[Biba69]

What is the 'comfort zone' for ECU's voltage (going to it)? And what will they tolerate? I would imagine temperature might well be a factor.

An easier question is - do they like lower or more likely a higher voltage (obviously up to a point on each end)?

I'm primarily interested in this for the Twin Spark's Motronic unit.

 

[Duk]

What is the 'comfort zone' for ECU's voltage (going to it)? And what will they tolerate? I would imagine temperature might well be a factor.

An easier question is - do they like lower or more likely a higher voltage (obviously up to a point on each end)?

I'm primarily interested in this for the Twin Spark's Motronic unit.

LOL! Why are you reading so much into it?
It is an automotive engine management system, it is designed to operate at automotive system type voltages. The sensors are fed a regulated 5 volts and the computer has an injector open time compensation map to allow for different electrical system voltage as this effects the injector's opening behavior. The processor too, probably has some form of regulated supply.
Most computers have internal protection against over voltage, but obviously there would be a limit to that.

If it's a custom installation, mount it inside the cabin and keep as much radiant heat from the exhaust away from the floor and firewall as you can. Make sure your alternator is in as good a condition it can be. I am fan of newer alternators that can deliver much more current at idle speeds than older alternators. This will give more consistent system voltage at low revs with electrical loads applied (head lights, ventilation fan, radiator fan etc.), but the computer should be able to handle lower system voltages. How low I couldn't say, but if you got a 75TS and switched on everything electrical while the engine idled and then measured the voltage at the battery, that'd be a good indication of what sort of voltages the ECU can work at and it could probably work at lower voltages still, to allow for dieing alternator/battery.
Though low idle speed voltages could be something to consider if you are chasing weird operating behavior (rich idle AFRs) in a custom installation.

 

[Biba69]

"Most computers have internal protection against over voltage, but obviously there would be a limit to that."

And that is the 64 dollar question. My 'current' (built in) voltage regulator only goes to 13.5V and charges my battery to 12.2V. I can get a 14.8V voltage regulator for my alternator. And should the difference remain the same after installing the 14.8V regulator, that would mean it will charge my newish Optima yellow top battery to 13.5V.

I have halogen lights/bulbs. 14V is the 'Rated life voltage' - though I'm not sure if that means that they will burn out quicker.

The TS is not installed, but I felt that if ECU's have a fairly narrow operating voltage, then that could be a problem down the line unless it/they like it on the higher side.

 

[Duk]

My 1988 Toyota MR2 had an alternator problem that had the voltage gauge go nearly full scale and all of the warning lights come on. What this voltage was, I couldn't say (regular charging voltage is/was 13.4v and less than half scale on the gauge), but it did this multiple times and the ECU never suffered from any problems. The problem was a slight oil leak from the distributor/CAS seal, that got into the alternator and somehow caused a problem.

At a guess, and I'd welcome any correction here, I would say 16+volts would be well within the over voltage internal protection range of the ECU.
Ask Festy about more specific for the 'system voltage correction' map inside the ECU for lower than ideal system voltage.

 

[festy]

I would say ideally keep the voltage between 12.5v-14.5v.
The ECU will work on lower voltages, but the coils and injectors aren't as happy to.
More importantly than the exact voltage is how stable the voltage is.
The ECU will log a fault condition if the voltage isn't stable enough, because it makes it very hard for the ECU to calculate dwell and injector pulse width if the voltage is a moving target.

I believe the ECU will refuse to operate below 6v (program logic, not electronic limitation), and I wouldn't recommend going above about 17v - but you'd have bigger problems than your ECU at either of those voltages

 

[Biba69]

festy, thanks. It is good to hear that there is a fairly wide voltage range.

As I've found out, my car won't start at anything under 11.4V (even on a mild day). In this case, I believe it is my MirelliPlex which doesn't like to be short changed, voltage wise.

 

[festy]

The lower the voltage, the longer the coil needs to charge (dwell) - and if the voltage is too low, then there isn't enough time to sufficiently charge the coil before it has to fire.

I've seen plenty of reports of rough running cars suddenly fixed by replacing a failing battery.
I don't believe it's specifically the ECU causing the problem in these cases, it's one of the more tolerant parts of the EMS considering it sources all it's power via an internal 5v regulator.
One theory I have on the failing battery scenario is if the battery has no charge depth, then the voltage drops sharply with load.
So when the ECU calculates dwell and injector pulse, it uses the 'unloaded' battery voltage which might be say 13v.
But as soon as it fires the injectors, the ~4A draw drags the battery's voltage down much lower than the ECU anticipates - so the injector opening time is nowhere near long enough to counteract the sudden voltage drop and the result is not enough fuel gets injected.
Similar with the coils - the dwell is calculated on the unloaded battery voltage. But as soon as the coil is energised, the ~10A load drags the voltage right down and the dwell time is not enough to produce a decent spark.
But as soon as the injectors close and the coil isn't charging, the battery voltage bounces back up, ready to fool the ECU for the next cycle

The alternator doesn't react fast enough to counteract these sudden loads - typical coil charging time is ~0.006 seconds and injectors at low load are open for an order of magnitude less.

Whereas a healthy battery will keep the voltage much more stable during these transient loads, so the ECU's calculations will be correct.

Well that's my theory anyway - not that it actually relates to your original question though, but might help explain why stable voltages and healthy batteries are important for EFI :sad:

 

[festy]

My 17v guesstimate was a bit optimistic, the twinspark ECU will consider any battery voltage over 16.0v a fault condition and go into a limp-home mode.
It must consider the sensor readings to be unreliable at high battery voltages because the coolant sensor, air temp sensor and CO adjustment trimpot are all disregarded and default values used in that situation.

 

[Biba69]

festy, I think I have a 14.7V regulator on order. So unless it fails in the on mode, I should stay well below that.