RPM drop and stall on load at idle

[thatdevsherry]

w.salam. I'll have to test my cold start as I haven't started the car cold after the starter repair. Mine wasn't "half-self" so that's a bummer, but it is still quick.

In the repair work they also changed the bush at the tip of the starter motor.

I doubt that it is a problem with the starter because the goal of starter is to spin the flywheel quick enough which it does just fine.

If the engine is rotating fast during cranking but it fires up late when hot, it points to a different issue probably related to fuel.

Last I googled (or AI'd) pointed towards the fuel system (check valve in fuel pump/pressure regulator/injector leaking pressure) due to the timing pattern i.e. issue because of heat soak.

On hot engine if I start after a few seconds of turning it off, it starts up instantly. But when left for around 30 mins it gives long crank.

I'm no expert but I'd advise against opening up starter again if it rotates the engine quick enough when cranking, Which it seems to do fine for your case (and mine).

[Xulfiqar]

Does the starter actually spin slower when the engine is warm?

[Xulfiqar]

Why not buy a gear reduction type starter? Aka automatic gearbux wala self.

It has much higher torque and is far more durable too.

[thatdevsherry]

I saw it in FSM but had no knowledge of them (being better and durable).

I'll try swapping that in once this starter fails as it seems to work fine after repair.

I'll check what the engine rpm is when it's cranking to get an idea of how fast it is. It lost it's original mechanical sound and now sounds like an rc car motor. So want to verify that it's contacting and cranking engine good enough.

EDIT: Cold crank was fast. ~1 crank I'd say. Hot cranks are also ~1 crank. Haven't yet noticed the delayed engine start when left hot for some time. Will keep an eye on that. The starter motor is spinning fast now so maybe issue was the starter motor cranking slow when hot and engine had a hard time firing up with that speed. Engine seems to fire right up now.

[thatdevsherry]

So I looked at the old bottle that I had filled up P/S system with, it was "type A". AI said it's ancient.

Flushed out the P/S with Kixx ATF DEX-III. Bought 2 bottles because I thought a good flush would use up more, but turns out 1 litre was all I needed as clear fluid came out very soon when I started the procedure.

P/S feels... same. Probably placebo that I feel the initial torque required to move the steering has lessened. But at least it now has what FSM says to use.

Although I noticed at hot engine that the PSP switch again has moved away from near center (90deg) to nearing full lock (~270deg). Not sure why it had come close to center that one time after I repaired starter.

Oh well, normal brake and steer has not yet stalled the car. It does drop rpm at standstill but I don't feel engine struggling to bring rpm up as it did before. Will try doing the killer "brake pump and steer" to stress it out. Pretty sure it'll die from it but I have noticed engine being better now with regards to everyday driving. Prev. the rpm would drop pretty consistently. Now not so much.

Will report back about the situation after driving for some weeks.

[thatdevsherry]

So I drove the car for some time today and I don't see engine coming to stall. I tried braking and making u-turns but the car did not stall. Rpm dropped but it didn't stall. So that's an improvement.

Although on full lock it takes a tad bit of time to bring rpm up. The switch activates but the rpm does not go up a lot, shoots from 650 to 750 first and then goes up a bit, or down a bit then up. This is when I'm holding the brake and steering is held at full lock. Otherwise rpm goes up fine. Maybe P/S pump is binding or it's something else.

But good that on normal driving it is not stalling.

However it did stall with the brake pump and steer test I did at idle. I'll make a video showing that because saying "brake pump and steer" is vague and useless.

Tested the same on cousin baleno with crappy P/S pump and it does not stall when doing the above. So my definition of this issue being resolved will be that this test passes for my car.

Can't say much until I do vacuum gauge test. But I now drive confidently knowing it won't stall (as of now).

Also I confirmed hot crank is delayed a bit. Starter cranks 3-5 times and engine fires up. Didn't notice before because starter spins fast now so I had to listen close.

I remember iacv rubber gaskit between the throttle body and iacv rotary shaft housing was torn and silicone was used as bandaid.

Will see if I can get the rubber gaskit and apply that dry. And also confirm the o-ring b/w iacv solenoid part and shaft housing is present. The 1% duty cycle test did show that air isn't leaking past iacv but better to do things right than apply bandaids.

[Xulfiqar]

Add some oil to the iacv to see if it helps

[thatdevsherry]

Will do. I noticed idle up behaves same with it without brake. Previous analysis was not correct.

On full lock (warm engine) rpm goes up to around 800-850 and then drops back to around 700 which is intended idle. Need to plug scanner to confirm values since isc duty cycle used to jump up so I don't see why it would try to come back to 700 idle with a high duty cycle. Could be stuck but I had put oil in it some weeks ago. Will try again.

[Xulfiqar]

Is the rotary valve surface worn out?

[thatdevsherry]

Are you referring to the shaft end that goes into solenoid case or the pieces that close the vanes in metal housing?

[Xulfiqar]

The whole rotary valve, its a zeolite bearing impregnated with oil, our desi mechanics love to clean it because “its trendy” and a quick grab along with throttle cleaning.

The oil is pulled out of the bearing and it then wears at an extreme rapid rate.

For your remedy, first make sure there are no makeshift gaskets anywhere including the main intake manifold one. No silicone or shellac or any other stupidity, just new genuine dry fit gaskets. Then go on resetting the idle, make sure your map sensor isnt lying either, it can also upset engine load signal.

[thatdevsherry]

Ah I see. Will replace the iacv gaskit as it does have silicone.

Intake manifold one I replaced a few weeks ago. It was the metal gaskit but had silicone applied around the ports so I replaced it and installed a new one dry.

Will swap map sensor from cousins baleno as a quick check and also check how map voltage changes when doing the full lock thing on both cars.

EDIT: wanted to clean map sensor as well but I'm not sure if wd40 electrical contact cleaner is good enough and won't damage it.

[thatdevsherry]

Did a desi experiment which might help narrow down issue.

I tried disconnecting sensors to see how engine responds.

Disconnect MAP: car doesn't go below 700 during steering or when i pump the brakes or both. Had to turn engine off and on again cuz disconnected map with any load introduced makes it go into a rpm up down cycle. Car never stalled no matter what I did.

Disconnect tps: rpm drops down when steering and braking. Car never stalled no matter how much load i put.

Disconnect iat: rpm does not go below 700 when i pump the brakes. on steering rpm drops a tad bit due to load (expected) but jumps straight up when psp turns on. Note that rpm after pumping brake does not go below 700. during pumps it goes up as normal but after holding last pump (or letting it go) the rpm drops down very nicely and slowly to 700. Car never stalled no matter what I did.

When I pump the brakes and let go, with all sensors connected, it drops and continues dropping below 700. But moving it to backup (or limp or whatever the mode it has is called) causes it to never stall. I even disconnected tps and iat and even though rpm dropped to 500 it didn't stall.

So I guess we can tick off the following because it should've stalled with sensors disconnected that make it do more oopsies:

• ignition timing
• P/S pulley binding

It very much looks like a sensor issue and nothing mechanical. Still including iacv as it's also a sensor. Not sure if disconnecting iat makes all params go to backup or just iat, but if just iat, i guess iat is reporting bad?

I noticed that in hot afternoon (30-35c ambient). fully warm engine, ecu was reporting IAT around 55c at idle on standstill.

It was hot outside and engine bay was also hot, but I have the stock cold air intake so I expected it to be lower. Maybe IAT sensor is reporting a higher value and throwing off density calculation (last i remember it needs map + iat to calculate air density or whatever the main thing it needs to calculate).

map also made rpm not go too low so maybe it might be contributing but doesn't feel like the main issue. Disconnecting IAT made engine work perfectly.

NOTE: IAT value that I mentioned is from my diy scanner. I used the same formula as the ECT sensor (which gives super accurate value). Pretty sure IAT also uses the same calculation hence confident in the reading.

EDIT: maybe iacv is good? If it was binding it would've still been slow to react when doing the above tests.

[Xulfiqar]

The iacv is not a sensor its a motor

Motor = takes input from ecu to do something
Sensor = provides ecu with a reading of something whether its temperature or speed or position

The iat will always show a bit higher on a running engine whilst idling, its the ambient air in the engine bay being heated up and warming up the casings

The iat will show true temperature on a moving vehicle as the engine bay is also cooled by exhausting its air and the airbox also has higher airflow as the throttle is open.

Keep it up, you will get there, it maybe a bad ecu too, some older Mitsubishi had this problem too, the ecu could not move the idle motor fast enough due to lack of amps.

[thatdevsherry]

Thank you for correcting the terminologies. Learning something new everyday.

I haven't got to the conclusion that it's a bad ecu. My cousin also overhauled his baleno engine and experiences somewhat similar issue i.e. stalling when braking and turning but it is rare than mine, as in I can stall at will on idle but on his car I can't. Chances are slim that both our ecu went bad right when we overhauled the engine.

His issue I somewhat identified is that when braking and clutch is pressed while car is still moving, the rpm drops to idle but then goes below it if brake and clutch is still pressed. And when idle is super low and we add steering input it vibrates and then either recovers or stalls but has a hard time when being at low rpm. Mine also dropped rpm when braking for a long time. However it doesn't seem linked to VSS.

Mine is also similar in that first the idle drops, and then the car stalls. Not like it dies straight from 700 to 0. It's the drop to 500 and then being unable to bear the load and stall.

Although you mentioned ecu not moving iacv fast enough, but why would it suddenly be unable to do after overhaul? Don't think it's that because of this.

We both got the iacv gaskit and o-ring so will install those to clear out iacv.

I got a used iat sensor just in case it shows different reading.

Learning quite a lot because of this issue xD although I lose sanity sometimes.

FSM says IAT value at idle at warmed up engine should be around ambient +15C, which at 35c would be around 50c so maybe we're good or maybe iat reports a slightly higher reading, ecu adds less fuel as it thinks air is hot and not too dense (but it is a bit), and we end up with a lean mixture and engine stalls.

Meh just a theory. Also with iat sensor disconnected the car didn't stall. If ecu is slow to move iacv, it would've stalled still, no? It always recovered fine without any hesitation or vibration. So I think ecu can control iacv fine. Some sensor is lying and I have to find out which.

For above context, with all sensor connected, pumping brake and then letting go makes rpm drop lower than 700. Like it keeps falling from 1000 (due to pumping) till 500.

But with iat disconnected, after the pumps it drops gracefully from 1000 and slows down near 700 in a controlled manner, never goes below that.

Now I instantly go to map when I see this because pumping is changing vacuum a lot and then the engine has to adjust. But first I'll try popping out iat sensor (which seems like a hard thing but fsm is like yeah just pop it out easy peasy) and then I'll test it in heating water (as per fsm instructions) and see the resistance to water temp. as per fsm graph.

i couldn't find a vacuum gun type thingy to test map otherwise I would've tested map with that too as fsm shows map voltage at different vacuum levels.

EDIT: In case it points something, stall only happens when engine is at operating temp. never before.

[thatdevsherry]

Checked both iat sensors and the resistance at 20c and 80c water is pretty much identicial (and conforms to fsm spec). Will just connect the other iat sensor in front of air pipe just to see if it stalls with the other sensor or not.

Can't take the existing iat sensor out without breaking it or deforming air pipe. Weird rubber.

Cleaned map sensor with contact cleaner just in case (couldn't find map specific cleaner).

EDIT: after cleaning iat and map sensor issue persists. however small that's new is that pumping brakes now does not make rpm go too high. goes to around 850 where prev. it would go to 1000. but rpm still drops after letting go. and adding full lock steering when rpm is low makes it vibrate and either recover or stall (mostly stall in my case).

[thatdevsherry]

Videos (brake pump intensity indication of frustration)

brake pump and steer stall. Note rpm dropping to below 700 to near 500 after pumping, and steering load at 500 stalls it. See below iat video where rpm drops a bit when steering before iacv takes over.

brake pump and steer without iat sensor. Note graceful drop to 700-750 and never going below 700. only goes below 700 when steering is added, and it instantly recovers. Maybe those millisecond delay for iacv is what causes stall when rpm is at 500? Although I believe rpm should never even go below 650 after pumps like in this video.

brake pump and steer without map (kinda useless)

EDIT: might help if i make engine param logs with the above tests. can see fuel injection duration etc (even though diy scanner calculation is wrong we can still see the difference b/w the tests).

[thatdevsherry]

A good update.

Opened iac valve fully. removed silicone. added new rubber gaskit. Used the iacv valve that came with throttle body (was using another one) and I soaked it in machine oil for 10 15 minutes (or i should say bathed it xD). Bolted it to throttle body. Cleaned iacv solenoid with contact cleaner and pen-pineapple-apple-pen.

O-ring b/w solenoid and valve body was missing in my iacv before. red flag

So added a new o-ring and started her up.

Used the expensive scanner to do actuation test. RPM dropped to 500. No idea why that happened since I had adjusted base idle before. Oh well I fiddled with the actuation test 2 3 times (because each time i redo it ends up at a different rpm) and ended up with base idle of ~675.

Got home, started logging the brake and steer and..
PSP switch started activating early (90-100deg) from center.

Tried doing brake and steer at full lock, car vibrated as if about to stall but it recovered. Did it 5 to 10 times and it didn't stall once. Behaviour similar to my cousin baleno that I also couldn't stall.

Don't know if it's related that the early psp activation is a sign that car won't stall at full lock. Because I'm pretty confident I can stall it as I did before. Now I couldn't.

Will check if psp switch behaves the same at warm temp again.

Thinking I should up the base idle further to 700 or in other words, making iacv flow duty be at or near the minimum level mentioned in fsm (5%).

Will update on situation on next drive.

EDIT: I would've preferred if it acted similar to the iat disconnected flow because that's just perfect. But guess its suzuki being suzuki. Sensors being too sensitive i guess.

Barometric and map seemed to be fine. Not sure if ecu has a barometric pressure sensor inside but when i turn ignition on, barometric is at 101.57 kPa. As soon as I crank (even a teeny tiny crank that doesn't start engine) the baro value changes to 95.71 kPa.

MAP at ignition on is 95.71, so maybe ecu first uses hard-coded one and then on crank uses the map value to set barometric pressure for iacv and fuel adjustment.

Screenshot of me doing the brake and steer today when car didn't stall. the purple line is at 500 when 0, so the rpm sometimes dropped to 400 as per the logs. Maybe better to add some noise filtering or average window in visualization logic.

Engine speed - ISC flow duty - PSP switch - MAP sensor voltage

Above seems fine as ISC flow duty goes up as soon as psp switch is activated, rpm takes a while to come up, maybe it's the lag in milliseconds that it takes for iacv motor to move the shaft.

Engine speed - ISC flow duty - PSP switch - battery voltage

Looks about fine. Goes to around 12.8V - 13V sometimes so not sure about that. But otherwise seems to remain in good zone (as per me... looks bad actually but self soothing). Though drops are momentary so maybe it's all good.

Engine speed - ISC flow duty - PSP switch - injector pulse width

Injector pulse width does go up, should check if it's any different with IAT sensor disconnected.

Engine speed - ISC flow duty - PSP switch - Ignition advance (calc. not confirmed but should be near)

[Xulfiqar]

Set base idle to 700 rpm then retest.

[thatdevsherry]

Is setting to 700 fine right? Not sure what FSM means when it says idle should be 700 +- 50 with isc duty of 3-30%. Most videos I saw were setting just below intended idle. FSM gives a range so not sure what's the best base idle to set for it.

I'll move to 700 while keeping isc flow duty >= 3%. Will retest.