VEMS Community Forum

Hi all.


I have now gotten the Subaru driving,
And would like to know, how to run the injectors, right now they run 1-3-2-4 , like the firing order,but the ignition is run wastedspark, 1+3 & 2+4.

And do i run divider 1 or divider 2, ?

I have the crank injection set to simultanius, as this get the car started easyer(i think)


Any suggestions ?


/Skassa

You should be injecting when the valves are closed.  Anything else results in making fine tuning difficult.

You should run sequential injection with divider = 1.  I worked out that you should fuel the cylinder that fired last event.  ie if cyl 3 is currently firing, then VEMS should fuel cylinder 1. 

I must find the scrap of paper I worked this out on so I can check - my memory is playing tricks on me this week :(

Quote from: Denmark on April 30, 2008, 11:05:44 AM
I have now gotten the Subaru driving,
And would like to know, how to run the injectors, right now they run 1-3-2-4 , like the firing order,but the ignition is run wastedspark, 1+3 & 2+4.

And do i run divider 1 or divider 2, ?

I have the crank injection set to simultanius, as this get the car started easyer(i think)

When ever the spark is firing you know that the inlet is closed - so you are firing onto a valve rather than directly into the cylinder.

So you set your h[0] table to the sequence:
0 -> 8
1 -> 4
2 -> 2
3 -> 1

And run divider 1

Firing all the injectors for cranking means that you get 4 times as much fuel, you can get better resolution going to individual and then setting your cranking pw to 4 times what it is now.  Stick with what you're happiest with

Holy crap,

You all got to much time now, i have not had the time to read for a few hours, and now all kinds of nice suggestions,
This is just great :)

I does indeed sound like iÃ,´m alsmost hydrylocking the engine before it starts, so i will change to alternate, and up the cranking pw,

Rob is that sequense right?,
The firing order is 1-3-2-4 , and not the regular 1-3-4-2.

My table is now like this:
0 -> 8
1 -> 2
2 -> 4
3 -> 1


When i get of work, i will make a log on my way home...

Thanks,
Skassa

I have made some logs, but donÃ,´t know how to upload them,
I have found out that i think there is a problem again, with the trigger and ignitions ,when the fan comes on,  i just thourght it was something else,but it misfires(like cracy), and i cant get the car moving, till the fan stops again, not great when going home today!.
this all happens when the p259CH6 is turned on, so there is something wrong, when the p259chip is used!.

Here is a link to the P259 output that i use,

http://www.vems.hu/wiki/index.php?page=MembersPage%2FSkassawiringSubaru

The Tacho is working great, with no trooble at all,
I have run out of inj fets, otherwise i could move it to one of them, but all 8 are in use :'(.

/Skassa

Got a spare IGBT?  That will get round the problem.

How far from the ECU is the common grounding point?

Rob

Yep, i got 4 extra ign fets, so i will try one of thoose , but i will not be before tomorrow,

I also changed to the VEms 300kba mapsensor, as the oe one was reading much out, when really 240kba boost, its in megatune stating 200kba,

So i will try and map it with this one tommorrow,

I use only the oe wiring, through the adaptor wiring,and i have made extra grounding all over, with a homemade grounding kit.

I did just try the "white paper behind the exhaust", and there is wet spots on the paper
this is when i have made the H0 table like you suggested,

0 -> 8
1 -> 4
2 -> 2
3 -> 1


I canÃ,´t feel a difference though,

I also noted that itÃ,´s runníng high Kba on idle, around 45kba, on the oe ecu is around 30kba, also for overrun fuelcut to to work, it have to be a 40kba.


I will give it ago again tommorow, but if i cant get it working, i will put a HKS fcd,and a dawes device on the oe ecu, on friday,(i got that in the garage,so will be easy to put on, and the dawes really works for boost)
so itÃ,´s ready for the long trip saturday, it would just be nicer to be able to save some fuel on the long trip, as the oe ecu is really running rich..

Thanks,
Skassa

You calibrate the map sensor in software.  If this isn't right, then it tends to make you think things aren't right when they really are.

As for the idle problem, it sounds like you might have too much iac and too little timing.

Yes i know i calibrate the map sensor in the software, but even thourgh i have done that,itÃ,´s off, the futher it getÃ,´s away from 100 kba.

But i have not tested the max boost on the oe sensor, i have just read it to be 228kba,
When i put this in megatune, 100kba is 22kba wrong.

""You should run sequential injection with divider = 1.  I worked out that you should fuel the cylinder that fired last event.  ie if cyl 3 is currently firing, then VEMS should fuel cylinder 1.""

So this will be: injection order 4-1-3-2

0 -> 2
1 -> 4
2 -> 1
3 -> 8


Does this look right??

Firing order looks right. 

I've never had a pressure sensor that wouldn't calibrate.  You need at least 2 points (preferably 4 or more!) of known absolute pressure to do the calibration, since you can alter both the scale and offset.

Hmm,

I have just been reading the car manual,and it states:

after engine varmup manifold pressure: -400 to -500 mmHG,
so the 45kba that the oe map was reading must be right,

I just dont get that my boostgauge is showing 1,4bar when megatune only shows 1bar, and on the oe ecu, it was running 0.85bar as specified,abd with a VOS chip that runs 16psi,the gauge showed 1.15bar,
but it can ofcourse be that much off, 

right now i will move the fancontrol to a ign fet instead of the P259 output.

And change the injector outputs.


thanks,
Skassa

Converting -400 to -500mmHG to kpa where kpa = mmHG * (101.325 / 760.0)
-400 * 0.1333 = -53.32kpa
-500 * 0.1333 = -66.65kpa

So it looks like things are reading incorrectly for both sensors.

When I speak of the common grounding point I mean the part where all the grounds are joined 12-15cm from the EC36 connector, is the common ground 12 to 15cm from EC36 and are all the grounds clustered there?

The reason that this common grounding point needs to be there is because it equalises the ground voltages and this means that sensor readings remain stable no matter what voltages are flowing to ground.

I have now gotten the car to drive decent,
There was a guy who have made a Vems manual, that have switched to wires to the IAC , so when this was corrected the idle, was made great within 5min,
I will offcourse not say names, of who it can be that, have writen that manuel ;D


well there seems to be another problem, and that is, that when the fan switches on, the car misfires like cracy, and i have made a 6mm ground wire, from the ground wires that are between the ecu connectors, and 20cm to chassis ground.

I thourght it was something to do with the output, but itÃ,´s not, i have removed the output to the wastegate control, and used that inj to run the fan, but the problem was still present,
I then removed the wire from the EC36 connector and grounded it directly,
and the same thing happen,
So i must be noise in the trigger wires, it goes away if i just raise the rpm a bit.
The same when cranking, there is lots of misifre just when it starts to run, but as soon as it get to around 1500rpm the problem disaper, and dont return til the fan comes in!

Suggestion to make this work during tommorow?.

Thanks,
Skassa

Might be worth putting better shielded wires to the triggers.  Subaru wiring isn't exactly top notch...
If you're after a software fix, then some experimentation with the filter settings may help.

Also - check the grounds to the fans - maybe these are playing up.

If the UserGuide has any incorrect information please tell me so that I can rectify the problem.

Rob

Quote from: [email protected] on May 01, 2008, 07:30:40 PM
If the UserGuide has any incorrect information please tell me so that I can rectify the problem.

Rob

ItÃ,´s the part with the duel iac valve, it says:

EC36 pin 6 close
EC36 pin17 open,

but itÃ,´s the other way around, pin 6 is open and 17 is close, iÃ,´m currently also helping gettin a celica ready for the ring trip, and that was also the wrong way around, so itÃ,´s not the subaru manual thatÃ,´s wrong,

When i switched the wires it worked right away, with the 20v settings  :)


Regarding the noise, i was thinking more of some board change, to filter out the noise,other ecuÃ,´s canwork off the shell with oe wiring, so it odd to be possible.


What can i try with the filtering?

Thanks,
Skassa

I'm not that experienced with the filter settings, other than 36-1 and 60-2.

Rob did tell me about changing a couple of capacitors on the board, but I don't remember the details.

I will try to mount a 10k pullup to see if this will help,

itÃ,´s kind of the same problem, i had with the 10v corolla,

This is just worse, as i cant run the fans, or i can, i just needs to be over 2000rpm, but that is really deficult to drive like that

Its a pull-down you need.

Measure the resistance of the sensor and use a pull-down of the same value.

We're discussing this with Peter at the moment.  Hopefully he'll have some more possible solutions.

BTW you can keep the IAC connected as it was and set the channel value to i6

Okay, thanks .


I will go to bed then, i hope for the car to be running, when i get up in the morning ;D.


Thanks,
Skassa

Are you expecting a visit from the engine elves?

It would be nice, but it have not happend :)

Here is a pic of the crank and cam sensor data from the manual:

(http://img.photobucket.com/albums/v41/Denmark/sensorohm.jpg)

Okay...
We really need the measurements from your sesnors so that we can work out the pull-down. 

Also Jorgen was telling Peter and myself about setting the LM1815 to Mode1 which is where pin5 of the LM1815 gets pulled high to 5V, this makes the arming voltage much higher and helps stop noise causing false triggers.  There is provision for this on the board as two pads are present on the board near each LM1815's pin5 that will allows a soldered link to be put in place.

The LM1815 is 'armed' on a rising voltage and triggers when the voltage falls to 0V - so its not on the peaks of the trigger but when the peak has returned to 0 that the ECU works from.

The LM1815 has three modes:
The first is the standard one they ship from VEMS, the adaptive trigger, which is where no voltage is applied to pin5, this means that the system will 'arm' on any input voltage over around 30mV. 
The second (we'll call it Mode=5V) is where 5V is applied to pin5 and the 'arming' voltage rises to over 100mv (check the LM1815 data sheet for the details). 
The final mode is where pin5 is pulled to ground and any voltage over 0v 'arms' the chip.

Apparently its possible to order Mode=5V with the completed units from the webshop!

Thanks alot guys,

We go the car running today,and the fan can kick in, without the car stalling, i tried before we put the blobs on the solderpads, that if i keep the rpm over 1200rpm, the fan will not cause trigger errors.

But what we did as a quick sulution was to make the blob, so both the cas and cam lm1815,  does not see the noise under 200mv(raised from the 45mv as oe),this was first tried and it helped alot, the problem is that the car was warm, and the trigger noise is more when cold, so i will not not fora sure til tomorrow morning.
We then made a pulldown on the second vr sensor 4,7kohm,but its not perfect, but my electronic friend would rather put a 15-20k ohm, on the other side of the 12kohm that is already on the board,
But that i dont no much about, but will get it better, when/if we get home from the Ring trip..

we did not have time to scop the signal today, as there was lots of other things to do before we leave at 6am tomorrow.

Thanks,
Skassa

well,

I got the car to the ring and some hard runs there (9.30min on my 6th round on the ring)and back again, without problems, the only trooble i have now , is the trigger noise, most bad when the car is cold in the morning, it talkes around 5min on and of cranking to get it running, were it humps and sputteres, then suddenly it burst to live, and runs without problems, strange, but iÃ,´m sure that itÃ,´s trigger noise, but i have to get this working with the oe wiring, to being able to sell/map the product here in DK.

So we will give it ago this weeekend again, hoping tofind out whatÃ,´s needed to filter the trigger input,
On sunday i have to map a pnp LinkECU , so i will take a look to see how the trigger input is filtered on that(to get a idea to make it work)

I still run only 0.9bar, but there is some more power and drivebilty over the oe ecu, and i run a average of 7,8ltr/100km, on the run home, that was with a speed between 140-160km/t, that is lambda 1.05 and 45 degrees of advance  :)


/Skassa

Just to recap - Are you using a pull-down resistor on each of the trigger inputs?

No ,i i now only have a pulldown on the second  trigger, but both lm1815 chip are in the mode, were they ignore whatÃ,´s under 200mv, instead of the 45mv default.

But we willl scop it this weekend to see if there can be figuered out how to make it perfect.


/Skassa

It would be a good move to put a pull-down on the primary trigger too.

In the cases where the VR sensor gets constant stimulation (36-1, 60-1 etc) false triggers are less likely, with the Japanese cars where the ground voltage can wander and is subject to interference you need to clamp things as much as possible.