G'day

So the other day I stopped face in to a shop, in the car, and in the glass I could see that one of the low beam head lamps was not working. Haven't driven the car much at night. How did I miss that?
Hmm. Believe it or not a few days later I pulled into my car port and I had reason to use high beam and one of the high beams (on the other side) was out. What!!?? Hmm, I'd better do something about this.
Well the fuses are there to protect the circuits in an emergency, like a crash or crushed wires. So if something hasn't happened then it shouldn't be the fuses. Hmm most likely the bulb.
First cab off the rank is the high beam bulb on the drivers side. I pull the bulb out and...it's fine! Hmm a blown fuse? Why? I don't know which fuse it is. So I pull both of them out (left and right high beam). Nope. Fuses are fine? Replace the fuses, hit the high beams and I have power and ground at the bulbs connector. What's going on? The bulb is a H1 type and the electrical connector is a bit strange. The center contact is fairly regular but the outer contact is a sort-of clamp arrangement!? Sliding the bulb in and out of the connector reveals that this clamping action is very light. Could it be a poor outer connection?
I use a pair of long nosed pliers to bend the contacts and install the bulb back in the connector and test. yes! The heat of the bulb probably softened the metal and it lost it's electrical contact.
Next the passenger main beam. Er is this going to be complicated too? Could this just be the bulb? I'm going to bet on the bulb.Two weird problems in a row? Hmm well it is me. Yep! Just a blown bulb.
Now, I have to have a go at the engineers for this.....again. Why are the high beam and low beam bulbs different? The low beams are H7's 55Watt and the high beams are H1's 55Watt. If they were the same and you lost a low beam bulb, you could swap out one of the high beam bulbs...in a pinch. You can do without the high beams but you can't do without low beams.
*Sigh*

I've been doing a bit more thinking and research. (oh no! You say)
I've been wondering why the oxy sensor from the "black" C3 is producing better performance than the original.
More exactly is there a way to do better? Well, I'll need to know how it works in the first place before I can perfect it! (Ha! Perfect it? What a fat head) Ok then, try and make it better or find out why it is better.
So here is what I found: First up the sensor is called a "narrow band" sensor. That is; it outputs large (relatively) voltages over a very narrow band of air fuel ratios.
It's response to the air/fuel ratio (and the oxygen by-products) are NOT linear! You can see it's sort of a back-to-front 'S' curve.
In practice that means that for small changes in the fuel (air/fuel mixture) you get large changes in the Oxy sensor voltages.
Think of it like driving a car with a steering lock to lock of a 1/4 of a turn!! That is 1/8 of a turn locks the steering all the way to the left or right. You would "wobble" down the street as you tried to keep the car straight!!
So the poor engine ECU is trying to do this "juggling act" with the Oxy sensor swinging wildly for every tiny change the ECU makes to the air/fuel ratio. BUT like all good jugglers the secret is to keep the system "predictably unstable".
Hence the engine ECU keeps it swinging backwards and forwards in a predictable manor. Like balancing a pool cue on your finger. It's easier if you wiggle your hand backwards and forwards to keep it from falling! It's not "stable", but it doesn't fall down! It's very difficult to keep the pool cue dead vertical and stable, but it's easy to keep it "predictably unstable"!
That's why if you monitor the output of the upstream oxy sensor it looks like a sine wave.
Over all the engine ECU is trying to keep the output voltage swinging around 0.45 Volts, the "sweet spot" between too lean and too rich.
You might wonder why they don't just make the output of the O2 sensor linear...well they do! It's a more modern invention but its called a Wide band O2 sensor. And no, you cant use them to replace a narrow band sensor, it doesn't work that way. We are stuck with what we have got!
As you can imagine if the shape of that backward 's' curve were to change, the output of the oxy sensor would not be correct. If that's the case then the poor engine ECU is working on duff gen (bad info)! It can only react to what it's getting from the oxy sensor...right or wrong!!
The engine ECU also looks at how much fuel it has calculated it has used (number of pulses X fuel per pulse). Then compares that to what the engineers said it should have used (as an average). This is called the fuel "trim". In percentage, and can be positive or negative. That is; the engine (actually) used more or less than was "calculated". It's rare that the value is zero, all engines are imperfect. Below 5% is acceptable, below 2% is very good.

Ok, next time I have some practical experiments testing my oxy sensors and I got some "interesting" results!

Well I promised a part 2 and this is sort of it.
My plan was to see if I could tell a slightly bad oxy sensor from a slightly good sensor. I mean a good sensor. Well spoilers...I failed. But I did find something interesting.
That fountain of knowledge, YouTube, said I could use a gas torch to test oxy sensors. Well sign me up!
The theory is that the heat of the flame will activate the sensor and because the flame will have consumed the oxygen you will remove the oxygen from the element of the oxy sensor too. And, as we learnt, minimum oxygen equals maximum voltage out.
First up the genuine sensor that I removed from the engine. This is the one that worked, but I wasn't happy with it's performance. Stick it in the vice...hit it with the torch. It takes a few seconds, but it eventually springs into life. The flame is too hot so I need to waive it over the sensor. I cant leave it in the one spot.
First up the voltage readings are all over the shop. It's clear I wont be able to detect subtle differences between any oxy sensors! But with the flame on the sensor it goes into the high 900 millivolts. Well this will test if a sensor works or not but not much more than that.
NEXT...a brand new, cheap Chi...er...knock off sensor. Takes a bit of coaxing but we get a result. As I said the voltage is allover the place, so is it better? Is it worse? Cant tell. It gets up well into the 900 millivolt range just like the other one.
Now I'll just remove the flame..... WTF!!! Those guys on the YouTube said that the sensor should react quickly! So when you remove the flame, the sensor voltage should drop very quickly.
It takes a few seconds for this one to drop back to about 0!
I didn't pay attention with the genuine sensor, so I put it back into the vice and lit it back up. Sure enough as soon as I removed the flame instantly the voltage drops to near 0 Volts.
Well I didn't expect that! That's a big fail on the cheap O2 sensor front. Odd, I put one of these cheap sensors in the "other" C3 that I sold and it seemed to run fine!? Bad quality control? Different batch?
Hmmmmmm....should I put it in the engine and see what happens? I could monitor the output and shut down the experiment if the mixture gets too rich.
Hmmm.

I have a bit of a retraction to make. A retraction or an addition, you decide.

It bothered me since yesterday, when I posted about the knock-off oxy sensor, why did the "other" (cheap) oxy sensor on the "other" C3 that I fixed, work...better?

Ok, to reiterate; I fixed another C3 a while ago and to get it running I used a cheap knock off upstream oxygen sensor. I bought it from Ebay. What did I have to loose? The engine was in a mess and I decided the oxy sensor in it needed replacing. A genuine NTK sensor was going to be expensive, If I could even find a new one.
After all that, the engine in the first car ran well, in fact when I got the second car up and running, the second car's performance was not as spritely (in my opinion) as the first one. Unfortunately I sold the first car before I got the second car up and running. So I could only rely on my memory as to how well it ran! I couldn't actually compare them. Anyway I was sure that the engine in the second one (the one that I am now driving) was lacking and should do better! So I have been on a quest to get a bit more out of the engine. I have changed this and changed that...to no avail!
Despite the original, genuine, oxy sensor "working", I was desperate to gain more performance out of the engine. I had just recently replaced the original oxy sensor with one out of my "black wrecker" C3 now the performance is much better!! Changing the sensor proved to be the secret!

Anyhow, It bugged me that cheap O2 sensor on the first car was apparently better than the genuine (used) one that I had been using in my current car.

So now we come to the "failure" of the cheap sensor that I outlined above. BUT was it just easy to say it was a bad unit? Was it that simple? The other cheap sensor in the other car ran well, but this one is a dud? What if it needed a burn in period, more heat...or something? So I put the sensor back into the vice and lit up the propane torch.

Initially the results were much the same; poor response both in removing oxygen (heating) and adding oxygen (cooling). But this time I thought "what have I got to loose"? If I melt it into a molten blob, who cares? So I kept up the cycles of heating and cooling. Getting hotter and hotter. When I got the interior of the sensor glowing things began to change. The swings between high voltages and low volts were bigger and changing more rapidly! Ah, you've got to get this thing hot! No, really hot! I have no qualms in saying that it's really hot in the engine exhaust! So I think my test was a failure not the sensor.

So why did the "old" genuine sensor perform better than the new "knock-off" sensor in the first run of my "oxy sensor Vs gas torch" test? Hmm, don't know. Some sort of run-in required on new sensors? Maybe these sensors need to get very hot and there is actually something wrong with the genuine sensor? When I heated the cheap sensor up I got much higher voltages than the genuine one (over 900 millivolts).

So I intend to put it into the engine and see what happens!? Stay tuned.

Well I have to eat my words...again!
The cheap knock-off sensor performed well! How long will it last? Dunno (but the one in the other C3 is still going). Do I recommend it? Sorry but it would be silly of me to "recommend" anything other than an OEM part. But... it worked for me.
A brand new oxy sensor is about £60 and I payed £10 for the knock-off. By the way a genuine one (for the TU3 engine) is an NTK OZA495-pg2 (A/01). You can get them from the usual suspects. That, is probably a lot cheaper than buying the OEM part from Citroen (which is exactly the same part)!! Just to re-cap; the original oxy sensor gave, at best, average performance.
The one I pulled out of the wrecker gave much better performance but the heater element went open circuit and caused an engine light.
Hmmm I might argue the knock-off sensor is a fraction bit better again.(!?) Again...only my perception.
The biggest problem with the original sensor was that from a slow or standing start the original sensor caused the engine to "hesitate" off the line. So if I came to a "round-about" and slowed down, but I quickly got a break and floored the accelerator, the engine would hesitate. Now this was only if you "mashed" the accelerator peddle! Slow applications were fine, but if you stepped on it, the engine hesitated. Power off the line is now quickly available.
The other thing was the power when going up a hill. Now, you can feel that there is more power available when going up a hill and the auto tends to shift down less often.
Does all this stuff make a massive difference....no, BUT, it is certainly more than £10 worth of improvement!!!
So from all of this I would say that the oxy sensors have a finite life. For the TU3JP engine I would recommend replacing the sensor at about 150000Km (or maybe 100000 miles?) If the performance is ordinary.
I don't know much about diesels but I hazard a guess that worn out oxy sensors might be just as much of a problem.

Just a quick update. The fuel economy improved just slightly from 6.8L/100Km to 6.4L/100Km. I don't know what that is in feet and inches . Not a massive improvement but it does prove that a well performing engine is also more economical!
My day-to-day driving includes highway and bumper to bumper driving, so I think that they are good, representative, numbers.