I think base timing should be about 5 degrees with the wire grounded... Can you not get it to come down?

 

well the thing is that I timed it already with a timing light and the connector grounded, so I wonder why its not okay if the timing was on the mark (5degree). Should I ajust it according to my datalogger by moving the CAS while the engine is running?

My knock sensor connector was broken (female side) so I replaced it. I read that the advance timing changes if the knock sensor is broken, but is there a way to test it... If you notice on the graphics, the car didn't have any knocks....

 

No, base timing should be 5 degrees with the wire grounded. Once you unground the wire, the computer will take over and bump it up to the 18 range at idle... (I don't know specifically what it is). Is the log showing advance with the pin grounded? If so, it is within the range. Your throttle position sensor, however, is not good. It should be a nice smooth line that increases and decreases as you press and release the gas pedal. Yours looks typical of one with a dead spot. Put a meter on terminals 3 and 1 of your tps and open the throttle very slowly (key on engine off). You should see the voltage go from .4-1volt as you open it. Watch for voltage just dropping off and back on again. If it does replace the tps. If it's nice and smooth all the way through the range of motion, check the connector and wiring to the ecu.

That throttle position sensor being out of whack like that is going to make everything go haywire...

EDIT: Actually that timing does look off. The numbers were really hard to read on my screen. But if the pin is UNgrounded, you don't have near enough advance.. If it is grounded, it is a little too much. It LOOKS like its between 6 and 8 but its really hard to tell on this screen. You should check into that TPS first though; it shouldn't look like that. Then re-check your base timing with the pin grounded and a timing light.

 

Thanks for the answer, nice and clear!

Actually, I did the ignition timing WITH the pin grounded and the Timing light on the 1st piston like it should be. It looked like the timing was on the mark. When I tried to ground the pin, the datalogger gave me a value of 22, the pin was well grounded, because I heard that the idle changed. Is it true that the knock sensor can untime the ignition if its broken?

I'll check for the tps voltage like you said and I'll keep it update!
Thanks

 

I would do it just like you did; pin grounded and light connected to #1 and use the mark on the crank. Maybe the ecu is giving it in something other than degrees before top dead center, I don't know... and yeah, I have seen broken knock sensors cause timing to be too far retarded, but nothing like you are experiencing... usually just loss of power and a little worse gas mileage. Your log shows NOTHING though, which it shouldn't... there should be somthing going on there I am pretty sure. The ecu may see no activity and go into some kind of failsafe mode, but I don't think it would do everything yours is doing... just the lack of power and using more gas. Or it may think there is no knock in which case its just going to follow the pre-programmed timing maps, which wouldn't cause your problems either.

 

***Update***
I tried a few combination to test the TPS, didn't have any reading. Key on, crocodile type wires plugged to my multi-tester. I tried 1and3 and put it on DC 40 just like I would test a battery, NADA, not even 0,01V, tried in MV, nothing... Tought maybe I misplaced the wire so tried 2-4 (wich may have been 1&3) DC40 or MV and nothing too!!

however, when engine is running I have a reading, wrong but still.... any clues?

 

I think its a bad tps. Change you meter to ohms and take a reading across 2 and 4. It should be smooth as you open the throttle from 3500-6500 ohms.

 

Your tps is all over the place. Also you injector duty cycle is very wavy. This may be because of the tps, or maybe more a mas issue, get yourself a new tps and do another log. Have you done a proper boost leak test. The factory boost gauge is a load gauge, it works off tps and mas air flow. Replace the tps it's obviously stuffed then keep looking.

 

Well, the boost leak test I did was removing the filter, putted a whistle plugged to my compressor with a airtight bag onto the whole and shot 10psi in the system. I found a few leaks at first, but fixed them and now, I can hear some air, but I can't find a leak so I tought it may be just the flow in the plenum ... should it be completely silent or what??

 

Check the ohms on the tps before you just replace it. That way you can tell if the tps is just bad or if you have an ecu problem. I had a car once that had a tps code and all the symptoms of a bad tps. I checked the actual sensor but it was good. So then I checked continuity to the ecu an it was good also. Finally, I hooked up a scanner and watched the live data stream. The voltage reported by the ecu was way different than what my meter was saying as it sat right next to the scanner. Replaced the ecu and everything was fine. I didn't want to scare you unnecessarily, but I was giving you the steps to try and rule out the ecu . And injector pulse width is all over the place also, but the tps readings would cause that. If we check the tps, then the wiring from that to the ecu, and it still is showing tps readings like your data log, I would diagnose it as an ecu problem. I would also put a voltmeter on the tps and watch a live data stream and see if they match. If the tps is smooth and the datastream isn't you could have a bad connection or a broken wire that is intermittently making contact. Move the voltmeter to the ecu pins for tps and see if it still does it. If its smooth there but the ecu reports choppy, its the ecu. I always try to double verify an ecu failure because they are normally rare (not on old mitsubishis, but overall pretty rare).

As for the knock sensor... I know it sucks, but more than likely if there is NOTHING on the log, you have a bad sensor. Those sensors tend to crack from the heat of the engine, and when they do, they don't give any readings just like yours is doing. The car I mentioned before that had the bad sensors was an Infiniti Q45 which uses the same type of sensor. That particular car would use a safer more retarded timing map when it didn't have a knock signal. I'm not sure if Mitsubishi does the same thing or not, but I highly doubt it would do what yours is doing... I was going to address that after you verified the tps because changing the knock sensor is kind of a pain and I don't think it is causing your problem.

As far as sensors necessary for running, besides MAF, tps has to be number 2. It could very well be causing every drivability problem you are having, so make sure its bad and that all the wiring is good before blaming the ecu though.

 

I did check the resistance in ohms of the tps sensor and I had a 4.5 reading on idle. but when I started opening the throttle, resistance went back to 0 until I put it back to idle and it shows 4.5kohms again. I don't know if this is the sign of a bad tps or if I just did it wrong, but I'll do it again tonight to be sure.

I'm not sure I understand how to test the ECU though I shoul plug the voltmeter to the tps harness or directly on the ecu pin? 7
thanks ill keep you update tonight!

 

Sounds like you won't have to worry about any further tests... That TPS is definitely bad... That's exactly what happens to them; they get dead spots like you are describing.

 

Ok thanks man you are really helpfull! About that knock sensor, I mentionned I replaced the female connector (the one going to the ECU, but I'm pretty sure I did it wrong, because I never crimped connectors before and mine was not quite beautiful loll. So how do I test it? Key on and voltmeter or ohms? what are the values I should get? I tried to take a look in the manual, but couldnt find anything!

Many thanks

 

Up!!

Just did the test of the TPS again, but I'm a little confused. Is the 4th pin the *highest* when you see it installed on the engine?? Because I had a test that gave me 0.76k0hms at idle and went up all the way to 5.25 smoothly....Another started at 4.65komhs and when I opened the throttle, it went down to 0 straight until back to idle. In each case, I thinks its out of specs that the tps is down to 0,75 ohms at idle... One of the test is 1&3 and other 2&4 but i don't know wich is what

thanks

 

If the little lock tab is on top, pin 1 is on the left.

 

To check the knock sensor you can check resistance between the side with the white wire and ground with the plug disconnected. I think if it's less than 500kohm its bad. This was the case with the Infiniti. The sensors LOOK the same, but I could not find a procedure for Mitsubishi either. I checked the FSM and Mitchell

 

OK so that would make number 1 the highest! redid the test and I had 0,75kohms at idle to 5,25kohms (throttlr at 30%). After just to see, I ajusted it with the feeler gauge and everythiing till my datalogg shows a 18% and redid the test (0,30kohms-4.67 when full throttle)

So its not a dead spot, but it came up smoothly the whole time. But specs are 3500-6500 so im way off... what do you think? bad tps still?

 

TPS

Injector duty Cycle

Injector pulse width

here is my tps on a new log I did on neutral... last one was on the road. The sensor had a good response to the gas pedal, I tried to maintain and it did, I tried to give some kicks and the reaction was there too! I don't understand why the tps looks good when on neutral but looks all over the place on the road. Even if the car was jerking, I maintained the gas pedal so why did he read like that? I don't know if it helps, but I had to turn the tps fully counterclockwise to have that 18% value on idle...

Injectors have a strange reaction though... I had them clean and tested the resistance, 3.5-4.2ohms each. I installed brand new connectors. before the ECU gave me a code 41, but it didn't reappear. Bad fuel pressure?

 

Now you have figured out the right pin try the voltage test again. Just swap it around to volts on the meter and test the same pins. With key on engine off, your the tps should read from around 0.5v at closed throttle to 4.5v at full throttle. From your log unless the scaling is off, the tps should go right up to 100% not 30%.

 

Just did exactly what you said, 2&4 to meter and put it on volts. Not even 0.1V at idle and not moving when I turn the throttle. It looks like its completely dead! And I turned the key on even if I don't know whats the purpus of the harness is disconnected?? And on my log I didnt push throotle to fully open on the lane I showed, but it went up to about 90 on a kick I did with the pedal fully in

 

If you are checking voltage and not resistance, the harness needs to be connected and the key on. My datalogger shows 13% at idle, not 18%. Also, the TPS shouldn't be in any extreme position... it should be about in the middle of its travel. Your next step would be to connect a voltmeter to the plug at the tps, and have someone watch it as you drive. Use your datalogger and see if your friend sees the dead spots that are in the log. If not, move the voltmeter to the ecu plug and try again. You are looking for a bad connection between the ecu and the tps. Strange as it sounds, you may have had both a bad TPS and a bad connection to the ecu.

The procedure for setting the tps is posted on 3si somewhere... I am assuming you are using that guide?

 

yes I am using that guide. I have a good feeler gauge and I did every steps. However, i don't understand how to test a bad connection to the ECU! What if it was the ECU itself that is the issue? How do I test that? I bought a new TPS, I'll set it and see what happens... If its a bad harness or a bad ECU how do i make sure before I buy a new ECU? Also have to change the knock sensor.... what torque is required to reinstall another?
thanks

 

Ok, you attach your volt meter directly to the sensor while it is connected to the harness, and run the voltmeter leads inside the car, then drive it. Have a friend watch the voltmeter as you drive and compare it to what your datalogger is telling you (which is what the ecu is telling it). They should match. If they don't, you could have a connection problem. In other words, if your friend watching the voltmeter says there are no spikes or dead spots, but your data log says there is, you could have a bad connection from the tps to the ecu. So, next, connect the the voltmeter to the ecu plug directly with everything connected as it should be. Have your friend watch again and see if your data log and voltmeter match. If they match, as in you see spikes and dead spots on both the voltmeter and the datalogger when before you only saw the spikes and dead spots on the datalogger, you just verified a wiring problem. If they DO NOT match, it is probably your ecu. If you are testing the input just before it goes into the ecu and it doesn't match what is coming out, you know the ecu is somehow messing it up inside.

I couldn't find a specific torque for the knock sensor so just make it snug. Don't crank it down and risk cracking it.

 

ok thanks!!I understand the steps, but to test the good wire While it is connected, do I need to skin a part of the wire and plug like a crocodile wire clip on it, then to my voltmeter?? I know I may sound noob but I really suck at electronics loll

 

I have clips that you just clamp onto the wire and little pins punch through the insulation... If you do it your way be careful they don't touch while your driving. Maybe tape them up real good with some electrical tape.

Here's the ecu diagram if you need it; I highlighted the tps wires.

 

Awesome man thanks a lot!!! That diagram is a killer! You said the pins I want are 2&3 right?But why are they in a different order on the diagram(1,2,4,3) than on the TPS itself? And is the way I plug the voltmeter important? I guess if I plug it the wrong way I'll just have negative values??

I'll receive my order monday I can't wait to test that! Hope this thread will be usefull for a bunch of people too!

 

No problem! Yes, 2 and 3, and they probably just changed the order to make drawing it easier (?) Because 4 is the ground and both the switch and the resistor connect to it, which would have looked sloppy if they put 4 at the end. And you are correct, if you plug the meter in backwards it will just show negative.

It has been a weird one, I'll say that! A verified bad tps, and wiring or ecu problems at the same time with the same component... I can't remember the last time that happened! LOL

 

Well for now I don't know if the wiring or the ECU is bad, I'll have to wait for my new TPS before I test the whole thing. I never tested the wiring since I didn't understand how to do it correctly before ordering the sensor LOL.

Just the TPS resistance that was way too low if really it should start at 3500ohms at idle to 6500ohms fully open. On my datalog I couldn't get it set under 18% too, so I hope that the reading of the datalog was good and the ECU or harness is not bad. Anyways I'll let you know whats up once I get it done!

Thanks again

 

No problem! Yeah, Mitchell says if it is not from 3500-6500 to replace. Period. So that is a verified bad tps according to them.

 

Little updtate, I think its a ECU or Harness... I calibrated my new TPS to my throttle body and I have the same values as my old one. Reverified the testing method and it says pin 1&4 you should have 3.5-6.5kohms (I have 5.33kohms) and when you plug on 2&4 it should go up smoothly as you open the throttle (I have 0.80-5.33kohms).... Didn't finish reinstalling everything yet, but I think I misstested the issue -_-"

If I can crimp that knock sensor connector well for once, I may be able to reinstall everything ans see how it goes. Anyone on this forum lives in Canada in the Quebec province so I can plug my ECU in their car to test it?? I'm willing to pay

Thanks

 

Yeah, sounds like the tps is fine. The BEST way to test this would be to put a voltmeter right on the ecu plug; ecu plugged in to the ecu and probes stuck in from the back. Then use a commercial scanner or something that gives you tps voltage from the ecu directly and see if your voltmeter and scanner match. I don't know if your data logging software will show the actual voltage, but if it does, that would work. This is exactly the problem I had with a GM ecu. Tps was replaced and calibrated but when you looked at the meter it had a completely different voltage than what the ecu was reporting. I replaced the ecu and it was fixed.

 

Thanks for the reply. Yeah I think I messed up with the testing, I guess I have a spare one loll. For the testing, I don't think my datalog gives me the voltage. What i see is the position of the sensor and not the voltage.

I'll have to check if there is a way to see the voltage from the ecu. It may be my harness tough, and that would be a pain to do! I'm looking for someone to put my ECU in his car, but didn't find yet!