Author Topic: Closed Loop VE Tuning  (Read 15654 times)

Offline rob@vems.co.uk

  • Hero Member
  • *****
  • Posts: 3115
  • BHP: 49
    • VEMS Forum
Closed Loop VE Tuning
« on: June 17, 2010, 06:04:55 am »
Jorgen's Guide to Open Loop tuning - from http://www.vems.hu/wiki/index.php?page=TuningSession%2FClosedloopTuning

Closed loop tuning guide
This is the closed loop version of this guide, the recent development of fast high precision wide band lambda sensors has made a number of more or less well developed closed loop tuning procedures to emerge. This is what we feel to be the most scientific of them all.

First of all, most people use far too many load bins. It will make the tuning time consuming. I suggest that you put load bins at: 20, 30,50, 70, 90, 102, 120, 140, 160, 180, 200 and 250 kPa. Note that this applies to boosted AND normally aspirated engines. Using all 12 load sites in the normally aspirated range will only increase your workload.

It's very important to understand that any WBo2 readings taken when the engine has missfires, bogs or just acting strangely is unlikely to be correct. For example a missfire reads lean! because the sensor only measures burned petrol, liquid petrol is ignored and only the air will be measured. Under conditions like this you need to read the plugs to get it right.

As this is the closed loop guide you need to enable closed loop lambda control: Settings->EGO(Lambda) control. Set Lean limit(%) to 25% and Rich Limit(%) to 1%.

The reason that you don't want to allow the ecu to enrich the mixture more then 1% is that you may run into missfires and such things during the initial tuning procedure, as you can read above a missfire reads lean and that would trich the ecu into enriching the mixture even furter which would usually lead to a very bad missfire condition that coul foul your plugs and possibly even your Wbo2 sensor!

You can also experience missfires from VERY lean conditions, reading the plugs will show you if this is the case. The best way to learn what is going on is by reading the sparkplugs.

We assume that the injector opening modeling parameters are ok, with small injectors it's easy to get it close enough but with large injectors it's fairly hard. With some luck you have injectors that we have documented the opening characteristics for. (As of December 2005 there are no properly verified entries.)

The closed loop tuning procedure
Make sure that the engine is properly warmed up, this includes that the oil temp is in the working range.

We start with the extremely low load sites.

Start off by letting the car idle, now tune the VE at the loadsite the engine use until the EGO correction is between -7% and 0%. You will probably need to blipp the trottle a few times to clear up the engine when doing this, expecially when going from a pig rich base tune.

Now increase the rpm somewhat, 2k or so is fine. Do the same here, tune VE until EGO correction is in the proper range, -7% to 0% applies here too.

Continue doing all low load sites you can reach in the same way.

It's now time to take the car for a spin, mentally prepare yourself to hold a few load sites and start the logger.

50kPa: 1100rpm, 1600rpm, 2200rpm, 2800rpm. (or the load sites that are closest to those!)

Now check the log and identify the areas where you have held the above load sites (and any other load sites that you have held accidently), correct the VE entries in these sites to get in the middle of the 7% to 0% range. For example if you have 20% EGO correction in a loadsite you decrease the VE entry at that loadsite with around 17%, trying to hit 3% correction. Repeat this for all load sites where you have held a steady load and take the car for a spin testing those loads again, repeat adjust-test-adjust until you are in the -7% to 0% EGO correction range on these load sites.

Now examine the small number of tuned load sites around 50kPa and verify that they have values in the 100-200% range. This should not happen but if they aren't something has gone wrong with the req_fuel calculation and you want to adjust req_fuel and start over with the tune. Maybe you used one of the old formulas that has been used for req_fuel, maybe the req_fuel doesn't match the way you drive your injectors or maybe you didn't know the correct injector flow. In any case it's ok to just change the req_fuel to get the ve in the right range. If the VE entries are too low you want to decrease req_fuel and if the VE is to high you want to increase req_fuel.

EX. You have found that you need to set VE to 50% at 50kPa and 3000rpm. To hit the middle of the 100% to 200% range (150% which is about right for that loadsite) you want a three times higher VE. That means you want to divide req_fuel with three. For a change this large you should first check that you aren’t injecting at a higher rate then is recommended in the manual. We recommend that you either inject in bank-bank mode (which by ours and most american manufacturers means that you have two injector banks that inject in an alternating fashion one bank per crank revulution.) or in sequential mode (which by all definitions means that there are one injector bank per cylinder and that each injector inject every other turn of the crank and that the injectors are controled in a sequential manner.)

When the 50kPa sites above is tuned fairly well you can go to the next three rpm sites, still at 50kPa: 3400, 4000, 4600.

Again repeat the 'test, adjust, test' cycle.

You are now ready to go to the next load level (70kPa) You need to be more careful now as problems can start causing engine damage now. Just let it run steady state for 2-3 seconds or so on each site the first time until you know that you are starting to home in on the settings. You also need to listen for detonation now and read the plugs for signs of detonation between the runs.

Repeat the 'test, adjust, test' cycle.

Continue repeating the above procedure for all the load sites until the car is tuned at all loads and rpms.

Jörgen Karlsson