starting a vehicle engine with a microcontroller

Anyone any experience of this?

I'd like to automate a car engine (petrol) driven air compressor, the engine has an ecu which controlls the injectors and ignition (from the original donor car).
I was thinking of just closing a relay for the ignition (wired into where the ignition switch was on the loom) then close another relay to bring in the starter, then measure the battery voltage with an a to d on a pic, when the engine fires the battery voltage rises by a quite a bit due to the releived load on the starter, and when this happens release the starter relay as the engine has started.
The battery voltage drops to around 6 or 7 while while the engine cranks, then comes up to around 10 when the engine fires (before the alty kicks in).
And I'd have to include a timer so the starter doesnt remain on too long, and on top of that maybe a number of tries counter resulting in an alarm if the engine wont go.

You can check if the engine has started by checking the frequency of the ignition coil or injector pulses. Many ECUs have a seperate clean 0/12v pulse to drive the tachometer.

Put a secondary cutoff on a seperate series relay for the starter motor - something independent of the main start circuit which will act as a failsafe if the starter is operated for more than 10 seconds (for example).

Also hook in a monitor for the oil pressure switch - if the engine oil runs low then you can cut out the engine.

All pretty simple stuff and a lot of it is already done by remote start car alarm systems.

Commercial vehicles with a fancy start button cut out the starter as soon as the engine fires, I'd assumed it was done on starter current, but I spose the starter current is directly related to engine speed, and its engine speed we go off when we turn the key on our cars every day.
I'll try that, maybe look for a 10% change in speed over a short time, the ecu doesnt have a pulse o/p I'll use an indirect pickup from the coil ht lead.
And good advice on oil pressure, I can just use the engines existing switch, it does actually have an oil level sensor, I can read this before the engine starts.
I dont really want to go to the expense of a battery isolator relay, those things are huge, however I can put 2 relays in line with the starter solenoid and just use one to operate the starter, and use the other if the starter relay doesnt disengage (welded contacts).
I've had a starter relay seize on before, not very nice.

The best test is the oil pressure switch... This is negative as soon as the pressure is above 1.8bar ( I doesn't get there if the engine hasn't started ) We use this method on cranes...

Now thats an idea.

I had a sierra that was in a state, often the oil light went out on the starter before she fired but it took a few seconds, however this is a better engine and starts willingly, I can lash that setup with a relay and see if the idea works.

Look at some generator starting panel control instructions. Oil pressure/Level is one sensor that has to be checked before and during running.

Water temp and coolant level is also monitored.

I THINK that most systems run the starter for a fixed time with a fixed number of trials with a break in between the trials. Oil pressure is checked to see if the engine started.

Look at pdf page 45 of this: https://www.electro-tech-online.com/custompdfs/2012/10/A030G173.pdf 3 start cycles of 15 sec with 15 sec of rest is the default.

Take a look to see what's appropriate to monitor control for the engine and what's appropriate for the compressor. Don't forget a lockout/ tagout possible emergency stop switch.

The question is, what happens to the starter motor if you don´t disengage it quick enough after the motor has started?
I know my mother managed to seize a starter by keeping it engaged for too long, so when the motor reved up the windings got stuck to the armature.

kubeek said:

The question is, what happens to the starter motor if you don´t disengage it quick enough after the motor has started?
I know my mother managed to seize a starter by keeping it engaged for too long, so when the motor reved up the windings got stuck to the armature.

Click to expand...

As an ex mechanic... I have seen ring gears worn to near smooth because solenoids have failed... The starter doesn't turn but the two gears run together ( very noisy ).
I have never seen an armature destroyed because of same....

They make starters on things like bobcats that use a high speed motor in the starter they will not last long if they don't kick out soon after the start of the engine I've seen them melt down with stater switch failure.

And there a lot of these on small engines

Thats my concern not disengaging the starter in time, and thats why I was thinking of monitoring the current to it, as soon as the engine begins to start the current will drop giving a very early signal to drop out the starter, probably a lot sooner than a person could.

The oil switch idea worked I tried it, allthough maybe a little slow at dropping out the starter, might be a slow switch or starter.

Motorbike starters have a one way clutch, as soon as the engine fires the clutch disengages, inappropriate here though.

Mr rogers... this is a bendix starter so the armature would be driven by the flywheel if it didnt engage, I spose if the engine revved to 3k which it does then the gear ratio would make the starter screm and theres a good chance of it exploding, like old dynamo's did.

These auto start systems have become very popular in the northern US. My sister has had one on her car for some time.

The oil pressure switch is a secondary indicator in that oil pressure has to build in the block. If you want faster response monitor the engine RPM. Both should work.

You mentioned a timer. This application screams for a micro controller so you can apply some logic to the process.

Yes either a 16f628 or a 16f876, and an rfmodule to talk from the compressor to the control panel.
I plan several lockouts, oil pressure, oil level, water temp, fuel level, over revs, low or high volts, estop button, high air pressure, broken drive belt and lack of comms to the panel are what I've thought of so far.
Over pressure is a switch hard wired to the ignition supply so isnt software dependant, and theres a mechanical over pressure releif valve.