Just put my car back on the dyno this morning as although flat out it is amazing, in traffic the idle to low TPS% transition the car is very jumpy.
Having added a number of low down tps values to ensure a smooth transition I'm still finding it has a very rough transition from 800rpm tick over (sat in traffic) to 1300-1700rpm (letting clutch out in 1st to move off) I find I have to dip the clutch and rev the engine to 2000rpm to get it moving without stalling. The VE table was set on a dyno holding each cell until lambda stabilised, as such I'm confident the it is correct, my acceleration enrichment is 100/40/10/0, is this appropriate for Alpha-N? or any other suggestions greatly appreciated.
This is the log file I took yesterday before taking it back to the dyno today: http://vems.hu/vemstune/sharingcenter/reports.php?cmd=view&key=EuAHqQ
Config from today's dyno session (the issue still exists, but I haven't logged it yet): http://vems.hu/vemstune/sharingcenter/reports.php?cmd=view&key=cfnWVY
Thanks
Alex
Try from: dtps 3/10/33/150 with values 0/25/55/100. Give most attention to second 10vs25 point. Position of this point is the key.
Another common problems with itbs is:
-improper linkage system. It should be constructed with variable ratio over range. At small openings pedal travel must be 3...6 times less sensitive comapring to WOT region.
- play between itb axle and TPS. Seems you have this problem. TPS signal is still 0, but Lambda already goes lean.
Also try to manage lower idle advance around 5deg if possible. It will give more power at very beginning.
Gints
Quote from: GintsK on August 10, 2012, 07:31:48 AM
- play between itb axle and TPS. Seems you have this problem. TPS signal is still 0, but Lambda already goes lean.
Gints
If this is the case then it´s the first problem to solve without knowing TPS has moved it´ll still just fuel along for idle.
Gints, you were spot on about the TPS, I could move the axle slightly without it registering a change in position. The fault is in two places, the 1st is the TPS didn't like small movement in its position, I rotated it about 15degrees and re-calibrated, the interface between the axle and TPS also had a bit of play, this is now snug; shimmed with of baking foil. Test drive will be tomorrow.
Thanks
Alex
New TPS fitted and calibrated, not the same scale as the old TPS so I went back to the rolling road again. The car is significantly smoother in traffic and small movements in the throttle are reflected in TPS signal.
I've taken another log which still shows the engine going lean on acceleration: http://vems.hu/vemstune/sharingcenter/reports.php?cmd=view&key=E3pwnF
Is now an appropriate time to set dtps 3/10/33/150 with values 0/25/55/100 as per recommendation?
You can try for sure.
But for me seems you have to tune VE map properly. In your log target and actual lambda never matches. And shape of VE map tells the same.
Gints
Quote from: GintsK on September 03, 2012, 06:25:11 AM
You can try for sure.
But for me seems you have to tune VE map properly. In your log target and actual lambda never matches. And shape of VE map tells the same.
Gints
I don't understand what is causing the lambda target and actual to mis-match, each cell was set on rolling road holding each one until they match yet when I log I don't get the same results.
Your injector dead time is not configured. It is zero. Means your VE table matches just at specific battery voltage and multipliers (lambda, MAT). You can ascertain by slowly changing target lambda near your idling. Actual lambda should follow 1:1. If real lambda is too sensitive your dead time is to small. If following is sluggish, deadtime is too high.
What about air temperature?
I guess you use MAT sensor from Ford company? Or? Not possible to get precise readings without board modification. Resistance at 25degC is far too high.
Here is IMO best sensor in terms of reaction time and heat soak issues: BOSCH 0 280 130 085
It is wise to check all this before rolling road. God is in the details.
Gints
Gints
Correct, I'm using a ford MAT sensor, plot points were configured using boiling water cooling with a thermometer in it, will replace as recommended though (ordered, do you have the MAT sensor plot points or are they one of the standard options?).
Turns out my TPS was also breaking down as the voltage would drop off with the throttle kept at the same opening, although new it was a pattern part rather than OEM, so that will be replaced as well.
So... on to injector dead time, at the moment I have "traditional" selected, this doesn't give an option of deadtime, do you recommend switching over to "simplified"?
I'm using Bosch injetors 0280156127 : Static Flow Rate: 42lb/hr @ 43.5PSI = 440 cc/min (+/-2%),Coil Resistance: 15.9 Ohms (+/-0.35 Ohms) / High Impedance / High-Z (No ECM driver modifications required), is there a way to find/calculate sensible starting values?:
(http://locostbuilders.co.uk/upload/Screen%20Shot%202012-09-03%20at%2018.52.12.jpg)
As I wrote: if you change the target, actual lambda should follow your changes. You can achieve it by both methods.
In addition you can check voltage compensation slope: actual lambda should stay settled while voltage drops/rises. Traditional method is good just for relatively narrow range. In simpyfied you can adjust dead time at various voltages or rewrite/adopt from somewhere published injector data.
For that Bosch sensor I use NTC 1900 256 patch. With BOSCH 17-point curve it gives couple degrees error what can be adjusted then.
Gints
Gints
Appreciate what you are saying about setting deadtime, I've tried both options; using simplified and the data from the source bottom of the page (converted to 43.5psi fuel) I used:
8v 2559
10v 1754
12v 1241
14v 1004
16v 820
I tested by turning main beam on at idle which caused lambda to rise from .90 to .96 as the volts dropped for 13.5 to 12.7(ish), I'm assuming the different boards and flyback resistor effects deadtime.
Although "simplified" I would end up having to try to find 5 different figures, so I moved back to traditional.
I used 200 for injector rampup as recommended within help files (is the correct figure within the data below? I couldn't spot it), and then proceeded to increase injector voltage compensation until lambda stabilised using the above test, this was 2000us, does this look feasible before I start logging and tuning again?
(http://locostbuilders.co.uk/upload/Screen%20Shot%202012-09-08%20at%2020.37.31.jpg)
(http://locostbuilders.co.uk/upload/Screen%20Shot%202012-09-08%20at%2021.06.40.jpg)
Slope of simplified curve seems OK. Values, you are right, varies depending from flyback voltage, pressure etc.
2000 theoretically is way too much. about 400 accordant to this datasheet. 400...1000 is usual range. http://www.vemssupport.com/forum/index.php/topic,1261.msg13418.html#msg13418
Do not leave open time @13,2V zeroed. Start with 900 there. Otherwise VEMS adds just [192/1.5]
[email protected] - too less . By finding open time @13,2V you can find right adder what makes multiplication results correct. By finding Injector voltage compensation - immunity to voltage drops (just small drops - in traditional case because correction is linear). Be careful - if you change current consumption you change also engine load with it.
Gints
so a year on....
I believe I have my injector slope sorted, I went back to the rolling road again about 3 months ago (this is getting expensive!) to sort out idle and top end, by torque curve has increased and it is better in traffic, yet it still isn't right.
It is booked in at the local rolling road for Friday 11th as I know that there is a some further room for improvement in the VE table following setting the balance on the throttle bodies with a Carbtune kit - this beats using a normal airflow meter for ease, speed and accuracy by a long way.
I have increased the accel enrichment to dTPS 3/10/33/150 with % 5/40/60/120, but I'm still seeing quite a swing from lean to rich on acceleration. I am also experiencing misfire under certain loads, I would really appreciate someone taking a look through my log/config and advising where I am going wrong, its getting closer, but I ran out of knowledge a long time ago and don't want to wreck the engine through a silly mistake.
https://app.box.com/s/1zgkw424tvtm3q37d7qx (https://app.box.com/s/1zgkw424tvtm3q37d7qx)
For reference I am running a 2009 2.0 Ford Duratec, all standard except lightened fly wheel, tubular manifold and a set of DICE throttle bodies with 440c blue injectors, it is putting out just shy of 220bhp at 7200!
I am in Bristol, if anyone is local and is prepared to help in person I will gladly pay a good hourly rate as my local rolling roads aren't familiar with VEMS.
Thanks
Alex
Throttle response is hardly dependant from injection angle. Usually 350...500 degrees gives best results. On past 1.2.0 firmwares lambda signal is fast enough to find best combo.
Quote
... a Carbtune kit - this beats using a normal airflow meter for ease, speed and accuracy by a long way.
Carbtune uses vacuum principle. But what is important to us for matched mixtures are air mass passing to cylinders. So your statement about accuracy is questionable.
At same vacuum airflow may differ. Especially on worn engines. For same reason we do not use Speed density for engines with ITBs: pressure after the throttles do not tell a lot about the actual air flow.
Gints
Gints
Appreciate the reply, but not really sure what to take from it.
I understood Injection angle can only be changed if using cam sync, otherwise the default is 720, I don't currently have cam sync; if it is worth moving to I'm happy to invest in a new cam cover with sender, will this improve the lambda consistency and reduce the misfire I am seeing?
Didn't realise my carbtune comment was potentially misleading; for me it worked and allowed me to set idle significantly smoother than using one airflow meter and moving it between cylinders - I don't have the luxury of having four of them.
Other than changing to cam sync is there anything else that needs to be addressed within my config?
Thanks
Alex
With ITBs my suggestion is camsync, yes. It makes difference for sure!
I just took a look to your config. Before accel you have to tune your VE table. You do not need a rolling road for VE table optimization. Just drive around on various loads, change throttle position as smooth as you can, and finally use VE tune by statistics under tools. From your log it already calculates more common shape of VE table. I think you can find some video on youtube how to.
Where your injectors are located? Does it sprays directly to valves? Or?
Cam cover: ebay here I come!
I have looked at tune by statistic, but I assumed that live VE tune was more accurate which is the approach I used on the dyno.
My injectors are in the throttle bodies, immediately after the butterflies, they are angled towards the head.
Live tune can not be more accurate. It changes just current cell while neighbour still have influence.
Tune by statistics do not corrupt table at one point and that means - error from ideal is smoother if initial table is smooth.
Just towards the head or exactly towards the valves? It is important. Spray pattern as well.
Interesting comment on live tune, I'll make the effort to get a good log and use tune by statistic. Out of curiosity, how do you tune a new engine from scratch?
Injectors are spray type "E" (dual spray), 20 degree alpha 50, a clocking angle of 90 degree. They are angled towards the head, not exactly at valves as they are around another 40 degrees away.
Even on dyno I use tune by statistics. Still MLV. First is raw tune. Second WOT runs - from them I find optimal places for rpm axes. Then fine tune.
BTW EGO correction nowadays is pretty fast. You can rely on it for WOT tuning. May be not on second gear. But third is OK.
If you want streetable/enjoyable car it is very important to spray directly to valve. Valve is way hotter than walls. If injector sprays on walls most of fuel just form huge liquid film. And you do not knew when this fuel going to burn. Ten cycles after. Or even never. That is worse than carburettor. At full load picture comes better because of air flow. Watching your datalog tended me to ask this question.
Gints
Great info and thanks for taking the time to look through my log.
I can easily move the injectors to the head, which I'll do at the next opportunity.
I'll also use ego correction, what is a sensible number of engine cycles before a change?
If you have place at head, there is still option to add second row of injectors and use staged injection. It is easy to tune, but you have to knew exact flow rates of both rows.
Cycles. 4 if sensor is very close to head. 10 if more than 1 meter.
One note: first value in accel enrichment should be zero: 0 40 60 120. Otherwise Ego correction might not work.
Once again thanks for the info.
The engine currently has 4 x 440cc injectors sat in the ITB that are active, it also has 4 x 240cc in the head that are unused.
The simplest option would be to move the 440cc to the head and map from there.
I'm expecting there to be a drop in top end performance, if its a couple of %bhp i can live with it, what would be your ballpark expectation?
If its a bit more I'll get hold of some injector clips and make a new bit of loom.
Mapping with staged is similar. You only have to determine when start to add fuel from furtherest injectors. Say 240 will be maxed out at 200hp. Somewhere at 70...80% would be reasonable to add spray from secondaries. Watch at what rpms/tps you reach 70..80% of your current max duty cycle - and choose those for switching. Switching point can be changed on tuned map.
From my understanding main advantage from distanced injectors is cooling. Vaporization of fuel from walls may reduce temperatures below environment. Cooler = more air = more power.
So if you want technical way - go for staged. If you want just drive with less building efforts - 440cc at head would be OK.
Quote from: se7ensport on September 29, 2013, 06:04:01 PM
I have increased the accel enrichment to dTPS 3/10/33/150 with % 5/40/60/120, but I'm still seeing quite a swing from lean to rich on acceleration. I am also
btw, for what reason those % are computed from the req. fuel (ms) parameter (engine set up) and so are independent of the VE values
VE values? VE values for speed density is almost constant... don't knew...
There is fuel film values what is PW dependent. Try those.
Progress, but settings confirmation needed:
I now have 215cc injectors in the head and 443cc in the throttle bodies, I have set the thresholds for secondary injection as 25% tps and 1700rpm.
What I would like to understand is how the actual fuel injected is calculated and should the thresholds used be the same as figures the VE table.
Example: Using only 440c injector the RPM/TPS table at 25%TPS the amount injected goes from 100 at 1500rpm to 110 at 1700rpm, and at any point between the two figures VEMS calculates the required fuel between the two i.e. at 1600rpm it would use 105.
Now with two set of injectors (215c primary and 443c secondary, which gives a total of 547c once second injectors are triggered) and the staging set to 25% and 1700rpm I calculate the table to become (roughly) 200 at 1500rpm and 80 at 1700rpm.
So what does VEMS do at 1600rpm??? My concern is that it would interpolate a figure of 140 but only use the primary injectors as the trigger threshold for the second injectors has not been passed.
Do I need to use threshold figures that match the X and Y axis on my VE table?
We'll I got the above totally wrong, VEMS manages the second set and the 'VE table has its regular shape, I had mistakenly thought the table would drop in values as the second set came in.