grade. Also, a brief run on a rolling brake dynamometer showed only about 275hp to the ground, and we're thinking that 350++ hp is more in line with "normal" for this engine, especially freshly rebuilt.
During the course of diagnosis, what we have found is that the DDEC computer is reading turbo boost pressure of only around 16psi at max boost. All the charts (and experience) say that's low, and it should be more like 22+ psi.
Since the turbo looked fine, we hooked a manual gauge to the blower intake. What we found is that turbo boost pressure is actually building to 22-23psi. The reading that the DDEC-II computer is getting is consistently 5psi or so lower than actual at max boost, and it looks like it is reading 25% or so low all the time.
Next we put a regulated compressed air of 20psi supply directly to the boost sensor. What the DDEC reads starts out at around 15-20psi, but rapidly tails off to zero in about two seconds. Now, mind you, this is with a constant regulated 20psi applied to the sender.
We've replaced the turbo boost sensor twice with the same results for all three senders. We've metered the harness back to the ECM and it looks OK, at least on an Ohmeter. There is a good +5v supply voltage between the ECM ground and ECM 5v supply leads at the boost sensor. We've even swapped ECM's with the same results.
Our speculation now is that the ECM is reading a low boost pressure, and so is not metering enough fuel to the injectors, thus the low power.
So here are my questions for the group:
1. Does this boost sensor behavior sound normal? Or should we read the actual boost pressure at the ECM?
2. What else can I do with the engine harness to verify that it is good (or bad)? We have a new harness on the way, so changing it is not out of the question.
3. Has anyone seen this sort of problem before, and, if so, how did you solve it?
4. Any other ideas? We're at the end of our collective ropes here...
Heelllllp!!!
-Sean |
niles steckbauer (Niles500)
Rating: N/A
Votes: 0 (Vote!) | | Posted on Monday, May 23, 2005 - 8:57 pm: | |
Sean - When Ace had a problem and replaced his Radiator surge tank sensor he continued to have a problem until he replaced the 'sending unit' - you may also have a relay there? - just a suggestion - Ace was somehow also able to bypass the DDEC at the engine switch and run it off the system - maybe you can do that and test the psi level at your manual gauge location - MY 8v92 DDEC II shows about 25 psi or a little over that on the dash at max boost (don't know if its even close to accurate) and she runs right out fine.
Niles |
Pete/RTS Daytona (Pete_rtsdaytona)
Rating: N/A
Votes: 0 (Vote!) | | Posted on Monday, May 23, 2005 - 9:21 pm: | |
I don't understand why your DDEC boost reading "falls off to zero" with a constant 20psi
Put your voltmeter on A-ground terminal and the B-boost signal terminal of the sender - and see if the voltage reading changes when the DDEC reading starts going down.
Pete RTS/Daytona
I'm no Diesel Mech - but I don't believe the DDEC reading or the sensor reading should drop to zero when a contant 20 PSI is applied |
Geoff (Geoff)
Rating: N/A
Votes: 0 (Vote!) | | Posted on Monday, May 23, 2005 - 9:47 pm: | |
I don't quite follow this story-- last week the problem was black smoke and low power, so now the problem is simply low power and the smoke has gone away?
--Geoff |
Geoff (Geoff)
Rating: N/A
Votes: 0 (Vote!) | | Posted on Monday, May 23, 2005 - 9:54 pm: | |
opps |
Chuck Newman (Chuck_newman)
Rating: N/A
Votes: 0 (Vote!) | | Posted on Monday, May 23, 2005 - 10:21 pm: | |
Sean,
Did you check the new ECM's programming as well as the old ECM. As you describe, it appears you have tested all the hardware links.
Are you still at PEDCO? If so, I'm in Long Beach and could drop by tomorrow afternoon (5/24/05) if you have the time. I'm going by there on my way to Ontario.
Chuck Newman
Oroville, CA |
Sean Welsh (Sean)
Rating: N/A
Votes: 0 (Vote!) | | Posted on Monday, May 23, 2005 - 10:42 pm: | |
Geoff,
Not sure if you're talking about the black smoke from over a week ago, or the smoke from Friday's dyno test...
The problem we came in here with is that the engine was dirted out. We did need a complete inframe rebuild -- liners, rings, pistons, blower, turbo, and heads. That part is done.
The smoke on the dyno turned out to be blue, not black (was my mistake in the transcription -- I was at the controls, so I didn't see the smoke). Virgil Cooley of Pedco was concerned that the smoke meant the turbo had puked, so Saturday they pulled the turbo back off (it was fine) and we even ran the motor with the manifolds disconnected to make sure no smoke was coming from oil in the heads or kits (it wasn't).
At the moment, the current speculation is that the blue smoke on the dyno was leftover oil in the muffler -- blow-by from before the rings and liners were changed. We're running with the muffler and exhaust completely out right now, just to make sure backpressure is not part of the problem, and there's been no smoke. (It's been interesting tooling around town with, essentially, a straight pipe. I can out-rumble the Harley guys now...)
So, yes, the only problem we are shooting right now is low power, and why the turbo boost pressure as seen by the DDEC does not agree with the (higher) readings we are seeing on a mechanical boost gauge.
Pete/RTS -- that's on the agenda for first thing tomorrow. I'll have to pierce the insulation on the wires leading to the sender to get the reading. Since the sender is right next to the turbo, I couldn't do it today because the whole thing was well over 200 degrees when we got to this point in the troubleshooting.
We are putting our satellite dish up "between road tests," so to speak, and I wanted to get as much input from this group as I could before we stow the dish in the morning for tomorrow's tests.
Thanks for any and all input.
-Sean |
Sean Welsh (Sean)
Rating: N/A
Votes: 0 (Vote!) | | Posted on Monday, May 23, 2005 - 10:54 pm: | |
Chuck,
Yes, we're still at PEDCO, and we have nothing but time, at least until this is fixed. You are welcome to come by.
When I said that we "swapped ECM's" -- we used an ECM that PEDCO had here from a 450 HP fire truck. It's definitely a different set of programming, but it's from an 8V92TA and it's DDEC-II, which means they are more alike than different. For the purposes of a quick test, it fit the bill. We were able to road test with this ECM, even though we had a constant Check Engine code for a water pressure sender we don't even have, and water temp read zero. The turbo boost readings behaved exactly the same, which leads us to believe the "tail off" problem is not in the ECM.
If the ECM needs to be checked further, PEDCO can't do it and we'll have to send it to Valley Detroit, which is the local factory-authorized dealer. That's on our agenda for tomorrow as well, if we can't find the problem through other means first. I suspect, given the current state of troubleshooting, that we will try a new engine harness before we bite the ECM bullet at Valley.
-Sean
p.s. Long Beach is a long way from Oroville... |
Chuck Newman (Chuck_newman)
Rating: N/A
Votes: 0 (Vote!) | | Posted on Monday, May 23, 2005 - 11:27 pm: | |
Sean,
You've swapped the ECM and sensor. You don't have a boost plumbing leak as I did (welded up a hole in the intercooler). That leaves only the harness. Disconnect at both ends and check connector. Probably won't see much due to molded connectors. Ohm the lines cold. Ohm them again with a wiggle test -- separately at both ends. Then heat up each end (hair dryer, etc.) and ohm again both ends with and without wiggling wire at connector. I don't have my books here, but I think the DD ECM is the same as GM, Ford, etc. in that the sensor ground wire returns directly to the ECM. It does not use the chassis and is not paralleled with other sensor grounds. Either or both wires can go bad (high resistance or partial short to each other or to chassis ground.v And it may be a crimped wire-to-pin connector that has become temperature sensitive. I've seen all the above anomalies at one time or another.
One last thing. As mentioned, disconnecat both ends of both wires and ohm to chassis ground, hot and cold. Should be open in all cases.
Have been in Long Beach for two days at telecommunications conference. Can't take it any more. Leaving early. Going up to visit son in Ontario. Hopefully see you tomorrow.
Chuck Newman
Oroville, CA |
Chuck Newman (Chuck_newman)
Rating: N/A
Votes: 0 (Vote!) | | Posted on Monday, May 23, 2005 - 11:44 pm: | |
Sean,
MAK took me to Virgil's, but I don't remember the address. Help a failing memory.
Chuck Newman |
Sean Welsh (Sean)
Rating: N/A
Votes: 0 (Vote!) | | Posted on Tuesday, May 24, 2005 - 12:03 am: | |
Chuck,
Since you're typing at 8:44, I gather you're missing the big opening night gala with all the booze...
We did ohm out all the sensor-to-ecm wires, including the wiggle test, and it was pretty hot at the time. Also, a truly bad ground, voltage, or signal wire should cough up a code, I think 33 is low volt and 35 is high volt, or something close to those. So the value the ECM is seeing on pin 432 is within the correct range.
I've seen weird things happen on these harnesses, though, and I am not ruling out some sort of high-Z coupling to a nearby circuit bleeding the signal off, or capacitive coupling, or some type of induction, etc. etc.. Which is why replacing the engine harness may be the next step. (I just don't want to buy a new harness, and pay to swap them, only to find out the problem lies elsewhere.)
PEDCO is at 9911 Norwalk, in Santa Fe Springs. Take I-5 north through Anaheim, etc., past Santa Fe Springs, and into Norwalk. Exit Norwalk Blvd (which is one or two exits north of Rosecrans) and make a right. In maybe a mile or so you will see a pedestrian overpass cross the road diagonally at an intersection -- this is Telegraph. Go through the next signal, which is Hawkins, and PEDCO is on the left just before the signal after that, at Smith. We are in the back, which is actually the Norwalk Blvd. Entrance (the front is on Smith).
You can also click the map link on our web page.
-Sean |
Pat Bartlett (Muddog16)
Rating: N/A
Votes: 0 (Vote!) | | Posted on Tuesday, May 24, 2005 - 5:55 am: | |
Sean when you take your local pressure reading are you using the same port that the DDEC is using? There is no way a pressure sending unit should fall off when a known constant is applied. First thing is to make sure you have an accurate gauge, some of these automotive gauges are just plain garbage! You could always T off of the port and hook both sensor and gauge up for a comparison! I'm not a DD mechanic but I am a instrument tech! I've been following your blog and it has been very interesting. Good luck!.....Pat |
Pat Bartlett (Muddog16)
Rating: N/A
Votes: 0 (Vote!) | | Posted on Tuesday, May 24, 2005 - 6:02 am: | |
Sean one more thing never assume because the sender is new that it is good, i've been bitten by that bug before, possibility that there could be a bad batch of sensors! One more suggestion you said you used a ohm meter to check the wiring, did you meg the wiring? Sometimes you won't see a problem until voltage is applied! Oh yeah make sure you disconnect wiring when you meg it! ...........Pat |
Geoff (Geoff)
Rating: N/A
Votes: 0 (Vote!) | | Posted on Tuesday, May 24, 2005 - 7:16 am: | |
A nearby shop had the same problem with a NJ Transit Eagle and after changing everything they could think of they ended up taking it to the Detroit Distributor where they found a bad ground connection in the wiring harness.
--Geoff |
Pete/RTS Daytona (Pete_rtsdaytona)
Rating: N/A
Votes: 0 (Vote!) | | Posted on Tuesday, May 24, 2005 - 7:45 am: | |
Before you get to the point of replacing the harness - why not
Cut the A/B/C wires from the sensor and run three new wires
A-to Y2 (circuit 452 - Sensor return)
B-to-P1 (circuit 432 - sensor signal)
C-to-W1 (circuit 416 - 5 volt)
Hmmm -
A- put your meter on the 'C' when you have the failure and make sure the "Sensor supply" is a constant 5volts.
B- externally/manually Ground the 'A' sensor point when you have the failure to elimate a flaky ground
this problem may be just outside the (high/low) DDEC error settings ??
again - I'm no diesel guy - but I'm learning |
Chuck Newman (Chuck_newman)
Rating: N/A
Votes: 0 (Vote!) | | Posted on Tuesday, May 24, 2005 - 12:12 pm: | |
Sean,
Pat and Pete have excellent points. Been there.
But based on what your telling us, I was thinking late last night (sometimes not a good thing) it sounds like voltage leakage to chassis ground causing a ground loop problem as Geoff described.
The ECM firmware I have seen is designed to generate error codes for extremes -- open circuit and short conditions. It is not smart enough to detect a little to much or to little voltage from the sensor on the "slider wire" from the sensor to the ECM input. Normally a small amount of distributed capacitance is not a problem, particularly in the length of wiring here. But a small leakage to chassis ground will change the input voltage to the ECM enough to give a bogus reading on the DDEC data stream.
Another thought: Did this problem start AFTER the engine overhual? If so, I would almost guarantee a wire or connector was hit/damaged/loosened, etc. during the engine work. In 40 years I have been there countless times. With only one exception, the new anomaly never was just a coincidence.
Lunch on me today.
Chuck Newman
Oroville, CA |
Sean Welsh (Sean)
Rating: N/A
Votes: 0 (Vote!) | | Posted on Tuesday, May 24, 2005 - 1:07 pm: | |
OK, Folks, here's the latest...
We measured the signal voltage coming back from the sender. First I measured at the sender, and then we moved to the harness connector on the ECM.
I had the negative lead of the meter on signal 452 (ground/return) and the positive lead on 432, which is the Turbo Boost Sender signal.
Ignition on, no pressure at sender, sender reads 1.07 volts. When we apply a constant 20psi to the sender, the signal reading goes up to 2.98 volts, and stays there, rock steady. This is at either end of the cable.
However, the computer reports only about 12psi of boost when the compressed air is first applied, and, again, the boost readings tail off down to zero over a period of about three or four seconds. Still, the sender is sending a steady 3 volts.
I'm stumped. We see this same behavior with two different ECM's. We are now trying to get an answer from someone as to what "normal" behavior is for this reading from the computer (although my gut feeling, and no one here will dispute it, is that a 20psi steady signal from the sender should result in a 20psi steady reading from the computer).
The fact that the signal is rock steady at 3 volts at the engine harness connector to the ECM suggests that the harness itself is not the problem. I'm wondering now if there is some other input to the ECM that causes it to moderate it's idea of boost pressure? Or perhaps the ECM "calibrates" atmospheric pressure by steady, unchanging readings over time, and thus only a changing reading registers as actual boost?
I'm stumped...
Responding to some posted comments:
Cutting the wires to the sender and running fresh wire to the ECM is problematic, because wire 416 and wire 452 are shared among many senders. So these wires run from the ECM connector to various junction points in the harness, where they are split off to the various senders. The only wire that runs direct is 432, which is unique to the boost sender.
Problems in the harness wiring on wires 416 and 452 would result in weird readings from multiple senders, and we're not seeing that.
With regard to whether the problem existed before the overhaul, I can't say. Clearly we had a power complaint coming in here, but bad rings, liners, and turbo would account for that. We did not start looking for a boost pressure problem until after the in-frame. So maybe yes, maybe no.
Still hunting for ideas...
-Sean |
Stan
Rating: N/A
Votes: 0 (Vote!) | | Posted on Tuesday, May 24, 2005 - 1:48 pm: | |
Sean: When you monitored the voltage on 432 at the ECM, were you measuring on the ECM board or is there still a possible high resistance from where you had the meter probe to the board foil? |
Sean Welsh (Sean)
Rating: N/A
Votes: 0 (Vote!) | | Posted on Tuesday, May 24, 2005 - 2:11 pm: | |
Update:
Well, one of the people PEDCO called to ask about this finally told us that, since we are measuring with the engine not running, the DDEC will, indeed, calibrate the boost back down to zero.
So we need to take the reading with the engine running. That will have to wait, since PEDCO is changing the injectors right now.
I'm eager to see what the results are when running. We know from yesterday's road testing that the DDEC is reading less boost than we see on the mechanical gauge. (To answer an earlier post on this: we did double-check the mechanical gauge calibration -- it's PEDCO's gauge, and it's a shop-quality brass job. And, yes, we are in to the same area on the blower intake with both the gauge and the sender.)
Stan -- We have not opened the ECM, but we get the same result with two ECM's, so a board problem is unlikely. I'm trying to avoid opening the ECM since the ECM sealing gasket is not available as a separate part (Detroit considers the whole ECM to be the FRU).
Since the symptom is consistent with lack of fuel, we are now pondering whether the main delivery line is large enough, and/or if one of the braided lines is crimped internally.
Of course, there could still be a problem with the boost reading, which would result in the ECM not metering enough fuel to the injectors.
-Sean |
Jim Shepherd (Rv_safetyman)
Rating: N/A
Votes: 0 (Vote!) | | Posted on Tuesday, May 24, 2005 - 3:13 pm: | |
Hi Sean. This is fun for the rest of us, since we can get our trouble shooting hats on.
On a mechanical engine, lack of fuel would cause a drop in boost. I would think it would be the same on a DDEC if the supply line was somehow blocked or restricted.
With the new Fuel Pro, it should be easy to tell if you have a fuel delivery problem or air in the line. That assumes you replaced both the primary and secondary fuel filters as they recommend. If the FuelPro looks OK, then the only delivery lines would be from the Fuel Pro to the pump and from the pump to the heads. It might not be a bad idea to change those out anyway, since I would suspect they are hard to get to.
When I installed my Fuel Pro, I had a heck of a time with fuel delivery. It turns out that the filler cap o-ring was missing. This let the pump suck air from the filter and you could not see any sign of the problem.
As I recall, you had a Fuel Pro 382 installed during the overhaul. Since that was a change, it could be a suspect cause. If you go to http://www.davcotec.com/fp382.htm you will find all sorts of information on installation and troubleshooting.
They are fairly adamant that the filter should be placed so that the bottom of the filter element is above the high fuel level in the tank. That should not cause the problem you are having, but it is a bit counter to normal thinking.
The Fuel Pro is a great device, but it caused me a lot of problems until I found the air leak.
Something new to think about.
Jim Shepherd
Evergreen, CO
’85 Eagle 10
Bus Project details: http://www.rvsafetysystems.com/busproject.htm (updated 2/17/05) |
ChuckMC9 (Chucks)
Rating: N/A
Votes: 0 (Vote!) | | Posted on Tuesday, May 24, 2005 - 5:17 pm: | |
"...no way would I trade my DDEC for a mechanical. When the engine has any sort of problem at all, I just look at a display and it tells me what's wrong..."
-- By Sean Welsh (Sean) (69.35.62.223) on Monday, March 07, 2005 - 9:11 pm
...did I really post this? That's the sort of troublemaking these other rapscallions do.
Sean, it will be rolling by THIS Friday. That's what you meant last week. |
Sean Welsh (Sean)
Rating: N/A
Votes: 0 (Vote!) | | Posted on Tuesday, May 24, 2005 - 6:02 pm: | |
Jim,
We pulled the FuelPro out almost immediately, just to eliminate it as a source of trouble.
We still have the secondary filter, because I don't have the little framistat to put on the filter holder to bypass it, and we didn't want to build a manifold to supply fuel to the ECM, etc.
Also, we have the 380, which is Detroit-branded. The instructions for the 380 (in contrast to those for the 382) say nothing about mounting higher than the high fuel level, which is good, because that's simply not possible in this coach.
I've got all the PDF's from Davco on my hard drive already...
Chuck -- I stand by my statement. Still no way I'd trade my DDEC for a mechanical. If anything, I wish I had a DDEC-III instead of a -II. Having those error codes available to me while going down the road is a real comfort.
I did, perhaps, overtstate the ECM's ability for self-diagnosis :-)
-Sean |
gg04
Rating: N/A
Votes: 0 (Vote!) | | Posted on Tuesday, May 24, 2005 - 7:22 pm: | |
Just a short one...If you do not have any senders..sensors hooked up that the computer thinks it should lhave (error codes) the sucker will not run like it is suppposed to it takes care of it self and cuts back on power until what it sees as a problem is repaired (error codes you mentioned) if sensor is important enough off goes the engine...had the same minor problems at first all gone now...gg04 6L71ta ddecII.. |
Geoff (Geoff)
Rating: N/A
Votes: 0 (Vote!) | | Posted on Tuesday, May 24, 2005 - 7:24 pm: | |
Chuck is welcome to quote me-- I swore I would never buy a bus newer than 1987 so I wouldn't have to have DDEC!
--Geoff
'82 RTS CA |
ChuckMC9 (Chucks)
Rating: N/A
Votes: 0 (Vote!) | | Posted on Tuesday, May 24, 2005 - 7:27 pm: | |
Hey my friend, hope you had a rueful grin with that.
With the most incredible coach I've ever seen, one of the smartest guys I've had the fortune to communicate with, a wife who loves to ride motorcycles, and the help of the smart guys on this board, you won't be held down much longer. |
Bill Gerrie
Rating: N/A
Votes: 0 (Vote!) | | Posted on Wednesday, May 25, 2005 - 10:15 am: | |
Sean
Just as I am getting ready to go to Alaska I noticed your posting. The biggest problem with any DDEC unit is the wires. As the problem is common to both ECU's you might want to check the pin 432 right at the connector on the ECU. These are only crimped on and not soldered. In the book it tells you how to remove the pins from the connector. You could have a high resistance connection there which will fool both ECU's. Just a thought as according to even DD the biggest problem is bad connections. Another thing about the Fuel Pro. My friend mounted it low and another one mounted it high and there is no problem with either one. You must eliminate the secondary filter as per Fuel Pro instructions. I know they sell a special piece to do this but everyone I know just replumbed them and did away with the secondary one altogether. Good luck with your problems and if I get a chance and find a computer I will follow your progress. Bill |
niles steckbauer (Niles500)
Rating: N/A
Votes: 0 (Vote!) | | Posted on Wednesday, May 25, 2005 - 2:20 pm: | |
Sean - Is there any possible way this problem is connected with the frame rubbing on the blower intake ? Specifically, did it do some damage by transference when the blower was torqued into the frame ? Really reaching here - Niles |
Cathy and Raymond
Rating: N/A
Votes: 0 (Vote!) | | Posted on Wednesday, May 25, 2005 - 10:55 pm: | |
You will not see a high boost just running around, need to pull a hill ect.
The ECM will not sead a reading to a Prolink
with air pressure put to the sensor, the ECM is not getting the info for all the sensors.
Has the timeing and injectors sensors been checked, they are adjustable and must be set right. They would not have been removed on the work done, but need to be checked (low power)
Under the ECM at the cam shaft pulley ( gets a lot of oil ect there) check the wires too.
I have allso seen the fuel line from the fuel pump on one of the heads put on the return port and then the return on the supply port "low power" |
Sean Welsh (Sean)
Rating: N/A
Votes: 0 (Vote!) | | Posted on Thursday, May 26, 2005 - 1:44 am: | |
The problem is solved and we are on the road!
Details here:
http://ourodyssey.blogspot.com/2005/05/were-free-were-free-free-at-last-from.html
-Sean |
|