VE Table
My VE table estimate using the estimator from the Megasquirt website:
[need to add table - it's not with me here.]
If anything, it's a little optimistic - which is no bad thing for a 1st attempt as it only means too much fuel.
etc...
- A Lambda table has also been created. This is important, even if the wideband sensor is not installed yet. It is used indeed (eg. lambda=0.95 is used as a *1.05 multiplier!). Note that this is the only sane operation: this way only lambda table needs adjustment when you want to change lambda (not both lambda and VE tables)
- Bought some bosses for the exhaust and lambda should go in this weekend.
Progress so far
Well, the car now runs, and it has taken me to work. Short 6 mile trip. Initial results are good. The fans come on when they should, the engine warms up at a sensible speed, hot start is excelent, the engine is reasonably smooth in operation, but could be better!
I got the wideband calibrated and turned on auto learning and this has done wonders for making the car work well. It's turned into a rather quick car now - but I still want more!
Bad points are:
- The coolant and intake temperatures are still a few degrees wrong (better than before, but still wrong) - Refined again now, and only 2 degrees "wrong" now at 60 degrees. I think I can live with this...
- The VE map is not at all refined, but is getting there with auto learning now turned on
- Acceleration enrichment is not right - but is getting better
- Cold start is better than it was, but not perfect.
note
Keep in mind that the lambda table is always used in fuel calculations, but indeed _also_ for supplying the ego controller with target lambda values, be sure to fill it with sane values.
The reference table you created with VE estimator from Megasquirt site might be VE * Lambda, while ours is pure VE, it can be needed to divide the table you created by the lambda table you put to get an accurate estimate, however during the tuning session it is easily fixed.
Mine is definitely pure VE, I did check that. But it's going to need tuning anyway...
1st Long Tuning Run - 05/02/06
Having got a reasonably set up car, I thought we'd take it to a Piston Heads meet near London so I could optimise the "motorway" bits in the map (ie high speed cruising). I shouldn't have bothered as it was a total nightmare.
We filled up with fuel, and the car wouldn't start leaving the petrol station. (Dead battery and poor terminals I think)
All was going fine after that - all settings on the laptop and LCD were perfect until the engine died without warning. I got it started again, but it was running seriously rich (lambda at 0.75 to 0.95) and very rough. Since we were parked on the M-way hard shoulder, we had to move, so I thought we'd be able to limp home - ot was only 7 miles (12km) or so. The rich condition got worse and worse until the back fires got so bad I thought we'd kill the cat. Then saw sense and got towed home.
Possible causes are:
- The fuel pressure regulator has failed. - I'm borrowing a pressure gauge to test this tomorrow.
- with an LCD and keyboard (and some experience using the menu), it's possible to change req_fuel to limp home (avoiding high RPM and high load of course)
- missfire or leak in the exhaust
- if max ego enrichment is set high (not recommended anyway, see TuningSession/VEtuning) and WBO2 sees fresh air from naughty sources like missfire (eg. a clam plug) or exhaust leak, it will enrich which just makes the problem worse
- injector failed: leaking
- possible but unlikely
- coolant sensor disconnected
- -40C will enrich the mixture
- VEMS has sulked and decided to pulse all the injectors for a LONG time. (Wiring was all tested just before the trip, as I recently installed a new fuseboard)
- dump the config with mcd and mct, save it to disk, and compare (diff) with old known good backup.
- do you happen to remember the EGO correction ?
- if it does, for some reason, on the LCD you'll see the injector pulsewidth (bottom left corner, 10usec resolution) and you'll most often see the reason too (MAP, CLT, EGO, etc...)
- the only possible cause (with a known good, released and previously working firmware and config) that ain't show up on the LCD would be the FETs killed because of disconnected flyback connection. (unlikely because you don't use PWMing and D100 protects from this)
On the plus side, there seems to be no engine damage - compression is good on all cylinders and the plugs are carbon fouled, but not melted or anything. All were even, so it's most likely not an injector problem. The sensors all give sensible values in VEMS too, so I don't think it's died either. It still works the fans and fuel pump, althouh I've not test fired the injectors yet. (A job for later)
Oh well, better write this one down to experience...
Back to: MembersPage/DavidBlades