25 MayMegasquirt Two Months Later…
Lots to report since the last post – the car has been in use as a daily driver once again, and there have been quite a few things I needed to iron out.
Firstly, I had trouble with the fuelling percentages – they weren’t repeatable. I maintain that the SD card logging feature of the MS3 is an absolute godsend, as I relied on it heavily to diagnose most of the starting, idling and running issues I’d had with the fuel injection and also the operation of components such as the idle valve.
It turned out that the MAT (manifold air temperature) sensor wasn’t working correctly. The datalogs showed the temperature randomly changing between either 30 degrees or 60 degrees. I hadn’t had chance to install it properly, and it was merely cable tied inside the engine bay. Since this was an open element sensor (used for faster response on turbo cars), I think perhaps dirt may have got inside and damaged the sensor. Even when you think you’ve disabled any sort of MAT correction in tuner studio, there are still adjustments being made to the fuelling based on air density predictions, so I completely unplugged the sensor – now the fuelling doesn’t suddenly go rich or lean across the board.
Another big annoyance was that the alternator seemed like it was on it’s way out, but I couldn’t be sure if that was the problem or the battery. I eventually realised the problem when the radio started cutting out as I was coming into work, and couldn’t get enough current to start it again when leaving later that night. After jump starting it 3 times on the way home, the battery and alternator were both clearly toast, so they were replaced with a new Numax 015/038 unit and a 75A alternator. To ensure the wiring could cope, I ran a couple of thick 28A brown wires from one of the alternator’s spare B+ terminals to the solenoid, as well as connecting the old Lucas 3 pin connector. They both work like a champ now, and can cope with anything I throw at them (heated seats, subwoofer amp, etc).
The idle valve hadn’t worked when first installing everything (Bosch 0 280 140 516 PWM two wire unit), despite using someone else’s published frequency and duty settings. After using the test mode I eventually got it to work, albeit at quite a high operating range (67% minimum at 99hz). I later found out that this was because the MS3X output still requires you to wire a diode between the two pins of the valve, so after installing one that dropped the range down to around 30-60% and made the operation noticeably smoother.
One other problem I hadn’t managed to solve was getting a solid idle when cold, and this was stopping me from using closed-loop idle settings on the Bosch PWM idle valve. Even warmup-only mode was rough when cold, and no matter how much I tried tuning the fuel map to smooth out the values around where it would idle, it was a pain in the arse getting it to idle and start reliably. I upped the MAP lag (basically this is smoothing the vacuum signal) value from 50% to 70% and reduced the MAP takeoff restriction from 2mm to 1.5mm which helped a bit, and also stuffed some rubber coolant hose inside the idle valve inlet/outlets to reduce their effective sizes (which now has limited the valve’s max opening to 2000rpm).
Ultimately though, this problem was only solved by changing the injection angle in the sequential injection settings. Moving it away from 360 degrees made the engine idle a lot smoother, and reduced the amount of variance in the MAP signal. I also discovered the cranking rpm was set too high (at 700rpm), I’m guessing this is something I’d messed with when I had trouble getting it to start when first installing the Megasquirt. Changing that to a lower value and fixing the injection angle means the starting and idling of the car is now as smooth as a modern car, and it actually runs so much smoother when driving the thing too. Result!