Tomas L

Can you please give me an explanation why a cam position sensor is required to retard an individual knocking cylinder?

RobNL

TT, now you're confusing me. First you tell that in order to tell exactly which individual cylinder is knocking and needs to be retarded, you need a distributor/cam sensor:
and now that Lorenfb says that in order to know which individual cyclinder need to be retarded, it needs a cam sensor, you're telling we don't need one, because we don't have one

So do you mean that our DME/KLR is able to point out which individual cylinder (1, 2, 3 or 4) needs to be retarded as a result of knocking?

Based on your statement:
I would say no. The DME/KLR only may know the pair it is retarding, but is it also capable to decide which pair needs to be retarded or does it just retard the next ignition signal to the coil when it detects knock and "listens" to decide what to do next?

mark944turbo

"And you get it wrong. Just the knock sensor input and KLR pin 24 to know when to listen for knock."
I am not going to fight, since you know the code and I dont, but I still find this very hard to believe. Looking at the unfiltered output of my knock sensor using my computer's soundcard, it is very predictable and stable at idle with the strongest signal at one frequency. As rpm increases, the signal basically goes crazy. By 6k it doesnt look anything like at 2000rpm, so it would just be stupid for motronic to treat the two situations the same way.

951 has no cam sensor, I figured it could tell which cylinder it was firing by counting teeth on the flywheel and using the ref signal to tell where it was in the firing order before it fired each spark. I will scan and post my source of information on Monday when I am back at school. It must be wrong, but it is from a technical presentation by porsche engineers to PCNA back in 1985. Maybe porsche changed its mind on this subject?

TT

Here I was talking about the 968. I don't know why it has a cam sensor since the 968 ignition still uses a distributor and not a coil pack. The main reason other DME's use a cam sensor is to know which cylinder is firing and gate the ignition signal to the proper coil in the pack.

I'm not saying that the 951 doesn't need one, the 951 doesn't have one.

Actually the KLR doesn't even have to be this smart. With one knock sensor, it boils down to just knowing the cylinder count. Be told to listen for knock, indicate/save the status, increment to the next variable set for the next listen event, listen for knock, indicate/save, ... Reset index after four events for the 951. It would then know the relative cylinder and its knock status. Retard the ignition for that cylinder based on the state variables for that cylinder. Thanks guys.

The signals the KLR looks at for knock detection are highly filtered and gated near the firing event so it can treat the situation in the same way. It does get noiser with the increase in RPM, but knock comes through like a bell. What you see at the actual sensor is too raw.

Lorenfb

"Funny, I don't see a PN in PET for the cam position sensor in my 951 and I don't remember seeing it when changing out the cam." - TT -

He has to learn to read. There's no implication in my statement that the 951 has
a cam position sensor.

I said: "This requires info from a cam position sensor (Hall sensor)."

"Some more guessing again! Do some reading, like the 928/964 manuals which provide
info about the Bosch knock system. They all function basically the same." - Lorenfb -

Again, learn to read! I said: "They all function basically the same." BASICALLY
So read the 928 & 964 manuals.

"I told you that you are not in Kansas anymore, Toto." - special tool -

No hope here!




*
TT
View Public Profile
Send a private message to TT
Send email to TT
Find More Posts by TT
Add TT to Your Buddy List

fast951

Hey Lorenfb, it seems to me that you are fishing for information much more than providing it. How accurate am I? Don't get all huffy, no harm intended!!!

Lorenfb

"I would say no. The DME/KLR only may know the pair it is retarding, but is it also capable to decide which pair needs to be retarded or does it just retard the next ignition signal to the coil when it detects knock and "listens" to decide what to do next?" - robnl -

He's correct. You only know which pair of cylinders. A system can use a cam position
sensor to selectively retard each cylinder. The others are also correct in that you
can store timing position counts for selective retarding. The key benefit of the cam
position sensor is for sequential injection (968), diagnostics, and a possiblity a simpler
and more accurate knock system. Also, if multiple knock sensors are used it allows
you to selectively "listen" to individual sensors based on what next cylinder will fire.

The key point about the 951 system is that there're 2 SEPARATE control units.
The DME is the only one which has the "count position" and not the KLR.
The KLR only "knows" a knock and retards the NEXT cylinder, and not which cylinder.
The DME doesn't know about the knock. Yes, the KLR can count "four" to the
next time that cylinder needs to be retarded.

The 928/964 can "remember" the knocking cylinder and continue to retard in
3 degree increments to a max of 9 degrees. If there's no more knock on that cylinder
the DME begins advancing in 3 degree increments to the normal timing.

TT

Reading the above, you state that individual cylinder retard would require a cam position sensor. Apparently not. Your citation states it does do individual cylinder retard and so does the 944 Turbo Service manual. I hope Bosch didn't lie to Porsche.

The scenario I mention would allow the KLR to perform individual cylinder retard. It would only require the KLR to know which cylinder out of a sequence of 4 ignition events had knock. Keep the information synchronized using the ignition pulses and the KLR can retard individual cylinders. It won't know exactly which cylinder (1, 2, 3, or 4) but it won't need to. The timing for the cylinder which has knock will come four ignition events later.

The 968 cam position sensor never made sense to me with the normal distributor setup, sequential injection would explain it.

hosrom_951

Then why does the 951 have two reference sensors? one for TDC and one for Crankshaft correct? the DME uses the TDC for timing, what about the crank sensor? what is it used for?

Lorenfb

"the DME uses the TDC for timing, what about the crank sensor? what is it used for?"

By crank sensor, you probably mean the diagnostic one off of the dampener pulley,
not used by either DME or KLR. The DME still requires both a speed and reference
signal off of the flywheel.

Tomas L

The KLR can count to four but you seem unable to att one and one together and get two??
That does not only contradict Porsches information handbook but you seem to contradict yourself here!?