Cranks, no start
I have looked at this vehicle at another shop. The vehicle came to the shop running, for head gasket replacement. When the intake manifold was put back on the engine, the wires to the cylinder 1&3 fuel injectors got pinched between the manifold and the cylinder block. This killed the electrical to these two fuel injectors. The shop repaired the pinched wires, which were the two trigger wires to the injectors. The repair was a good repair.
At this point, the engine would not start, the shop told me there was no injection pulse to the two injectors, which was tested with a noid light. The shop decided the electrical short had killed the PCM, so they sent it off to a rebuilder and had it repaired. With the rebuilt computer in the car, it still would not start.
When I got called in I found there was power to the injectors, but the PCM would not pulse the 1&3 injectors. I also found there was no spark to these two cylinders. Questioning the shop, the ignition coil had been replaced twice with a coil from AZ. I found there was current flow to both coils, but no spark output to cylinders 1&3, so a new coil was installed, which got good spark to all 4 cylinders. The engine still would not start.
At this point, I have tested the CKP/CMP for proper operation and timing. diag.net/file/f7dcsb3zv…
Scoping the injectors, I found that all 4 injectors are pulsed when the key is turned to the start position, then when the engine starts to turn, the injectors are pulsed sequentially. diag.net/file/f5onc8daa… When this waveform was taken, I had unplugged the cylinder 1 injector to verify the current flow of the #3 injector. I am using two current probes, hooked around the trigger wires at the PCM inside the car. Each injector draws around .9 amps of current, so when two are fired, the scope shows close to 1.8 amps.
The scope capture shows that the drivers are capable of firing all 4 injectors, although once the engine starts turning, it will not pulse the #1 & 3 injector. I tried unhooking the battery and shorting the cables together, which did not change anything. At this point, the shop sent the car to the Subaru dealer who installed and programmed a new PCM, which did not change anything. I have thought about an injector that draws too much current shutting down a driver, but I did not find anything there.
So, my question is, what kind of strategies are built into the PCM to keep it from pulsing those two injectors? What can cause the PCM to shut the drivers down?
Perform a resistance check on the injectors.
I did not use an ohm meter on the injectors, although I did a dynamic ohm test AKA current waveform.
Have you verified the integrity of that crank sensor signal? I have never seen one that looks like that.
That waveform I posted was taken during cranking, that is why the change in amplitude. I compared that waveform with two others taken from a good running engine, taken both running and cranking, All the waveforms matched.
Have your load tested and resistance checked the wiring from the injectors to the pcm. Did you check their wiring repair and see if maybe they messed up other wires. Sounds like the pcm sees incorrect resistance or maybe to much current and goes into safety. I'd definitely recheck all their work and see if they rewired it properly or if maybe it pinched at another spot.
I did take the harness apart where the repair had been made and verified the wiring was hooked to the correct injector. I did not load test the circuits, since the injector current, when the engine was first cranked looked the same as the current on the other, Cylinders 2 & 4 did. I have not tested an injector with an ohm meter is several years, but this might be a good time to fire the…
I remember a case study from a class many years ago with similar issue of dead cylinders. After a head gasket job the injector wires got switched on one side and were spaying on the wring cylinder. May not be related but have you checked the wiring to see all injectors are connected to the correct injector?
Commonly referred to as cross connections.
The correct cylinder has been verified. All cylinders are hooked correctly. I don't think that having the two injector plugs crossed would cause the PCM to stop pulsing the injectors though. There is something lurking in the wings that the PCM doesn't like.
Albin, How fast is the engine idling? I have had this problem several times over the years with Subaru vehicles. Every time it has been caused by a out of control idle. If the PCM can't lower the engine rpm at idle it will disable the injectors to slow the engine down. Cylinder one is the first to get shut off and then if it still isn't low enough cylinder three follows. I know the first time…
This engine will not start. When I first started on this, I had a scan tool hooked up, and I do not remember looking at the engine speed pid. I did look at the CKP & the CMP pids, which turned to yes when the engine started cranking. I will check the RPM when I get back to the vehicle.
I missed the no start in your original post. I've never seen a Subaru that didn't run on two cylinders. Have you tried giving it some throttle while cranking to see if it starts? Does it even attempt to start? I think you have another problem and the injector issue is secondary. It has never ran since they put it back together? I'd be checking their work over closely and treating it as a…
On earlier models sometimes you'd have to coax them to run on 2 cylinders and on even earlier models, depended on which timing belt was broken :O) Later models with throttle by wire, different strategies are at play. Come to think of it, it's been sometime since I've experienced a 2 cylinder running event... Now that the crank reluctor has been found damaged as stated below in the update, this…
With that one tooth broken on the reluctor, this one would try to start but would not make it, since the ignition timing was not always correct. Trying to run it on propane would result in fire being blown out the intake and explosions in the exhaust. There were only two ways to find this problem, either take the timing components off and inspect, or use some waveforms to trace the problem back…
>>or use some waveforms to trace the problem back to its source. Prudent and safe. Thanks for the leg work, the next time I get one of these 'No Starts' from another shop I'll know what to look for.
That injector current waveform looks like its shorted . Try another injector or scope the other known good injectors to compare waveform. You can also do a ohm check to verify if there is a short. Most ECU's will shut down an injector if it draws too much current to protect the ecu driver etc. There are some strategies that will intentionally shut down injectors if there is a misfire in that…
Thanks for the thought. When I get it back in, I will get out the ohm meter and see what it has to say.
Why does your green amp trace time line up with the voltage but the yellow shows more time ?
I'm wondering why the injector current is a square wave and not a ramp like a traditional port fuel injector. I'd also check fuses if there has been a short to ground, goofy things happen when current flows where it shouldn't.
I also think the injector has a steep ramp. And on your capture, the blue trace doesnt move, but shows it is marked voltage like the red. Was it not hooked up or am I missing something. I was thinking it was the other bank in sequential fire. And then also. Your average isn't matching the command from the red trace. Just a couple odd things, on a side note, I'm not s suby guy, so these all may…
On the blue trace, I had the injector unplugged so I could get a current ramp of the #3 injector only. Thanks for your reply.
Hmmm, verify there are no BIU (Body Integrated Unit) Security codes. Verify ECM has a RPM input from the CKP. Look for anything in the data PID's that would cause, contribute to or indicate Fuel-Cut. Make sure ECM is commanding fuel pump on. Look in the MAF for anything that shouldn't be there.
Hi Albin, Perhaps look at every ground wire that is mot deformed or loose in the engine compartment or a connector is not fully seated or the connector locks are not engaged. I had a vehicle this week that the lock was not engaged because the connector was not fully installed on a ignition coil causing a misfire.
Albin, How about wiring #1 and #3 piggy back on the other drivers? I am in agreement with the others who think there is a deeper problem and the missing #1 and #3 coil driving is a symptom, not the root cause. Normally I would have gone down the same route as you, verifying the key inputs for injector firing. Will this thing run on propane? It should if the coils are firing properly. If you…
Agree. Take the injectors out of the equation and see if it runs on spray or propane. This will prove out the integrity of the mechanical end of the engine work and pinpoint the actual area to concentrate on
I don’t see any inductive component to any of the injector waveforms. The current and voltage waveforms only show resistive components. Also the on time for the injectors is about 25ms. That seems too long but I don’t have a cranking to compare.
Looking thru my files, I only have this 09 Imprezza 2.5 cmp, ckp, cyl 1 sec, B2 VVT sol on start
Not trying to be overly simplistic. And I speak diesel way more than gas. But troubleshooting is relative. Too many times over the years of troubleshooting it is almost always the smallest detail that we put in our head that can't be it and skip it. When all it is said and done that small "cant be it" is what turns out to be the issue. So these days I fix everything no matter how small related…
Hi Albin just wondering did they try to start the engine with the injector wiring shorted.
Yes they did. That is why the original PCM was sent in for repair. After the no start, the shop used a noid light to check for injector pulse and they found none. The wires that were repaired were the negative- trigger wires to the #1& #3 injectors. If these wires were shorted to ground, all that would do is to turn on the injectors 100% of the time, and would have no effect on the injector…
If they tried for long enough the oil pan may have more gas than oil in it, does it make any attempt to start or at least fire. If the injectors were run at 100% duty cycle they may be damaged enough for the ECM to shut them down?. The Pico waveform as far as I could see has a very long injector pulse. Look forward to your update.
The injector .psdata file shows ChC and ChD ...but not ChA or ChB. Also, ChC and ChD appear to show opposite polarity like when a current probe needs to be flipped in order to display positive current. I agree with Moonsamy, Mark S, Steve P, ...regarding current ramp shape and lack of normal induction. The injector image file shows an inductive kick for cyl 2 injector voltage, but it doesn't…
Have you physically looked at the crank trigger? To me this looks like a tooth got chipped. If 1 tooth is chipped it will fire up to the teeth that are correct. The missing tooth arrangement does not look what would be correct for any manufacturer I know. Most are 12-1, 24-1 or 58-2. I would take off the cover and look at this. I have had a Subaru loose a small tooth. They are only 1mm wide and…
That is an odd looking pattern but i found one on another sit that matches the both the cam and crank, peak for peak. The RPM variation during cranking looks like the starter is working really hard but I don’t have another to compare.
Can you run it at all on an alternate fuel source. Like propane? This would at least rule out other issues, like more wires having been crushed. That is a strange fuel injector current waveform. I don't see any ramp, even on the unaffected injectors. Plus the spacing is uneven on your screenshot with the injector unplugged. I think more wires were damaged or something is plugged into the wrong…
So this is the crank trigger for a 2010 Subaru. I see roughly 15 peaks on your waveform. This trigger looks to have more then that.
The 15 is 180 degrees. The missing tooth, 15 , double skip tooth and 15 back to the missing tooth are 360 degrees. Quite odd looking.
The first GIF is the 2010 Forester 2.5 ckp cranking and the second GIF is from the 09 Imprezza 2.5 cranking
Does this help ? (1) Crankshaft position sensor (2) Crank sprocket (3) Crankshaft half-turn
The 2010 Forester 2.5 cranking and the 09 Imprezza 2.5 idling both have the same ckp tooth count.
I pulled up a picture from world pac that shows 2 "flat" spots on the crank timing gear. Which corresponds to what we see here. And in Ray's GIF. The part number is 13021 aa141. I cant figure out how to link that picture here. I seem to recall that trigger wheel you showed on older models.
Hi Albin, how about you go in cylinder with a ignition trace and also crank and cam? Maybe the ignition system is not firing at the right time. Is the timing done by gears or a belt?
Well Ladies & gentlemen, the problem is found. This should be a lesson to me to be more prudent in looking at waveforms. Today the car was hauled back to my shop. Step #1 was to use an ohm meter to test the resistance of the injectors. The resistance specification is 12 Ohms for the injectors, they tested out 11.93, 11.88, 11.83 & 11.93. That is close enough to 12 for me. I then…
Well I still feel like this was a good experiment :) I don't think I would have found it. Since your waveform is cranking, the lower amplitude seemed appropriate. Now I know it means something is wrong. Thanks
The amplitude of the pulses was normal and correct. There is 1 pulse missing. The slightly wider gap between the shorter pulse, where Albin’s arrow points and the next, taller pulse is where the pulse should be. The point count is 15 including the taller pulse. It should be 16 as in Ray’s capture and counting the points on the sprocket in Hollis’ post.
Thanks for clearing that up for me. I saw your reply to Tim Suggs where you state 15 pulses; but, I never compared the number of pulses to Ray's known good.
I messed up badly on that post. I counted the one set of pulses and assumed that the other set was the same count. Had I counted both, I would have found that there is 13 on the other half. If I had followed up on why they were not the same count, I MIGHT have found the missing pulse. That’s probably a stretch. It was a lot easier to see what was wrong after I knew what was wrong.
What kind of mechanic does that kind of thing? I've changed dozens (hundreds?) of those and I just can't see how you would break one off. Who would be stupid enough to put a broken one back on? I suppose the same guy that broke it? I guess I should start downloading and comparing wave form files before relying from now on. It would probably save me some typing. :-)
Eric, Some people are very good at quickly disassembling and reassembling things even if they have no clue how they work. He may have even noticed he broke the tooth prying the sprocket to do the crank seal, but figured it didn't matter. I almost did that once, but decided the sprocket would stay on, instead. I crossed-off the seal and the .2 from the order. After that I bought a ton of crazy…
Two Cheap-O flat tip screwdrivers, place in vise, heat with torch, bend tips 90 degrees, there's your crank gear remover reluctor tip savers :O) Subie crank gears have two existing holes, tap with 8Mx1.25, use steering wheel puller if the gear is stuck and the Cheap-O Pry-Drivers don't work, there's your 0.2hrs back :O)
I just slide my big pickle fork/tie rod remover behind the gear and pound it off with a big hammer. Haven't broken one yet. :-)
Torch?!? Is that a mainland word? Tapping the holes, is indeed the best trick I know, plus I found a few special tools in the close-outs of 2009. IDR what they look like or where they are at this point.
Torch as in Oxy-Acetylene to heat the screwdriver's to bend them to make the tool's. McTool-Gyver stuff :O)
Like this?; hdwallpapers.cat/wallpaper/hawa….jpg
Yeah Glligan, that should do nicely :O)
This deeper zoom shows it a little better but I think I would have been unlikely to notice it. The code certainly helped! Thanks for posting, Albin. I wonder if someone used a pry bar through the CKP hole to turn the crank a little and just broke off the top of the tooth that way. I'd be sore tempted at this time to "close up the gap" by whatever means to get the amplitude back up and I'm…
I think the tooth was broken off when the sprocket was pried of the crankshaft when the seal was changed. The tech told me he had taken the timing belt off two times to verify everything was set up correctly. This shop has its best teck laid up for a while. The tech was doing some work with a tractor and the tractor tipped over and broke some of the guys ribs. Looking back at this mess, it…
Wow that is really hard to see. Does the manufacturer give a spec for how much voltage peak to peak it needs to trigger the signal to the computer?
These are some known good ckp sensor measurements while cranking for an 09 Imprezza 2.5L
I once saw one where there was no picket fence appearance.It just appeared to be floating.
The one that this happened to me broke when reving it to redline. Was in shop for basic maintenance and I took it on a usual hard test drive and quit running after it went to redline. Was doing same thing and firing on few cyclinders and I had checked everything. I scoped crank thought it looked weird and pulled cover to find one missing tooth. Glad I could lead you to the solution!
Thanks for that zoom Phil. I see it now.