Honda Civic Forums banner

1 - 20 of 24 Posts

·
Registered
Joined
·
17 Posts
Discussion Starter #1
Hi

I've got a 95 1.5 16v coupe with a D15B7 engine number.

It failed it MOT on Emissons with a CO 3.31% & lambda of 0.901, the HC is fine @ 148ppm.

I've changed the Lambda probe already and its made no difference, I've also removed the back box as the MOT guy reconned it had calapsed internally and with causing a large back pressure.

In fact the numbers are all over the place, we did put it back on the machine with the new probe and it came in but by the time we started the actual test it back up to 3% again and today it was right up to 8% with 0.8 lambda.

Am i right in thinking that it isnt the Cat as only the CO reading would be out.

There no codes coming up, although i did have 41 and 3 but i think there due to me fitting an incorrect probe and the MOT guy disconnecting the MAP sensor with it running earlier today. I've since cleared the codes and taken it for 10 minute run, retested it and there are now no faults.

It runs fine although there did seem to be a small change in engine tone a few weeks ago but i've put that down to the exhaust.

Any idea's, all i can think of is a wiring fault or the ECU? or would a faulty MAP sensor cause this sort of fault?
 

·
Registered
Joined
·
7,629 Posts
would not change a map sensor, did you change the oxygen sensor then get the emisions tested again if not you need to take the car for a hard drive to rejuvenate the cat as the oxygen sensor has been killing the cat with high co content,
if that doesnt work you may have a cat gone down ,
but my next port of call because of your random readings would be the engine tempurature sensor this will give the ecu wrong reading effectively richening the car because the sensor will richen the co content if its reading cold for example.
 

·
Registered
Joined
·
17 Posts
Discussion Starter #4
Hi
Thanks for the reply

i've had it retested a couple of times now, and given it a few hard drives + the MOT guy is quite good and did the test for my a few times giving it a few hard revs between tests.

If the cat was gone as well i would have would have thought the lambda would now be ok and CO still out but they seem directly corelated lambda goes down CO goes up. It all points to a lambda issue so unless i've been sold a duff probe i'm thinking wiring or ecu.

I've been taking another look at things, ive hooked up a meter directly the loom (after the sensor plug) and am getting about 0.38V at idle, it does rise a bit when i rev it but it a bit late in evening to be reving cars up especailly now i've got a straight through pipe instead of a rear box.

The thing is i would expect the output fultuate more if it was working properly in closed loop, its as if the ecu is set to open loop, but i must be getting an output from the sensor as it changes when i rev the car but its sems like its not getting back to the ecu.

I think i need to test the sensor output at the ecu and also test the continuity of the signal wire.

its also weird that the figures all came in fine for a short period then went completely out again, all points to an intermittant wiring fault i think, or maybe i'm completely off track.


I've now been back to the MOT station 4 times just to test the CO, i know them quite well but i think there getting a fell up with me now so i really want to find the fault before i take it back again.


Rob .
 

·
Registered
Joined
·
17 Posts
Discussion Starter #5
i was thinking temp sensor as well.

I think i'll see if can test the outputs of that and the map sensor tomorrow evening. Haynes explains seems to have a resonable sections on testing engine sensors for a change.

If all this fails, does anyone want to buy a cheap coupe that failed its MOT only on emissions :(
 

·
Registered
Joined
·
7,629 Posts
rob said:
If the cat was gone as well i would have would have thought the lambda would now be ok and CO still out but they seem directly corelated lambda goes down CO goes up. It all points to a lambda issue so unless i've been sold a duff probe i'm thinking wiring or ecu.
Lambda is a direct reading of the oxygen content in the exhaust and fluctuates with the co because it has to ,
as a direct relation to the co content higher co reading- lambda reading goes down , high lambda means only that there is to much oxygen in the exhaust system , even a slight blow in the exhaust will cause high lambda reading as there is more oxygen in the exhaust than there should be and its being drawn in before the tail pipe,
so basically there is no such thing as a lambda issue related to the oxygen prob ,
this is a common mistake as the oxygen sensor is sometimes called a lambda probe but thats not what it does
it measure fuel to air ratio and adjusts the co in relation to the readings it gets ,
also if your going to test the oxygen sensor the voltage readings should go up and down on idle and fast idle any where in a 0-1v range and 0-5v range dependant on sensor most are 1v sensors,
hope this helps clear things a little more any ?s just ask.
 

·
Registered
Joined
·
17 Posts
Discussion Starter #7
So if the cat is clogged the lambda reading would be genuinely high due to the backpressure as its a pre cat sensor?

If i take it for a hard drive it should burn the crap off the cat and it will clog if left idleing? YES AND PLODDING ROUND TOWN ALL THE TIME

Also i was seeing very little flutation in the voltage reading at the sensor, if it was working properly in closed i would expect a constent flutation in reading between 0-1v as the sensor tracks the fueling, right? YES

My reading seems constant only changing if i blip the throttle. As if the fueling is working in open loop. NO IT SHOULD FLUCTUATE ALL THE TIME , THEY DO ALTER ON A BLIB BUT SHOULD STILL BE IN A WAVE PATTERN WHEN AT REV AND IDLE.

Sorry if i'm not getting this right, i normally only work with carbs but i understand closed loop systems as i'm an electronic engineer. NO PROBLEM ANY TIME.
 

·
Registered
Joined
·
17 Posts
Discussion Starter #9
Hi

Thanks for the reply.

i've had another play with it tonight and it definately doesnt seem to be going into closed loop.

I disconnected the coolant and put a dummy load of 300ohm across the connect to simulate a warm engine but i'm still getting 0.38v at the sensor, a also disconnected a vacum hose and the output goes right down to 0.1v but still no tracking.

i reconnected the hose and took it for a hard drive then measured again and the output was up 0.6v then came down gradually as i let it idle.

I then removed the dummy laod and reconnected the temp sensor and as soon as plugged it in the fan cut in, i'm assuming thats becasue the engine was hotter than my load resistor was indicating. this would indicate that the coolant sensor works ok, i would test i directly but it really hard to get a meter acorss it in situ.

i also disconnected the lambda sensor and i still get 0.38v on the loom side of the connector indicating the ECU is producing a bias voltage which is normal for use when the probe is still cold. This indicates i have a connection to the ECU.

I'm beginning to run out of ideas, all i can think is the ECU itself is knackered?

If i've got any of this wrong please comment?

Thanks
Rob
 

·
Registered
Joined
·
7,629 Posts
hi
its the probe its self that produces the voltage not the ecu the only voltage going to the probe is to power up the heater
sorry if you already know that , what i am saying is its not the ecu that would stop you getting a wave voltage ,
are you spiking the right wire (WHITE) sorry again if you are ,attached diagram with wiring colours on...
 

Attachments

·
Registered
Joined
·
17 Posts
Discussion Starter #11
Hi paul

I have spiked the signal wire.

I've done some reading up, as you know a sensor is effectively open circuit until it reaches operating temp. To overcome this a bias voltage it set or ouput from the ecu (i cant remember where i read this or if it was talking in general) but it seems to be true as if i disconnect the sensor i still get 0.38v on the loom which is constant i.e. doesnt change at all even if if rev the car or remove the vacum pipe.

The bias voltage is to effectively set a safe rich condition until the probe has a low enough impedance to take over at which point the ecu will detect this and change to closed circuit operation.
 

·
Registered
Joined
·
7,629 Posts
right forget about the ecu for a minute ,
disconnect the o2 sensor from the loom warm the engine up and measure the signal wire from the sensor you should still get voltage there in wave form 0-1v this way you are taking the ecu out of the equation a voltage should still be produced as the sensor produces the voltage if it doesnt its buggered , did you buy a cheap one or good quality eg, bosch , beru , ntk, etc
 

·
Registered
Joined
·
17 Posts
Discussion Starter #13
i should have done that but i forgot, i will have to do this tomorrow.

i would guess the sensor was a fair chep one. It was £55, the genuine honda part was £170.

It is a "Fuel Parts Uk Ltd" part.

http://www.fpuk.co.uk/main/index.asp" onclick="window.open(this.href);return false;
 

·
Registered
Joined
·
17 Posts
Discussion Starter #15
maybe i've got a duff probe then, i'll check the output with it disconnected tomorrow. i think it is at least doing something as it drop right down giving a really lean reading when i pull a vacum hose off (as i would expect).

I will also check the old sensor as per this artical, if the old probe is faulty its likely the new one is as well.

http://www.pallant.f2s.com/rover/tech/lambda.htm" onclick="window.open(this.href);return false;.

I'm also wondering if its not going it closed loop (oscillating) becasue i'm testing at idle and probe is not hot emough although as its a heated probe i would have thought this wouldn't matter.
 

·
Registered
Joined
·
17 Posts
Discussion Starter #16
I've just tested the old probe as per the procedure give on the link above and it seems to work perfectly?

I.e. heat with blow torch and it goes +0.9v remove heat and it drops to 0.1v in about 1.5 secs.
 

·
Registered
Joined
·
17 Posts
Discussion Starter #18
Hi Paul

I think i will refit the old sensor, i doubt i'll be able to return the new sensor assuming is working, i'll test in the same way. But i brought it from a local motor factor.

I'm still not sure i understand why lambda would be out if the cat is knackered.

would a faulty cat cause me to see a constant voltage at the the sensor?

I would have thought a faulty cat would have just stopped the C0 being turned in CO2 without having a huge effect on the on the lambda output.

As we discussed eariler if the cat was clogged it may cause a CO build up and an increased mixture but since i removed the back box the car is extremely free reving, in fact it runs better than its ever run.

Sorry i'm doubting you but i'm just reluctant to buy a new cat to test out the theory out this is especally true as the car only cost £350 in the first (a bargain price i know as it was from a mate). Also everyone else i've spoken to, a couple of MOT inspectors etc recon only the Co and hc reading change significantly when the cat fails and lambda reading is an indication of the cars fueling so normal the sensor and a failure of the feedback loop (i.e. wiring or ECU).

maybe i'll either have to get rid of the car or bite the bullet and by a cat.

Rob
 

·
Registered
Joined
·
17 Posts
Discussion Starter #19
well progress, maybe??

i check the sensor ouput withit disconnected from the loom and it seemed to do what i thought it should I.e. sat at about 0.4 drop to 0.1 when i took a vacum hose off.

so i fitted the old sensor back again, i also took the plugs out and they definately need changing, the gap were huge about 2.5mm so closed them down to 1.1mm, they a good light brown colour.

But as soon as i started the car it when striaght into closed loop with good fluctation between 0.2v and 0.8v. I then did the same out of car test on the new sensor and got hardly anyout.

I'm going to pop the car into the garage tomorrow if a get a chance and get it back on the machine after a hard run down the bypass.

I'm hoping the fault was a clogged rear box after all but i had already changed to a bad sensor when i fixed it.

fingers crossed !!
 
1 - 20 of 24 Posts
Top