trying to diagnose a fuel pump or FPR problem

[thechuckster]

hi folks,

am still trying to get my 18RTE to idle with some stability - but am having a ongoing battle with it stumbling and nearly stalling every 7-8 secs (unless i pump the thottle).

My logs show the O2 reading high (deliberately tuned rich), then it start to taper towards stoich, then suddenly plummet to extremely lean - this would indicate that the engine is running out of fuel.

But... if i pump the throttle - this suddenly increases the pulse-width - the engine stays running but with a bit of coughing and popping at much lower rpm... the AFR doesn't rise instantly - it seems to take a few goes at enriching things before the engine catches and we're back to a fast idle (~1200rpm with 6-8% throttle open)

... idle then stablises and we repeat this all again in about 8secs.

If i dont touch the throttle, the engine will stall about a second after the O2 bottoms out.

My query is: would a dodgy fuel pump have this kind behavior, or the FPR?

I can email the log files of those interested - you'll need megalogviewer to make sense of them.

petrol: i don't have a surge tank - but have just put 20 litres of new ULP into an empty tank - it was doing the same when tank had about 40+ litres of 2-year-old BP ultimate.

The guage on my FPR does show static line pressure at what i think is 38psi with the pump on - once the engine is running, it moves about in tandem with MAP. Pressure does drop away when i turn the pump off (is this normal?).

I do have a an issue with the alternator, but voltages at the ECU are ~11.8 Volts when running so i dont think it's a problem - it's not affecting pulse widths and sparkage.

Cranking is slow (~300-350rpm) but it catches easily.

It's quite hard for me to get the car started and me out to the engine bay to watch the fuel pressure... it invariably dies before i can get to the throttle body and give it a bump...

It did have a stable idle previously - but am not sure whether changes in timing and fuel maps have killed that stability... was too many ECU changes ago ... ;(

Prior to this i've:
-dumped contents of tank (to remove risk of it being old, crap petrol)
-put new loom from ECU to engine bay
-checked all sensor wiring eleventy-trillion times - i can safely say these are all good now!
-replaced wires to TPS and O2 with shielded cable
-changed dwell from 4ms to 3 ms (was thinking that the spark might be not healthy enough)
-pulled timing back (at this fast idle, it was at ~30 degrees, is now about 24 degrees)

For the curious, the ECU is a MegaSquirt running MSnS-Extra, v.024s code, 22RE VR sensor, bosch generic electronic ignition transformer coil, 4agze injectors with resistor pack from a FWD toyota, malpassi rising rate FPR, ex-RB20 turbo, external bosch fuel pump using carbi tank pickup.

suggestions most welcome - wow - sorry for the long post too

cheers and tia,
Charles.

[chris davey]

Doubt it will help but 24deg timing at idle? Mine is 15.

[slide86]

you shouldnt loose fuel pressure when you turn the engine off. it should keep pressure in the rail. some pumps have an anti return in the pump to stop the fuel from running back into the tank when the pump is not on.

also you say that there is only 11.8v when the engine is running. it should be more than this shouldnt it? most charging systems are round the 13.5 to 14.2 mark.....

one test you can do is to clamp the return line with the engine running and observe the fuel pressure. it should max out at idle, indicating that the pump is still strong. if it doesnt then you may have a weak pump.

[thechuckster]

thanks guys!

chris: yep - is a bit high stiull - but remember this is a low-compression engine (7.4:1) so it can sort-of cope with high advance when off idle

slide: good idea - one more thing to try before i resort to removing the fuel pump (not an easy job - much piping and zip ties in the way). yep - the alternator isn't charging - but there's enough volts there to run things.

[YelloRolla]

It could be;
That the ignition curve is too steep? ie idle advance is 10 deg @ 1000rpm at small load, and then off idle is going too high. OR sometimes the value placed at below idle is too low. With the Wolf on my car I have the ignition at 500rpm to be about 18deg (from mem) this will help stop the car from stalling by "bumping the rpm up" if it goes that low, and then off idle the ignition value (at <6% load) is not much higher than whatever the idle value is (as the engine is under no load) - once the load goes up though - the ignition advance goes straight to 40deg. This kind of thought process needs to be considered on your fuelling as well, maybe there is a missing value in the fuel table at or close to the idle rpm/load causing the oscillating.
I personally think that your idle ignition advance is too high and should be around 10-12deg.

How big is the plenum? Have you tried adjusting the throttle stop to hold it open a shade and then wind your idle ignition back a bit to get the idle RPM down? If not, then try it. If this is successful, but your throttle opening is too big (ie the TPS shows a part throttle open condition), then you can drill a small hole in the throttle plate.

Can you sus the MAP readout while it goes lean?

With fuel pressure - make up a temporary set up with fuel hose from your regulator to the gauge and tape the pressure gauge to your windscreen so that you can see it from inside the car. I am almost dead-certain that it is not your pump or relay.

HTH

[BeRad]

my old fpr used to drop pressure once turnd off, wasnt a problem with mine. so long as its holding the right pressure when running it shouldnt matter.

my old celica had this prblem after i had the engine out, it had dirty fuel in it, it would idle a bit rough and would stall after 5 or so seconds if u didnt keep the boot into it, turned out the dirty fuel had put shit on the plugs and they wernt firing strong enough to keep it running, just replaced the plugs and ran sweet

[Norbie]

Charles: I have a mechanical Autometer fuel pressure gauge sitting around at home. You're welcome to borrow this for diagnostic purposes if it helps. Plug it into the fuel rail and you'll know straight away if there are any issues with the fuel pump or regulator.

[thechuckster]

thanks guys - very helpful!

i'll put up some pics of the timing map tonight - and try to get a screenshot of the log file showing the stumble. From memory, the first thing to drop is the AFR, the map then starts to rise (is in KPA) then the pulse width increases to try to get the engine back to idle.

i'll also see if i have a set of plugs somewhere. i think i have enough bits of plumbing shit to remote the gauge

i have a small mechanical guage attached to the side of the FPR ...

norbie: thanks for the offer anyway!

plenum is undersized - probably 1.5 litres plus the runners which are about 500ml each

i'll have a check of the stop-screw on the throttle - from memory it was open "a bit" plus i have a small air-valve (ex-camira) that provides a bit more air on idle. TB is from a RB?? 6 cyl holden (~70mm OD from memory).

i saw an example of the 'bump-stop' approach to timing maps on msefi.com - you have a valley of low advance numbers for idle across a wide MAP range with raised numbers at very low rpm to stop it stalling

[brett_celicacoupe]

how is the fuel pump triggered to be on?


maybe you could wire a test light in parallel to see make sure its on while the engine is running?

edit; similar circumstances to the 1G the other day when the battery went flat. it lost it's memory of how to idle the engine so it would naturally idle it at about 300-400rpm. only problem was that the AFM wasnt open enough to trigger the FPR. i had to baby it for a few hours to stop it shutting the FP off.

it would run for about 5-7 seconds after the fuel pump shut off before it would splutter, then die.

[oldcorollas]

i saw an example of the 'bump-stop' approach to timing maps on msefi.com - you have a valley of low advance numbers for idle across a wide MAP range with raised numbers at very low rpm to stop it stalling

i also used that with a bit more fuel at 500rpm when i was testing an MS2...
other thing is to check your ignition sense at those rpm.. what is the idle setting at (below which it thinks it is cranking)?? if you go into crank mode, it could get whacky and stall perhaps?

i would start by bumping up idle fuel (start with a little and try a lot) at all points around idle, and remember that when it goes to stall, the MAP jumps up significantly, so that has to be taken into account also.... if the high MAP fuel is too high or too low, then it won't be able to recover stall... same goes for the spark...

i'd be interested in the log file.. what rate is it sampling at (5-20 points a second?)

[thechuckster]

hi stewart - there's about 430Kb of log file on msefi.com - plus my MSQ file.

thread is: http://www.msefi.com/viewtopic.php?t=21303

not sure about sampling rate - i thought that was a fixed rate?? where do you change it in MSnS-Extra?

(I'll have to load shots of the log graph from home - attachments size limits are too small)

brett: fuel pump is controlled by ECU - does a 2 sec prime when the ECU powers up then turns on as soon as i get a cranking signal. I think i have enough fuel - it's just the loss of pressure when power is off that bothers me.

ta,
Charles

[brett_celicacoupe]

shouldnt be pressure loss......i remember pulling a hose off a pump on a commodore 1/2hour after it was run......fuel went freaking everywhere. there was a lot of pressure in the line.

what fuel pump are you using?

[thechuckster]

from memory bosch 040? an external pump as used by eleventy-trillian EFI commodores?

[oldcorollas]

hi stewart - there's about 430Kb of log file on msefi.com - plus my MSQ file.
thread is: http://www.msefi.com/viewtopic.php?t=21303
not sure about sampling rate - i thought that was a fixed rate?? where do you change it in MSnS-Extra?
(I'll have to load shots of the log graph from home - attachments size limits are too small)
brett: fuel pump is controlled by ECU - does a 2 sec prime when the ECU powers up then turns on as soon as i get a cranking signal. I think i have enough fuel - it's just the loss of pressure when power is off that bothers me.
ta,
Charles

sampling rate is in comms?
yours is sampling about 10/sec.. fast enough (i think max for MS1 is about 15samples per sec?)

as the rpm start to die, the crank angle is decreasing, and the MAP is increasing.. the O2 is rich until a certain rpm, then goes lean (by which time it is too lean anyway) (and when rpnm too low, misfiring makes the reading lean anyway?)

first thing i would do is make copy of current map.. then alter the 500-1500 rpm range to have flat spark and VE for the MAP values you can see there.. sayy 40-70kpa.
40kpa is a healthy idle vacuum, so it should be ok on that front...
there aren't any weird codes when it stalls.. just.. falls flat ... when it really trys to stall, the spark falls wayyy down..
can you chack the real spark angle vs what is indicated? it could be a few deg out, which makes all the difference for semantics

set a flat VE of about 30, and spark of?? no idea.. maybe 20??
see if it idles....
then change spark in the rpm and MAP range up by 5deg then down by 5deg.. see how it changes...

then, go up in fuel by 5VE, then down by 5VE..

however.. at 30VE, you have PW 2.1, and at 50, 3.5.... these are pretty small pulsewidths..
(oh, can view msq in notepad ok... kinda )
your REQFUEL is only 4.3, which seems a bit on the low side... then again, it is turbo engine... for NA, i'd aim for req fuel about 12ms.. so about 6ms for 1 bar boost?
any reason why VE1 and VE2 are different?

how many squirts are you running?

i guess, try and use flat tables in that region, then move up down, see if either the idling or the stalling (during) get better, then try to have a liner change (ie ramp up) across all MAP values for a given RPM... which kind of takes MAP fluctuations out of it....

sounds like.. fun

[thechuckster]

had to go out so am using computer at sister-in-laws house - they think i'm totally geeky...

will try that stewart... i will try your method out on tuesday nite, tho i think i should also getthe alternator fixed up so there's no voltage correction happening...

fuel is on alternating squirts - simultaneous gave me even shorter pulses. Am pretty sure the ReqFuel is at 8.5 or similar - perhaps the 4.3 is per injector?

Am not using 2nd table so not sure what is in VE table 2

if i had the chance, i should have put a pressure sensor on the fuel rail... oh well... plus i really need a wide-band sensor.

cheers,
Charles.