Crankshaft Position Sensor Diagnosis

I posted in a separate thread most of the symptoms that I still have (see https://rennlist.com/forums/996-foru...placement.html ). It appears that I had two separate issues. First was a bad alternator, and that appears to be addressed (after two alternators). But the remaining symptoms are still there: very jerky after the engine gets fully up to operating temp. It feels like the rev limiter kicks in but at rapidly lower RPMs. It first starts after 15-20 minutes of stop-and-go driving and I turn on the AC and rear window defroster to make it happen sooner. The RPM is at around 3k when it first starts. But within 2-3 minutes more it kicks in just starting from a stop, like 1k RPM and shudders throughout the range. Another couple of minutes and it's completely undrivable. No problem at all when cold. Durametric shows that the DME consistently throws P0336, "ref mark sensor", the crankshaft position sensor (CPS), when this happens. When it gets really bad it also throws various cylinder misfires (from both banks) and sometimes even the ABS/PSM warning lights go on. It's a whole engine phenomenon, not one bank or one cylinder. So the logical culprit would appear to be the CPS. There are numerous reports on this forum and others about heat sensitivity of a failing CPS. I spent the day today researching and measuring my CPS. I am not sure how to interpret the results and wonder if anybody here has gone through this, especially looking at the waveforms. First, the wiring diagram shows 3 pins: 1. The signal 2. Ground 3. Shield (protect against electromagnetic interference) At room temp the resistance between 1 and 2 should be between 800 and 1000 ohms according to Porsche diagnostic manual, and I measured 935. So that's OK. I also measured resistance between the harness pins and they were either 45k or 7k; no dead shorts and no opens, there are no specs here but I guess those are not obviously bad. I heated up the CPS in a toaster oven (protected by foil) to around 190 F and it measured about 1.2k; a little higher as you'd expect but not an open circuit or short or something. Then I measured the expected distance from the tip of the sensor when mounted in the engine case to the flywheel teeth using a depth gauge. As far as I can tell it ended up about 0.8 mm (30 mils) which seems rather small. One thing I tried was to add thin spacers to increase the distance; I'm currently running with spacers that roughly brought space to 1.5 mm (60 mils), and it is hard to know if it's doing better or not; in any case the difference is subtle. More importantly I have some waveform images. First is an image of what the Porsche diagnostic manual says the waveform of the sensor pin 1 should look like. Porsche CPS Waveform.png We see that there is one full cycle up and down for each tooth of the flywheel ring, and a gap where one tooth is missing. Wikipedia and other sites indicate that CPS units that produce waveforms like this are inductive sensors. A general car diagnosis web site I found had a very similar waveform. Both show the waveform about 6, 8, 10 V peak to peak. The DME unit detects crank position by finding the waveform's zero crossing and it knows where the missing tooth on the flywheel ring is by the gap. But now here is the waveform from my CPS. I got this by pinching a very thin wire into pin 1 of the CPS connector and running that into an oscilloscope, with ground taken from the engine case nearby. Note: Wide variation in voltage, more than a factor of 2 Peak-to-peak voltage only 0.6 - 1.2 V! (Note the scope is at 0.2 V/div) Also when I zoomed in on the time axis, it looked like the gap had its own zero crossing or two instead of being a smooth, slower transition. Note that this photo unavoidably has a bright part of the scope scan where the camera shutter was open, and the rest of the waveform is shown dimly from the leftover phosphor glow. The picture shows slightly more than 2 full flywheel rotations. IMG_0808.jpg So the question is, is my CPS bad? Does anyone know what these waveforms really look like in practice? So I can imagine that if my CPS is producing a small fraction of the voltage it should be, heat, low voltage, or other issues may prevent the DME from reliably detecting the waveform's zero crossings.

Hi Mike, the sensor is an inductive sensor like you said and it has a magnet inside. It generates ac voltage between the 2 pins as the teeth of the flywheel approach and leave the sensor. The third pin provides a shield connected to the ground. The proper way to measure the sensor output signal therefore is to set the scope for ac coupling and measure the voltage generated between pin 1 and pin 2, and not pin1 and chassis of the car. Also, be careful measuring the resistance of the harness side with the DME still connected as you are sending electricity to the DME. The full diagnostic of code P0336 should call for removing the DME and check the continuity between pin 1 of sensor and pin 32 (plug III) of DME, then pin 2 of sensor and pin 46 (plug III) of the DME. Also check resistance between pins 1&3, then 2&3, of the sensor. They should read infinity. If those check out, replace the sensor.

Btw, I agree the sensor seems bad if it can only generate 0.6 to 1.2v peak-to-peak. Too bad my car is on the lift now and can't be started for a few days. I would have checked what mine looks like.

Quotes source posts: 11457815, 11457815

Ahsai, thanks. I should have done that but ended up doing the single-sided measurement because the wiring diagram shows DME pin III/46 labeled GND and using a brown wire, while III/45 is labeled SIGNAL (see image below). If the sensor produces AC (which it certainly should) then I would presume that what I measured was half the AC voltage, offset by some DC. However..... CPS DME pins.png OK I should do the full set of diagnosis steps. I've never taken the DME out before so it's a new thing. There are actually five continuities specified: III/32 to pin 1, III/45 to pin 3, III/46 to pin2, III/32 to ground, and III/46 to ground. All of these should be in the range 0 - 5 ohms. Note that the pin numbers are reversed from the sensor connector; I guess the harness is numbered the other way around?? Anyway if this is correct and pin 2 is actually ground (and the shield is grounded at the DME), then I got the full voltage measured. Here's the page from the diagnostics manual: P0336 diag.png

If you suspect the CPS why not just replace it? While generally I do not like to throw parts at a symptom if the part is not very expensive or hard to get at sometimes it is quicker to just throw in the new part. BTW, I am not convinced the sensor is at fault. They generally work or they don't work. The behavior still reminds me of an electrical problem, the electrical system -- battery, alternator, wiring -- is not able to provide the engine and the rest of the car with sufficient electrical power when the engine is running. I know you have been through 2 alternators and I believe also replaced some very questionable wiring/cabling, but I still think the symptoms are electrical power related. However, you are certainly free to suspect the CPS and continue your diagnosis with the CPS as suspect #1.

Quotes source post: 11458405

I will be getting a new CPS this week and try it out. Yes it is possible it is still an electrical power problem but I would be at a loss to figure out what it is. Belt slippage when hot? It's a new belt and supposedly the tensioner is an either-or failure, not by degree. Maybe I should watch the 12V supply and see if there are glitches. I measured voltage drops all around the charging system -- but not under heavy load -- and saw all very low drops. Ground between alternator and battery was only 25 mV. Worst B+ drop was alternator to battery, 150 mV.

Quotes source post: 11458398

Ok, my pin numbers were for 7.2 and I just checked for 7.8, your schematic is correct so lets go with yours. Don't know why the pin numbering are reversed at the sensor but you can just use the wire colors to tell them apart. Basically DME pins 45 and 46 are for the signals from the sensor and pin 32 is for the shield. Removing the DME is very easy. You just need to remove plug III. There's a lever lock to unclip and remove in one motion. Just make sure you disconnect the battery and you should be fine. I'm curious to see what you find.

Quotes source post: 11458405

When I agree with you that in general they either work or don't work, the issue is the DME keeps complaining about it, consistently. Usually the DME doesn't complain for nothing (and in this case something very specific) so I still think it's a good idea to get that diagnosed first. I also agree that the drivability issues may or may not be related to this code. Btw, Mike can you measure the B+ jump start terminal (since it's easy to access) and the ground chassis for the ac output of the alternator? It should be way less than 1v p-2-p.

OK, back from errands and have the rest of the day to work on this. Plan of attack: 1. Prep: * Disconnect harness at DME and CPS, inspect, measure all possible resistances between 3 CPS wires and ground at DME * Reconnect CPS, measure same resistances at DME with CPS in * See if I can squeeze two wires in to CPS connectors at DME for scoping 2. Reconnect and run engine until hot and shows P0336 * Monitor AC voltage throughout (from cold to hot) at jump terminal, look for glitches * Monitor CPS signal, hopefully at DME * Let's see if I can trigger scope on B+ glitch and see if CPS signal looks odd * Turn engine off, disconnect DME, measure CPS resistances with CPS connected BTW is there any way to bump the throttle from outside the car? Durametric?

OK, pretty sure I have the smoking gun. It's the CPS. I have the following measurements: 1) At the DME connector, with CPS in: Pins III/45 - III/46: 932 ohms (resistance of CPS windings) All other combinations of pins III/45, III/46, III/32, and ground are infinite 2) Between DME connector and CPS harness: III/45 - 1, III/46 - 2, and III/32 - 3: 0.4 - 0.5 ohms All other combinations are infinite OK so there are no shorts, no opens, the harness appears fine, the connectors all look good, no burnt wires or insulation (I opened up the big wiring mess under the center of the rear between the DME and the relay panel to look). BTW I had to remove all the DME connectors because the shield goes to another connector, IV I guess. I left in a couple of very small wires (wire wrap wires for those of you who remember) sticking out of pins III/45 and III/46 at the DME to hook up the scope. I hooked up the other scope input to the alternator as well as a DVM measuring AC as Ahsai suggested. Engine is cold. I started up the car and looked at the scope. That's the first image. Everything looks great! Top trace is CPS at 5V/div: CPS is outputting between 7 and 15 V today (some discrepancy could definitely have been because I was measuring against ground yesterday)! Bottom trace is alternator output at jump point, in AC mode, 0.2 V/div. AC on the DVM starts at about 100 mV and when idle settles to 680 RPM, AC drops to 50-60 mV. I turn on AC and rear defroster and let it warm up. Still looking good.... go inside for a quick chat with the spouse (ummm, no such thing, never quick). Come back out because the sound seems a little rougher. Oh my. The CPS signal has dropped down to 1/10 of what it was before. This is what I saw yesterday. See the second image. I am amazed the DME can make sense of this signal! To confirm that I didn't just lose the signal somehow, I will let the engine cool an hour or two and look at the CPS signal again. DC voltage at the jump point while idling ranged 12.2 - 12.9 depending on when (cold was higher) and how many accessories were on. It varied and rose and fell, was not constantly dropping. CPS-AC-scope-cold-1.jpg CPS-AC-scope-hot-1.jpg

Excellent diagnostic with very satisfying results and findings Would be great if you can leave those connections there and retest with the new CPS but I'm pretty sure you nailed it already.

Good job. Now you know for sure.

Thanks guys for the help and guidance. Now maybe I can actually drive my car for real after shipping it off in January! Replacement CPS is arriving under "warranty". I actually thought about leaving the connections but they were really intended to be (very) temporary and not disturbed. Still, it wasn't that difficult to put them in. Did you know you can easily disassemble the DME connector III into separate 26-pin connectors and have easy access to the wires and female connector pieces?

Yes those black and grey inserts that house the pins just push out after you pry the locking tabs on the outer housing gently. Once you remove the inserts, the pins can be removed by pressing down the small triangular part of the pin with the smallest eyeglass screwdriver. Press the triangular part and pull the wire gently then the pin will catch again at the second locking slit. You press the triangular part again to clear that. Super easy. 0622141815 (Medium).jpg 0622141816a (Medium).jpg

Btw, you said your jumper terminal voltage varies from 12.2 to 12.9 at idle? That sounds a little low even after fully warmed up. The instrument cluster voltmeter is reading the same?

Yes, jump point voltage went down as low as 12.2 with the DVM attached, and (at least based on a glance through the back window) the dash voltmeter agreed. I've checked the dash voltmeter against the battery voltage a number of times and found it is very accurate, maybe within 0.1 V when I can get that resolution against the lines on the gauge. The low voltage is of some concern, but it was 12.2 only for a short period right after I turned on the rear defroster. It mostly stayed around 12.5 while loaded with AC and defroster. When throttle was blipped it always rose to something over 13, maybe 13.2. In the back of my mind I am thinking I may not be out of the woods yet even with a new CPS. However, the CPS failure (very low output) occurred over a range of voltages and RPMs even when I blipped the throttle. But what could cause a persistent low voltage with those excellent drop measurements is beyond me. Of course, the correct data will always show.

BTW Ahsai, you say you have a 997 shifter. We got my wife an '05 Boxster last year which had a B&M short shifter installed and have the 987 original shifter. I've done some research and it appears that this shifter is compatible with 987, 997, 996, and 986. If so I was thinking about maybe using it for my '02 996. Would you recommend that? One of my cats appears to be gone (something I discovered at my indy's last week as he ran the PIWIS on this issue). I may need a replacement and Gert's option seems like a good one. Gert supplied my car's X74 as well, although to the PO, not me. The X74 is just completely awesome.

As long as you have 13v or higher most of the time, you should be fine. The battery won't be charged at anything below 12.8v. What's the p/n of the 987 shifter? If it's the same as the 997 shifter, I will highly recommend it if you're looking for much tighter shift and slightly shorter throw. I love the 997 shifter and I had the 996 porsche ssk before. I didn't like the notchiness of it. Oh yea, the x74 is just awesome, totally transformed the car. If you get Gert's sport cats and the Fister's mufflers, our cars will be like twins down to the side skirts

Quotes source post: 11459632

I think I'm OK with the voltage. With all this diagnosis the car has been idling for hours over the last week and it starts up with a full force crank no problem at all. That was not at all the case before the last alternator replacement. I think it's fine but once the new CPS is installed and I drive for a week or so I will know. Re: shifters, everywhere I look I see that 986, 987, 996, 997 shifters are all compatible with each other. The B&M short shifter kit box lists all the above models as compatible. Somebody on the Pelican Parts discussion forum also answered with an unequivocal "Yes" when asked the same. The 987 shifter I have does not have a p/n on it that I can find. Anyway, I think I'll plan on making this update when I get a chance.

Quotes source post: 11460530

I checked for you on www.autoatlanta.com Indeed the shifters are exactly the same across those models you mentioned. You should be good to go.

Just to add a data point, these are my ('03) camshaft and camshaft waveforms. My crankshaft signal is 10-15v p-p. Yellow is the camshaft waveform and cyan is the crankshaft waveform. Note the camshaft waveform actually is not a uniform square wave pattern but rather the whole cycle is like short, short, long, long. IMAG001.BMP IMAG002.BMP IMAG003.BMP IMAG004.BMP

New CPS arrived overnight shipment on Tuesday morning and I installed it within 30 minutes. Now it's Friday and the car has been running flawlessly. I think the problem is fixed. Ahsai, great images there, and good reference points. I hope this thread shows up on a variety of searches for similar kinds of symptoms.

Superb work, guys! Great learning here! Merci beaucoup!

Quotes source posts: 11470966, 11471111

Congrats on fixing the problem. Yours and mine seem to be the very first Porsch 996 CKP and CMP waveforms in the wild Hope it will help future users. Thanks. It's fun probing the DME haha

Update. While I thought the new CPS fixed the issue, it didn't. Each time I have improved the car somehow, the misfiring problem receded somewhat further into the background, but would come back with a vengeance under certain circumstances as I described early in this thread. Heat or time, or some combination, would bring back the massive misfiring on all cylinders and the P0336 code (Ref mark sensor, i.e. crankshaft position sensor) thrown. This problem having been present from the moment the car came off the truck from Flat6, Jake Raby has been extremely helpful, well beyond reasonable expectation. We have had probably 100 emails back and forth; Jake even brought Tony Callas in to help. The theories were all the way from "what, a fourth CPS?" to fuel tank venting problems to shorted CPS wiring harness to fuel pressure problems to low voltage due to an under drive pulley installed by Flat6 to help out with track use. But none of these checked out after further investigation. A dozen Durametric logs were taken and forwarded to Jake and Tony, and there was no smoking gun. This weekend Jake and Tony were teaching an M96 101 class nearby and they asked me to bring the car up Saturday evening. Imagine me driving, Tony in the passenger seat wrestling with Durametric, Jake in the back seat, and the car jerking us all around with the misfires. Every once in a while I'd stop and Tony would jump out and adjust the gas cap or the EVAP regen valve or something. It was probably pretty hilarious. Tony concluded pretty quickly that the misfiring was so violent that it was not a fuel, leak, or vacuum issue, but had to be something like a sudden ignition adjustment or momentary ignition cutout. After some time of complete befuddlement it was either Jake or Tony, I don't remember who (I do but the other guy is too embarrassed to let this out) thought of a scenario that would explain it. I decided to leave the original flywheel and clutch in because they were in good shape but of course partly worn (the car has 42k miles). The car was then placed on the truck for the ride back, 2300 miles. If the trucker left the car in first gear then it is possible that the bumping along the ride stressed the flywheel enough to damage it so the CPS loses the signal from the flywheel teeth, the ECU gets confused momentarily, and the engine bucks like crazy. And the P0336 code would get thrown. Jake has arranged for a mechanic to help with this in a creative win-win way :-). I will get a new flywheel and we will carefully evaluate both old and new to see if this is a likely cause. Further updates when I know more. Hats off to Jake for sticking with this even though his responsibility ended when the car was put on the truck. We all hope this is finally the right diagnosis. One lesson: when the DME throws a code, it's usually on target, or really close to it.

Mike is a trooper, and a very smart guy. Tony and I are both baffled by the symptoms and the inconsistency of the occurrences. We'll learn from it.. Its not everyday that a guy works on a car 2300 miles from his shop :-)

Sounds like some good progress. For that code, the flywheel is the next logical suspect anyway after ruling out the CPS and wiring. One thing I don't understand though, the flywheel teeth are not touching anything so why would towing the car damage them? Or we're talking about the springs inside the flywheel damaged by towing? Very curious to see what you guys find.

Quotes source post: 11521700

You must not clearly understand how a flywheel can create misfires.. As a flywheel ages the second mass loses control and will allow radial fluctuations that can create this issue. I have a special tool from Luk that is uesd to measure the radial movement of the flywheel to determine it's condition. With the car being shipped in gear for over 2,000 miles the movement on the truck was banging the flywheel and could have easily killed it or taken it far enough out of range to give this issue. Leaving a car in gear during transport can kill all sort of components and can even make the engine jump valve timing. We go to great lengths to ensure this doesn't happen, but it still does from time to time. I have an attendee of my class this past weekend who is in Mike's area and will pull the flywheel and ,easure it, then replace it. he will then send me the flywheel and I will measure it with my same special tool to see how much is has deviated from when my work was done. Dual mass flywheels can go bad like this and cause misfires even without aggrevation. We solve phantom issues with a flywheel replacement all the time. With Mike's car nothing made sense as it was perfect when I loaded it on the truck and it was horrible immediately after it was unloaded. The only real common denominator is the flywheel and the potential shipping damage. Its all on my dime, several thousand dollars worth. I do whats required.

Quotes source post: 11521741

That's interesting info to know and kudos to your dedication to get to the bottom of this and pay for it too. A question about the construction of the flywheel. Are the timing teeth, starter gear ring, and the center flange (shown in this photo) welded together as a single piece and bolted onto the crankshaft? https://rennlist.com/forums/11360135-post3.html Then the second mass is on the other side of the flywheel coupled to the first mass via springs?

Quotes source post: 11521741

I did some research on these two points and found out: 1) It appears to be normal practice among auto transport guides to tell the driver to put the car in gear, generally first gear. Presumably the trucker is more concerned about making sure the car physically gets to its destination than about damage that might occur internally. It is difficult to get references on this and I am not sure I got representative ones; I found 3 web pages from transport instructions in the UK, a page on ehow.com (which in my experience is hugely unreliable), and an Abu Dhabi transport company. Although Flat6 insists on keeping the car in neutral during transport, that may be against the trucker's policy, insurance, or whatnot. 2) A dual-mass flywheel is, as mentioned here, actually two flywheels rotating together but allowed to move radially from each other slightly by springs between them. These springs absorb rotational shocks from the engine before they are transmitted to the transmission, wheels, and chassis. However, these shocks, which are the changes in angular velocity of the crankshaft, are picked up by the ECU from the CPS signal. As each cylinder compresses or fires in turn, the crankshaft slows down or speeds up. The ECU, through the CPS signal, detects an orderly sequence of these rotational speed changes as correct cylinder firings. If they do not occur as expected, the ECU starts throwing misfire codes (P030x), and if the signal's integrity is questionable or far from expected, throws a sensor code (P0336). As Tony was able to narrow it down, my ECU was throwing P0336 almost instantly after clearing it once it was in bad shape. When run long enough under these conditions the P030x misfire codes were also thrown (they are accumulated for a longer period so don't show up instantly and only when really bad). The car drives perfectly when cold, and when on the freeway at higher and consistent RPMs (and airflow). It starts acting up after coming down to idle for a while when hot. Whether the trigger is heat, time, or idle is not known. As Jake described, idle is an unstable state for the car and could provoke failure of a marginal part that is sensitive to sudden changes in angular velocity. This all lends credence to the theory that the flywheel itself is at fault. It could be damaged in shape or position, springs broken, weakened, or acting inconsistently, or something else. We shall see.

Quotes source post: 11522141

Here is a thorough description of Luk (Porsche OEM) DMFs: http://www.schaeffler.com/remotemedi..._zms_de_en.pdf although it has almost no mention of where a sensor ring is attached. From the workshop manual I've attached an image below of the sensor and starter rings and my annotations are a guess as to which is which. This would mean the sensor ring is closer to the engine than the starter ring. The Bentley manual shows a few assembly photos that show that the sensor ring is closer to the engine than the starter ring. If that is true then both starter and sensor rings are on the primary mass attached to the crankshaft. So a plausible scenario for an errant misfire detection is as follows. Engine running is roughest at idle so the springs need to do lots of damping work and the primary and secondary masses of the DMF are frequently moving significantly against each other. If I happen to engage the clutch to accelerate at a time when the secondary mass is ahead of the primary mass, i.e. at a compression phase, the DMF will try to absorb the maximum rotation. If the spring is weak or broken it won't do so gently but will kick the engine side of the DMF backwards. The sensor ring is on the engine side, and it may detect this kick and sudden rotational slowdown as a misfire. There are probably several ways to test this hypothesis. One is that when the car is acting up, can I reduce the problem by clutching very very gently? I already know part of this: the problem sometimes occurs at idle (when it's really bad) but far more often when transitioning out of idle into gear, and when transitioning from deceleration to acceleration in gear. It seems not to happen at all when cruising. flywheel.png

Quotes source post: 11526610

Yes, your explanation of the starter ring and sensor ring is also what I understood. I'm just curious how the flywheel got damaged and what tripped that P0336 code. My understanding is the code is tripped because the DME cannot detect the correct number of teeth and expected reference teeth pattern after a number of revolutions of the flywheel. Obviously if one or more of the teeth are damaged, this code will be tripped but that doesn't explain why it works fine at other times and also can't imagine how towing can damage the sensor ring as it's not touching anything. So let's say the springs got damaged by towing (car was in 1st gear and handbrake was NOT engaged), your explanation of the second mass kicking back the primary mass which messes up the teeth counting is plausible. However, I think during idle (neutral) the engine is acting up too? It's hard to imagine the second mass itself (plus the clutch and transmission idle shaft) will kick back the sensor ring enough to cause it to miscount but I guess that's possible too. Also, how the temp/time trigger the symptoms? There are only 4 things: DME, wiring, CPS, flywheel so I'm pretty sure is one of the 4

BTW, I forgot now but did you ever measure the new CPS signal at the DME to confirm the signal's amplitude looks normal when the engine is normal and when it's acting up? Or it's like what you observed with the old CPS that the CPS signal was heavily attenuated when the engine is acting up?

Quotes source post: 11526686

No I didn't and now I am kicking myself. I still have the probe wires attached to the DME connector so I can still do this. Will see if I get some time.

I am cross posting this to the couple of threads I started while chasing this issue. It appears to be FINALLY resolved. To recap, after receiving the car back from Flat6 after an IMSB-caused repair, the car violently misfired after some time. It was perfect when put on the truck at Flat6, and really bad right off the truck. It has taken 4 months to get to the bottom of it. I am indebted to Jake at Flat6 and Tony Callas of Callas Rennsport for providing a huge amount of initial help, and especially to Clint Pfeifer at Euro Trends in Ceres, CA who finally diagnosed and chased this down. In the end it had nothing to do with the rebuild, but may be a typical problem for cars with more than a few years on them, possibly especially from northern or other climates where corrosion can occur. The bottom line was multiple layers of electrical issues: wiring harness and connector corrosion, slowly deteriorating electrical components, resulting in lack of spark when hot and intermittent CPS sensor. P0336 (CPS) fault code always thrown; P030x fault codes (misfires) thrown when severe misfiring occurred. Also learned: the ECU's fault codes are excellent starting points. If it throws a code, that is probably the problem, or very close to it. Don't start looking at other things until the entire diagnosis for the fault code has really been done, all the way down to the wires. Specifics: 1. 3 key pins of the X59 connector pair in the engine compartment were corroded: two CPS sensor wires and the main V+ wire from the ignition relay supplying power to the hot sides of the coils and injectors. As the engine compartment heated up, the resistance of these connectors increased to the point where spark failed, causing violent misfires progressively getting worse as engine compartment temperature increased (usually while idling or after stopping for gas or an errand). 2. Numerous electrical components were slowly deteriorating and not supplying as much power as new parts: alternator, alternator/starter cable, corroded pins on X59 connector, engine/chassis ground strap. Ignition coils looked fine, but replacing them improved the issues. An additional ground strap was added from transmission to chassis. 3. The sum total of voltage issues eventually crossed a threshold where coils did not supply enough voltage to spark. The dominant problem was the V+ supply to the coils losing voltage at the X59 connector, but it was not the only one. As each component above was replaced, the issues improved but never completely went away until the last link in the chain was replaced. 4. CPS cable in engine harness is compromised somewhere along the line. Replacing the X59 connector pins for the CPS wires reduced the P0336 codes, but did not completely eliminate them. The final fix was to install a completely separate cable from CPS into the DME. This nailed the P0336 codes. The car has about 150 miles on it since the final fix, under hot operating conditions, and has been running perfectly. Why the transportation from Flat6 to home tipped the scale, we don't know and probably never will. The problems were there all along.

Congrats!

Great thread! I'm working through some misfiring and a new code of P0355 after installing a set of supercup mufflers. My first thought was to immediately replace the mufflers, which I'm planning to do with Fisters. I recently replaced the starter & y-cable to address slow/no start when hot. The issue is *mostly* resolved in that no start isn't occurring and slow start is occurring pretty infrequently and with less severity. But, the issue is not gone 100%. Last night after installing the new exhaust I took the car on a spirited highway drive. A-ok, running very well & started up quickly in 85-90F outside temps. High speed shifting & acceleration was smooth on the highway. Once I got back into town I felt what seemed like a clutch slip, but soon realized it was some type of firing interruption - fuel, voltage, I'm not sure. As I drove towards home the problem became worse & threw P0335-Crankshaft Position Sensor A Circuit Malfunction. I stopped to read the code & experienced slow start followed by misfires when trying to accelerate in anything but a smooth, gradual pace. One data point that's different than the OP is voltage. I've consistently observed 14v after starting and while running with negligible variation. So next steps... I'm not sure that I have the degree of electrical issues that OP has given my car's history in the southwest (New Mexico 2 years, SoCal 11 years, Utah now 1.5 years). I did observe notable corrosion in the old y-cable (dark and variable coloration vs. new copper unit), and realized there's a similar cable that runs from the front of bank 2 back through the engine bay (this cable is likely corroded as well and could contribute to the lingering hot start). That's where I am. Worst case scenario I'll take it to my indy. Even with that approach I'd like to give them a good overview to improve the chances of success & reduce exploratory research & parts replacement. The car is sitting until I can get a clear idea on next steps. First order is to investigate CPS replacement as the misfires haven't thrown any codes but I have observed a rare ignition interruption or two in the past few months. My thinking before today was fuel pressure regulator & potentially a fuel pump, but after reading this and combining it with my recent slow start work, I'm less confident the fuel system is to blame.

OP here. It has been nearly 30k miles since the thread and the engine has been flawless after the fixes I described. Well except I found some minor throttle positioner issues a little after posting my last thread (bad throttle position readings, and poor idle) but I forget the codes right now. Given my experience, I focused right away on the throttle electrical connector, and knowing that I had disconnected and reconnected that connector a couple of times during my previous diagnosis efforts, I sprayed some DeoxIT DN5 in both sides of the connector and bingo, problem has been gone ever since. So I would recommend as a first step taking apart any connectors that have recently been touched, spray DeoxIT DN5 in them, let them sit a while and put back together.

Good to hear & encouraging for me as I start this little journey . Thanks for your contributions, they'll help guide my troubleshooting.

Did a little work today & got some reasonably close shots of the CPS. The code P0335 comes on during moderate to heavy throttle when the car is at the upper end of operating temps, say when coolant is >210F. The impact is sputter in power and the CEL stays solid, no blinking. Here are some shots of the CPS & incoming wiring. My next step is to buy a new CPS and I'd like to replace the up/downstream wiring to try and remedy the issue. An oscilliscope seems to cost around $100 minimum, so I'm comfortable taking some clearly aged parts & throwing money at it while the costs are relatively low. Any idea if the Bentley Manual includes a diagram of this data, or if anyon has one handy? I'll keep searching in parallel with inputs this thread. endpoint of the CPS a bit closer shot of the input into the housing looking up at the first connection to the CPS wire a different angle of the first connector

Finally have a new CPS installed. The awkward angle plus me being a dolt complicated the removal & install, so I delegated an hour of labor. After the first drive it seems P0335 is gone. It hasn't come on the last several drives, and pushing it this morning was met with no stumbling and consistent acceleration through the range. If the condition recurs I'll look into replacing the downstream wiring that looks to be nearly due for replacement. I used a matching Bosch part # for what it's worth. the offending, 15yo sensor:

mine was actually missing a piece of plastic from the top! car would buck on cold start up and not start at all when it was hot.

Fun times Did you just replace the CPS or do more thorough electrical updating similar to the OP?

Replaced the CPS and just recently replaced the starter and battery cables to it. That fixed my slow hot start issue. Before I replaced the CPS, I had to wait a bit before even trying to start it once it was hot. When the car was back on the road, I saw the bucking and no hot start issues were gone, but I could tell the starter was slow when hot.

Interesting. I'm on the fence right now on replacing more battery/starter cables. The y-cable is new, along with the starter (done last month). If the issue persists with slow start when heat soaked, I'll have to scope out a more thorough wiring update. If only it was easier to get to those parts, the work itself isn't very bad!

How many wire CPS is that, 2 or 3 ?

3 wires. 2 for the coil and one for the cable shield.

CPS issue is gone. Snuck in two canyon drives for the first time in a month or two.

Quotes source post: 14346505

OK, thanks, so it is not a hall sensor.

Quotes source post: 14347558

Nope, it's just a simple coil (with a magnet next to it) for robustness, reliability, and simplicity. The cam sensors are hall effect sensors.

Quotes source post: 14347562

Thanks again, I was thrown for a bit because of the shielding on a A/C output sensor, on the diesel side we never had any shielding on that type sensor, only on data communication wiring and throttle position wiring, back in early 2000's the A/C output sensors were popular but failures were common, all speed/position sensors in use with today's engine are hall effect sensors with far fewer failures.

OP, I am diagnosing the sensor as well. What is the voltage from the 3 wires when engine is not running? I know one is ground. When I measure the other two, both are giving me 2.5V....Do you happen to know?

Oh my, it's been a long time. I looked over the thread and didn't see where the DC voltages of the CPS signal wires were ever measured (may have missed it though). Anyway, when the engine is not running, the two signal wires should be at some equal DC voltage since the CPS is a magnetic pickup and will only force a voltage when it's picking up movement, i.e. a changing magnetic field. It's probably likely that the DME biases the wires to some value convenient for detecting an AC signal, and I'd guess 2.5 V which is half the 5 V the DME runs on. So, both being 2.5 V sounds, at first glance, like the right condition. BTW since fixing this issue and a similar throttle positioner issue (poor electrical contacts), the engine has accumulated about 45k miles and it has been running perfectly.

Quotes source post: 15771396

Thank you for prompt reply. My problem is very intermittent, random misfire on both banks without another codes. Still don't know what condition causes the issue. So my guess is the cps. Check fuel pressure and it was good. Also replaced coils and plugs, but issue still exist. Vacuum is good too. Will be looking at scope this weekend and see if I have any luck. When there is no misfire, car runs strong.

MK, according to wiring diagram, there is a ground shield, does it suppose to touch vehicle frame? I accessed the DME, connector III and found 980ohm to the crank sensor on pin 46 and 32, which indicate that the connectivity is good for the sensor coil. But the ground wire pin 45 is open and doesn't go to the vehicle ground. And when I measured the connector side under the car next to the sensor, the ground does connect to vehicle ground. Thus I think it means there is disconnect somewhere? I guess my other question is, does the DME pin 45 goes to vehicle ground or it should be insulated?

If your car crankshaft position sensor get faulty during your journey there are some temporary fixes which help you a lot. I find this post very useful for such situation. Read Here

Quotes source post: 18706559

Thanks for sending this tip. Sorry hadn't had a chance to report the final fix. After so much diagnostic, the issue revealed itself when I decided to rebuild the entire wire harness in the engine bay. It was the darn wire harness!. The wire was shielded and the shield was in bad shape. That explained why car misfired randomly only at lower rpm, as the voltage from the sensor is proportional to rpm. The ecu couldn't get a clean signal when rpm was low without a good shield. I had to route a new shielded wire (from local harware store) to bypassed the original harness from cps directly to the ecu. It worked perfectly and no misfire for 2+ years now.

Always nice to have a happy resolution!