Pipe sizes.

[LPGC]

Your take on the adjusters is correct.

The stepper motor position (or inline manual valve position) makes more difference to fuelling off idle than at idle. The sensitivity screw makes difference at all loads. So if the idle bypass is closed and the reducer is good but it's rich at idle but lean under load you might have the reducer sensitivity set too sensitive and the inline valve set too closed?

[Gilbertd]

The way I normally set up a closed loop system from scratch is to put a manual valve in set at mid way open. Close the idle bypass completely then adjust the reducer bias to get the mixture correct at idle. Open the throttle to around 2-2,500 rpm and see if it goes rich or lean. If it goes lean that is usually a sign that either the mixer is too big or the reducer is worn (as they will usually go rich at idle when worn), if it goes rich, then the idle bypass needs to be open slightly at idle. Once you can get it pretty close with a manual valve, then fitting the actuator just fine tunes but keeps the default around the same opening.

[Hairyloon]

So if the idle bypass is closed and the reducer is good but it's rich at idle but lean under load you might have the reducer sensitivity set too sensitive and the inline valve set too closed?

Wind it in (clockwise) for less sensitive?
"Inline valve"? Is that not the actuator?

[LPGC]

Usually you wind them in for less sensitive but some work the other way around.

Inline valve is the actuator unless you've fitted a manual inline valve (for testing purposes etc as Gilbert described) instead of the actuator.

Even if you fit a manual valve for testing purposes you could still see lambda readings so at least know if it's slightly lean/rich but you'll never get it to consistently flick lean/rich without the actuator fitted.

I seldom fit a manual valve as Gilbert described but having the inline valve/actuator half way open can be useful for initial setup/testing purposes. Instead of fitting a manual valve in place of the actuator I sometimes lock the default position of the actuator to near halfway open and then set the range of allowed actuator movement at idle rpm to zero and range of allowed actuator movement at above idle rpm to zero, which achieves much the same thing. Not sure if that's possible with your setup though, and not sure if you have a manual valve either. Should still be able to set it up if neither is possible though.

[Hairyloon]

Tried it today, just because I wanted to move the van. Had terrible trouble getting it started. Or rather, the trouble was in getting it to keep running after firing up.

I know that it's typically harder to start on gas, which is why folk start on petrol, but it started well enough before.
Maybe there's a knack, or maybe it'll be better when I've finished tuning it, but I mention it mainly in case it's a clue.

I seldom fit a manual valve as Gilbert described but having the inline valve/actuator half way open can be useful for initial setup/testing purposes. Instead of fitting a manual valve in place of the actuator I sometimes lock the default position of the actuator to near halfway open and then set the range of allowed actuator movement at idle rpm to zero and range of allowed actuator movement at above idle rpm to zero, which achieves much the same thing. Not sure if that's possible with your setup though, and not sure if you have a manual valve either. Should still be able to set it up if neither is possible though.

I have a manual valve, but I took it out when I put the alternator in. Don't see any need for it, and I'd heard that that pipe wants to be as short as possible...
It looks like I can set the actuator to any position: I can set maximum and minimum to the same, or maybe 1 point different, so as to be sure not to confuse the system.

Screenshot_2022-01-03_20-02-24.png (23.96 KiB) Viewed 4829 times

Not sure what the "Optional default lock" is, nor the "Full throttle option", but they may explain in the manual if I look again.

[Gilbertd]

Optional default lock sets it at a set opening. Ordinarily, the default will fine tune itself so it is always in the centre of the opening it is using to keep the mixture correct. So lets say the default is at 120 and the actuator is moving between 110 and 130 when running, all is happy and the default stays where it is. As things change, air filter getting a bit dirty for instance, the actuator will then need to open less to keep the mixture correct (less air means it needs less fuel) so it will be moving between maybe 106 and 126 when running. That will cause the system to learn the mid point and adjust the default to 116 instead. What can happen if the system isn't set correctly is that the average opening at idle is different to the average opening when running. Do a long journey, or let it idle for a long time, and the default opening will learn to suit whichever condition it has been used in meaning it is wrong when the condition changes. By setting default lock you prevent it from learning and lock it at one place with the available steps above and below. Does that system have the option to set different max and min openings at idle or out of idle? Although if you have told it you don't have a TPS it may not be showing the options as it doesn't know whether it is at idle or open throttle.

Full throttle option allows you to set a fixed opening at wide open throttle but again, you need a TPS for it to know when it is at wide open throttle.

For starting, mine is set to start on petrol and switch to LPG when decelerating past 1,200 rpm, so it starts, the revs rise to above 1,200 and it changes over as the revs drop down to idle (650 rpm when hot or around 1,000 when cold). That is with an injected engine though so I don't have to wait for a carb to empty. If I start like that it always starts on the 5th compression stroke, the first compression stroke with fuel in the cylinder (V8 but with a wasted spark ignition system). If I make it start on LPG, it cranks for around 3 times as long as it has to draw the gas in but then fires up and runs normally.

[Hairyloon]

Optional default lock sets it at a set opening. Ordinarily, the default will fine tune itself so it is always in the centre of the opening it is using to keep the mixture correct...

Hmmm. I wonder if that could be used to aid starting? Lock the default to whatever setting it starts best at and unlock once it's running. Would be a bit of a faff to set up, but most of that faff is on the todo list anyway.
That said, it started OK today, but I do need to go back and mess with the tuning again... With a laptop this time.

For starting, mine is set to start on petrol and switch to LPG when decelerating past 1,200 rpm, so it starts, the revs rise to above 1,200 and it changes over as the revs drop down to idle (650 rpm when hot or around 1,000 when cold). That is with an injected engine though so I don't have to wait for a carb to empty...

Nor for it to fill. Mine has a mechanical fuel pump, so starting on petrol will be a real nuisance. Though if I'm going to have to do that, I'll probably find an electrical pump for it.

[Gilbertd]

It will always start at the default as there is a delay, or at least there is on mine, for the lambda sensor. As most won't give an output until they are warmed up, you set the delay (I think the default is 10 seconds), it starts and runs at the default and then adjusts to the lambda output once the delay period has ended. If the default is locked you may need to set the max and min settings to give it sufficient headroom to keep the mixture correct when running.

[LPGC]

If you intend on starting on gas (which would be a very good idea particularly as your petrol system is a carb and you don't want to wait for carb float bowls to fill or empty etc) then it can be beneficial if the idle bypass is open a little and you turn the ignition on/off a few times before attempting to crank the engine. This primes the pipe from the reducer to the mixer (and inside the air inlet ducting) with a little gas and can prevent lean mixture (and backfires) at initial cranking.

No idle bypass set and there won't be any gas flow from the reducer until there is a vacuum signal from the mixer. Switching the ignition on/off a few times is because the ECUs usually open the solenoids for a moment when you first turn on the ignition (if set to start on gas) but without an rpm signal from a running/cranking engine the solenoids will very soon be closed, so switching on/off a few times gives more than one short priming pulse.

If an engine doesn't start soon during cranking on gas don't keep cranking... There will be some gas being pumped through the engine and into the exhaust even if the mixture is too lean or rich for the engine to fire and there can come a point where if a cylinder does fire then when it's exhaust valve opens a flame hits the gas in the exhaust and likely blows all the exhaust boxes apart (or inlet ducting / air filter). A guy once blew the bonnet off his Ford Exploder (Explorer) after trying to use a mixer system on it and I was once called out on Boxing Day to a guy who'd had a mixer system fitted on his Landcruiser with plastic inlet manifold (big hole in his manifold).

[Hairyloon]

If the default is locked you may need to set the max and min settings to give it sufficient headroom to keep the mixture correct when running.

My thinking was to lock the default only for starting (if it turns out to be helpful), then unlock once it's running.
That would involve finding out what the control codes are to do that, and setting up some kind of microcontroller to send them... or reprogramming the ECU itself, but unless they've released the source code, I can't see that happening.

If you intend on starting on gas (which would be a very good idea particularly as your petrol system is a carb and you don't want to wait for carb float bowls to fill or empty etc) then it can be beneficial if the idle bypass is open a little and you turn the ignition on/off a few times before attempting to crank the engine. This primes the pipe from the reducer to the mixer (and inside the air inlet ducting) with a little gas and can prevent lean mixture (and backfires) at initial cranking.

That reminds me, I'd been meaning to ask about this, but keep forgetting: the ECU doesn't come on until I crank the engine.
That doesn't seem right at all. I don't have to crank it far: literally a fraction of a second. I'm guessing it is the ignition pulse that wakes it up?

No idle bypass set and there won't be any gas flow from the reducer until there is a vacuum signal from the mixer. Switching the ignition on/off a few times is because the ECUs usually open the solenoids for a moment when you first turn on the ignition (if set to start on gas) but without an rpm signal from a running/cranking engine the solenoids will very soon be closed, so switching on/off a few times gives more than one short priming pulse.

I think that mine doesn't open the solenoids until you start cranking, though it seems a little slow to close them again when the engine dies.

If an engine doesn't start soon during cranking on gas don't keep cranking...

Good advice. Though mine is firing well enough, the difficulty is in keeping it going more than a couple of seconds.
I may go have a session with it shortly: I keep forgetting to take a laptop with me...

[Brian_H]

That reminds me, I'd been meaning to ask about this, but keep forgetting: the ECU doesn't come on until I crank the engine.
That doesn't seem right at all. I don't have to crank it far: literally a fraction of a second. I'm guessing it is the ignition pulse that wakes it up?

Definitely not right, where do you have the switched live being picked up from? Obviously you don't have the option of picking it up off a petrol injector if its on a carb engine. Thats what should wake the system up, the majority of its power comes from the battery feed.

The one I've brought (not got any further than connecting it to a battery and pc) wakes up as soon as both lives and the ground are connected to the battery.

Needs to be somewhere thats on with both ignition switched on and whilst cranking if you want it to run off it from cold I'd expect.

[LPGC]

Locking the default can be a good idea on some engines/setups but probably isn't as useful for starting. The most common reason I lock the default is because I want to set a range of movement for the stepper, e.g. if you got a setup that causes idle rpm surging due to 'overshoot' of the stepper motor causing the stepper to move a little too far toward (say) the lean position then if you don't also lock the default there might not be much point in limiting the range of stepper movement at idle, because if the default changes to a lower position the stepper might again move to a position that causes lean mixture at idle and the surge occurs again. The real problem might be that the lambda sensor is too far back in the exhaust or mixture might change a bit due to temperature/density changes but regardless of what causes such surge you want to dial it out where possible.

Some ignition switches (even on carb engines) turn off electrical loads (lights etc) during cranking the engine, to be able to start on LPG the ECU's switched live needs to be connected to a live that's present with the ignition in the normal run position and also in the crank position. There is a wire in the engine bay that does the same, the coil positive wire, but the coil may have a ballast resistor and even if it doesn't have a ballast it's not a great idea to connect directly to coil positive due to the voltage spikes that are likely to be present. But you can avoid such drawbacks by using a relay to switch the switched live from coil positive.

There may be a difference between the system being awake and talking to the laptop and the system opening LPG solenoids though... These systems are not designed for carb setups, they are designed for older fuel injected vehicles, they may not necessarily pulse solenoid coils until there is an rpm signal. When the 175's are set to start on gas they will pulse solenoids for a moment before closing them if there is no rpm signal but this ECU might not do that.

[Gilbertd]

My thinking was to lock the default only for starting (if it turns out to be helpful), then unlock once it's running.
That would involve finding out what the control codes are to do that, and setting up some kind of microcontroller to send them... or reprogramming the ECU itself, but unless they've released the source code, I can't see that happening.

Or you could just wire a pushbutton to apply 12V to the blue wire so you poke that for a couple of seconds before starting the engine.

[Hairyloon]

My thinking was to lock the default only for starting (if it turns out to be helpful), then unlock once it's running.
That would involve finding out what the control codes are to do that, and setting up some kind of microcontroller to send them... or reprogramming the ECU itself, but unless they've released the source code, I can't see that happening.

Or you could just wire a pushbutton to apply 12V to the blue wire so you poke that for a couple of seconds before starting the engine.

You and your sensible, simple suggestions...

[Hairyloon]

That reminds me, I'd been meaning to ask about this, but keep forgetting: the ECU doesn't come on until I crank the engine.
That doesn't seem right at all. I don't have to crank it far: literally a fraction of a second. I'm guessing it is the ignition pulse that wakes it up?

Definitely not right, where do you have the switched live being picked up from?

Good question... it was weeks ago that I did that, so I can't remember, but I would expect it's from the live side of the ignition coil: it seemed the easiest place.

The one I've brought (not got any further than connecting it to a battery and pc) wakes up as soon as both lives and the ground are connected to the battery.

We replaced a diode in the dead unit and it's working now, so maybe I should set that up on the bench to play with...

Needs to be somewhere thats on with both ignition switched on and whilst cranking if you want it to run off it from cold I'd expect.

The ignition wire should do that.

There is a wire in the engine bay that does the same, the coil positive wire, but the coil may have a ballast resistor and even if it doesn't have a ballast it's not a great idea to connect directly to coil positive due to the voltage spikes that are likely to be present. But you can avoid such drawbacks by using a relay to switch the switched live from coil positive.

I was thinking about fitting a relay, in fact I got as far as mounting it before I realised that I hadn't got any spade terminals. Another option is the shutoff solenoid in the carb, especially as it doesn't actually seem to do anything...

There may be a difference between the system being awake and talking to the laptop and the system opening LPG solenoids though...

When I said "Awake", I mean that the lights come on.

[Hairyloon]

Ok, I just had a little session with it and a laptop, and I'm puzzled: if it's running rich, why wouldn't it close the valve down?

[Gilbertd]

It should as long as the lambda sensor is showing it is rich (going red and voltage at or near 1V). Unless you have set the limits too tight and it won't allow it to close down enough. From your screenshot it appears slightly different to the other AEB units which have a number of steps above and below default setting. If that is what the display above is showing, rather than a minimum opening, then it will only allow the actuator to close down 20 steps below default.

[Hairyloon]

It should as long as the lambda sensor is showing it is rich (going red and voltage at or near 1V). Unless you have set the limits too tight and it won't allow it to close down enough.

When I say it's rich, I'm basing that assessment on the lambda display. It had plenty of scope to close it down further than it was doing, and had done at some other point in the session.
Could it be running rich to deal with the low revs?

From your screenshot it appears slightly different to the other AEB units which have a number of steps above and below default setting. If that is what the display above is showing, rather than a minimum opening, then it will only allow the actuator to close down 20 steps below default.

No, I think it can close down to 20 steps above completely closed. Presumaly I can change that if I want.

Something else that was odd was that it seemed to be stuck on a default of 221. I say "Stuck", I didn't try very hard to get it to change: once I'd noticed, I set it to 121, and it seemed to behave somewhat better.

[Gilbertd]

It will have learnt the default so if it thinks 221 is correct going from the lambda sensor voltage, then that is what it will have set. If you set it lower, then the lambda sensor will show lean (green with a voltage below 0.5V), so it will open up until the lambda starts to go rich. It should then wander up and down either side of the default while the lambda output will flip between lean and rich. Just had a look at the manual and the minimum is, as you say, the number of steps over fully closed so with the settings as they are it should have the full range of movement.

[Hairyloon]

It will have learnt the default so if it thinks 221 is correct going from the lambda sensor voltage, then that is what it will have set.

Yes, but it learned that from an incorrect setup. It's surely better to start it off from the middle to start learning afresh.

With the settings as they are it should have the full range of movement.

Except for the last 20 steps down to closed...