964 DME fault code - help needed
07-01-2003, 12:44 PM
#31
Your assumptions were correct. The 964 DME
does retard each cylinder in 3 deg. increments
to a max. of 9 degrees. The is per the Porsche
tech manual WKD 495 121 pg 2-08.
See, everyone learns when challenged!
Loren
does retard each cylinder in 3 deg. increments
to a max. of 9 degrees. The is per the Porsche
tech manual WKD 495 121 pg 2-08.
See, everyone learns when challenged!
Loren
07-01-2003, 12:51 PM
#32
Your assumptions were correct. The 964 DME
does retard each cylinder in 3 deg. increments
to a max. of 9 degrees. The is per the Porsche
tech manual WKD 495 121 pg 2-08.
See, everyone learns when challenged!
Loren
does retard each cylinder in 3 deg. increments
to a max. of 9 degrees. The is per the Porsche
tech manual WKD 495 121 pg 2-08.
See, everyone learns when challenged!
Loren
07-01-2003, 02:15 PM
#34
Addict
Lead Rennlist
Technical Advisor
Rennlist
Lifetime Member
Lead Rennlist
Technical Advisor
Rennlist
Lifetime Member
Dear Loren,
The data contained in the manual you are quoting (which is sitting beneath my feet by the way) needs to be quoted carefully. Having learned this the hard way. It details the very early 1989 C4 set up and is not always quite what is fitted to the real 964. I am sure you have seen many of the discrepancies. Page 2-09 contains a major one with the reasonance valve.
I am not disagreeing with you at all except the quote on controlling sequential fuel injection is a little inaccurate. The German version of the same manual does not say this. I have no idea why the English version does. I got into real trouble with my publisher over these Porsche words. You notice the title of these pages, fuel system and ignition.
I believe a short summary is in order.
The 964 uses a L-Jetronic fuel injection system which has been adapted to the Bosch Motronic DME.
The fuel injection system uses pulsed or sequential fuel injection.
The Hall sensor is used to provide the position of No 1 cylinder (TDC) to the systems inside the DME which need this information. The timing circuit.
The installation of a knock sensor system allows the use of more advanced timing maps.
The knock sensor system can only retard the timing of two cylinders a combined maximum of 9 degrees. If more retardation of timing is required the knock sensor system will switch off and the entire timing will be retarded 6 degrees. You will not find this in the service information manual.
The engine management system of the 964, compared to other systems of the same period is a relatively crude hybrid and both Porsche and Bosch have never allowed the data on how this system works internally to be released. I also point out that major changes were made to the firmware of the 964 DME from 1989 to the end of 1990. There are two series, 911 and 964. They are only one way interchangeable. A 911 part number DME will not work correctly if installed onto a 964 series harness.
The main reason for this post is to simply summarise the very basic operation of a small part of the DME.
Ciao,
Adrian
911C4
The data contained in the manual you are quoting (which is sitting beneath my feet by the way) needs to be quoted carefully. Having learned this the hard way. It details the very early 1989 C4 set up and is not always quite what is fitted to the real 964. I am sure you have seen many of the discrepancies. Page 2-09 contains a major one with the reasonance valve.
I am not disagreeing with you at all except the quote on controlling sequential fuel injection is a little inaccurate. The German version of the same manual does not say this. I have no idea why the English version does. I got into real trouble with my publisher over these Porsche words. You notice the title of these pages, fuel system and ignition.
I believe a short summary is in order.
The 964 uses a L-Jetronic fuel injection system which has been adapted to the Bosch Motronic DME.
The fuel injection system uses pulsed or sequential fuel injection.
The Hall sensor is used to provide the position of No 1 cylinder (TDC) to the systems inside the DME which need this information. The timing circuit.
The installation of a knock sensor system allows the use of more advanced timing maps.
The knock sensor system can only retard the timing of two cylinders a combined maximum of 9 degrees. If more retardation of timing is required the knock sensor system will switch off and the entire timing will be retarded 6 degrees. You will not find this in the service information manual.
The engine management system of the 964, compared to other systems of the same period is a relatively crude hybrid and both Porsche and Bosch have never allowed the data on how this system works internally to be released. I also point out that major changes were made to the firmware of the 964 DME from 1989 to the end of 1990. There are two series, 911 and 964. They are only one way interchangeable. A 911 part number DME will not work correctly if installed onto a 964 series harness.
The main reason for this post is to simply summarise the very basic operation of a small part of the DME.
Ciao,
Adrian
911C4
07-01-2003, 02:47 PM
#35
Track Day
Join Date: Apr 2003
Location: Norfolk, UK
Posts: 15
Likes: 0
Received 0 Likes
on
0 Posts
OK, if we're into quoting carefully then 'pulsed' and 'sequential' injection are not the same thing.
A pulsed injection system (from an electronically controlled injector) can be one of three types:
1) Full group, all injectors are pulsed together once per crank revolution.
2) Phased, injectors are pulsed in banks corresponding to B.D.C. once per rev.
3) Sequential, each cylinder's injector is fired individually at a calibrated crank angle once per cycle (2 revs).
Clearly, modes 1 & 2 do not require the phase of the engine to be known by the EMS. But mode 3 does and this is one reason why the dizzy sensor is required. If this input fails then either mode 1 or 2 will be reverted to (or if they have lazy code then it will "guess" the phase and be wrong 50% of the time!)
This is all quite different to non-pulsed (continuous) systems like the earlier Bosch K-Jetronic systems.
The timing of all this lot, the coil dwell/spark advance and the knock window etc etc is done using the information from the high resolution crank sensor input (=many "teeth" on the flywheel (?is it 58 - 60 with a double missing - on the 964 as with many Bosch systems?). NOT by the dizzy sensor. Whose function is, as Adrian says, to indicate TDC #1.
Cheers
Simon
A pulsed injection system (from an electronically controlled injector) can be one of three types:
1) Full group, all injectors are pulsed together once per crank revolution.
2) Phased, injectors are pulsed in banks corresponding to B.D.C. once per rev.
3) Sequential, each cylinder's injector is fired individually at a calibrated crank angle once per cycle (2 revs).
Clearly, modes 1 & 2 do not require the phase of the engine to be known by the EMS. But mode 3 does and this is one reason why the dizzy sensor is required. If this input fails then either mode 1 or 2 will be reverted to (or if they have lazy code then it will "guess" the phase and be wrong 50% of the time!)
This is all quite different to non-pulsed (continuous) systems like the earlier Bosch K-Jetronic systems.
The timing of all this lot, the coil dwell/spark advance and the knock window etc etc is done using the information from the high resolution crank sensor input (=many "teeth" on the flywheel (?is it 58 - 60 with a double missing - on the 964 as with many Bosch systems?). NOT by the dizzy sensor. Whose function is, as Adrian says, to indicate TDC #1.
Cheers
Simon
07-02-2003, 02:11 AM
#36
I do not have the info that Adrian has, so I
must assume that he is correct about how the
knock system is programmed. I have not personally
tested the knock system.
I agree with what Simonh says also. I think
it was previous understood by ALL that the
timing comes from the ring gear sensor (speed
sensor) and the notch which sets TDC. The missing
value (compression vs exhaust stroke) comes from
the Hall unit, that was not at issue. When you
write posts, some meanings get left out.
Here's what I know for sure as I can run the
964 DME in a bench test setup:
1. There are six independent injector drivers
which can fire separately.
2. There is only one ignition output to the
coil drivers for the upper and lower plugs,
so they both fire at once.
3. The Hall unit is not necessary to run in
the basic mode (Adrian-Limp) where all the
injections are fired at once or in banks.
The distributor handles the spark distribution
because it's linked to the cam.
4. The knock system can operate w/o the Hall
because it can effect all cylinders at once,
i.e. a common retard value.
5. This Limp mode, if you will, is basically
how the 3.2 Carrera Motronic runs, since it has
only the speed and ref. off the ring gear.
Note: The 964 eliminated the ref. sensor and
removed a tooth for TDC. If you integrate the
speed voltage (the basic average) and compare
this value to the voltage generated from each
tooth, you can find the TDC at the missing tooth.
So, I think all would agree that the 964 DME has
all the data (inputs) it needs to run a sequential
fuel injection and selectively retard each
cylinder independently. The real unknown is to
what extent Porsche/Bosch implemented those.
I leave that to Adrian, since my info/sources
are limited and I don't really want to spend
time running bench tests.
Good Luck to ALL
Loren
must assume that he is correct about how the
knock system is programmed. I have not personally
tested the knock system.
I agree with what Simonh says also. I think
it was previous understood by ALL that the
timing comes from the ring gear sensor (speed
sensor) and the notch which sets TDC. The missing
value (compression vs exhaust stroke) comes from
the Hall unit, that was not at issue. When you
write posts, some meanings get left out.
Here's what I know for sure as I can run the
964 DME in a bench test setup:
1. There are six independent injector drivers
which can fire separately.
2. There is only one ignition output to the
coil drivers for the upper and lower plugs,
so they both fire at once.
3. The Hall unit is not necessary to run in
the basic mode (Adrian-Limp) where all the
injections are fired at once or in banks.
The distributor handles the spark distribution
because it's linked to the cam.
4. The knock system can operate w/o the Hall
because it can effect all cylinders at once,
i.e. a common retard value.
5. This Limp mode, if you will, is basically
how the 3.2 Carrera Motronic runs, since it has
only the speed and ref. off the ring gear.
Note: The 964 eliminated the ref. sensor and
removed a tooth for TDC. If you integrate the
speed voltage (the basic average) and compare
this value to the voltage generated from each
tooth, you can find the TDC at the missing tooth.
So, I think all would agree that the 964 DME has
all the data (inputs) it needs to run a sequential
fuel injection and selectively retard each
cylinder independently. The real unknown is to
what extent Porsche/Bosch implemented those.
I leave that to Adrian, since my info/sources
are limited and I don't really want to spend
time running bench tests.
Good Luck to ALL
Loren
07-02-2003, 03:29 AM
#37
Gentlemen:
I thought I would throw in my opinions regarding what appears to me to be an area of confusion regarding this problem.
The DME, for all purposes, uses TWO hall sensors. One is what other car manufacturers refer to as a CRANKSHAFT POSITION SENSOR and it can detect the actual physical position of TDC on the crankshaft via the flywheel. The other picks up a signal from the distributor wiper to tell the DME where the position of the distributor wiper is inside the distributor. UPDATED ON 7/3/03: WHAT I SAID ABOVE IS NOT CORRECT! THE CRANKSHAFT SENSOR IN A 964 IS NOT A HALL SENSOR.
The check engine light DOES NOT REPORT A CODE FOR A FAILURE RELATED TO THE CRANKSHAFT POSITION SENSOR. The reason for this is simple: if the DME CANNOT tell where TDC is, there's no reason at all for it to even ATTEMPT to turn over or keep the engine running....it's a dead car. I believe the HAMMER can detect this, but the check engine light system won't bother reporting it. Why would the engineers at Porsche bother to code an error into a check engine light system on a dead car? It obviously needs servicing (unless someone simply wants to abandon the car and let it turn to rust!).
The other "hall sensor", detects the position of the distributor wiper (sort of, anyway...I think it just detects when the wiper is firing piston number 1) and is used in conjunction with the knock sensors to modify the timing in the event that it needs to be retarded due to knock. THIS is the culprit in the distributor, and it's referred to formally as the "hall signal", NOT the "hall sensor". This is the guy that is generating the fault code, NOT the crankshaft position sensor. If this goes bad, the car will still run, and the fault will likely NEVER even be reported until the engine is pushed to the point that knock appears and the DME attempts to reference the position of the distributor wipers only to realize that the reference isn't there.
While I'm at it, the reason the DME appears to be so "conservative" with it's timing is that here in the U.S. the engine is spec'd not for 93 octane fuel but 91 octane fuel.
Have at me...... <img border="0" title="" alt="[Wink]" src="wink.gif" />
<img border="0" alt="[cheers]" title="" src="graemlins/beerchug.gif" />
I thought I would throw in my opinions regarding what appears to me to be an area of confusion regarding this problem.
The DME, for all purposes, uses TWO hall sensors. One is what other car manufacturers refer to as a CRANKSHAFT POSITION SENSOR and it can detect the actual physical position of TDC on the crankshaft via the flywheel. The other picks up a signal from the distributor wiper to tell the DME where the position of the distributor wiper is inside the distributor. UPDATED ON 7/3/03: WHAT I SAID ABOVE IS NOT CORRECT! THE CRANKSHAFT SENSOR IN A 964 IS NOT A HALL SENSOR.
The check engine light DOES NOT REPORT A CODE FOR A FAILURE RELATED TO THE CRANKSHAFT POSITION SENSOR. The reason for this is simple: if the DME CANNOT tell where TDC is, there's no reason at all for it to even ATTEMPT to turn over or keep the engine running....it's a dead car. I believe the HAMMER can detect this, but the check engine light system won't bother reporting it. Why would the engineers at Porsche bother to code an error into a check engine light system on a dead car? It obviously needs servicing (unless someone simply wants to abandon the car and let it turn to rust!).
The other "hall sensor", detects the position of the distributor wiper (sort of, anyway...I think it just detects when the wiper is firing piston number 1) and is used in conjunction with the knock sensors to modify the timing in the event that it needs to be retarded due to knock. THIS is the culprit in the distributor, and it's referred to formally as the "hall signal", NOT the "hall sensor". This is the guy that is generating the fault code, NOT the crankshaft position sensor. If this goes bad, the car will still run, and the fault will likely NEVER even be reported until the engine is pushed to the point that knock appears and the DME attempts to reference the position of the distributor wipers only to realize that the reference isn't there.
While I'm at it, the reason the DME appears to be so "conservative" with it's timing is that here in the U.S. the engine is spec'd not for 93 octane fuel but 91 octane fuel.
Have at me...... <img border="0" title="" alt="[Wink]" src="wink.gif" />
<img border="0" alt="[cheers]" title="" src="graemlins/beerchug.gif" />
07-02-2003, 05:02 AM
#38
Track Day
Join Date: Apr 2003
Location: Norfolk, UK
Posts: 15
Likes: 0
Received 0 Likes
on
0 Posts
Bill are you definitely sure that the crank sensor is a hall effect device too? (The easy way to tell is by the number of wires on its connector, a hall effect device has 3 but a VR 2)
It's not common for the crank sensor to be hall effect, these devices being better suited to lower speeds than variable reluctance (VR) sensors.
I can believe that there is no check light fault for a failed crank sensor, as you said you can't start the engine without one. Where I work, our current systems can detect crank sensor failures as long as the engine is already running when the sensor fails, but if it's dead when you key on then there's no chance. (We can limp home on the phase sensor -just!) And these systems are 15 years further advanced from the 964 DME!
I very much doubt that it only looks at the dizzy sensor when there's a knock event and it needs to know something. All the synchronisation is done at startup. Basically the signal frame from the crank is recongised first, so it knows the current crank position. Then it looks at the hall effect sensor in the dizzy whilst passing TDC to indicate whether this is TDC#1 or not - output high for no metal present, low if it's there. (Unlike VR sensors, hall effects indicate the presence of metal continuously even when stationary so are ideally suited to this task). From then on the system never needs to check the hall effect sensor again until the next startup. It is likely that it is monitored continuously for diagnostics purposes only. Simply it must switch states (From high to low and back) each TDC - if it doesn't then it's knackered!
Cheers
Simon
It's not common for the crank sensor to be hall effect, these devices being better suited to lower speeds than variable reluctance (VR) sensors.
I can believe that there is no check light fault for a failed crank sensor, as you said you can't start the engine without one. Where I work, our current systems can detect crank sensor failures as long as the engine is already running when the sensor fails, but if it's dead when you key on then there's no chance. (We can limp home on the phase sensor -just!) And these systems are 15 years further advanced from the 964 DME!
I very much doubt that it only looks at the dizzy sensor when there's a knock event and it needs to know something. All the synchronisation is done at startup. Basically the signal frame from the crank is recongised first, so it knows the current crank position. Then it looks at the hall effect sensor in the dizzy whilst passing TDC to indicate whether this is TDC#1 or not - output high for no metal present, low if it's there. (Unlike VR sensors, hall effects indicate the presence of metal continuously even when stationary so are ideally suited to this task). From then on the system never needs to check the hall effect sensor again until the next startup. It is likely that it is monitored continuously for diagnostics purposes only. Simply it must switch states (From high to low and back) each TDC - if it doesn't then it's knackered!
Cheers
Simon
07-02-2003, 06:38 AM
#39
Addict
Lead Rennlist
Technical Advisor
Rennlist
Lifetime Member
Lead Rennlist
Technical Advisor
Rennlist
Lifetime Member
Dear Loren,
I do have a huge amount of information but the trouble is a lot of it contradicts itself. The Porsche service manuals never have the same explanation twice. They are also badly written and or badly translated because they were really meant to be training manuals. The 1989 C4 one you and I both have was actually written based on the prototype/pre-production models and not the actual production model.
At the moment I cannot add much more because the thread has taken another turn. It is also getting very deep and away from the original aim of trying to help somebody fix their problem.
As Fox Mulder says "The truth is out there".
Ciao,
Adrian
911C4
I do have a huge amount of information but the trouble is a lot of it contradicts itself. The Porsche service manuals never have the same explanation twice. They are also badly written and or badly translated because they were really meant to be training manuals. The 1989 C4 one you and I both have was actually written based on the prototype/pre-production models and not the actual production model.
At the moment I cannot add much more because the thread has taken another turn. It is also getting very deep and away from the original aim of trying to help somebody fix their problem.
As Fox Mulder says "The truth is out there".
Ciao,
Adrian
911C4
07-02-2003, 08:54 AM
#40
Track Day
Join Date: Apr 2003
Location: Norfolk, UK
Posts: 15
Likes: 0
Received 0 Likes
on
0 Posts
I think the original aim was to answer the 4 question posed by Ingo at the top of this thread. Hopefully I've helped with numbers 2 and 3 with my posts above explaining how the EMS interprets and uses signals from the crank and distributor sensors. Sorry I can't offer any help with questions 1 & 4 as I don't know the answers.
Hopefully by closer investigation and pooling the collective knowledge of the members of this forum we will get to 'the truth'.
Cheers
Simon
Hopefully by closer investigation and pooling the collective knowledge of the members of this forum we will get to 'the truth'.
Cheers
Simon
07-02-2003, 01:04 PM
#41
Adrian, I totally agree. Also, I saw your web site
and you have my sympathy about your loss.
SimonH is correct. The speed sensor is NOT a hall
device but a magnetic pickup unit as used on most
all Bosch systems to produce pulses to count for
crank angle position.
A hall unit is a semiconductor device implemented
as either a diode (2 leads) or a transistor
(3 leads). The hall conductivity is affected
by a magnetic field. There are no magnets
on or near the 964 ring gear. See page 2-11
of WKD 495 121 (964 Tech Manual) for a discussion
of the inductive (magnetic) sender used on the
964.
As SimonH & Adrian imply, let's move on.
and you have my sympathy about your loss.
SimonH is correct. The speed sensor is NOT a hall
device but a magnetic pickup unit as used on most
all Bosch systems to produce pulses to count for
crank angle position.
A hall unit is a semiconductor device implemented
as either a diode (2 leads) or a transistor
(3 leads). The hall conductivity is affected
by a magnetic field. There are no magnets
on or near the 964 ring gear. See page 2-11
of WKD 495 121 (964 Tech Manual) for a discussion
of the inductive (magnetic) sender used on the
964.
As SimonH & Adrian imply, let's move on.
07-02-2003, 01:38 PM
#42
Addict
Lead Rennlist
Technical Advisor
Rennlist
Lifetime Member
Lead Rennlist
Technical Advisor
Rennlist
Lifetime Member
Dear Loren,
Thank you very much.
As you say the reference mark and speed sensor is an inductive sender. The reference mark signal is sent 84 degrees before TDC. There are 60 teeth on the motronic timing ring. On page 2-12 it clearly states that the TDC sender system has been omitted. It is no longer required. Time to move on and get back to fixing problems. I hope that nobody is too confused.
Ciao,
Adrian
911C4
Thank you very much.
As you say the reference mark and speed sensor is an inductive sender. The reference mark signal is sent 84 degrees before TDC. There are 60 teeth on the motronic timing ring. On page 2-12 it clearly states that the TDC sender system has been omitted. It is no longer required. Time to move on and get back to fixing problems. I hope that nobody is too confused.
Ciao,
Adrian
911C4
07-02-2003, 01:43 PM
#43
And to go slightly off-topic, here is one of the major problems facing repair shops; incorrect or misleading technical and repair information.
If even the manufacturers (and this applies to virtually all of them) cannot supply or print correct repair information, then to expect shops, both dealer and independant, to achieve 100% repair satisfaction on diagnosis of complex problems is not realistic.
This situation is not going to improve with the continuing increase in electronic systems, some of which barely make to the end-of-warranty period.
If even the manufacturers (and this applies to virtually all of them) cannot supply or print correct repair information, then to expect shops, both dealer and independant, to achieve 100% repair satisfaction on diagnosis of complex problems is not realistic.
This situation is not going to improve with the continuing increase in electronic systems, some of which barely make to the end-of-warranty period.
07-03-2003, 02:30 AM
#44
SimonH:
I wasn't meaning to imply that the crank sensor is ignored during knocking. If I gave you that impression then I wrote it up incorrectly. As to whether or not the crankshaft sensor is a Hall sensor, I reluctantly have to admit that I'm possibly confusing the 964 with something that may not even be automotive related. These things happen when you get senile. <img border="0" title="" alt="[Eek!]" src="eek.gif" />
My previous post should probably be ignored with respect to the crank sensor.
<img border="0" alt="[cheers]" title="" src="graemlins/beerchug.gif" />
I wasn't meaning to imply that the crank sensor is ignored during knocking. If I gave you that impression then I wrote it up incorrectly. As to whether or not the crankshaft sensor is a Hall sensor, I reluctantly have to admit that I'm possibly confusing the 964 with something that may not even be automotive related. These things happen when you get senile. <img border="0" title="" alt="[Eek!]" src="eek.gif" />
My previous post should probably be ignored with respect to the crank sensor.
<img border="0" alt="[cheers]" title="" src="graemlins/beerchug.gif" />
07-03-2003, 05:09 AM
#45
Addict
Lead Rennlist
Technical Advisor
Rennlist
Lifetime Member
Lead Rennlist
Technical Advisor
Rennlist
Lifetime Member
Dear Bill,
Do not worry about such things. Happens to all of us. The old brain box can be filled with so much information that it can get muddled up. I am doing the 911SC book at the moment. I thought I had learned CIS when I did the 964 turbo section of my book. Nope got to start again and I get them mixed up.
Dear Jon,
I think the problem with your chip is in the suppliers modifications to the codes. I suspect that cylinder engine temperature and air temperature input considerations may have been missed. Not everything is linear as I am sure you are aware. Small change here seems to work but causes big change here and nobody notices because it was never tested. This type of failure is quite common in reverse engineering projects. Porsche and Bosch do not provide the address codes. The EPROM is just one chip among a large chip set. They all have to work together and I know by the end of 1990 Bosch had worked out with the 964 series of DMEs how to detect the aftermarket chips.
Ciao,
Adrian
911C4
Do not worry about such things. Happens to all of us. The old brain box can be filled with so much information that it can get muddled up. I am doing the 911SC book at the moment. I thought I had learned CIS when I did the 964 turbo section of my book. Nope got to start again and I get them mixed up.
Dear Jon,
I think the problem with your chip is in the suppliers modifications to the codes. I suspect that cylinder engine temperature and air temperature input considerations may have been missed. Not everything is linear as I am sure you are aware. Small change here seems to work but causes big change here and nobody notices because it was never tested. This type of failure is quite common in reverse engineering projects. Porsche and Bosch do not provide the address codes. The EPROM is just one chip among a large chip set. They all have to work together and I know by the end of 1990 Bosch had worked out with the 964 series of DMEs how to detect the aftermarket chips.
Ciao,
Adrian
911C4