Public-domain OBD1 PCB board - ORDER HERE
#183
Originally Posted by E6MPB
Ray,
I believe the knock sensor output on pin 11 provides a instant high to low transition when the ecu detects a knock signal with the correct frequency components. I use this signal output into a simple binary counter cct to determine the number of knocks detected over a period of time. The pulses are short and sync'd to cylinders. It is not transmitted with any protocol around it. I do not know about the counting system within the ECU that is mentoned above. If the counter escalates up I take my foot off the accelerator. Saved me I'm sure a few times with the SC. Compensating for knock with the w/a injection helps alot and kills the knock detected. Now left with a little transitional knock but working through that by retarding the injection as it gets on boost. When I did get a sitution when I could hear the knock (a single ping)the counter was really going for it so I think it is a realistic measure.
BR, Mark.
I believe the knock sensor output on pin 11 provides a instant high to low transition when the ecu detects a knock signal with the correct frequency components. I use this signal output into a simple binary counter cct to determine the number of knocks detected over a period of time. The pulses are short and sync'd to cylinders. It is not transmitted with any protocol around it. I do not know about the counting system within the ECU that is mentoned above. If the counter escalates up I take my foot off the accelerator. Saved me I'm sure a few times with the SC. Compensating for knock with the w/a injection helps alot and kills the knock detected. Now left with a little transitional knock but working through that by retarding the injection as it gets on boost. When I did get a sitution when I could hear the knock (a single ping)the counter was really going for it so I think it is a realistic measure.
BR, Mark.
I assume this was a prior to 993 model? 993 diagnostcs has long explanation of timing retard that occurs with knock sense to prevent damage.
From your explanation, I wont see anything with scope in garage. Since California doesn't have proper octane gas I could put a pulse monitor on pin 11 and accelerate uphill, to see this pulse. I've been wondering if octane booster is useful for autocross.
Ray
#184
"I do not know about the counting system within the ECU that is mentoned above."
The DME ECM "knows" which cylinder caused the knock and selectively retards that cylinder
in 3 degree increments to a max of 9 degrees. Once the knock stops, the DME ECM advances
the timing back in small increments. This causes a lag effect in the utility of using knock
data, i.e. a less significant knock may exist but may not be indicated because the timing
is retarded beyond the subsequent knock point. Thus, the knock data is more of an average/overall
evaluation of the knock conditions versus an instantaneous one. That's why the data presented
via the testers is based on the 10K cylinder ignitions interval.
The knock data are stored by the main DME ECM u-p once received from the DSP
implemented other u-p. Thus, each cylinder's knock data is there, although the DME
ECM diagnostics doesn't output this data which has much more utility than the simple
Yes/No info.
As I mentioned, using a scope should easily provide the pulse shape if one
triggers the scope for the appropriate transition, assuming the scope is fast
enough. Again, the conditions for a knock must be met, i.e. load & etc.
The DME ECM "knows" which cylinder caused the knock and selectively retards that cylinder
in 3 degree increments to a max of 9 degrees. Once the knock stops, the DME ECM advances
the timing back in small increments. This causes a lag effect in the utility of using knock
data, i.e. a less significant knock may exist but may not be indicated because the timing
is retarded beyond the subsequent knock point. Thus, the knock data is more of an average/overall
evaluation of the knock conditions versus an instantaneous one. That's why the data presented
via the testers is based on the 10K cylinder ignitions interval.
The knock data are stored by the main DME ECM u-p once received from the DSP
implemented other u-p. Thus, each cylinder's knock data is there, although the DME
ECM diagnostics doesn't output this data which has much more utility than the simple
Yes/No info.
As I mentioned, using a scope should easily provide the pulse shape if one
triggers the scope for the appropriate transition, assuming the scope is fast
enough. Again, the conditions for a knock must be met, i.e. load & etc.
Last edited by Lorenfb; 09-23-2006 at 03:47 AM.
#185
Ray, Loren,
this is from my 993 obd1 RoW. I had to use a pulse detector rather than a scope to get any reading at first as I couldn't carry mains with me and didn't have a battery powered storage scope. Then I recorded the signal into my laptop to see what the transitions looked like . Then I built the counter cct with a reset some 4s after the last knock detection just to save me having to clear the cct all the time. Its basic but it does the job of telling me instantaneously that there is knock detected. Yep understand the diagnosis explanation but don't know the process to retrieve the numbers from the ecu via the K/L lines.
For me any knock is bad as I am under boost at the time so I tend to use the pin 11 output as a way of controlling my right foot . I may go the next step and relieve pressure at a preset detected level but got to work that bit out yet.
Cheers, Mark.
this is from my 993 obd1 RoW. I had to use a pulse detector rather than a scope to get any reading at first as I couldn't carry mains with me and didn't have a battery powered storage scope. Then I recorded the signal into my laptop to see what the transitions looked like . Then I built the counter cct with a reset some 4s after the last knock detection just to save me having to clear the cct all the time. Its basic but it does the job of telling me instantaneously that there is knock detected. Yep understand the diagnosis explanation but don't know the process to retrieve the numbers from the ecu via the K/L lines.
For me any knock is bad as I am under boost at the time so I tend to use the pin 11 output as a way of controlling my right foot . I may go the next step and relieve pressure at a preset detected level but got to work that bit out yet.
Cheers, Mark.
Last edited by E6MPB; 09-23-2006 at 04:03 AM.
#186
Actually, what Mark has done is the basis for a system John Speake (developer of the Spanner,
a 928 diagnostic tool) developed to optimize the EZK ignition maps on the 928. The 928 EZK
outputs the knock data on pin 31. Check out the 928 thread on John's unit.
Here's the link: https://rennlist.com/forums/928-forum/300589-new-ez-k-tuner-software.html
It indicates the analytics really needed for a robust evaluation of the knock data.
Good thinking for both Mark & John!
So without some external circuitry (use of I.C.s) , as Mark has done, using a PC
to sample the knocks will not be possible given the data rate of the knocks
and the PC's serial/USB ports. The knock data are output from P1.6 on the
main u-p thru two inverters (I.C & discrete).
a 928 diagnostic tool) developed to optimize the EZK ignition maps on the 928. The 928 EZK
outputs the knock data on pin 31. Check out the 928 thread on John's unit.
Here's the link: https://rennlist.com/forums/928-forum/300589-new-ez-k-tuner-software.html
It indicates the analytics really needed for a robust evaluation of the knock data.
Good thinking for both Mark & John!
So without some external circuitry (use of I.C.s) , as Mark has done, using a PC
to sample the knocks will not be possible given the data rate of the knocks
and the PC's serial/USB ports. The knock data are output from P1.6 on the
main u-p thru two inverters (I.C & discrete).
Last edited by Lorenfb; 09-23-2006 at 04:41 PM.
#187
Board done... software does not connect.
Sigh.
I get keystrokes bounced back in hyperterminal, I see I have power on various items, but the software does not work. I will triple-check everything tomorrow and see if I did something stupid...
Runnng the software on an old Win2k box. It should work, should it not?
I get keystrokes bounced back in hyperterminal, I see I have power on various items, but the software does not work. I will triple-check everything tomorrow and see if I did something stupid...
Runnng the software on an old Win2k box. It should work, should it not?
#189
Partial progress... any ideas?
Well, I tried mirror-imaging the pinouts on the DB9 connector and I can now access Climate Control, ABS, and Alarm. I still cannot access Motronic or the Air Bag function.
2 questions:
(1) Does the fact I have the DB9 pinout correct for the 3 functions I have, mean that the DB9 pinout is correct for all functions?
(2) Is there something specifically I should be looking at that is common to the 2 functions I do not have working yet?
Waiting with bated breath... and slight asphyxiation...
thanks
Thaddeus
2 questions:
(1) Does the fact I have the DB9 pinout correct for the 3 functions I have, mean that the DB9 pinout is correct for all functions?
(2) Is there something specifically I should be looking at that is common to the 2 functions I do not have working yet?
Waiting with bated breath... and slight asphyxiation...
thanks
Thaddeus
#190
Originally Posted by Thaddeus
Well, I tried mirror-imaging the pinouts on the DB9 connector and I can now access Climate Control, ABS, and Alarm. I still cannot access Motronic or the Air Bag function.
2 questions:
(1) Does the fact I have the DB9 pinout correct for the 3 functions I have, mean that the DB9 pinout is correct for all functions?
(2) Is there something specifically I should be looking at that is common to the 2 functions I do not have working yet?
Waiting with bated breath... and slight asphyxiation...
thanks
Thaddeus
2 questions:
(1) Does the fact I have the DB9 pinout correct for the 3 functions I have, mean that the DB9 pinout is correct for all functions?
(2) Is there something specifically I should be looking at that is common to the 2 functions I do not have working yet?
Waiting with bated breath... and slight asphyxiation...
thanks
Thaddeus
For DME switch to DME position. For anything else put in other position. The SRS switch works the same way.
Ray
#191
Originally Posted by raycm
Did you install the 2 switches? One switch Has DME position. The other has SRS position.
For DME switch to DME position. For anything else put in other position. The SRS switch works the same way.
Ray
For DME switch to DME position. For anything else put in other position. The SRS switch works the same way.
Ray
The only thing I did not install were the LED's, I could not find the right size. I jumpered all the LED holes with wire. Could that be causing the issue?
#193
Thaddeus,
I haven't personally had this problem, but the 964 guys have. They have checked that their car voltage is good (might want to run the engine while connecting so you get ~14V if you aren't doing that already), and that their serial ports are getting enough juice (notebook plugged in instead of running on the battery).
Do you have continuity between the pins on your OBD plug and the traces going to the switch? I suspect a bad solder joint.
I haven't personally had this problem, but the 964 guys have. They have checked that their car voltage is good (might want to run the engine while connecting so you get ~14V if you aren't doing that already), and that their serial ports are getting enough juice (notebook plugged in instead of running on the battery).
Do you have continuity between the pins on your OBD plug and the traces going to the switch? I suspect a bad solder joint.
#194
I reheated the solder joints, and checked for contnuity. I run the car (cough cough) and I'm using an old desktop PC. No luck.
Seems like the wires that carry the signal for Motronic and airbag are 7 and 3, is that correct? Maybe I have a bad switch, I will test it with a jumper instead. Fortunately I have switches that have a neutral (both poles off) position so I can do this without too much hassle.
sigh...
You might have to put me down for a completed board... this may be beyond my ability (cough cough) to troubleshoot.
Seems like the wires that carry the signal for Motronic and airbag are 7 and 3, is that correct? Maybe I have a bad switch, I will test it with a jumper instead. Fortunately I have switches that have a neutral (both poles off) position so I can do this without too much hassle.
sigh...
You might have to put me down for a completed board... this may be beyond my ability (cough cough) to troubleshoot.
#195
Thaddeus,
First you said hperterminal echo'ed but can't talk to car. Then you found the RS232 was backwards. I'd say hyperterminal could not echo with that connector wired backwards.
Now you say with both switches left side CCU, etc work. Actually one switch left and one right is the positions to get your results.
I can answer one question for you. Pin 7 is only Motronic. Pin 3 is all others. Pin 15 is wakeup for all but SRS.
These suggestions are wild *** guesses.
Make sure diodes are installed correctly. If you have alarm with RF transmitter in key read the earlier posts about this issue.
Ray
First you said hperterminal echo'ed but can't talk to car. Then you found the RS232 was backwards. I'd say hyperterminal could not echo with that connector wired backwards.
Now you say with both switches left side CCU, etc work. Actually one switch left and one right is the positions to get your results.
I can answer one question for you. Pin 7 is only Motronic. Pin 3 is all others. Pin 15 is wakeup for all but SRS.
These suggestions are wild *** guesses.
Make sure diodes are installed correctly. If you have alarm with RF transmitter in key read the earlier posts about this issue.
Ray