LPG ignition 1

[andrebrito88]

Hello all,

I'm here to take some knowledge about the LPG and ignition timing.

My car got 2 ignition maps...one named RON91 and another named RON98, RON98 its more advanced than 91 (that is very safe), in normal conditions RON98 its used, and if a knock its detected will happen an ignition retard in the direction of the RON91, this process its called knock adaptations and i can check and see that its an RPM vs LOAD map and got the number of degrees of retard that its happenning in that especific RPM and LOAD so i can reduce the RON98 map in that zone.

My especific question is, can the knock sensors detect the knocks on LPG like in petrol? Or the lpg just dont knock? (Like the E85).
Is the knock the maximum torque power in lpg too?

Because i see another fuels like E85 that we can't use this method (next) because the knock doesn't happen or if it does its already too late for the engine, once the lpg got very high octane value too happen the same thing or we can trust the knock sensors/adaptations to tune the ignition on lpg?

Because normally on those cars if we are on petrol we advance the ignition and wait for the knocks to appear in the table and after that we reduce that amount in the ignition table and boom, optimized petrol ignition.

Thx for the answears in advance.

 

[GregLocock]

LPG does knock (autoignition of end gas). Sorry i don't know what the optimum advance is.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[Lou Scannon]

There are too many variables to give any pat answers online.
To go about this properly you need an instrumented engine on a test stand, with high speed cylinder pressure data acquisition, and the means to process and interpret the data. Failing that, at least an already calibrated accelerometer-based knock detection system, which it seems you have. But then you really need to be recording and processing the raw data, so you can compare knock signatures on different fuels and make sure the time window for knock detection is actually capturing knock on specialty fuels such as E85 and LPG.

"Schiefgehen wird, was schiefgehen kann" - das Murphygesetz

 

[GregLocock]

Wouldn't the knock signature be the same regardless of fuel, more or less? It is a function of the location of the sensor and the structural properties (resonances) of the head, and the cylinder size, isn't it?

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[andrebrito88]

''But then you really need to be recording and processing the raw data, so you can compare knock signatures on different fuels and make sure the time window for knock detection is actually capturing knock on specialty fuels such as E85 and LPG.''

The ECU will record all the data for the individual cilinders so i can check how much retard its needed for each, my ecu its coded for fuels from RON91 to RON98, RON91 its like the minimal ignition timing, and RON98 its maximum ignition timing, if it detect an knock it start to reduce the timing from RON98 to 91 and vice-versa if it doesnt detect any knocks, its kinda good system actually.

The real question its about the LPG fooling the knock sensors, i want to make sure that the knock sensors react to the LPG knock like the Petrol knock, or is it a diferent sound and it doesnt detect it?
LPG does knock when over-advanced like petrol?

 

[andrebrito88]

''Wouldn't the knock signature be the same regardless of fuel, more or less? It is a function of the location of the sensor and the structural properties (resonances) of the head, and the cylinder size, isn't it?''

Thats the real question, can the LPG knock like Petrol if over-advanced? And does the knock sensors sense it like a knock?

 

[Lou Scannon]

Dunno, if the LPG knock comes at a different crank angle, for instance, then the background (mechanical) noise might be a little different. Furthermore, as I alluded, a modern knock sensing system surely "looks" for knock in only a very narrow crank angle window; knock outside of this window will not be detected.

A premixed charge of any fuel knock under the right conditions.

I suggest you find and join an enthusiast's forum for your particular engine. There should be a lot of knowledge and experience in such a venue to draw from.

"Schiefgehen wird, was schiefgehen kann" - das Murphygesetz

 

[andrebrito88]

On the groups from my car/engine there are a few with LPG but not optimized in terms of ignition/afr and etc.

But yeah thats what i need to know, if the lpg does knock or not.

"Furthermore, as I alluded, a modern knock sensing system surely "looks" for knock in only a very narrow crank angle window; knock outside of this window will not be detected."

So we can agree that an knock outside that narrow range isn't mechanical dangerous, since the OEM tune doesn't look for it? What u think?
That range its tunable too in my ecu.


As i know too, enrichment of the mixture doesnt cancel the knock in vapor fuels but instead it gets worse (this is tested my me).

 

[EdStainless]

What others have alluded to is that the knock signal from different fuels can be different (timing, rate of rise, etc).
E85 (and 100% etoh) certainly does knock, and people use sensors to pick it up but they are filtering the signal differently than for petrol.
This is why it was suggested that you look at the raw signal so that you can see the earliest signs of detonation.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, consulting work welcomed

 

[andrebrito88]

Thx for the advices.

Lookin around at my ecu maps i see that my ecu looks for knocks from 5 ATDC to 75ATDC.

Do you think checking for knocks before the TDC too will be a good thing?

 

[andrebrito88]

This is what i love when share ideas,

Diferent opinions, and points of few give diferent ideas...i have never check when the ecu looks for knock until today.

Nice point of view

 

[EdStainless]

What does your system consider a knock? A certain number of peaks, at what size?
It may be that you need to be looking at more sensitive data for LPG.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, consulting work welcomed

 

[andrebrito88]

I know that the system amplifies the received signal in order to make a better filtration.

And i know the system converts an signal that is generated by the knock sensor(voltage) into a factor, everything after that defined factor its considered an knock.

For example this factor its much higher in the part-load area because of the more advanced ignition (MPG related).

In theory if it was a knock it generates an vibration, the knock sensor its piezoelectric.

 

[GregLocock]

" theory if it was a knock it generates an vibration, the knock sensor its piezoelectric."

No. Knock sensors are cheap, they only operate over a limited frequency range. When we develop the location for them we use proper accelerometers (that was my job), once the correct site is found an appropriate sensor is selected. Knock is detected by comparing the amplitude in a certain window with the average level.

The best knock detector is a length of brake tube and a coffee cup, that's how knock sensors are calibrated.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[andrebrito88]

Well thats nice to know thank you for your knowledge sir.

So the million question...the sound generated by petrol its the same as the lpg?

Did someone in this forum ever tune an ignition map based on knocks?
Old school guys tune the ignition by advancing until can ear the knock and after that reduce few degrees before it happen, is this applicable to knock sensors specially with lpg as juice?

 

[gruntguru]

I doubt the onset of knock will get you close to optimum timing on LPG. I would want to map it on a dynamometer.

You can use the same knock strategy but re-map it first.

je suis charlie

 

[MikeHalloran]

I worked for an engine dealer for a while.
They sold big natural gas engines for the oil patch, and rebuilt them to zero time.
I asked the old guys why the engines under 'test' seemed to misfire continuously despite the presence of super-zapper ignitions and expensive plugs and such.
The response was that they all misfire with no load, there's nothing you can do about it, and they smooth right out when you add a serious load like a generator.

That's about as much as I learned there.



Mike Halloran
Corinth, NY, USA

 

[Lou Scannon]

What I would do in your (OP's) predicament, failing access to the raw knock sensor signals, is at least find out what knock detection DSP chip is on-board the controller, then try to dig up a datasheet on the chip and learn the theory of operation. At least then you'll know how it works and its calibration parameters.
In several previous jobs, I was deeply involved in knock detection and abatement subsystem development and calibration, and this was critical knowledge to even get started on calibrating for a new application (and I consider a completely different fuel as a new application).
BTW, if you're resourceful, you can certainly acquire the raw signals externally to the parent controller, but you will need some know-how on electronics, high speed data acquisition, and data processing.
There do exist several aftermarket complete aftermarket engine management systems that include knock detection and abatement capability, but that's about as much as I know about that particular subject, so I can't vouch for their ease of use or effectiveness; but for sure the better ones have websites and user forums that will have a lot more information than you're likely to find on your OEM controller.

"Schiefgehen wird, was schiefgehen kann" - das Murphygesetz

 

[dgallup]

There are quite a lot of aftermarket companies selling CNG and bifuel kits. LandiRenzo seems to be one of the biggest. I think for the most part these systems use the factory ignition system including maps and knock sensors. They have an add on ECU that that controls fueling only. At least that's the way it is on the Ford pickup we converted.

----------------------------------------

The Help for this program was created in Windows Help format, which depends on a feature that isn't included in this version of Windows.

 

[romke]

The octane number of LPG will be somewhat higher then gasoline (around 102 - 104). So, you should not experience knock when the engine is set to operate at RON 98 gasoline. However, there are quite a few caveats. First: the engine will run somewhat hotter, because no heat is needed to vaporise the fuel. So, you will need the extra RON. Another problem is that the mixing quality of LPG and air in most LPG fuelled is not very sophisticated, meaning that mixture strength (lambda) may vary, resulting in a varying octane requirement depending on engine load, speed and ambient temperature. LPG can certainly knock, when running the engine too lean. A third complication is that because LPG is somewhat slower burning, you need a somewhat bigger spark advance - another complication increasing the octane requirement of the engine. All in all there is no simple way to get it right and optimal. Most LPG fuelled vehicles run on a rather rich mixture - thereby avoiding knock, but ending up with a non optimal fuel consumption. Finally there is the composition of the LPG itself. LPG roughly is a mix that may vary between 60%propane/40%butane and 40%propane/60% butane, depending on season and supplier. Getting it right thus can be very complicated if you want knock free running together with a decent specific fuel consumption. There is a reason that car manufacturers selling cars in areas where LPG is cheap shy away of designing a conversion kit themselves....