Performance Chips & Knock Sensors
 

NO CHIP WARS please......
<br>
Can a aftermarket and/or performance chip disregard the input from knock sensors and still allow the motor to run?
<br>
I recently test drove a turbo body with Vram engine transplant.
The exhaust was really quiet and I could hear the car knocking under load from 4.5K up to 6K. I was driving next to a concrete wall. I couldn't believe the timing wasn't retarted by the DME to compensate. If the exhaust would have been any louder, you would not have been able to hear the knocking at all.
<br>
The owner of the car denied being able to hear anything, I know what I heard.
<br>
The car had a Cyntex chip.

Todd

Maybe the car had no knock sensors at all so the ECM had no inputs. Might or might not give a CEL for that.
Microphone type knock sensors aren't worth much anyway. They can detect and cause retard on sounds other than knock and miss on real knock.
Current high tech EFI engine management (Delphi etc.) uses ion sensing across the plug gap to detect knock in the cavity itself. In other words the spark plug and wire acts as the knock sensor itself. It is highly reliable and cannot be disabled other than by running wires with the incorrect resistance.

It had the sensors, I saw the leads coming out of the shroud. There was no CEL (as in most 3.6 transplants).
<br>
I don't know how effective the Porsche 993 knock sensors are??
I just know that knocking / pinging is NOT a good thing with these engines.

The inputs from the knock sensors can ALWAYS be disabled. They can be ignored
just like any other DME input. It's all a result of how the firmware (EPROM data)
is modified. Typically ignoring the knock sensors is how the performance chips
can advance the timing without the DME overriding the new advance curve of the
chip. Advancing the timing is really what provides the potential for a performance
increase.

I was hoping that you would chime in here. I suspected this was the case. Isn't disregarding the knock sensor inputs a potentially dangerous situation, especially for a 993 engine?
<br>
One problem is that many of the exhaust systems (especially on the 3.6 transplants) are so loud that you can't hear the knocking / pinging.
<br>
It was an 84 degree day when I test drove the turbo bodied varioram car. It was pinging like wild on 93 octane. The owner denied being able to hear it...

Todd

I can't believe any reputable chip designer would intentionally ignore/disable the inputs from the knock sensors.

I guess you could ask Cyntex.

Those inputs are very simple, you get a knock read you drop one to three degrees from the advance map, you get it again you drop again until it goes away.


edit:

the problem with the microphone type is the ECM may get bad reads on other noises or no reads on knock when it should.
That's why new technology in ion sensing is vastly superior as you are detecting in the cavity itself with the plug gap resistance and you can then even adjust advance by cylinder.

For you software weenies, here is a screen shot from my Delphi Automotive software for my 1.8 liter EFI ECM.

Notice I have marked the "disable knock sensors".

As I said before however, no "reputable" software designer would do this and give the EPROM to a customer. But some software does allow the disabling of knock control as a tuning tool.

I know why but I'm not telling. :)

http://forums.pelicanparts.com/uploa...1084979905.jpg

"I was hoping that you would chime in here. I suspected this was the case. Isn't disregarding the knock sensor inputs a potentially dangerous situation, especially for a 993 engine?

One problem is that many of the exhaust systems (especially on the 3.6 transplants) are so loud that you can't hear the knocking / pinging.

It was an 84 degree day when I test drove the turbo bodied varioram car. It was pinging like wild on 93 octane. The owner denied being able to hear it...

Todd"

You're right on Todd. I couldn't have expressed it better! It's the combination
of heat, the extra timing advance, and the lack of the knock sensors. That motor
will probably need rings and pistons before long.

"I can't believe any reputable chip designer would intentionally ignore/disable the inputs from the knock sensors. "

Please! Get real. Those guys do a lot of things, many of which they are
unaware of that cause other problems.

You talkin bout me man??? Those is fitin words!! :)

"As I said before however, no "reputable" software designer would do this and give the EPROM to a customer. "

People do all kinds of things to sell products. What planet do you come from?

Look two screens up Loren. If I disabled the knock control in the program and sent you a new calibration or eprom to load and you blew up your motor with detonation, I don't think I'd do much work again.

But as I said the key word is "reputable".

Gotta go work on a guys program now. A Calli guy!!! :)

Iread someplace that Saab is or has stopped using their Trionic (p?) knock detection. Too bad, it was a neat system.
-Chris

Loren,

Can you start the engine and disconnect either or both of the knock sensors and notice any change?

How could you detect that they are/are not being utilized by the DME?

As always, thanks for your input.


Todd

I don't want to start another Loren / Steve wong exchange, but...
1.) certainly anyone who does chips may do so unscrupulously and ignore the knock function.
2.) SteveW is rather unique in that he can /does design a form of "poor-man's " knock sensor for the 3.2's ( which never came with knock control), by designing a basic ignition advance / fuel profile, then captures intake air temp readings and uses that to retard advance in a sort-of open loop fashion for hot running conditions. Seeing that he's not overly aggressive with advance timing in the first place shows this to his credit for conservatism.
---Wil

Yes, I don't want to either Wil. I am looking at this from the 993 varioram engine perspective. This engine came with and needs the knock sensors to properly and safely run.

On the Vram, the hall effect sensor basically reports the ignition advance to the DME and depending on knock sensor input retards or advances the timing to optimum. What happens with the Cyntex chip is the timing is advanced without regard to the engines knock conditions.

Todd Embs

http://forums.pelicanparts.com/uploa...1085002555.jpg

That is a question for Cyntex.

If they are "piggybacking" on the stored DME ignition map then they could be adding advance when the knock sensor is saying to retard. (editorial: piggybacking is bad design by the way).

If they have replaced or in fact updated the stock DME ignition map then the knock sensors can serve their intended purpose and retard on the replaced or updated ignition map.

"2.) SteveW is rather unique in that he can /does design a form of "poor-man's " knock sensor for the 3.2's ( which never came with knock control), by designing a basic ignition advance / fuel profile, then captures intake air temp readings and uses that to retard advance in a sort-of open loop fashion for hot running conditions. Seeing that he's not overly aggressive with advance timing in the first place shows this to his credit for conservatism."

Someone swallowed it "hook, line & sinker". Never heard so much B.S. It just
keeps on coming! It's laughable. The naive just suck it up!

To check out the knock sensors on the 993, you really need the Porsche
Hammer or ST2 tester.

Tooo funny Loren! I couldn't understand that either!!!! :) An ignition map using RPM and IAT as the X and Y axis? If you got intake temp why not put in a MAP sensor and do it right? Oh well. Too much for tired minds me thinks!

Ignition maps use RPM and the artifical load signal of the AFM to lookup the base ignition value. There is no reason that these base values cannot be trimmed by other factors such as the engine temperature, intake air temp, rate of rpm change, and in this thread, knock sensors. Think about it.

Perhaps some of you should stick to what you do instead of making statements about which you do not know.

Could you give some specific examples of chips that do that i.e ignore the knock sensor? Do you know of any that don't?
-Chris

I said the knock sensors could be disabled. I've seen a lot of 964s & 993s
where pinging occurred. I'd have to assume the knock sensors were disabled
since the pinging stopped when the stock chip was re-installed. As you know,
the 964/993 DME retards the timing in 3 degree increments to a max of 9
on EACH cylinder individually. It also slowly advances back to the original
timing setting.

Obviously, without dis-assembly of the hex code of the EPROM, one could
never really know. Something has had to happen to the knock sensor data,
since pinging occurred in the 964/993 cars I've been involved with.

I was hoping you had a list of bad apples but that info is still helpful.
-Chris

If I may add,....

I cannot say for certainty that the chip tuners I know disables knock-sensing on purpose. Without hard data, that would simply be conjecture.

I will say that some chip tuners are too aggressive with their timing maps and these cars will detonate under hot condtions and high loads. Further, some chip tuners do not test and development their products with the locally available octanes of the customer's.

One thing about the 3.6's,....if the internal distributor drive belt is broken, it can ping like hell no matter what gas is in it. Its caused by the lower rotor coming to rest pointing at a set of contacts and igniting that cylinder at the wrong time.

There are many reasons that these engines can detonate and the PST-2 or Hammer will display any unusual knock sensor activity.

Hahahaha... what a clown!!!

Show me the AFM software you use and I'll show you MAP.

Heck I'll show you mine anyway ..... but you can't have it.

Nanananaa!!!! :)

Or are just you one of these "theoretical guys"??

I already showed you how I can disable knock sensors. :) Now I'm showing you basic ignition maps as well.

LOL!!!


http://forums.pelicanparts.com/uploa...1085062765.jpg

Since this thread is supposed to be about knock sensors and how they may or may not work, here is a nice picture showing how they work.

Using vehicle ECM data logging.

I have the cursor (verticle black line) at data point 1774 and I am recording knock retard on the front and rear banks of between 4.5 and 8 degrees. See data section and graphics.

The total advaance thus has been pulled back to 26 and 20 degrees vs the master ignition map at that specific point.

The TP (throttle position sensor) is 100, the MAP is 98 and the RPM is at 5012 for this specific data read.

The EPROM calibration for this vehicle is factory.

AND it needs tuning. Badly!!!


http://forums.pelicanparts.com/uploa...1085063906.jpg

Now the tuner (me on this one) goes to work on the EPROM tables and works to eliminate that power robbing knock retard via adding fuel at the appropriate points and or retarding portions in the ignition map.

Here is a data log of the exact same vehicle after several iterations of tuning focusing on fuel. The ignition map did not require edit and the knock retard was fully eliminated in this case.

Note RPM is 5045, TP% is 100 and MAP is 98. Essentially the same data run. the cursor is at data point 380

Let me tell you that vehicle owner could FEEL the difference when his ignition was no longer getting pulled 8 degrees!



http://forums.pelicanparts.com/uploa...1085064332.jpg


Now no "reputable" software tuner is going to disable the knock sensor retard capability of good software without the owners consent. "Reputable" that is!!!

But what do I know,,, right???? :)

edit: spark advance is now the 29 degrees specified in the ignition advance table cell MAP=100, RPM = 5000 with no pull back.

tsuter,
Did you reverse engineer the ECM to get the info your software is displaying (like Steve did) or are you just a "theoretical guy" when it comes to ECM algorithms? ;)
-Chris

Chris, mine is displaying the actual ECM EPROM tables in the software windows and I edit them first off line after polling data then directly update (reflash) the EPROM tables with the new tables via the OBD port.

It ain't theoretical.! :)

edit: the factory ECM tables in the EPROM are completely overwrittten and unretrievable so best is to socket a new chip and keep the factory chip.

Ain't this an ugly mess, thanks to tsuter and loren, some honest facts are being discussed. You bas-turds should be banned from the BBS, no more truth telling or I am going to report you. I demand only lies from this point forward.
<br>
Back on topic, I was worried that on especially the 993 engine during a hot day, chips that disregard the adaptive function of knock sensing may be causing major damage to the engine over time. It appears this is true. So you 3.6'ers better tune in.

Re: Performance Chips & Knock Sensors
 

Yeah Todd you are right. I guess the simple direct answer to your question was... yes they can and yes the motor will still run.

But who likes simple??? :)

I don't question your table values. I was just teasing you for infering that Steve's only a "theoretical guy" when he actually got his hands dirty and did some reverse engineering to acquire his knowledge of ECU internals and algorithms.

A "clown" would be someone who bought tuning software that doesn't even work on a 911 trying to school Steve on ECU algorithms. Of course that wouldn't apply to you. Of course.

-Chris

Chris, I'm not the one who said this:

"Perhaps some of you should stick to what you do instead of making statements about which you do not know."

And I'm still waiting for a real ignition table with other than RPM and MAP axis. One with about four variables and especially temperature. Hahaha!!

I think I'll be waiting a long long time!!!


edit: but maybe he meant Loren.... who knows..... who cares. :)

Young Padawan: you have much to learn. These are not GM motors. Porsche Motronics use either an air flow meter or a mass air flow sensor to determine engine load, not MAP sensors. The software we use is proprietary and custom written by us, there's no reason to try and display all it's functions, as it's not for sale. Here are screenshots from one of the programs we use displaying the base maps of a stock 89 Carrera. As you can see, with the x-axis representing the rpm points, the y-axis being the load from the AFM or MAF sensor, the Y-axis represents the fuel or ignition value. Using these base values, they can be trimmed and adjusted via the warm up tables, head temp sensor, intake air temp, rate of rpm change, knock sensors etc. Trimming these base values allows adjustment for such factors as high ambient temps to maintain a predesignated knock margin - especially for the 3.2 motors which do not come with knock sensors.

http://forums.pelicanparts.com/uploa...1085070824.gif

http://forums.pelicanparts.com/uploa...1085070861.gif

I will hook up a scope to the knock-detect signal of the diagnostics plug on my 3.6 soon. This is not the input to the DME but an output that "talks" to the Hammer. I don't think one needs a Hammer to display the signal. I'll keep you posted.

A chip for a 3.6 is essentially executable code for a 8051 microcontroller plus a data section. All that is stored in a non-volatile memory chip called EPROM (Erasable Programable Read Only Memmory) The computer (DME) has inputs (crank sensor, head temp, CSP, AFM, region plugs, oxygen sensor, etc) as well as outputs (FI, Ignition, etc). The program that runs processes these inputs and does numerous calculation with stored values in the data section of the ROM to generate outputs. Some of these data sections are aranged as tables. This is what people talk about when they talk about maps. Fuel maps and ignition maps for part throttle work like a lookup table. Based on two inputs (RPM, load) you get a numerical result. Many forget that this result is only one factor that goes into the equation of producing the final electrical output to the engine.

For the 3.6 there are different chips out there with different levels of modifications/approaches:

1. Stock code with modified maps for part throttle and full throttle. If these mods are not too agressive you can consider this a "safe" chip. Ususally one map is modified and all pointers to the different map point to this one (disabling the region plugs). Be aware of the gas quality the chip asks for.

2. Stock code with modified maps and modified base values for RPM limit, fuel cutoff, and other goodies. Caution, here you start to get into uncharted teritory. The rev limit had a reason and if you raise it the added stress will result in $$$$ if you haven't prepped the engine properly (rod bolts, valve train)

3. Completely different executable code. This is no mans land. You surrender your engine to whoever thought out this code. I would stay away from it for two reasons. This chip could have been made by a dumpass or it was made for a monster engine with all the goodies that comes appart after every race weekend. Go figure what that can do to your stock 3.6.

Ingo

Great points Ingo, few really understand what they're putting into their cars
and how it may be affected. I've seen #3 on a 993TT where the car wouldn't
pass emissions because the OBDII diagnostics & readiness states were gone.
This was a '97 where the memory had been flashed and no original code
was available. The customer had to buy a new DME ($1200 - $1500).

And YOU are gonna teach me about the merits of MAF over MAP??? That's OT for this thread sonny boy. :)

But I'm still waiting for more than two variables you claimed in that ignition table. Come on now show us the ignition trimming you are claiming for head temp, IAT, RPM delta...... shadow tables.... please???? In addition to knock sensors I guess????

No cold start idle advance or warm up enrichment tables now... we are talking about multi variable ignition trimming.

It is nice theory anyway!!! LOL


Oh and nice 3D by the way too!!!

http://forums.pelicanparts.com/uploa...1085078501.jpg

I'm surprised I haven't heard this one (like the air temp):

Our chips optimize the timing based on the octane being used. We are able to
determine the octane level by use of a proprietary algorithm. We basically
calculate the piston's acceleration, which is proportional to the octane level,
by taking the derivative of the speed sensor signal. Knowing that value, we
are able to select the appropriate timing map for the octane being used.
As a result, our chips exhibit no pinging for a wide range of octane levels.

Sounds good to me. Whatever it takes to sell a product.

I run GIAC chip software (Garrett) on my race VW1.8T with their IBE Chipswitch software. Utilizes a key fob to switch from pump(91-93) to race(104), valet, to kill programs... up to five programs. Unfortunately they don't yet offer direct port programming like the Delphi Automotive software pictured above but they will eventually as DPP is what the market wants.

GIAC's IBE chipswitch is not currently available for all Porsche applications - only VW and Audi.

APR V-Tune is direct port programming but also only currently for VW/Audi.

As far as running Delphi on a 911 there is a thread on CIS to EFI converisons and if I ever converted mine (which I won't), I would only convert to some form of DPP software.


edit: For Steve W. Yes my VW uses a Bosch MAF sensor not a MAP and yes I know the difference!!! LOL.

If I may, I would like to add some $0.02 on knock sensing here:

Knock sensors are basically microphones. Some are (mechanically in manufacturing) tuned to specific sound frequencies that are created when an engine knocks. Others are broad band and the knock sensor interface in the ECU detects only those frequencies experimentally determined as knock signal for a specific engine.
The knock frequency is determined by:

f = M*c/pi*B

f is the typical frequency for the knock signal.
M is a 'Mode constant' for the different resonance modes in the cylinder. Depending when in the stroke knock occurs the Mode constant might change.
c is the speed of sound in the cylinder.
B is the AVERAGE bore size (taking into account the shape of the combustion chamber).

The speed of sound is calculated by
c = square root of (k*R*T)

where k and R are constants (1.4 and 247 respectively).
T is the average gas temperature in the cylinder at knock time in degree Kelvins.

So if the engine (or fuel) is modified by changine the combustion chamber shape, fuel type, deck height, fueling and so on, the temperature or knock position changes and knock (even if there) might not be detectable by the ECU, simply because the characteristic knock frequency for the engine changed. A lot of ricer kids blow up their engines with Nitrous because of knock. Their ECU never retarded because it was tuned in to the wrong knock frequency (gas temp difference).

Knock sensors are a very limited method of detecting knock and work well enough only for an unmodified or mildly modified engine. But all bets are off for a seriously build engine.