Vauxhall Owners Forum banner
1 - 20 of 26 Posts

·
Registered
Joined
·
30,330 Posts
If the knock control has recognised knocking combustion for any one cylinder, the CPU will adjust the dwell angle by 3° in a retard direction for the next phase. The dwell angle of the other cylinders is not affected by this measure (cylinder?selective control). The dwell angle adjustment in a 'retard' direction is repeated for every combustion which is recognised as knocking and for each cylinder selectively (individually). If no more detonation is sensed, the ignition is adjusted by 0.75° in direction 'advance' after 20 to 120 knockfree combustions (approx. 2 seconds). This is repeated until the base map value is reached again or until knocking combustion is registered. The knock control only affects the dwell angle in an engine speed dependent load range in which knocking combustion is to be expected.
 

·
Registered
Joined
·
1,561 Posts
Or a simpler way is its a microphone which outputs a voltage when it detects knock.

The quote Oddball has selected about individual cylinder trim is crap, pure and simple.

If knock is detected then the timings pulled back on all cylinders.

Injection systems are not really that sophisticated.

Chris
 

·
Registered
Joined
·
9,212 Posts
Hi Chris,

Sorry, but the quote Oddball has used is straight out of the motronic operation manual. I have the PDF somewhere.

It is capable of doing cylinder based knock, it simply samples the knock sensor during a window of time around ignition to check for knock on that cylinder.

The rest of the operation is as per Oddballs quote. Only for the Bosch motronic system though, don't know about other ECUs.

I'm trying to find the doument, but coming up blank at the mo. I know it's here somewhere.....
 

·
Registered
Joined
·
3,963 Posts
Chris_H42 said:
Or a simpler way is its a microphone which outputs a voltage when it detects knock.

The quote Oddball has selected about individual cylinder trim is crap, pure and simple.

If knock is detected then the timings pulled back on all cylinders.

Injection systems are not really that sophisticated.
Thats very simple! It's outputs a voltage of the noise of the engine block that the sensor is tuned to. That may be knock or may be another noise. The signal goes through a bunch of filters and signal processors, the ECU only 'listens' when a cylinder is in the combustion phase. Then the current signal is evaluated against a previous averaged signal of what the ECU thinks is no knock.

Any yes cylinder specific knock control exists. It's perfectly simple to do with a SFI system as you already have information on cam phase to evaluate with the knock signal.
 

·
Registered
Joined
·
1,561 Posts
Hmm I have to say I seriously doubt that.

its like injection, the SFI when the cam sensor plays up resorts to batch (or gang if you prefer) fire and it resorts to batch fire at higher revs as well. The time period is to short.

Also the ECU's for individual cylinder trim and so on are very bloody expensive. It takes a lot of sensor input and so on to allow individual cyl trim and a very impressive ECU unit.

If the cars ignition is coil packs then chances are its wasted spark so at least 2 cyls will be retarded.

I might be wrong here but if I am then I will be very surprised indeed.

Chris
 

·
Registered
Joined
·
3,963 Posts
The amount of money Bosch and manufacturers spend on ECU developement puts aftermarket ECU manufacturers to shame. To mass produce stuff like that is not going to add cost as you would think if you know how many units your talking about the amount of units Bosch shifts.

Motronic 2.7 manual. Just uploaded it, 1.7MB
 

·
Registered
Joined
·
1,561 Posts
I will read the manual (downloading it just now) but there simply isn't enough sensors to do individual cyl trim. It also needs a lot of dyno time to do it.

One way to look at it is if the stock ecu is so good why is there a market for upgraded ones?

If the ecu is so good then the engine iwll be makign max power all the time while remaining emmisions friendly and returning good fuel consumption.

Have a gander at one of the bosch injection system books, lot of info and it may seem like heavy going to some but it explains things nicely.

Chris
 

·
Registered
Joined
·
9,212 Posts
I have a couple of Bosch motronic books waiting to be read, but lots of material before that.

The system is easily explained in the bosch manual you are downloading. It doesn't require that many sensors, and the processing is done via a seperate processor to the main one, so it takes up no processing power.

The ECU is very powerful, much more so than most of the standalone ones that come as aftermarket replacements. Most aftermarket ECUs ditch the knock sensor altogether!! nice!

Most won't do much with an oxygen sensor either! Most people move to "better" ECUs because they can control the Ign and fueling of the engine much easier. If it were possible to have all the functions of the original ECU, and retain all the sensors, and safety functions at the same price, everybody would be choosing that system.

:beer:
 

·
Registered
Joined
·
15,617 Posts
I would like a job with engine ecu's :(

a lot to learn though.




VW VR6 Motronic 2.9

Two knock sensors are used. A knock sensor works like a microphone to
"listen" for spark knock or detonation.

When knocking occurs, the ignition timing is retarded until the knocking
is eliminated. Since the knock limit differs from cylinder to cylinder
and changes within the operating range, knock regulation is done
cylinder
selectively.

* Signal usage: *

Knock regulation does not occur until the engine coolant temperature of
40° C (104° F) is reached. Knock sensor I (G61) monitors cylinders 1,
2,
and 3. Knock sensor II (G66) monitors cylinder s 4, 5 and 6.

With the aid of the Hall sender signal, the ECU can determine which
cylinder is knocking. The ignition angle of the knocking cylinder is
retarded in steps until the knocking stops up to a maximum of 12°.

If spark knock is still detected, the ECU will retard the ignition
timing 11° for all cylinders and record a fault.

* Substitute function: *

If a knock sensor fails, the ignition timing angle of its assigned
cylinders is retarded.

* Self-diagnosis: *

The ECU recognized an open circuit if no signal from knock sensor I
(G61)
or knock sensor II (G66) is received by the ECU at an engine coolant
temperature above 40° C (104° F).
 

·
Registered
Joined
·
3,963 Posts
The spark knock period is only from around TDC to 60-90 degrees later on the cylinder that is on the power stroke. That allows for one sensor to be used for all cylinders as no two cylinders can knock at the same time.

The electronics, as brian says, is done by a purpose designed processor. Thinking in terms of computing power is not really the right way. Purpose designed processors (like graphics card processors) can do things much easier than a general purpose processor doing the same task.

As for why it doesn't allow max power the answer lies in the measuring system. Knock sensors and most of the electronics are crap basically. First off most systems usually only look at the first harmonic (found on a per engine basis through many thousands of pounds worth of research). This as you know will have other noise that present in it. The system will work well at lowish revs (below maybe 4000rpm) and tail off afterwards. And this will only work on a standard engine, you chage something major, you change the harmonic frequency of knock!

More modern systems use DSP chips to analyse the signal. This gives better processing power allowing you to analyse the higher order harmonics. This increases the accuracy and rev range of the knock control system. A good paper on this is Engine Knock Detection Using Spectral Analysis Techniques With a TMS320 DSP.

Knock sensor based systems can not really be done aftermarket as the amount of research needed to be done on the engine to fine what the knock 'sounds' like on it is immense.

Spark-Knock actually creates a pressure spike and oscilation in the cylinder. A knock sensor measures the noise of this. Even better system measure the actual cylinder pressure through the use of fiber optic pressure transducers or spark plug ion-current sensing. This is the most direct way you can measure it. Ion-current sensing systems are comming out on cars now. In future this would allow fully closed loop igniton timing!

The pressure measurement systems however can be implimented relatively easy in aftermarket. As the systems improve and costs fall they should filter down.

Anyone still awake? I really should be doing some real work!
lmao
 

·
Registered
Joined
·
1,561 Posts
I downloaded it eventually (modem speed impresive! lol) and had a quick look at it and I was right about it only doin git at a certian rev range. Damn I am good! lol

As for aftermarket ecu's, depends on where you go, full on race engines won't have that tripe like knock detection etc, they will be built properly and mapped properly so have no need for it.

I will read the whole thing in more detail when I can be bothered.
There are some very impressive ecu's about, ones which have speerate egt's and so on allowing individual cylinder trim of both injection and ignition, not cheap and not the quickest to set up either!

Chris
 

·
Registered
Joined
·
3,963 Posts
*Shrug*

The rev range problem is inherant to all knock sensor based control systems and not just cyclinder specific ones. This is not because of processing power.

We don't drive race cars. Our engines suffer from lot more variables that affect the way they run than a race engine. I personnally think these advanced knock systems are bloody impressive and can't wait for them to filter down into the mainstream.

Other things like Saabs variable compression engine and that are going to change the simple old IC engine quite substantially. Although we are still nowhere near the most efficient compression ratio of 16:1.

I wish I chose Automotive Engineering and not Computer Systems and Networking for my degree... :rolleyes:
 

·
Registered
Joined
·
9,212 Posts
There seem to be a lot of IT/Electronic engineers that seem to get a craving for engine performance at the moment. It seems to be because of the inclusion of electronic systems, all of a sudden things seem to make more sense when you have software, injectors, and other electronic sensors.

Carbs, and all that shebang never really did interest me much when I was young.

A friend of mine has just gone to work for a motorsport company, and he is a software engineer by trade! go figure...

...anyway. Is this rev range thing a real issue? I mean, I thought knock would be most prevelant around max torque, thus on most engines that would be fine in the <4000rpm range when we are detecting knock?
 

·
Registered
Joined
·
3,963 Posts
Yeah, the electronic systems are cool. I think I may be a geek... :D
That paper on the DSP has a bunch of programming, there is a big overlap of all things computers really.

Spot on with the torque and knock thing. But my point is mainly that these systems are far from flawless. They will not save an engine with problems, they will only allow you to have some slight adaptation in the ECU.
 

·
Registered
Joined
·
9,212 Posts
Spot on with the torque and knock thing. But my point is mainly that these systems are far from flawless. They will not save an engine with problems, they will only allow you to have some slight adaptation in the ECU.
Totally agree. The ECU only has a certain range that it can change it's settings by, after that, the engines on it's own!
 

·
Registered
Joined
·
9,212 Posts
0ddball said:
I think if TOMMCF ever returns to this post he will most likely commit suicide or spontaneously implode from inter-cranial pressure.
yep, sometimes these threads go off on a bit of a tangent... lol

why do we call wheel nuts, wheel nuts. When in fact, they are bolts!?
 
1 - 20 of 26 Posts
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top