astra 1.4 8valve lpg turbo.
Turbo Mini Forums

That's how I understand it.

Saul Bellow - "A great deal of intelligence can be invested in ignorance when the need for illusion is deep."
Stephen Hawking - "The greatest enemy of knowledge is not ignorance, it is the illusion of knowledge."

hi paul, thanks for chiming in ,

my calc is at 18 I think its in the msq back a page .pw at tickover is around 4.5ms .

however ,my numbers in the ve map are in the 45 range at tickover ,so I was going to drop calc to 10 then increase map to increase resolution .also I was going to try playing with simultaneous 2 and 4 squirts ,which will half my pw so then I may get into that sort of problem .

I beginning to think I need as many squirts as possible to counter the very low mass of the gas injected .and the possibility I may be blowing away from the inley .inj timing would also affect this .

ohhh so many options !!lol

Medusa + injection = too much torque for the dyno ..https://youtu.be/qg5o0_tJxYM

I would start by getting the REQ_FUEL right. It is used for calcs other than just the main pulse widths, so best get it right and stick with it.

Then I would put your 3.3mS injector opening time as your "dead" time.

Then play with your VE at 2 squirts until you get it running OK.

Then change to 4 squirts. If it does not run as well as with 2 squirts, then start bringing the dead time down until it does.

You should then be in the right ball park.

Saul Bellow - "A great deal of intelligence can be invested in ignorance when the need for illusion is deep."
Stephen Hawking - "The greatest enemy of knowledge is not ignorance, it is the illusion of knowledge."

From what I've seen on Jean's forum - where people have been experimenting with P&H drivers for LPG - the LPG injectors are horribly slow.

Yes, the "dead" time is open and close added, and yes, if changing the number of squirts changes the overall fuelling it means the dead time seting is wrong.

With petrol injectors the recommendation is to physically measure what it actually is but I can't think how you could do that unless it remains liquid after the injector (never seen one so don't know but I doubt it).

The simplest way would probably be to experiment with dead time until you get near equal fuelling (AFRs) on either two and four squirts.
EDIT - Paul has just said that already...(end edit)

You might also want to check the battery voltage calibration as I suspect they are more sensitive to battery voltage than a petrol injector especially if you are driving them through resistors (which I seem to think you are but I may be wrong).

Edited by Rod S on 24th Jun, 2014.

Schrödinger's cat - so which one am I ???

Edited by Paul S on 24th Jun, 2014.

Saul Bellow - "A great deal of intelligence can be invested in ignorance when the need for illusion is deep."
Stephen Hawking - "The greatest enemy of knowledge is not ignorance, it is the illusion of knowledge."

We'll have to disagree then but, from the manual,

With 'Multiply MAP', without 'Include AFR target'
PW = DT + (ReqFuel * MAP * VE[RPM,MAP] * AirDen * BaroCor * corrections)

With 'Multiply MAP', with 'Include AFR target'
PW = DT + (ReqFuel * MAP * Stoich/AFRtarget * VE[RPM,MAP] * AirDen * BaroCor * corrections)

In other words the dead time (DT) is added as a fixed value to all the other parts of the calculation.

It's a bit of a fudge as there will be some fuel flowing as the injector starts to open and begins to close (hence why the recommendation is to actually measure it) but the electrical pulse width (which is what shows in TS) is computed fuel pulse width required plus the dead (open/close) time.


EDIT - spelling

Edited by Rod S on 24th Jun, 2014.

Schrödinger's cat - so which one am I ???

It is a bit of a fudge. The more I think about it, the more complex it becomes.

If you add both the opening and closing times to the calculated pulse width, then you will have too much fuel.

The MS manual says to calculate the dead time by subtracting the closing time from the opening time, but then this is only an approximation.

Best actual build a test rig and measure it, but I doubt Rovert can do that with LPG.

EDIT: Found this on the MegaSquirt website:


However in Rovert's case and long opening and closing times, the method introduces a lot of error. He needs to experiment as stated above.

Edited by Paul S on 24th Jun, 2014.

Saul Bellow - "A great deal of intelligence can be invested in ignorance when the need for illusion is deep."
Stephen Hawking - "The greatest enemy of knowledge is not ignorance, it is the illusion of knowledge."

Schrödinger's cat - so which one am I ???

hi chaps,


well that has opened up a can of worms ! ,

that diagraM OF PAULS AGREES WITH HOW I THOUGHT OF IT ,BUT IT SOUNDS LIKE ,oops, a bit of experimentation with 4 and 2 squirts is my afternoons fun .


paul, the problem with the req fuel is I have no idea of the flow rate of the injectors .so my plan as to aim for 10 ,and as good a resolution on the ve map as I could get ?

few observations from this morning .

I reduced req fuel to 10 from 18 then multiplied the map by the correct percentage to compensate .this gave autotune greater resolution ,and helped a lot .

I mapped it in neutral for a while ,and had it revving nicely and pretty cleanly.

tickover ,a bit high at 1000 rpm ,had climbed to 1400 ,a good sign , but it was still a bit rough .

left it on autotune for a while then after it had settled at around 74 ve , I turned the gas pressure down until the tickover was smoothest , autune was now at 98 ve ,and gas pressure 10 psi less at around 5 psi .

I wonder if the injectors have a problem under high vacuum 20inhg and higher gas pressure with opening cleanly ,maybe due to the resistor pack . and lower battery voltages .maybe have a decko at the voltage compensation map as per rod .

oh rod , im on the resistor pack you did your da vinci sketch for .


robert

Edited by robert on 24th Jun, 2014.

Medusa + injection = too much torque for the dyno ..https://youtu.be/qg5o0_tJxYM

Is your gas pressure regulator reference to intake vacuum (MAP)?

If not and you are running low gas pressure, then your VE values will be heavily scewed to compensate.

Saul Bellow - "A great deal of intelligence can be invested in ignorance when the need for illusion is deep."
Stephen Hawking - "The greatest enemy of knowledge is not ignorance, it is the illusion of knowledge."

yes paul referenced to inlet vac , vac at tickover is 20 in hg. gas pressure on full throttle is 12 psi .

Medusa + injection = too much torque for the dyno ..https://youtu.be/qg5o0_tJxYM

Did you ever run this engine on the MS with petrol?

If you did, then the VE used at idle should help you get to somewhere near the correct REQ_FUEL setting on LPG.

If you did not then I would adjust the REQ_FUEL to work with around 50% VE at idle.

From the above info ie. 98% VE on a REQ_FUEL of 10 and 12 psi WOT fuel pressure, then I would have though a REQ_FUEL of 20 would be about right.

Edited by Paul S on 24th Jun, 2014.

Saul Bellow - "A great deal of intelligence can be invested in ignorance when the need for illusion is deep."
Stephen Hawking - "The greatest enemy of knowledge is not ignorance, it is the illusion of knowledge."

no I didn't paul , it has not run petrol fro about 5 years I think .

I am at 72 in the tickover area , and I have now tried:


4 squirts ,adjusting inj open time from 1.1 to 3.3 ,and allowing the tune to adjust .

results were ,around 2.3 to 2.5 worked best ,but tickover would not rise above 1000rpm ,and the adjustment did not cure that .ve numbers were anywhere from 30 to 45 . I think I may have a theory ,and that is that the injectors are too big ? and so going unstable at such small openings .


back to 2 squirts ,and instantaneously up to 1400 rpm tick over as before .

then turned the tick over down to 1000 rpm ,all good. pulse width 4ms ,and req fuel 10.

then went for a drive ,allowing it to auto tune , very smooth , bit hesitant until adjusted but all in all very good . map is populating with numbers in the 70 to 100 range . id prefer the 100 to 140 range but hey it works ..

two things , one is a hesitancy on first second of accel I can't seem to get rid of with the accel adjustment,may be lag in the fuel pressure ..

and the other is if I rev it ,then let it drop it will for 10 or 20 seconds run in the 12:1 afr instead of the set 13.5:1. I adjusted the shut throttle overrun down to 50 % but no effect , I wonder if this is also lag in the vapouriser not dropping the pressure fast enough ..so I have shortened the signal pipe by half .


oh I also checked timing and was 1 degree out so adjusted that too.


all in all ..progress!

Medusa + injection = too much torque for the dyno ..https://youtu.be/qg5o0_tJxYM

ms tip in issues are something i had issue with for ages. when they intriduced the hybrid accel enrichment it helped tonnes. as i could tweak mapdot and tpsdot. changing to 4 squirts/cycle helped massively too. the icing on the cake was to make the cells on the VE table above where the car rests at idle slightly richer than necesary. it ment on cruise and transitions at ccruise could get a little ugly. with the injectors being so "laggy" i dint think youll be able to completely eliminate the throttle tip-in issues. :/

turbo 16v k-series 11.9@118.9 :)

Denis O'Brien.

that's sounds good denis , I only have it on tps enrichment at the mo ,that doesn't seem to do much immediately the pdals pressed , I will have to have a look at the map related one too .this used to be called asynchronious enrichment years ago !

weirdly... looking at my data logs.. shows me my air temp sensor is stuck on 21 c..

I have 5v at the sensor and the sensor works so this could be a problem .

I started it today and let the warm up enrichment auto tune do its thing .seems to work ,I have no idle valve working at the mo , so idle went from 500 to 1000 as it warmed up .

tomorrow I will try to work out how to activate the idle valve dupher,its a stock astra one so should be easy ... haha.

Medusa + injection = too much torque for the dyno ..https://youtu.be/qg5o0_tJxYM

Using 4 squirts/cycle is a kludge that is useful when using batch injection: that gives you more pulse width updates for a more up-to-date fueling. The real solution is full sequential injection where you get the same number of pulse width updates but with a more efficient use of the injectors.

Jean

http://www.jbperf.com/

turbo 16v k-series 11.9@118.9 :)

Denis O'Brien.

bit of afternoon fun in the batcave ...

comparing the first dyno run with the last ,there were several runs in between..












and a summary of the data ..





this was with a room temp of 28c and a inlet manifold air temp of 60 c ....bit hot for such low boost . should be around 30 to 40 c on the rd .

last run was a long loaded run to give the time to map it .


happy !

Medusa + injection = too much torque for the dyno ..https://youtu.be/qg5o0_tJxYM

Fantastic!

what boost was that on?

joe, its at 5 psi up to 4500rpm, then climbing to 8psi at the top rpm .( due to a small wastegate flap).this give a nice level trq curve so im leaving it that way .oh yes .

Medusa + injection = too much torque for the dyno ..https://youtu.be/qg5o0_tJxYM

well thats a pretty decent figure then.

yup not bad , but then it does have a 38mm inlet valve ...as standard !

I relly want to do a run with a nice cold inlet .I figure 150 maybe possible .

oh I cured the non working inlet temp sensor by bending the terminals in the plug so it contacted .

and I spent a fair while on the stepper motor idle control , with no luck , the pinlte moves in and out a little upon key turn to 12v , and if I change the steps in the map it moves in and out a little ,but no luck with it having any effect on the idle speed while its running and I change the steps ! so that's a bit confusing .

Medusa + injection = too much torque for the dyno ..https://youtu.be/qg5o0_tJxYM

Good result Robert.

In the 13's at last!.. Just

thanks gary ,looking forward to driving it today .I think its going to be wheelspin central

Medusa + injection = too much torque for the dyno ..https://youtu.be/qg5o0_tJxYM

You know, I have read all this and I think I even understand that last bit !

140bhp sound like a lot of fun in an Astra that looks "normal"