CPS Failure. 1990 S4. Intermittent start issue solved. Lessons learned
05-05-2016, 03:12 PM
#1
Pro
Thread Starter
Join Date: Dec 2010
Location: Palgrave, Ontario, Canada
Posts: 700
Likes: 0
Received 2 Likes
on
2 Posts
CPS Failure. 1990 S4. Intermittent start issue solved. Lessons learned
I've been battling a intermittent, no-start issue with my '90 S4 Euro. Intermittent issues can drive you crazy, since you do some work, the car appears fine and you think you've solved the issue, then it recurs. Very frustrating.
The bottom-line, it was caused by a failing crank position sensor (CPS). There were some interesting characteristics that I thought I would add to the knowledge base here.
The no-start first occurred in the Fall. Mostly the car would start when cold, and not as often when warm after running, but not always. And the frequency of no-start events slowly increased. While I was not driving the car over Winter, I worked on it.
I could always tell if the car was going to start as soon as I turned the ignition on, even before I cranked the starter. The diagnostic feature was that the relays for the two engine computers would click on then off alternatively, and furiously, before eventually stopping. If I then turned the key to Start, the engine would crank, but fail to start, and most tellingly, the tach would either remain pegged at '0', or bounce around crazily to between 0 and 3K or more. This behaviour is likely very diagnostic, so I emphasize it here.
My initial assumption was that the issue was electrical. Some people recommended checking spark and injector action, but I was pretty sure that diagnosis would not help either way.
So starting with the easy things first:
- I checked the fuses
- I started cleaning grounds, and eventually cleaned every ground except the one under the steering column. Especially important were the ones on the bellhousing that ground the computers (as kindly suggested by John Speake after I contacted him for advice)
- I checked and eventually replaced the appropriate relays, cleaning their sockets with Deoxit
- Checked the plugs for the LH and EZK
- I checked the 14 pin engine harness plug and socket, since I had been working near there when the issue first occurred
- I replaced the negative battery terminal after it broke when I went to straighten it while cleaning the connection
- I replaced the rear knock sensor (according to the WSM this can cause a no-start) since the mount and plug end were cracked
- I swapped in the LH and EZK from my '88. John Speake had already told me that when the computers fail they do so totally, not intermittently, but it was easy to test
- I checked the resistance of the Temp II sensor at cold (A full diagnosis should also check resistance at other temperatures, but since my car would run fine if it started, that did not seem relevant)
- I also used the 928 Diagnostic Tool that I borrowed from Rob Peace. There were no stored error codes and no findings when I ran the diagnostic tests as the car idled
Still no joy.
As noted, I was also running through the WSM diagnostic table. I had checked just about everything relevant to my situation listed - except the crank position sensor (CPS). The RL discussions about the CPS suggest it seldom fails, but several people had posted that it had in their case. A full diagnosis requires an oscilloscope, which I don't have. The CPS is not cheap, but I decided to buy a new one, since there was not much else left to check.
So removing the air intake and filter, MAF, etc. I was back to to the state I was in when I checked the bellhousing grounds (and replaced the U-shaped fuel hose). There are horror stories about efforts to remove the CPS where the cap breaks off and the barrel remains jammed in the block. There is a Post by Stan here on RL, where he describes his recommended procedure. A critical step is to first rotate the cap. After soaking it is penetrant for a while I was easily able to do this. With a bit of wiggling I was able to move the sensor up a bit, then get an angled pick/puller under the cap close to the barrel (aiming to reduce the leverage that might break the cap off). Stan used mechanics wire wrapped around the barrel, but I didn't need to do this as the sensor pulled up with relatively little effort. Phew!
BTW, the Greg Brown U-shaped hose I had already put in seemed to give me easier access to the CPS. Also, while I was 'in there' this time I also replaced the fuel hose to the fuel cooler - the old one looked fine, but for peace of mind...
Anyway, once I'd pressurized the fuel system with my switched jumper and checked for leaks, then reassembled, I held my breath and turned the starter. The car fire up immediately. I turned off the engine and tried again - still worked. Then I took it for a drive to warm up, came home, turned it off then tried to start it again - success! I still hold my breath every time, but so far it works.
In hindsight, a failing CPS makes perfect sense. If the LH can't get a good signal, it likely reboots. And since the EZK is booted by a signal from the LH, it shuts down and the cycle continues. At least that is my non-expert hypothesis. Also, in talking to a mechanic friend, he told me that in many vehicles a failed CPS does not result in error codes being recorded in the engine computer(s) since the car doesn't actually think it is trying to start without an engine rotation signal.
Of course I could have come up with a rationalizing explanation if any of the other things I tried had worked.
There was one other clue I had missed. Ever since I got the car, the idle fluctuated slightly (+/- 50 rpm) - a kind of loping note. I assumed a small vacuum leak or ignition issue. It was on my list to investigate, but I hadn't gotten to it as it was very minor. Well now with the new CPS, idle is rock steady...
The bottom-line, it was caused by a failing crank position sensor (CPS). There were some interesting characteristics that I thought I would add to the knowledge base here.
The no-start first occurred in the Fall. Mostly the car would start when cold, and not as often when warm after running, but not always. And the frequency of no-start events slowly increased. While I was not driving the car over Winter, I worked on it.
I could always tell if the car was going to start as soon as I turned the ignition on, even before I cranked the starter. The diagnostic feature was that the relays for the two engine computers would click on then off alternatively, and furiously, before eventually stopping. If I then turned the key to Start, the engine would crank, but fail to start, and most tellingly, the tach would either remain pegged at '0', or bounce around crazily to between 0 and 3K or more. This behaviour is likely very diagnostic, so I emphasize it here.
My initial assumption was that the issue was electrical. Some people recommended checking spark and injector action, but I was pretty sure that diagnosis would not help either way.
So starting with the easy things first:
- I checked the fuses
- I started cleaning grounds, and eventually cleaned every ground except the one under the steering column. Especially important were the ones on the bellhousing that ground the computers (as kindly suggested by John Speake after I contacted him for advice)
- I checked and eventually replaced the appropriate relays, cleaning their sockets with Deoxit
- Checked the plugs for the LH and EZK
- I checked the 14 pin engine harness plug and socket, since I had been working near there when the issue first occurred
- I replaced the negative battery terminal after it broke when I went to straighten it while cleaning the connection
- I replaced the rear knock sensor (according to the WSM this can cause a no-start) since the mount and plug end were cracked
- I swapped in the LH and EZK from my '88. John Speake had already told me that when the computers fail they do so totally, not intermittently, but it was easy to test
- I checked the resistance of the Temp II sensor at cold (A full diagnosis should also check resistance at other temperatures, but since my car would run fine if it started, that did not seem relevant)
- I also used the 928 Diagnostic Tool that I borrowed from Rob Peace. There were no stored error codes and no findings when I ran the diagnostic tests as the car idled
Still no joy.
As noted, I was also running through the WSM diagnostic table. I had checked just about everything relevant to my situation listed - except the crank position sensor (CPS). The RL discussions about the CPS suggest it seldom fails, but several people had posted that it had in their case. A full diagnosis requires an oscilloscope, which I don't have. The CPS is not cheap, but I decided to buy a new one, since there was not much else left to check.
So removing the air intake and filter, MAF, etc. I was back to to the state I was in when I checked the bellhousing grounds (and replaced the U-shaped fuel hose). There are horror stories about efforts to remove the CPS where the cap breaks off and the barrel remains jammed in the block. There is a Post by Stan here on RL, where he describes his recommended procedure. A critical step is to first rotate the cap. After soaking it is penetrant for a while I was easily able to do this. With a bit of wiggling I was able to move the sensor up a bit, then get an angled pick/puller under the cap close to the barrel (aiming to reduce the leverage that might break the cap off). Stan used mechanics wire wrapped around the barrel, but I didn't need to do this as the sensor pulled up with relatively little effort. Phew!
BTW, the Greg Brown U-shaped hose I had already put in seemed to give me easier access to the CPS. Also, while I was 'in there' this time I also replaced the fuel hose to the fuel cooler - the old one looked fine, but for peace of mind...
Anyway, once I'd pressurized the fuel system with my switched jumper and checked for leaks, then reassembled, I held my breath and turned the starter. The car fire up immediately. I turned off the engine and tried again - still worked. Then I took it for a drive to warm up, came home, turned it off then tried to start it again - success! I still hold my breath every time, but so far it works.
In hindsight, a failing CPS makes perfect sense. If the LH can't get a good signal, it likely reboots. And since the EZK is booted by a signal from the LH, it shuts down and the cycle continues. At least that is my non-expert hypothesis. Also, in talking to a mechanic friend, he told me that in many vehicles a failed CPS does not result in error codes being recorded in the engine computer(s) since the car doesn't actually think it is trying to start without an engine rotation signal.
Of course I could have come up with a rationalizing explanation if any of the other things I tried had worked.
There was one other clue I had missed. Ever since I got the car, the idle fluctuated slightly (+/- 50 rpm) - a kind of loping note. I assumed a small vacuum leak or ignition issue. It was on my list to investigate, but I hadn't gotten to it as it was very minor. Well now with the new CPS, idle is rock steady...
Last edited by martinss; 05-06-2016 at 09:51 AM. Reason: revise title
05-05-2016, 07:18 PM
#3
2nd Gear
Join Date: Aug 2012
Posts: 2
Likes: 0
Received 0 Likes
on
0 Posts
Yes I had a similar situation trying to start my car Melvin after replacing the fuel hoses, however mine was an easy fix, the 14 pin engine plug, the hard plastic plug had cracked right where the yellow starter pin was and the starter was getting intermittent power to start. I used some epoxy to glue it back together, it's fine for now. I assume the plug is plastic, it could even be Bakelite.
Good right up Martin.
Good right up Martin.
05-05-2016, 10:21 PM
#5
Great write up! It's nice when folks take the time to document their symptom(s) and repair(s).
It would be nice if you had the model/year of your car in the title...to make searching it out easier for those with similar problems in the future.
Brian.
It would be nice if you had the model/year of your car in the title...to make searching it out easier for those with similar problems in the future.
Brian.
05-05-2016, 11:01 PM
#6
Rennlist Member
Great contribution to the knowledge base!
Congratulations. Well done detective work!
Frank
Congratulations. Well done detective work!
Frank
Trending Topics
05-06-2016, 05:42 AM
#8
Rennlist Member
Martin,
An interesting post and a thorough write up thereof. No idea how you came up with the notion that the CPS is not a common failure item as many of us DIY hacks regularly steer investigation in that direction [amongst others]. The LH is an obvious target for mysterious intermittent problems and seemingly so too is the CPS- as you now know! I tend to think either they work or they do not but clearly not always the case.
I also note your last sentence about the small oscillation in the rpm's- never heard that one before but just goes to show the power of these forums. I have a small issue like that at the moment so your note is lodged in grey matter.
I would give you a small friendly slap on the back of the head for seemingly not posting earlier in the event - apologies if I missed anything. I always take the point of view that a problem shared is a problem halved and you are among friends!
I would also suggest you modify the thread title to read: CPS failure plus the rest of the title line to make search engine access more pertinent- posts like this are real treasure.
Regards
Fred
An interesting post and a thorough write up thereof. No idea how you came up with the notion that the CPS is not a common failure item as many of us DIY hacks regularly steer investigation in that direction [amongst others]. The LH is an obvious target for mysterious intermittent problems and seemingly so too is the CPS- as you now know! I tend to think either they work or they do not but clearly not always the case.
I also note your last sentence about the small oscillation in the rpm's- never heard that one before but just goes to show the power of these forums. I have a small issue like that at the moment so your note is lodged in grey matter.
I would give you a small friendly slap on the back of the head for seemingly not posting earlier in the event - apologies if I missed anything. I always take the point of view that a problem shared is a problem halved and you are among friends!
I would also suggest you modify the thread title to read: CPS failure plus the rest of the title line to make search engine access more pertinent- posts like this are real treasure.
Regards
Fred
05-06-2016, 07:43 AM
#10
Rennlist Member
Join Date: Aug 2014
Location: Adirondack Mountains, New York
Posts: 2,398
Received 309 Likes
on
161 Posts
Perhaps successes like this should always have their own summaries like this excellent example, with a standard title "928 Problem solved - [theProblem] - [theSolution]", with a link to the main thread, if any.
People could write them long after the fact. A suggested format could be put in the "New visitor" thread.
(For the record, I changed the CPS during my intake refresh. It came out easily, but I am lucky to own a car that is largely free of crusty stuff.)
People could write them long after the fact. A suggested format could be put in the "New visitor" thread.
(For the record, I changed the CPS during my intake refresh. It came out easily, but I am lucky to own a car that is largely free of crusty stuff.)
05-06-2016, 10:00 AM
#11
Pro
Thread Starter
Join Date: Dec 2010
Location: Palgrave, Ontario, Canada
Posts: 700
Likes: 0
Received 2 Likes
on
2 Posts
Title amended per suggestions. I'm glad post this could be useful.
I recall several posts that I read which specifically said the CPS seldom fails, and others that had the failure.
The intermittent nature, plus the fact that I was not driving the car for the last 6 months, are the reasons it took so long to fix. Really there were only a couple of days of work, and all the things I did (grounds, fuel hoses, battery terminal, relays, etc.) have long-term preventative value.
There is great help here, and several members assisted me directly as noted. I did make some posts to other threads, but did not start my own. Suggestions were made around spark and ignition checks. As a supporter, I always post my work on the 928 OC forums, but readership there is much lower unfortunately.
I recall several posts that I read which specifically said the CPS seldom fails, and others that had the failure.
The intermittent nature, plus the fact that I was not driving the car for the last 6 months, are the reasons it took so long to fix. Really there were only a couple of days of work, and all the things I did (grounds, fuel hoses, battery terminal, relays, etc.) have long-term preventative value.
There is great help here, and several members assisted me directly as noted. I did make some posts to other threads, but did not start my own. Suggestions were made around spark and ignition checks. As a supporter, I always post my work on the 928 OC forums, but readership there is much lower unfortunately.
05-06-2016, 12:11 PM
#13
Rennlist Member
Title amended per suggestions. I'm glad post this could be useful.
I recall several posts that I read which specifically said the CPS seldom fails, and others that had the failure.
The intermittent nature, plus the fact that I was not driving the car for the last 6 months, are the reasons it took so long to fix. Really there were only a couple of days of work, and all the things I did (grounds, fuel hoses, battery terminal, relays, etc.) have long-term preventative value.
There is great help here, and several members assisted me directly as noted. I did make some posts to other threads, but did not start my own. Suggestions were made around spark and ignition checks. As a supporter, I always post my work on the 928 OC forums, but readership there is much lower unfortunately.
I recall several posts that I read which specifically said the CPS seldom fails, and others that had the failure.
The intermittent nature, plus the fact that I was not driving the car for the last 6 months, are the reasons it took so long to fix. Really there were only a couple of days of work, and all the things I did (grounds, fuel hoses, battery terminal, relays, etc.) have long-term preventative value.
There is great help here, and several members assisted me directly as noted. I did make some posts to other threads, but did not start my own. Suggestions were made around spark and ignition checks. As a supporter, I always post my work on the 928 OC forums, but readership there is much lower unfortunately.
Heat and time gets to these things as they do with many other items almost irrespective of mileage or so it would seem. Things like the CPS, ISV, breather hoses, ignition leads, Hall sensor, water pump, timing belt etc etc etc have a trouble free life of about 10 years and after that it is simply a crap shoot as to what fails and when. An astute owner who depends on the car as a daily driver would systematically replace these - if like me you do not go too far and drive the car as an interest one can probably afford to be a bit slack on some items apart from the TB/WP which i probably over do.
Rgds
fred
