ECU tuning - difficult hot soak restarts?

[joshuaho96]

Curious to know if anyone else has run into this issue before. I helped a friend tune an EFI conversion a little recently and an issue we ran into is that after getting coolant temp into the 90C range, after shutting the engine down and letting the whole engine bake for 5 minutes it wouldn't crank over anymore. Adjusting the cranking fuel pulse width did nothing. Surprisingly what solved it is setting a long fuel pump prime. He's running an MS2 or something like that so it wouldn't allow him to prime for longer than 3 seconds but keying on and off three times to get about 9 seconds of priming has reliably allowed it to restart now in this hot soak scenario. Anyone know a good explanation for exactly what's happening here? My working theory is that the fuel in the rail has boiled from heat soak and the pump needs to cool off the rail to have functional fuel injection again but I'm curious to know what others think.

[Dose Pipe Sutututu]

26 minutes ago, joshuaho96 said:

My working theory is that the fuel in the rail has boiled from heat soak and the pump needs to cool off the rail to have functional fuel injection again but I'm curious to know what others think.

This is my understanding as well, I generally overcome this in the Haltech by setting a very high injector prime pulse, i.e. the very first injector pulse when the ECU powers up. 

It only seems to be an issue (with my experience) with larger injectors on E85. I've only ever experienced this issue with my own car which has 1480cc (often called 1650cc) injectors. Most of the cars I've tuned in the past generally have the run of the mill 1000cc Bosch injectors.

[Murray_Calavera]

Just checking, is there a check valve in the fuel pump or installed in the line?

[MBS206]

Being that this is a conversion, there could be so many things going on here.
You could have a bad pump, wiring issues, it could be how the rail has been plumbed, where the Fuel Pressure Regs are located instead, etc etc. 

Without detailed explanation, from vehicle, pumps, wiring, how the whole fuel system is as a whole, the engine of choice, etc etc. Pretty hard to diagnose from online.

As this has been what you've just been tuning, you could be low on fuel, pump may not be able to properly build pressure etc, could be an internal leak inside the fuel tank, and hence when its hot it might be struggling to push fresh fuel through.

[The Bogan]

Does it have engine fault codes?

I had similar with a bad CAS or CPS

Started fine when cold, but wouldn't restart at temp

[joshuaho96]

4 hours ago, Murray_Calavera said:

Just checking, is there a check valve in the fuel pump or installed in the line?

Supposedly yes, but I have a theory that both are suspect. Haven't been able to get a pressure gauge on it to monitor what it's doing after shutdown as I got busy with other things since then but I suspect it would be interesting.

7 minutes ago, MBS206 said:

Being that this is a conversion, there could be so many things going on here.
You could have a bad pump, wiring issues, it could be how the rail has been plumbed, where the Fuel Pressure Regs are located instead, etc etc. 

Without detailed explanation, from vehicle, pumps, wiring, how the whole fuel system is as a whole, the engine of choice, etc etc. Pretty hard to diagnose from online.

As this has been what you've just been tuning, you could be low on fuel, pump may not be able to properly build pressure etc, could be an internal leak inside the fuel tank, and hence when its hot it might be struggling to push fresh fuel through.

This is a series 3 Jaguar E-type convertible. Prior to this it had Stromberg constant depression carbs which were absolutely horrific between that and the bizarre early 70s emissions vacuum retard distributor. I definitely agree that the conversion nature makes it hard to actually disambiguate what's going on but at the time of the hot soak we had come back from a fresh fill so the tank was full. The whole fuel system was just redone for the EFI conversion so I would be kind of surprised if it's leaking inside the tank but anything is possible.

21 minutes ago, The Bogan said:

Does it have engine fault codes?

I had similar with a bad CAS or CPS

Started fine when cold, but wouldn't restart at temp

At the time of the tuning session it still ran the factory distributor, actually the ECU didn't have any engine position so there's no concept of anything like fuel injection timing or ignition sync. Verified the spark plugs were in fact firing too.

[The Bogan]

Distributors have a resistor thingie that can cause a similar fault

I had a old Ford Bronco with a 351, once hot if you shut it down it wouldn't restart

The heat would separate the connection, it contracted when cold and would work again 

[Kinkstaah]

I had similar (though probably red herring issues) here with my LS1 on hot restarts, but it wasn't fuel related as much as it was IAT related. The IAT/density model is too aggressive at pulling fuel out when the air temp increases, so it was hyper lean for the first few seconds and cranking.

You could solve this by adding a ton of fuel (as above) in hot conditions, or by correcting the IAT curve or the calculations or whatever is available in the ECU.

[Dose Pipe Sutututu]

This is how I get around it, probably not the most ideal. Also because the fuel temp sensor is part of the flex sensor, you're not able to get the correct fuel temperature at the rail so the entire fuel density vs. fuel temp becomes irrelevant when the fuel soaks in all the heat from the rail. If I wanted actually properly use the fuel density table, I really should install a fuel temp sensor on the fuel rail itself.

And this is my fuel density table vs. fuel temp vs. ethanol concentration (used BP data and blended it with available E85 data found on the net).

 

[MBS206]

4 hours ago, Dose Pipe Sutututu said:

This is how I get around it, probably not the most ideal. Also because the fuel temp sensor is part of the flex sensor, you're not able to get the correct fuel temperature at the rail so the entire fuel density vs. fuel temp becomes irrelevant when the fuel soaks in all the heat from the rail. If I wanted actually properly use the fuel density table, I really should install a fuel temp sensor on the fuel rail itself.

And this is my fuel density table vs. fuel temp vs. ethanol concentration (used BP data and blended it with available E85 data found on the net).

 

On a setup where the fuel reg is at the other end of the rail, and hence returning to tank from the engine bay, Apart from at restart, fuel temp at the rail, should be so close to fuel temp at your flex sensor, it would be negligible.

Reason being, the amount of fuel flowing past at speed through the main item (fuel rail) that will heat it up, and pulling heat away.

By the time you've primed the fuel pump for 3 seconds, and then gone straight to crank, for say another 2 seconds, that fuel is matching pretty much the temp of your flex sensor. The rail may be hotter, but the fuel is barely increasing due to the speed it's going past at.

The part where you get "fuel heat soak" is when you don't have enough fuel in the tank, and just keep pumping it past a hot thing really quickly, and it's picking up small amounts of heat more than it can drop off. Our tanks are plastic these days, not great for heat transfer out...

Think of it like turning the kitchen tap on and waiting for the cold water to come through in summer. It gets through and rapidly changes from hot, to cold. This does take a while in some houses, as the water that's truly cold, is a LONG way away from your tap. Even if you held an Oxy Acetylene torch to your tap, the amount of water rolling through won't make the outlet temp change much at all.

 

You also need to be careful where you install a temp sensor, and how much you're relying on that temperature. Temperature sensors have a first order response, which means they can't be super quick. Secondly, most I've seen are a metal sensor, in a metal housing. You screw that into a metal item, you're more likely to be measuring the metal temp you screwed into, than the liquid you want to measure. IE metal could be hotter from other factors, or could be getting more cooling and hence reading lower.

 

Even from your data, from 10c to 90c, that's only a 7% variation when on 100% fuel. At startup, I bet you could easily remove 15% fuel from your cranking table and it would still start up, especially when warm.

 

A returnless system obviously does have more heat soak type issues in it, as fuel isn't leaving the rail, except through the engine...

[wardiz]

How does look like your fuel post start correction ?

Is it really usefull to have the temp in y axis ? 1 litre of fuel weightinh 775g @ 40C is going to weight the same @60C, isn't it ?

Edited August 21, 2023 by wardiz

[Dose Pipe Sutututu]

1 hour ago, MBS206 said:

Apart from at restart, fuel temp at the rail, should be so close to fuel temp at your flex sensor, it would be negligible.

Ahhh you see, my flex sensor is on the feed side of the main tank to surge tank, thus only is sampling the main tank's temperature. Once running and everything is flowing around as it should, then there's probably negligible difference in temperature, however on start up there would be a difference.

59 minutes ago, wardiz said:

How does look like your fuel post start correction ?

Is it really usefull to have the temp in y axis ? 1 litre of fuel weightinh 775g @ 40C is going to weight the same @60C, isn't it ?

I'll post that up when I'm back on my tuning laptop.

Liquid density changes as it gets warmer, suppose the 20°C difference isn't going to make a huge difference however more data would provide for accurate fueling so it doesn't hurt to have (plus I just copy paste the same table into all other tunes I do).

[wardiz]

1 hour ago, Dose Pipe Sutututu said:

Liquid density changes as it gets warmer, suppose the 20°C difference isn't going to make a huge difference however more data would provide for accurate fueling so it doesn't hurt to have (plus I just copy paste the same table into all other tunes I do).

The change is probably small enough to have no impact ( the tuner tutorials I follow never mention that). However the temperature mais have a greater impact on the atomisation of the fuel than the change of density.
 

But my main though is that more fuel may be needed at startup than regular gazoline. Thus do you have a kind of compensation table for the fuel post start offset relying on the fuel composition ?

[joshuaho96]

23 hours ago, Kinkstaah said:

I had similar (though probably red herring issues) here with my LS1 on hot restarts, but it wasn't fuel related as much as it was IAT related. The IAT/density model is too aggressive at pulling fuel out when the air temp increases, so it was hyper lean for the first few seconds and cranking.

You could solve this by adding a ton of fuel (as above) in hot conditions, or by correcting the IAT curve or the calculations or whatever is available in the ECU.

I think he ran into similar issues, once it actually kicks over he experiences lean running for a little bit before it corrects. I suggested just papering over that problem with after start enrichment at high coolant temps but the whole thing is getting redone with a crank sensor because he's sick of fighting the distributor which was constantly misfiring at idle but weirdly not at cruise or full throttle. I'd see something like a full point of AFR lean-out in response to a misfire so I disabled all closed loop corrections even after I was done messing with VE tables. He's also using one of those old style AEM controllers that blasts the heater the second it gets power from the ignition switch so the sensors aren't going to last very long regardless and there's no real failure detection in those controllers either to my knowledge.

Once he gets the new setup running I'm sure I'll have new and exciting issues to learn about.