Mazda 6 Forums banner

1 - 7 of 7 Posts

68 Posts
Discussion Starter #1
After spending many hours tracing a fault with my car I decided to post up some information from what I learned during the process.


Series II cars to meet Euro V emissions rules all use a combined diesel oxidation catalyst + diesel particulate filter in the same shell.

There are four sensors in this housing; Three EGT sensors, one pre cat (nearest engine) one post cat pre DPF (middle of Cat-DPF) and one post DPF at the back of the Cat-DPF housing.

Then there is the O2 sensor in the pipe leaving the Cat-DPF.

While these three EGT sensors have different part numbers, they are all electrically the same.

They are a high temperature NTC thermistor. The different part numbers account for the different plugs and lengths of cable.

From the WSM the resistance should be somewhere between 82 & 137 k OHMs at 50 deg C to be declared within specification.

I characterised the sensors up to 250 deg C and at 50 deg C mine were all at about 100k

To do the characterisation I used a saucepan full of canola oil and I heated it up slowly with a digital meat thermometer in the oil also.

I used oil so I could co above 100 deg C, otherwise water would do.

Here is what I recorded for the middle and back:

20 280000 260000
50 10000 10000
100 37000 35000
150 16000 15000
200 8000 7000
250 4300 4100
300 2600 2600
350 1800 1800
400 1300 1300
450 1000 1000
500 770 770
550 625 625
600 518 518
650 438 438
700 377 377

Note anything over 250 is an estimate based off the typical NTC thermistor curve for the type.

I tested one with my heat gun at about 500 deg C and its pretty close.


OK so why would I want to do this?

My car has been modified to have the DPF taken out and there has been modifications to the fuel maps for extra torque ~40% more.

One thing I found is that the engine always seemed to go into some sort of a torque limiting mode at high loads.

It bugged me for nearly one year. I got no help from the modifiers and I didn't even bother with Mazda as they would not want to have any involvement wit modifications, I figured I would be on my own.

For another background
; The DPF in these cars performs a second function, it is also a NOx converter otherwise known as an LNT. (lean NOx Trap)

To perform this function, the temperatures need to be strictly controlled.

And so while the car is operating, the ECU tries to keep the DPF entry temp and exit temp between a narrow temperature window so that the DPF can reduce NOx.

The problem being that with the fuel map modifications in my car, my post cat EGTs would probably be alot higher. Also without the DPF core being in place, the last EGT sensor would be seeing temperatures alot higher than expected thanks to no thermal mass damping plus NOx reduction being endothermic, no associated heat loss.

The result; the ECU would start reducing torque at high loads thinking the LNT function was out of range.

The 3rd EGT sensor also exists to ensure the DPF doesn't melt during a regen process, something that has been coded out of my ECU but the resulting DCT codes are now also no longer available. This might be useful for cars with operational DPFs; The last sensor performs a vital function in controlling the completion of the regen cycle.

Why is all this important?

If your vehicle is unmodified the ECU will go into limp mode if any of the EGT sensors go out of range. You can test yours with the information provided above.

If your vehicle is modified and you have a cat but no DPF you car may not like the reading the ECU receives from the 3rd EGT as it will go out of range without the DPF in place.

Depending on what changes where done in the ECU by the tuner concerned, these EGT values could be ignored or just the DCT codes ignored.

What did I do to fix it?.....

I wish I had the time to read the PID values from the OBDII port and log it, I purchased the software and had the hardware, what I didn't have is time and someone to assist with reading the values out of the laptop while I was driving.

And so I decided to take an educated guess to make the ECU think that the temperature on EGT 3 was lower that was what actually there.

To do this I connected a 220R resitor in series with the sensor. Being an NTC thermistor this extra resistance will cause the ECU to treat it as normal its just that the max temperature has been biased lower.

I made the change near the sensor itself, but this was painful as the wire being high temperature does not take well to solder and I used glued heatshrink to give the joins mechanical strength.

How does it drive: The problem "seems" to be fixed, I will only know after a long drive at high load at high ambient temps.

Its coming into winter here now and so a full test may not happen for a while.

In summary:


68 Posts
Discussion Starter #2

I took the car for a test drive down the freeway where I have experienced the issue before.
I loaded it up in 6th gear and did a 30km loop and it would appear that the problem has either diminished substantially or gone completely.

Now that it seems ok I am going to take it on a 2000km trip next week and so I will know for sure after that.

There still is some strange behaviour where the idle speed hunts after coming to a stop immediately after a high speed/High load run.

It might not be related to the other fault but seems to be heat related; by pressing the clutch pedal in for 3 seconds and releasing it stabilises that issue, but it did before I made this change to sensor #3 .

It would seem that a different section of the ECU code operates with the clutch pedal down as not.

68 Posts
Discussion Starter #3
more updates

Another update

It seems that the car works better since the mod but the issue is not fixed.

I did a 2000km round trip and the same sort of problem surfaces; after a few hours of driving the torque starts to drop away.

Its not as bad as it was and so maybe my guesstimate of the series resistor value for sensor 3 was off the mark.

I am going to take a step on the high side of what I think it should be and so I have increased it to 1k ohm, if there is an issue now then sensor 2 will need to be modified.

Maybe I should have spent the time logging what happens but I didn't and still don't have the resources and so trial and error it will have to be.

The other option is that I get the ECU mods taken back to factory, but I don't want to do this just yet, not only will it cost, but I like the extra torque that is available, it is more about making the car run properly

68 Posts
Discussion Starter #4
Yet another update

Modifying the series resistor on sensor #3 hasn't done a lot but there is a slight improvement but not what I had hoped for.

So I've now inserted a series resistor into sensor #2 in the same way. I have selected 470 Ohms (I wanted to make it less but didn't have a suitable resistor)

The theory here is to let the ECU believe that cat has still reached light-off and so it won't attempt to dump extra fuel into the exhaust at idle, yet it will disguise any temperature excursions.

I expect the cat will physically handle the higher temperatures, its just they would be sub optimal for the particle filter element which is now no longer there.
Before DPFs came about a lot of cars never had post cat EGT sensors so I am not expecting any real world issues.

I guess we will see how it goes over a longer trip but already the idle speed hunting issue seems to have gone which I thought was probably an interaction with the throttle plate which activates to reduce air flow through the engine to keep the cat warm at idle.

I'll have to see whether there is any fuel economy penalty for this change.

With my driving style the car does high 6s (L/100Km) City and high 5s (L/100Km) on the freeway at a constant 115km/h cruise control on.

It is also curious to note that the fuel economy is a lot better with cruise on!
I think I am reasonably good at maintaining a constant speed without.

Maybe the ECU runs a different algorithm with cruise turned on, it would not be a surprise if it does.
My best fuel economy was 5.2L/100km at 115Km/h, I don't remember what the weather conditions were like that day, there could have been a slight tail wind!

68 Posts
Discussion Starter #5
Another update

After the change to sensor #2 I had a long drive up one of the nearby test freeways and where the problem would have surfaced before it didn't and so I would say the problem that was created through adding of extra fuel and keeping the cat minus DPF has been largely resolved.

I was able to keep up with traffic easily in 6th. The day was about 18 deg C and I did a 270km round trip without issue.

The real measure would be with higher ambient temps (30+ deg C)

But I have decided to sell the car and this will mean putting everything back to stock.
The original ECU code will be loaded in, the exhaust changed back to a stock system (which I just purchased off a used wreck)
Plus I will convert the brakes back to original items.

Its been a fun journey with this car. I still like it as the hill climb champion, but it doesn't suit me anymore for the sort of driving that I now do.

68 Posts
Discussion Starter #6
Yet another update......

I haven't sold the car as yet but had the opportunity to test it again on my 30km freeway test loop, this time at 28 deg C ambient and with the A/C on an a load on the roof racks.

Within 10km of driving in 6th, (1750-2000rpm) the ECU went into torque reduction mode and the car was basically undriveable below 3000rpm. I stopped to let it cool down for 10 minutes and then drove back on the same piece of road in the other direction in 4th ~2800rpm where there was no issue.

And so what I have done has gone part way to fixing the problem however, If I were keeping the car I don't think raising the series resistor value anymore than what it is would be a fix. My estimate is that the cat would already be at about 700 deg C when the ECU goes into torque reduction mode. I am not sure that the cat would have much longevity operating at higher temperatures and so the choices would be not to drive it at high loads below about 2200rpm which is a shame as that's where its most fun. Otherwise return the fuel maps back to stock and live with stock torque and no issues.

It would seem that there is either too much fuel and not enough air in the exhaust between 1750 & 2000 rpm, so either the fuel needs to be reduced or injected a bit earlier so as to make sure a more complete burn.

This is all a bit academic now as I am committed to selling the car as soon as the original code can be loaded in.

I wonder if a water meth injection system would have helped keep the EGTs down and the cat temp under control.....maybe someone else can run with that idea from this way forward.
1 - 7 of 7 Posts