|
|
|
Aug 29, 2012 12:20:42 GMT
|
I have a tough one for you today; The car is a 1987 subaru xt turbo, the problem is a non steady idle (the engine shakes a bit) and the inability to keep a steady rpm whilst driving, I'm guessing it's the same "stumble" at idle as well as whilst driving. I've replaced and checked loads, but today I finally put a scope on the injector signal and was a bit surprised. It looks like the injectors either get a long puls or a short puls, without me touching the accelerator or anything else changing. The short and long pulses differ a factor 4 and it's consistent, everytime exactly the same difference (2mS and 500nS). All injectors do this and at every (sensible) rpm. This wouldn't be so strange if it weren't for the fact that the ECU does not have active idle control (no stepper motor or anything) and also does not have an O2 sensor. So eventhough it might look like the ECU is actively adjusting to something, I don't see how since it has no input on what to do. So my question is (sorry for the book) is it normal for the injector duty cycle to have this behavior? And if not, what could cause it? I already swapped ECUs, injectors, TPS, CTS, MAF, checked fuel pressure, checked supply voltage (no ripples or noise) and fuel pressure regulator. This is a shoddy phone picture of my oscilloscoop showing the signal, 2 distinct pulses can be seen.   |
| |
Last Edit: Aug 30, 2012 7:38:03 GMT by joostvdw
|
|
|
|
andyborris
Posted a lot
Freedom is just another word for nothing left to lose.
Posts: 2,161
|
|
Aug 29, 2012 18:30:36 GMT
|
|
Do you mean that every 4th pulse is wrong and which is the correct pulse, 2mS or 500nS?
And from what you've swapped over already, it can only be a wiring issue! No O2 sensor, so it can't be an air leak...might be a leak in the vacuum pipe to the pressure regulator?
|
| |
|
|
gruss
Part of things
Posts: 242
|
|
Aug 29, 2012 19:15:20 GMT
|
|
Has the vehicle got aircon? Have you load tested the powers and grounds for the Ecu? How is the idle controlled on the engine?
|
| |
Daihatsu Mira TR-XX
Suzuki Alto Works
|
|
|
|
|
|
|
andyborris: I meant to say that the difference in length between pulses is a factor of 4, 4x500nS=2mS. The actual behavior is quite irratic so I don't know what the correct pulsewidth length is. I've checked fuel pressure (with an analog gauge) and it rises according to manifold pressure, so that doesn't seem to be the problem. Your wiring suggestion sounds plausible since I have a feeling it's some sort of interference. gruss: no aircon (sadly). How do I load test the powers and ground of the ECU? It already has a new ground from ECU directly to chassis. It has a crude type of choke which consists of a bimetallic valve which slowly closes while the engine warms up, but it never worked in the 7 years I've had this car. There is 1 screw on the TB which sets the idle by creating a small air leak and that's it. The idle does vary with engine and outside temp (high idle with hot engine in summer and low idle in winter, probably because it doesn't have a "choke") I must say however, this car always had a slight flicker in the (digital) dashboard and a leaky sunroof, so the grounding could be a problem. Changing alternator and battery didn't change anything. I'll clean the grounding from engine to chassis to battery tomorrow, I've done this before, so I have no hope in solving anything by this, but it can't do any harm... Thanks for the replies so far, there is hardly any specific information about this car (like injector duty cycle signals...) so I'm really looking for some general knowledgeble people.
|
| |
|
|
andyborris
Posted a lot
Freedom is just another word for nothing left to lose.
Posts: 2,161
|
|
|
|
|
I'd suggest fixing the idle valve, (not too pricey on Ebay and any of the Bosch valves should suit), because that seems to be the only way of controlling the idle.
You may then find that the throttle screw is open too wide.....at least it'll be one less thing it could be!
|
| |
|
|
|
|
|
|
|
It's a good suggestion, but I've copied the part about the "auxiliary air valve" from the factory service manual, it better explains how it works than I ever could  You'll see though that it's not a generic 2-wire PWM controlled IAC...  |
| |
Last Edit: Aug 30, 2012 7:38:31 GMT by joostvdw
|
|
gruss
Part of things
Posts: 242
|
|
|
|
|
You can load test a number of ways. With the circuit loaded ie key on engine off carry out some voltage to voltage checks using a multi/meter on both powers and grounds. Disconnect both ends of the harness and provide your own power then measure using a test lamp on the other end. I personally use load pro leads and help me out no end at work. Google them for more info.
Is the aux valve PWM? resistance check of the coiled winding when the circuit is off. Is the Ecu reacting to the valve and adjusting the injector duty. Could you scope the aux valve in relation to the injectors?
|
| |
Daihatsu Mira TR-XX
Suzuki Alto Works
|
|
|
|
|
Aug 30, 2012 13:50:52 GMT
|
|
I can definitively say it's not PWM; I've just put the scope on it and got a steady signal. I've also cleaned up all contacts and got 0,3ohms between chassis battery and engine. The AAV doesn't seem broken, it shows 80ohms, but it's probably stuck. I already tried to clean it once to no avail, it's also a sealed unit so can't really do anything else then replace. Does anybody happen to know of any other car with this sort of "choke"?
|
| |
|
|
|
|
|
Aug 30, 2012 18:55:50 GMT
|
|
Aux air valve has no connection to the ECU - just connected to 12v with the engine running (quite often supplied in parallel with the fuel pump) and heats a bi-metallic strip which closes a valve. Used on loads of early injected cars including K-jet (Golf 1 GTi and early Golf 2 Gti, XR3i, Volvo 240, various Porsches, early injected BMW, even Ferraris. In UK the easiest source is Volvo 240s though they are getting quite rare in scrapyards now.
Note that there are quite a few part nos for what looks like identical parts. This is beause the shape of the slot inside (and maybe the heating rate) varies from model to model to give a different extra air profile. The Volvo one is pretty short. Golf GTI one is longer. These are the only two I have direct experience of.
Volvo ones have a coil resistance of 50 ohms. You can test by removing and looking down the hole. When cold you should see daylight through some kind of slot/half-moon in the disc. If you poke a thin screwdriver down you should be able to move the disc against a light spring. If stuck you could try a does of carb cleaner and some more wiggling.
If it moves then you can try connecting it to a 12v battery and should see it close over a period of 30 seconds to 2 mins.
As regards the stumble and pulse width glitch, what type of air-flow meter does this have? I'd be wondering about this or the wiring to it. Quite often the inlet air temp sensor is also in the flow meter and this (or more likely the wiring to it) could also be suspect.
Cheers
Nick
|
| |
1967 Triumph Vitesse convertible (old friend)
1996 Audi A6 2.5 TDI Avant (still durability testing)
1972 GT6 Mk3 (Restored after loong rest & getting the hang of being a car again)
|
|
|
|
|
Sept 3, 2012 12:41:40 GMT
|
Thanks for that mm post Nick, I will hit the yards soon and grab a similar AAV from an early injected car. I'm guessing any AAV is better than none at all The car has a hotwire MAF with integrated air temp sensor. Strangely enough the scope doesn't pick anything up at the MAF but that might be because it can't trigger on a constant voltage, the DVM should yield some results. I could however disconnect the air temp sensor from the plug (either at the ECU or at the sensor) and just see if the duty cycle changes... Keep the ideas coming! |
| |
|
|
|
|
|
|
|
Sept 5, 2012 12:12:12 GMT
|
New developments; Unplugging the MAF makes the car idle slightly better so that gave me something to look at. Secondly the MAF has 2 ground wires and 1 of them had almost 7k ohms resistance to ground so a jumper wire to see if that changed anything was quickly in place. The injector pwm didn't change however, but it can't be good to have a bad ground on the MAF so it was nice to find it but it wasn't THE solution I was looking for. The fluctuation did seem to be slightly less then before, more of the wide signal and less of the small. Then I had a second go at using my oscilloscope on the MAF signal and got this:  Which coincidentally has the same timebase and period as my previous measurements on the injectors. It fluctuates quite a bit as you can see, so this might explain the injector fluctuations. Where do we stand? Looks like the MAF is generating that signal, but I have no idea why it would do this? I could swap back to my other MAF but since I didn't notice a change the first time around I'm guessing it won't do any good this time either. |
| |
Last Edit: Sept 5, 2012 12:13:19 GMT by joostvdw
|
|
andyborris
Posted a lot
Freedom is just another word for nothing left to lose.
Posts: 2,161
|
|
Sept 5, 2012 17:44:44 GMT
|
|
Wondering is the power source for the MAF is breaking down, is the MAF a hot wire, because it does look very much like the hot wire is cooling then heating!
|
| |
|
|
|
|
|
Sept 5, 2012 21:40:35 GMT
|
|
It is indeed a hot wire MAF, but I'm not sure it has a separate power source, there is 1 live feed going to it. Perhaps there is some regulator inside the MAF controller the power (and heat) of the hot wire? Any way of finding out (beside measuring the hot wire temp)?
|
| |
|
|
andyborris
Posted a lot
Freedom is just another word for nothing left to lose.
Posts: 2,161
|
|
|
|
|
More thinking dodgy connection, pretty sure that the hot wire runs on a fixed output, the airflow is measured by it's rate of cooling, no 12V means a blip.....shame you can't get one from a car known to be good.
|
| |
|
|
|
|
|
Sept 6, 2012 18:35:00 GMT
|
|
It's a shame I can't measure current with my oscilloscope, that would show if the hot wire heats up in the same matter.
Any other theories or things I can check?
|
| |
|
|
|
|
|
Sept 6, 2012 20:58:31 GMT
|
|
If you still get the duty cycle change with the MAF disconnected - then it isn't the MAF causing it.
Nick
|
| |
1967 Triumph Vitesse convertible (old friend)
1996 Audi A6 2.5 TDI Avant (still durability testing)
1972 GT6 Mk3 (Restored after loong rest & getting the hang of being a car again)
|
|
|
|
|
Sept 7, 2012 16:23:17 GMT
|
I decided that the AAV needed to work first, if it doesn't solve this problem it's always nice to have it functioning, I'm pretty sure that's hurting mpg when driving short trips. So I took it off the car and took it apart, wasn't really meant for disassembly, but nothing a drill, screwdriver and vice can't fix Turns out it wasn't stuck at all (!?) but the heating element/bimetallic spring was only mounted with 1 pin on 1 side, I guess it's hard to explain, but the effect is clear; instead of opening the valve it moved around in it's housing. I coudn't get the tiny roll pin out so I decided to slap some PU glue on the outside, glueing the connector solid to the body of the valve. I also bend the tip of the bimetallic spring some to gain some more movement but not sure yet if that's necessary, time will tell.  |
| |
|
|
|
|
|
Sept 11, 2012 9:16:56 GMT
|
|
All glued together and fitted in the car I now have a proper high and low idle. Which is nice, but it didn't fix my problem...
I also checked CO @ 900rpm which is 1.0% according to the gunson thingy, it's probably not the most accurate way of measuring but it will do, atleast I know it's somewhat oke. Did find out that it doesn't respond to adjustments of the CO screw, so that's a whole new problem.
Also managed to shock myself, but atleast I now know that the coil is good and it runs on 4 cilinders. I already knew both, but double checking never hur... wait, it did hurt!
|
| |
|
|
|
|
|
Sept 11, 2012 16:04:04 GMT
|
|
Unplugging the MAF didn't make the signal go away, the engine did run much smoother. Swapping in another MAF made it run a lot worse, hunting etc. so that MAF is definately not oke, the signal from it was also very noisy.
There is 1 last thing to do and that is permanently fix the bad ground on the MAF, but jumpering it to a known good earth didn't change a thing...
I'm kind of at a dead end right now, I've tried and checked everything I could possibly think of and the car is still not running right and also returning terrible mileage.
Could there be a mechanical trouble making the intake air surge, the MAF signal swing and the injectors jump accordingly?
|
| |
|
|
|
|
|
Sept 11, 2012 19:20:29 GMT
|
|
You said it doesn't have an O2 sensor, but maybe it should have one?
I thought terrible fuel mileage was normal on Subarus that age?
Can't think what else to suggest - you already changed/checked just about everything! Can you try powering (and earthing) the ECU direct to the battery to eliminate interferance from some other thing? Or maybe scope the supply and other input signals and see if anything shows up?
Nick
|
| |
1967 Triumph Vitesse convertible (old friend)
1996 Audi A6 2.5 TDI Avant (still durability testing)
1972 GT6 Mk3 (Restored after loong rest & getting the hang of being a car again)
|
|
|
|
You'll see though that it's not a generic 2-wire PWM controlled IAC...
