Megasquirt, enhanced
 

After living with my Megasquirt(s) for 4 years now, I've decided to correct some issues I want.

I made a custom circuit that does the following:

* Digital alternator control, trimmed using the IAT to charge depending on air temperature (IAT)
* Battery voltage light (steady on wjhen engine not running and voltage > 11V, blinking slowly when engine off and voltage <11V OR engine running and voltage < 13.5V, blinking rapidly when engine is running and voltage > 15.5V.
* A/C idle up
* A/C both fans turn on
* First fan turns on at 96C, second fan turns on at 101C
* CEL flashes slowly when coolant > 101C, flashes rapidly when coolant > 104C
* Power steering idle up (when parked and turning the steering wheel)
* Fans turn off when VSS>120Km/h, unless seriously overheating (coolant > 110C)
* Idle valve completely off unless speed < 15km/h (helps engine braking, avoids early overrun fuel cut ending)

Obligatory video showing the A/C idle up function (prevents stalling and idle oscillation). Note that the battery light is blinking - indicating low voltage while cranking:



Dimitris

yet again nice!

Must be a lot of load being drawn on the NB motors. I can run PID idle code without my a/c idle up function on and it has no issues recovering idle, it just dips down 100rpm, and slowly speeds back up. With the a/c idle up code, i have it set so it just blips the rpms up before the compressor triggers on and settles right back to my target rpms.

MS3 also add VSS triggers to the idle code as well.

Tell us more about the circuit.

It uses a microcontroller, just like my VVT stim. Basically it runs a program in C, that checks certain conditions and reacts to them, much like the MS itself.

Oh, I get to dump the PWM board, as my circuit raises the idle valve frequency from 30.5Hz to 488Hz (which is a native hardware PWM frequency when running at 16MHz).

Dimitris

For the record, when it idles without A/C at 1000rpm, the MAP is 28-30kPa and the idle valve is running at 32% duty cycle. With the A/C again idling at 1000rpm, MAP is at 40-42kPa and idle valve duty cycle is ~40%.

What is going on with the airbag light?

I have bucket seats installed, so the airbag computer is complaining because it can't find the passenger side seat sensor.

Are you not able to fake the signal to the airbag computer?

With all of the electronics work you have done, that must be driving you crazy :crx:

Jimitris,
Will the added functionality of the MS3X allow you to do some of these functions without offloading them to a separate board?

Or is it a question of configuring the outputs to do what you want?

Also for the AC idle up, could you have your board intercept the AC request signal from the switch and raise the RPM (idle duty) and then after a few hundred milliseconds engage the AC?

Hopefully, yes, it will allow not just me but everyone else too.

I am doing that already. If you notice carefully, you will see that in the video, I click the A/C button and turn the cabin fan switch to "1" (you can hear the clicking sound) yet the RPMs don't react immediately. First I add some duty, then after half a second I turn on both fans (coolant & A/C) and then after another half a second the A/C compressor.

Having a seperate target RPM with the A/C off (850rpm) and on (1000rpm) is not currently possible with the MS2, so I'm always idling at 1000rpm (the engine likes it better, too).

The reason why the RPMs actually fall by 100 and don't go back to 1000 is that I've set a hard limit (max) on the output duty cycle (40%) so that if some calculation is wrong somewhere, I won't be left with the idle valve stuck open. With the A/C and both fans on, I'm easily reaching that 40%, so I need to raise it up a bit, to 50-52%.

The only thing that's left now, is to first close the idle valve and then after a while, disengage the A/C and fans (otherwise the idle suddenly jumps to 1500 and stays there until the MS2 closed-loop idle code brings it down.

Dimitris

I'm going to remove all airbag-associated electronics and the airbags too, with OMP HTE seats and OMP 5-point harnesses I don't think they will do much anyway.

But really, it would be easier if I just removed the light bulbs in the gauge cluster. :)

Dimitris

I did that in my '90 and it started beeping instead :vash:

So I wrapped the bulb in heat shrink and put it back in. :giggle:

Awesome work with the MS, I'd love to have even the slightest clue as to how to do any of that.

This is what the first prototype I made looks like:

Initial design:

http://users.forthnet.gr/ath/twikicore/dsc_2456_1.jpg

Some components placed on the board, you can see the microcontroller board on the top left:

http://users.forthnet.gr/ath/twikicore/dsc_2443_1.jpg

Connected to the DIYPNP with a couple :D of wires:

http://users.forthnet.gr/ath/twikicore/dsc_2566_1.jpg

This first prototype only has 7-8 wires because it doesn't do:

* IAT based voltage correction
* VSS-based fans and idle control
* Clutch and neutral-based idle control
* Main (coolant) fan control

It's far more crowded on my latest version!

Dimitris