Jump to content
jetape

At a loss with fuel map tuning

Recommended Posts

I'm running a G4+ based vipec on a jetski (3cyl 1.5L supercharged).

I've struggled all of last season to make it do what I want/expect.  I did a rebuild  over the winter and thought maybe a fresh start this season would make things more sane, but it's not the case.

No matter what I do a cannot get adjacent RPM columns to tune without affecting each other to the point I can't get them to follow my AFR target by 1-2 AFR points.  I think I tried everything, doing it live on the jetski with a tablet, logging, mixture map, everything.  I can spend hours riding around logging, stopping and applying corrections based on mixture map to cells that have 1000+ hits.   Ride some more and log then again do mixture map corrections and it will break previously tuned areas.  I did live tuning with a tablet and blutooth keyboard and adding/removing values from cells as I hold a specific RPM.  Move to next column and do the same, then try to go back to previous and it's all messed up with 1-2 AFR shifts in the transient RPM range and the final cell.  I tried smoothing, interpolating, nothing seems to be working.  Countless hours trying things, big changes small changes it always ends up going in circle of tuning one spot and breaking another and not being able to correct it without breaking something else.

This can't be this hard, I've done flash tuning before where adjusting the fuel table actually made sense in the results.  I'm lost, I really don't want to dump this ECU because I like how much control it provides but I can't spend another season tuning every outing :(

 

Anything obvious that could be causing me so much grief?  Only thing I can think of is maybe my injector data is wrong but I copied the settings from one of the base maps which is supposed to have used the same Bosch 0280158040 injectors.

Attached is a map/log I just tested where I was trying to tune 2000-3000 rpm. 

3.pclr Log 2019-06-2 3;19;15 pm.llg

Share this post


Link to post
Share on other sites

I dont think you are going to be able to tune it with the lambda continuously bouncing around like it is in this log.  Either there is an air leak or one cylinder isnt working or the wideband or wiring is faulty etc.  If Inj PW, MAP & TP is reasonably constant like it is in the screenshot below you should have a reasonably constant Lambda to match.  

 

y8NDXiP.png

Share this post


Link to post
Share on other sites

All 3 cylinders are definitely firing just fine.  AFR is definitely bouncing all over the place while I'm riding and holding RPM. I'm using LC2 as the controller which sends signal to the ecu and a stand alone gauge.  I know some people don't like these controllers but they can't be this bad can they? 

Share this post


Link to post
Share on other sites

It's a short ground to the battery.

It doesn't feel like it's a display issue, I can be holding the throttle dead steady and afr will bounce up and down as will the rpm by 200-250. I'm not running closed loop so what LC2 sends should not be having any effect on it's own.

Share this post


Link to post
Share on other sites

What injectors are you using and where did you get the deadtime + short pw adder values for them? There might be something bad around your o2 sensor as others are suggesting which would throw off any feedback-based tuning, but looking at your injector numbers, there are some pretty savage numbers in there and if they arent correct, it would be throwing your AFR around a fair bit at low load as the ECU will have completely the wrong idea about how much fuel its putting in. The short pulse width table especally i'd want to verify - its all negative, which seems really unlikely and as an example, at the ~1.4ms effective shown in most of your log, its removing ~0.3ms from the commanded injector time (over 20%), and based on your pw adder table, you are well into the non linear range at this point, so if that table is wrong, its going to be making it behave REALLY wrong.

Share this post


Link to post
Share on other sites
2 hours ago, cj said:

What injectors are you using and where did you get the deadtime + short pw adder values for them? There might be something bad around your o2 sensor as others are suggesting which would throw off any feedback-based tuning, but looking at your injector numbers, there are some pretty savage numbers in there and if they arent correct, it would be throwing your AFR around a fair bit at low load as the ECU will have completely the wrong idea about how much fuel its putting in. The short pulse width table especally i'd want to verify - its all negative, which seems really unlikely and as an example, at the ~1.4ms effective shown in most of your log, its removing ~0.3ms from the commanded injector time (over 20%), and based on your pw adder table, you are well into the non linear range at this point, so if that table is wrong, its going to be making it behave REALLY wrong.

Injector data was copied from the "G4+ Storm Black Sample" basemap which is supposed to be using the same Bosch ev14 040 injectors as I'm using.

Share this post


Link to post
Share on other sites

In the screenshot I posted above your actual injector PW only varies by 5%, yet your measured lambda varies by 30% over the same region.  Because actual Inj PW is constant you can say deadtimes and short PW adders should be having very little effect on the amount of fuel injected.  The question you need to ask yourself is why are the fuel and air flow related parameters all showing constant but AFR is all over the place?  Either there is something causing an error with the measurement instrument (i.e air leak, bad injector, leaking intake valve, etc) or the measurement instrument itself is the problem.  Based on my experiences with the innovate products I would lean towards the later, but thats where you need to focus more before you will be able to get a decent tune into it. 

If it were me and I had made myself happy there were no air leaks etc then the first thing I would do is get a different wideband onto it. Have a look at the 14point7 spartan2, USD$75 without the sensor so not a big investment.

Share this post


Link to post
Share on other sites

I agree with Adam here.  I would ditch the LC2 and get something else for wideband measurement.  I have used Innovate products in the past on customer project as well as personal and every time was left with poor experiences at some point along the way.

I would honestly recommend something with CAN output to eliminate any chance of a ground loop/offset, which is what i believe youre experiencing here.  Either that or just a terrible amount of noise from something such as a alternator or ignition.

Share this post


Link to post
Share on other sites
5 hours ago, Adamw said:

If it were me and I had made myself happy there were no air leaks etc then the first thing I would do is get a different wideband onto it. Have a look at the 14point7 spartan2, USD$75 without the sensor so not a big investment.

14point7 is actually a local guy in my city, I looked at them before but went with LC2 due to the two configurable analog outputs I can use for the ecu and standalone gauge at the same time.

I would say it's not an air leak because it runs fine with stock ECU, it would most likely have a high idle if there was a leak with stock ECU.  I think.

 

2 hours ago, Brad Burnett said:

I agree with Adam here.  I would ditch the LC2 and get something else for wideband measurement.  I have used Innovate products in the past on customer project as well as personal and every time was left with poor experiences at some point along the way.

I would honestly recommend something with CAN output to eliminate any chance of a ground loop/offset, which is what i believe youre experiencing here.  Either that or just a terrible amount of noise from something such as a alternator or ignition.

Can't use CAN :(  This marine version of G4+ has only a single canbus module and that's used up to make the build in jetski functionality work.

No alternator,  ignition noise I'm not sure how would one test for this?

Share this post


Link to post
Share on other sites

Disconnect the alternator completely  electrically and then Run the engine with logging turned on.  Also, try disconnecting any other potential sources of noise such as ISCV etc.  

Also, try connecting the wideband output and sensor ground to a multimeter (with it disconnected from the jetski) and see if the output still fluctuates. 

HTH, 

 Richard. 

Share this post


Link to post
Share on other sites
14 hours ago, Richard Hill said:

Disconnect the alternator completely  electrically and then Run the engine with logging turned on.  Also, try disconnecting any other potential sources of noise such as ISCV etc.  

Also, try connecting the wideband output and sensor ground to a multimeter (with it disconnected from the jetski) and see if the output still fluctuates. 

HTH, 

 Richard. 

this is one thing i noticed from the LC2 manual i briefly looked at yesterday, the power ground and sensor ground are the same wire on the module.  There is no dedicated sensor ground for the analogue output.  Im pretty sure this may be part of the problem if there is not a mechanical issue.

 

Share this post


Link to post
Share on other sites

From looking at the log, the fueling seems to follow the map signal.  I would try switching the fuel load from MGP to TP main and setting up a 4d table to account for boost.  

I have had trouble with 3 cylinder motors in the past with very bumpy map signal due to the pulses in the motor.  I even went as far as adding a decently sized vacuum chamber that the map sensor feeds from to dampen the pulses before ultimately switching to TP as load for fuel.

Share this post


Link to post
Share on other sites
3 hours ago, Brad Burnett said:

this is one thing i noticed from the LC2 manual i briefly looked at yesterday, the power ground and sensor ground are the same wire on the module.  There is no dedicated sensor ground for the analogue output.  Im pretty sure this may be part of the problem if there is not a mechanical issue.

 

Yes that's true, I've dealt with ground offset last season and the actual AFR readings track perfectly between what the ECU sees, my stand alone gauge shows and the Innovate software shows right from the LC2.

 

3 hours ago, Brad Burnett said:

From looking at the log, the fueling seems to follow the map signal.  I would try switching the fuel load from MGP to TP main and setting up a 4d table to account for boost.  

I have had trouble with 3 cylinder motors in the past with very bumpy map signal due to the pulses in the motor.  I even went as far as adding a decently sized vacuum chamber that the map sensor feeds from to dampen the pulses before ultimately switching to TP as load for fuel.

Something to try, I've spent so much time trying to make sense of the VE table reactions last season its crazy.  It takes a lot of effort and time to tune these since not like a car you can't just hop into it and drive around whenever you want.  Can't even tune idle without being in the water since there's no load on the motor while on trailer.

Also had a suggestion on hpacademy forum to lower the response time of LC2 a bit which I'll try as well. Going from MGP to TP as load as well.

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...

×
×
  • Create New...