KALOS 1.2 SE stalling/tuning problem sudden acceleration

[polarKALOS]

(Note added 28 november 2012 - Problem now solved, adjustment of tappets, stalling and acceleration and power and mixture problems cleared and confirmed
History of the problem for your comparison as below . . . . . .)

Good morning, I found an invitation to try out this site posted on a Chevrolet forum. I've read the guidance topics and hope to contribute rather than only seeking help.

A french speaker asks me to see if I may post here about a problem, which perhaps will help others searching or gather a response. All useful experience I hope, including for Aveo owners. He has bought a used 2008 model KALOS manual transmission with petrol/gas combination that behaves erratically. I hope the technical progress helps other readers.

When he drives (engine warm) and slows down shifting gears and declutched the engine rpm drops from, say 3000 to 1800 then swiftly surges up again once or twice to 2500 then drops like a stone and usually stalls.

If he stays in gear to slow down, foot off the gas pedal, the car slows a bit then suddenly accelerates toward the traffic in front requiring hard braking to prevent. The control of the throttle position seems miscalculated by the computer.

When cold starting it often stalls or during the first two or three parking movements. Reading others' experiences the indications are of an airleak somewhere. Here's what he's already done (with help).

Replaced the MAP manifold air pressure sensor with spare part. No change.

Test drive without the MAP connected. Engine control movement becomes smooth and normal and the car a bit more powerful too. It can stall yet far less or often not at all. Engine alarm on yet the tech detail in english on the CD mentions that without the MAP the computer chooses TP Throttle position mode and can still drive. There's no detail to say if the TP mode is only looking at the throttle lever angle to calculate a crude injection of fuel or whether it's some useful backup data on pressures due to throttle position and engine revs and temperature (some sort of intelligent estimate). No idea of the long term effects on emissions or performance/economy so he continues to investigate for a leak.

Switching to GPL liquid gas system gives the same results so probably not a fuel pump nor fuel pressure issue.

Investigated the EGR valve. Quite clean, zero alarms and he added a cover plate of thin metal to block it off completely to exclude any leaks from that circuit for a short test drive. No change.

Checking for airleaks. Used lighter gas canister (to see if engine speed changes with extra gas sucked into leaks). Inspection all tubes, hoses and connections up to each accessory (including brake assistance servo). Check for airleaks into the inlet system, including the few non-return valves and all seems airtight - so only the plastic air manifold itself and it's bolting to the cylinder block behind remain untested. Overall the engine running makes a powerful draw of air with good braking too. No change to the original problem.

Dismounting the throttle body reveals a bit about how it works yet it were clean already (as if new or replaced). Air filter body and inlet tubing inspected apparently super clean and airtight.

When testing the throttle action inside the car, choosing 2500 rpm (MAP connected, normal configuration) it matters not whether he keeps the gas pedal at a fixed position (same thing using the hand), the throttle controller forces a reduction bit by bit (you can feel the cable click or jumping slightly as the engine butterfly control motor does different to the requested setting) and slowly but surely the engine speed drops then stalls.

No other alarms nor problems indicated, no idea yet of fuel consumption although he's doing a test now.

Spark plugs seem new and perfect colour/combustion and when the car idles it stays steady at about 700 rpm.

The learning procedure for the idle for use in a garage (ignition on 5 secs, off 5 secs, start up, warm up, aircon on 10 secs, off 10secs, on 10 secs, off 10 secs, stop engine, restart). On a few occasions during that procedure the engine revved up high to 1500 rpm then reduced yet after the procedure the idle at rest is correct with aircon off or on, as if the system has learnt correctly a stable speed and control.

For the moment he drives with the MAP disconnected in what daewoo call the TP mode to have a safer experience in most traffic.

thankyou for reading; I hope complete information is appreciated and the owner will get an emission test without MAP in TP mode in the next few days. Considering an airleak can be a lengthy investigation he hopes to have made some efforts before suggesting a garage starts replacing parts all over.

By the way there's indication the vehicle did hit something in the past (radiator bent, slight damage under front wings) and that the master brake cylinder were replaced (looks brand new compared to the very slight dustiness of the rest of the engine space) and if someone were to have attempted to brake, either it's plausible they heard nothing (because the KALOS is very quiet and engine stall is not much different in noise levels) had no servo assistance without realising that were the cause; in other words it seems the brake servo has been replaced after a minor accident without investigating the real cause before.

In TP mode it's less than risky to drive - seems quite good actually. Any comments or encouragements appreciated.

regards
PolarKALOS

[PrecisionBoost]

Certainly sounds like a very dangerous problem.

I'm not familiar with the 1.2, but I know some of the newer Aveo (kalos) used a drive by wire system where the computer controls the opening and closing of the throttlebody.

Based on your description of the throttle body moving when the driver keeps a steady throttle position it sounds like it is a drive by wire system

Are you able to determine which type it is ?


Do you have a scanner?

If you hook up to the OBDII port you may be able to scan it and look at the Throttle pedal value ( voltage or percentage ) to see if it's behaving as you would expect ( steady change in reading as pedal is depressed)

[polarKALOS]

thankyou for the response

This 1.2 has the control-by-wire concept, at least for the first 20 degrees of movement of the throttle according to the CD, which is also the last 20 degs when decelerating and foot off the pedal. The CD mentions the computer observes as you describe, before controlling the butterfly valve based on all the sensor information using a small motor and gear which we can see.

I have mentioned to the owner the usefulness of getting a scanner tool or USB cable type to hook up to a PC and look at the data of sensors to know if they're working at all or linear. He's getting used to the idea. I hope he'll get one.

There was a diagnostic done on the car a few months ago by a garage for the previous owner; zero detail, no numbers, just a conclusion "mixture too rich". I don't think they were looking at the actual control positions in the detail necessary yet did propose changing the throttle body completely in an estimate two months later. The exhaust colour is very black rather than the better grey, so something to consider also.

thankyou again

[PrecisionBoost]

I think an OBDII scanner is the only way your going to figure this out, it makes sense that the car is stalling due to rich conditions on deceleration, I've had that problem a few times when the air fuel ratio dropped down to 9:1 due to a small problem with tuning.

I would look at the pedal positioning sensor as a possible culprit, the only way to see it is with a scanner or by simply replacing the sensor ( which might be a waste of money if the original sensor is good )

[polarKALOS]

Good afternoon

Without a scanner yet with the service CD info it were possible to put a couple of fine wires (insulated) just inside the rubber sleeves of the cables at the sensors and use a multimeter:

Throttle position sensor gives the volts inverse to the guidance notes yet that seems correct and that the Electronics swaps the reading over in it's signal receive circuit. If so, subtracting the values from 5V, the TPS gives 0.5V at zero angle mechanical stop (expected value 0.4 to 0.8V) and 3.4V at throttle fully open. Seems OK as it changes value smoothly as the pedal is pressed or moving the throttle cable sector manually. It's a bit tricky to visit throttle closed position because the activator insists on about twenty degrees open throttle as soon as volts are applied leaving the manual cable position behind (part of how the idle control is done auto without the adjustable stops). That gives 1Volt which is outside the expected value suggested in the techn notices. It seems the CD was written for a throttle body with two connectors, one four pin for the motor (to be disconnected) and a separate three pin for the position sensor. This car has all on an eight pin plug with seven pins connected so unable to separate the function. When running the engine however, the activation motor can and does close off the throttle a bit so moving below it's startup position with a chance to measure the value more closed nearly shut so to say.

Oxygen sensor test. It's a titanium dioxide type apparently (internet info) and should flip and flop at volts much higher or lower than 450mV and nothing inbetween.
The CD mentions high voltage means rich mixture indication and low means lean mixture.
The CD suggests it should follow quickly when testing the engine and hop up and down as the controller arrives to a mixture average.

First the heater coil impedance was tested at 6 ohms which gives 2 amps and seems correct. Apparently if the current doesn't flow the computer realises and makes an alarm which it never has so far.

The output of the sensor was 450mV at rest (the ECM's control circuit seems correct).
The output went high to 900 mV when accelerating the engine and dropped like a stone to near zero when closing the throttle suddenly with a quick response/follow.
When selecting steady 2000rpm or thereabouts the output went high and stayed at 900mV (mixture rich) for about thirty seconds, which seems a loooooooong time, then the sensor reports the value decreasing. The throttle little electric motor does try to close the throttle, to oppose any hand or foot effort to keep the throttle open, a real physical torque coming back through the shaft (perhaps there's a spring inside to prevent something breaking). The engine slows very slightly anyway (fuel reduction?) and the sensor reports fewer mV then flips to below 450mV and some cycles begin causing the value to flip above then below every two seconds or so on average. The average of the cycles seems to be even and therefore around the 450mV mark (the use of a multimeter smoothes the value even though it's less useful than a graphic).

What we don't know is if such sensors can flip and flop, around 450mV as required yet due to a faulty measure of the oxygen (i.e. working yet out of calibration). No way to tell unless someone has a spare or a V6 V8 engine with two sensors to do a comparison or a test bench with oxygen variation. It's working in principle however and may be quite alright. Some things work correct or not at all.

What's strange is that the exhaust pipe is obviously black sooty rather than grey no matter the distance travelled. A previous diagnostic for previous owner states "mixture too rich". So I'm thinking now about how the engine control could continue to add too much fuel most of the time, if the oxygen sensor flipflop is doing the High/Low changeover correctly. The difference is that we can test in front of the house vehicule at rest when it might be very different the rest of the time when asking for power from the system either on the flat or up hills.

thinking somewhat . . . . . any ideas what to check next? No rush because the TP mode driving is safe even if the system is contaminating itself Also, running rich at least the engine stays cool running for the moment.

[polarKALOS]

progress

The Throttle Position Sensor is in the MTIA throttle body around the shaft of the butterfly valve and today it was removed and opened up. The potentiometer is easily accessible and visible, as are the four wipers that give volts out for the cable throttle position and another signal to show how much the stepper motor has moved the shaft. It was reasonably clean yet the circuit was lightly wiped both sectors with an electrical/electronic contact cleaner for cars. Put back in - no change to the problem. The volts were already measured as probably OK and the control electronics can idle the system.

With the MAP reconnected a lot of effort to try to find any vacuum leak (a problem that seems to fit most of the diagnostic ideas for high idle speeds).

What's noticeable now is that if the engine speed is increased (car parked, in neutral, no load on engine) when releasing the throttle the rpm reduces to 1500 and stays there. The throttle valve position is, however, already at bottom in the usual position for idle.

There are three ways to get it to jump back down to normal idling.
Switching ON the air-conditionning does it immediately and the idle is stable at 700 rpm..
Driving the car and changing gears restores the idle from 1500 to 700 rpm.
Disconnecting the so-called "positive crankcase pressure" tube (the small tube from the air inlet just before the MTIA) then placing the thumb over it briefly (so causing a greater vacuum for a second or two) causes the idle speed to drop (as if a leak elsewhere has been closed off). Keeping the tube blocked for more than a few seconds causes a squealing sound from somewhere difficult to trace. It's loud, sounds like trouble; so it might be the leaky joint leaking more, or a component somewhere that is sensitive to greater vacuum or it might be something else inviting new damage. That positive crankcase pressure tube goes directly to the inlet manifold (nothing near the crankcase) with a small T junction to accept traces of fumes from the camshaft housing. The amount of air it lets pass from the airfilter side (diverting around the throttle body) into the manifold seems very significant for correct vacuum function.

Going to look again at everything to try to find a leak

Without knowing where is the vacuum leak the conclusion is it seems to be worst when the engine is running without driving the car and much less or intermittent when driving on the road with gear changes.

will keep you posted, for the record too.

[PrecisionBoost]

I did not see any mention of resetting the ECU ( disconnect battery power for a period of time )
I'm curious if there are/were any differences before and after.

Did your Kalos owner have this problem from the start, was the car running fine when he first purchased it ?

I know this may sound strange, but your problems seem very similar to a time where I swapped an ECU from an automatic into a car with a 5spd gearbox.

I did it on purpose because I was working on a project where I was modifying sensor voltages, I didn't want to risk screwing up my original ECU so I used the spare Auto ECU I had.

The only way to get the idle reasonably close to normal was to turn on the A/C or drive it for a short period of time.

There were also times where the engine would push forward in gear accelerating when I did not want it to.

If this problem with the Kalos persisted from day one of ownership I would seriously consider the idea that someone has done a repair and put in the wrong ECU type.

Even if it were a situation where the car was fine for months/years then suddenly started these problems, I would still consider buying a used ECU from a scrap yard, it is an expensive part, but your Kalos owner could always sell the ECU on EBAY if the replacement ECU did not solve the problem.

At least by replacing the ECU you would be able to rule out a large possibility in respect to control of the engine


The measurement of rich conditions for over 30 seconds seems strange, based on memory of my Lacetti (Optra5 ) the engine typically runs lean for 2-15 seconds but rarely did it push over into the rich side for more than 2-3 seconds

That is to say the car consistently attempted to run an average air/fuel in the range of 17:1 to 14:5 in low load situations

Under high loads it would push rich ( 12.5:1 ) for periods of up to 10 seconds, but it would pull back closer to 13.5:1 most of the time under higher throttle levels.

Obviously it is very hard for you to measure air/fuel from a narrow band sensor, the only way to get a good feel for the true ratio is to install a wide band O2 sensor into the exhaust as a temporary measure.

You should be able to find a used Innovative LC-1 somewhere on the internet, again, after you use it your Kalos owner can then sell it as this item will be very easy to resell ( perhaps even at a profit if the owner is smart buyer )

Sounds like your very detailed with respect automobile diagnostics, hopefully you figure it out.

[polarKALOS]

Hi

During the first checks of the MAP sensor the battery was disconnected and reconnected and, at first start up it was noticeable that the idle took some time to relearn. Using the learning procedure including 10 seconds of aircon the idle was then stable with either aircon on or off. Driving away there was still very little power passing the first two gears and the car stalled the same; i.e. no change. That routine were done four or five occasions. I've asked the owner to repeat battery disconnection to complete the picture since the throttle body was cleaned up.

The owner had this problem from the beginning (purchased two months ago) - and he tells me the previous owner told him about the stalling yet not the sudden accelerations.

Thankyou for the comparison about the ECU. I think it is a big clue. On the internet there are ECU's with photos from the same model, nearly the same year and with all the same numbers. One in Spain is marked only one letter different - E07E instead of E07F (which may be the country code Espagne/France or slight changes to firmware version 7) and another in Germany has a 3BOP instead of 1BOP for the type and is described specifically as coming from a manual transmission car. So, at first glance there may be differences; it may be the wrong ECU here. According to the CD the ECU has a port for communication with the automatic transmission (if there is one) and were capable of changing it's parameters if the port gives no response (it were capable of concluding it's a manual vehicle) yet there's no specific statement in the CD if the ECU will do that choice to change performance or needs to be taught it. Auto on KALOS vehicles is hydraulic/electronic without connection to the air system, unlike some older Volvo cars that used some sort of pneumatic servo. I hope you have a moment to recall whether the automatic of the two vehicles you spoke about used air for some part of it's function?

In any case I'll post what I can discover about the KALOS ECU numbers for future readers (not yet cleared up).

Thanks for the Lacetti overview. This owner goes quiet when the need for a simple scanner connection is suggested, is disinterested in measuring anything to find the facts and intends only to attempt the compulsory overall vehicle test (it's now four years old so due it's first inspection). There will be an emissions test and, if all goes right . . . it'll fail. Then the owner will take this more seriously (I hope). Thanks again for the detail.

I've owned three injection vehicles in the past yet this is the first occasion I'm looking at understanding it. Finding the CD on the net (Service Manual) was the only way to learn anything at all about the KALOS, then it were possible to look up selective bits and pieces on the net such as how the oxygen sensors work. I think his car were best handed over to a main dealer mechanic now, with the progress written out so that a technician there can learn from it (presuming he or she can proceed not only by changing parts). In either case I'll post how it works out, cause of the problem and so forth.

[PrecisionBoost]

I believe the model of the automatic is the same ( at least I think it is here )

With regards to the ECU, I believe the main issue was the firmware difference.

If one were to update my "automatic" ECU with firmware from a "manual" it is likely the problems I mentioned would go away.

I didn't bother trying, I only wanted to see what difference I could force with the ECU through additional hardware.

I was modifying the intake temperature sensor and map sensor to change ignition advance and fuel, in the end this ECU resisted the changes and did it's very best to keep the air fuel ratio on par with it's design specifications.

That is to say it did not respond particularly well to a piggyback management system, some ECU's are easier to fool than others, the unit I had ( made by Delphi ) was particularly stubborn, it was reliant upon O2 sensor very heavily.

The main dealer mechanic will be able to tell him right away if the ECU is wrong and the diagnostics should be able to pin point the problems fairly quickly.

Good Luck

[polarKALOS]

Thankyou

The vehicle has finally been to it's inspection and failed the emissions test CO (very high level).
The owner can still drive it to investigate problems. MAP disconnected, TP mode invoked works reasonably well so he can drive safely with a constant MIL alarm.

Dear readers, there might be some delay until the next update and detail about how it's resolved because the owner chooses to put off any further work until the camshaft timing belt has been changed (still the original on the vehicle 100 000 km) and after the valve tappet clearances have been checked/adjusted. The 1.2SE STEC motor 2008 still has manual screwdriver and locking nut adjusters. The owner agrees he might, maybe, perhaps get a 50 euro scanner after that. I have a feeling he's wasting my time and enjoying it.

Thankyou for the forum; I can read and learn and perhaps help one day!

[polarKALOS]

Progress one month later

The stalling problem and hesitation surging problems are solved. The tappet clearances were checked and found to be nil, zero; valves up against the rockers. Previous owner and this owner presumed hydraulic auto adjusters because cars from the eighties usually have those. According to internet, more and more manufacturers are counting on improved metals/design reducing wear while removing hydraulic tappets so for the Kalos 1.4 2008 they still are hydraulic but for the 1.2 they're mechanical manual adjust. This engine running on petrol no longer surges nor stalling and shows a bit more power. This "leak", however small were capable of changing the vacuum draw in the inlet system during compression strokes and combustion so putting the MAP data off, a weaker draw pressure indicating throttle wide open, causing the ECU to inject more carburant yet it's still not confirmed as the tappets may have been done at the very last chance before major leak or valve burn out. Recap only 100 000 km.

Why The car is running mostly on LPG (kit from new) and user experience indicates the LPG running hotter and cleaner causes valves to recess more, often leading to problems before 70 000 miles. It seems a reasonable candidate; the maintenance schedule indicates check tappet clearances every 10 000 and this were neglected by previous owners.
http://www.amrautos.co.uk/index.php/lpg-systems

The same website has a list of cars vulnerable to this longterm problem of which the Daewoo KALOS is not yet on the list while many other vehicles are, whether they use additional lubricants or not; some destroy the cylinder head. So this vehicle may have valves damaged or just saved by adjusting the tappets or just a bit more used than average or perhaps be in the clear for the life of the vehicle. Nobody knows until such cars get beyond 200 000. Without an endoscope we're unable to inspect the valves from the inside. After 200 miles the tappet clearances will be done again in case there's some dirt/crust still to come off. Without backfiring nor combustion in the inlet manifold it seems the car will be OK for the short term. Here is the link to the list of vehicles at risk (according to PINS kit supplier or owners or garage technicians or both, including older vehicules). Note again, the KALOS is a car with LPG kits not yet reported for problems although more valve wear is to be expected. If this is the first time this car has had adjustment there'll be plenty more "play" to use up before any trouble and from now on the increments in the tappet gaps are being recorded.
http://amrautos.co.uk/database_problem_cars.pdf

CO emissions and rich mixture problem not yet solved Before concluding that problem it's necessary to wait until the timing belt has been changed and do a test. To learn more about which symptoms can be caused by which problem I looked around and found a simple summary in a service notice for LPG system installers. Very useful tips and hints.
http://www.go-lpg.co.uk/Hints.html

and much more in this brilliant diagnosis manual from Teleflex which also serves an introduction to the finer points about problems
digidownload.libero.it/storyteller/varie/dm_sgi_type1_01.pdf

Pages 43 to 46 are about maintenance yet describe the factors and consequences to be able to understand it (Chapter 5). Highly recommended and easy to read.
Page 46 lists 9 mechanical or control factors with the effects on the tailpipe emissions whether running on LPG or petrol i.e advanced general knowledge.

Still to do Check if the CO problem is still there, consequences of the rich running for a year, check if the catalytic converter is still serviceable.

and one more subject . . .

Some owners find that leaking petrol injectors can cause additional rich mixture when operating on LPG. When the LPG system is without lambda sensor (like this car) it notes the normal petrol ECU calculations for injector timing and copies them out just a bit longer without checking emissions. If the petrol injectors leak under their pressure the total mixture goes high. The inverse is also possible but the petrol computer is aware of lambda values and can reduce it's petrol injection timing if there's an LPG injector leak. One Land Rover owner I read has fitted a switch to do similar to the emergency cutoff of the petrol pump (for accident impact) and he depressurises the petrol circuit by switching it completely OFF while motoring on LPG. He noticed a dramatic improvement in petrol mpg on his usual routes with the usual/same use of the LPG/startup so proved the problem that way. It seems this car may be without the difficulty yet it will be checked because the petrol injectors have very little use, may be fouled up in some way and may need to be cleaned.

half way there (maybe)

[polarKALOS]

Note: The link is

digidownload.libero.it/storyteller/varie/dm_sgi_type1_01.pdf

yet it's necessary to put that text in a google search then click on the link there at the top of the results. "http://www.digidownload" etc seems to disfunction.

[PrecisionBoost]

the valve tappet clearances have been checked/adjusted. The 1.2SE STEC motor 2008 still has manual screwdriver and locking nut adjusters.

That's interesting, I don't know much about the 1.2L engines, I'm surprised to hear that they do not have hydraulics

[polarKALOS]

Hi precision boost

I were surprised too. Getting to the point of checking the tappets were due to surprise with the articles about LPG cars. Actually doing them were due to "nearly perhaps having checked some other things first". It seems some owners and this one are getting stubborn that they believe they've bought a maintenance-free car with a maintenance-free battery too! Well, the Daewoo service manual mentions to check everything before changing out anything (sensors and modules) yet little or near no mention of tappets because, of course, they assume the pdf troubleshooting is being done after the minimum regular servicing has attended to wear and tear i.e only due to an actual breakage somewhere. Then, I believe the CD is really rather good, well written, polite and precise. The previous cars I used (VW Golf/Rabbit Gti then GLS then Opel Corsa, Audi 100 quattro, Opel Vectra, VW passat) - I passed to garages for a reason of travelling for work (independent confirmation of readiness) so haven't been doing much maintenance myself for a while (except one cylinder head).

Thankyou again for your encouragement. A scanner were to have been the right way to approach the subject to know everything else were working first as soon as possible and still is. With hindsight someone can observe and remark "tappets are easy", which they are, yet that's after the fact. In other cases it may be something more subtle including a minor airleak. The car's still not finished and it were one month of doubt and danger before posting here and getting on with it. Thank goodness the ECU TP failsafe mode were available and safe!

By inspecting, one air hose has been found chafing, half worn through on a "nick", slightly misplaced after the GPL installation it seems, now given some additional protection/re-routed.

Without a scanner it's unclear which way to check performance/mixture/emissions next, until the owner feels sick or smug with the consumption figures. I'll be asking around to find a scanner or tailpipe detector (wideband) before that!