## How to measure a Bosch LSU 4.9 wideband lambda sensor

Everyone loves a happy ending. Share your diagnostic success stories here.
Fat Freddy
TwoWaves
Posts: 318
Joined: Fri May 04, 2012 4:52 am

### Re: How to measure a Bosch LSU 4.9 wideband lambda sensor

THE BASIC, EASY, LEAST ACCURATE WAY

Hi Kim

Thanks for the interest to look at the thermocouple outputs, something I would like to see myself. Most of what I have figured out is from rummaging around on the net. I still have things to check out and compare (I'll use the AD595 for that as it's accurate to 1*C).
Time will tell whether I can find the specification for that sensor
If the sensors are truly 'K' type their output should be a given. Roughly for a 'K" type is 41uV/C or 0.000041V/C. So your probe could simply be C=A*24390 (This should work, I'm not sure if this was what I had before).
This figure is the difference between the heated joint of the two wire's and the cold end (cold junction). So not the true temp.
So to get that you have three options.

I) Stick the cold end in a ice bath. There fore the cold end is 0*C and the voltage output would be equivalent to a conversion true temp.

II) Build another circuit with a thermistor to measure ambient temp and add that to the calculated out put from the thermocouple. This is what is normally done on equipment using thermocouple's to measure temp.

III) Or the third option (which is what I did). Add the ambient to the above formula. This has to be done each time it is used. So the formulae will be (off the top of my head ) C=(A*24390)+ambient.

This method is not entirely accurate probably more indicative. Although the K type is almost linear, it is not quite. As I mentioned it is hard to check the accuracy. Here's a link on thermocouples
http://www.thermometricscorp.com/thertypk.html
And a vid.

EDIT:
And here's another good one that points out points of error that may need addressed.
http://www.picotech.com/applications/thermocouple.html
http://www.picotech.com/applications/temperature.html
This does bring up the point of cable extension (setting up other unwanted extra TC junctions). Which should be taken into consideration. But at that same time there is some contradictory arguments about that (which seem to have some logic to it). Anyhoo's it something that has to be remembered.
All these things need to considered because of the small inputs that are being worked.

Note: If you buy a cheap thermocouple from the electronics store you can chop it up into many thermocouples.

HTH
FF
Last edited by Fat Freddy on Thu Oct 02, 2014 11:39 am, edited 3 times in total.

Fat Freddy
TwoWaves
Posts: 318
Joined: Fri May 04, 2012 4:52 am

### Re: How to measure a Bosch LSU 4.9 wideband lambda sensor

THE MORE ACCURATE MORE EFFORT WAY.

For improved accuracy, I wouldn't use cheap thermocouple. Badly made up alloys will give poor performance.

Also in your probe you want to make up a look up table for the conversion and this table is here: www.omega.com/temperature/z/pdf/z204-206.pdf

This will give the voltage/*C conversion from -250*C to 1370*C. But that has a lot of work in it.

Or as I hope to, is use selected points and interpolate. It should be fairly accurate. But either way you will still have to apply the cold junction temperature.

Some folks seem to have problems getting an output, lordy only knows why. My suggestion if you want to give it a crack is to wire in a cheap TC, set the scope to the lowest voltage measure and heat the tip with a lighter. Then work from that.

Let me know how you go. Or if you get stuck.

HTH
FF

KimAndersen
TwoWaves
Posts: 213
Joined: Fri Nov 30, 2012 2:53 pm
Location: Denmark

### Re: How to measure a Bosch LSU 4.9 wideband lambda sensor

Hi FF

I must disappoint you, it's not K-type thermocouple sensor that volkswagen uses in this car. After further search for some documentation regarding the exhaust gas sensor (G235) fitted in this VW TDI 2.0 Pumpe Duse, i found that Volkswagen used a different type of EGT sensor.

This exhaust gas temperature sensor is a RTD (Resistance Temperature Detector) which i think are fitted to all newer Volkswagen where the exhaust gas temperature are measured.

Here is small quote from a dataset that describes different temperature sensors and there functions -

" The Resistance Temperature Detector, or the RTD, is a sensor whose resistance changes with temperature. Typically built of a platinum (Pt) wire wrapped around a ceramic bobbin, the RTD exhibits behavior which is more accurate and more linear over wide temperature ranges than a thermocouple. "

What that means to my project is that it makes it simpler to do the voltage reading, now that i have the right datasheet.

I can see from your documentation over the k-type sensor, there are much more things to account for, before you get the correct result, but it's definitely an interesting project with your K-type sensor.

Regards
Kim

Fat Freddy
TwoWaves
Posts: 318
Joined: Fri May 04, 2012 4:52 am

### Re: How to measure a Bosch LSU 4.9 wideband lambda sensor

Hi Kim
What that means to my project is that it makes it simpler to do the voltage reading, now that i have the right datasheet.

I can see from your documentation over the k-type sensor, there are much more things to account for, before you get the correct result, but it's definitely an interesting project with your K-type sensor.
I think we are both dealing with very similar issue's/hurdles. I'm not sure 'simpler' is a word I'd use.

But I would love to see what you come up with. I'm almost set to do another K type test run and do have a probe almost for the K type ready to go but I want to run test's along side the AD595, so need a probe for that to.
Although this is fairly low on my priority list I will post up the results.

Cheers
FF

KimAndersen
TwoWaves
Posts: 213
Joined: Fri Nov 30, 2012 2:53 pm
Location: Denmark

### Re: How to measure a Bosch LSU 4.9 wideband lambda sensor

Hi

I'm back !!!.

I have been busy with my exhaust gas temperature sensor project. The project are based on the same car from my case study which is a VW JETTA 2.0 TDI PD (103kW) equipped with a diesel particulate filter.

The exhaust gas temperature sensor (G235) on this car are mounted before the turbo and this is the sensor i want to measure.

This sensor are known as a RDT 200 sensor and stands for (Resistance temperature detectors) which means that this sensor has a resistance at 200 ohm at zero degrees celcius. With a increase in exhaust gas temperature the resistance in RDT 200 sensor increases as well. Before we can measure this sensor we must check some basic thing in regards to making a lookup table for that particular EGT sensor.

This picture shows how the sensor are connected to ECU. Seen with my eyes is this a voltage divider and this as simple as it gets. Simple or not - things can still go wrong if you haven't got the right datasheet with the right resistance table or if there are a voltage drop to the sensor.
So far, have i seen datasheets from three different manufacturers and the resistance table are the same on the RDT 200 sensor. Although there are different standards - do I believe that RDT 200 seems pretty similar in terms of the resistance in the sensor.

Before you can make the lookup table and thereby do the calculation, you must measure voltage supply from ECU to sensor - in my case, was it 5 volt precise and i mean 5.00 volt PRECISE. Next step is to find resistance of the pull-up resistor inside the ECU, which I did by putting a 330 ohm resistor in series with the pull-up resistor and then measure the total power consumption through the resistors.

It turned out to be a 1000 ohm resistor there were inside the ECU. That fits very well, what the manufacturers are proposing. Now you are ready to make the calculation by using ohms law.

Now that i have studied my waveform capture from the test run today and here am I thinking about test run where the exhaust gas temperature reaches beyond 900 degrees Celcius - i didn't expect that it go that high, although it's a hard acceleration, i would have expected a temperature around 850 degrees Celsius.
Maybe, there are a issue with sensor at that temperature - I doubt that it is showing correct - just a feeling that i have !!!

It could have been nice with reference to a new exhaust gas temperature sensor (G235) to prove or disprove the precision of this sensor reading.

What's really fun about this sensor is to watch how quickly it reacts to changes on the acceleration postition sensor (APP) during the test run.

Regards
Kim

Valhalla
TwoWaves
Posts: 66
Joined: Sat May 18, 2013 10:22 pm

### Re: How to measure a Bosch LSU 4.9 wideband lambda sensor

Hi Kim,
Interesting results. If you are sure that the pull-up resistance is 1kOhm (and I'm fairly sure that you are right), then your sensor is over-reading the temperature. Assuming that the sensor is measuring the average temperature of the exhaust gas stream, and not a direct exhaust port temperature, then the peak temperatures seen during the accelerations in the lower gears are too high. You are much closer to the truth when you suggest 850celcius as a working maximum - this has to be the case because turbo exhaust housing material would be specified to around 830celcius for a steady-state maximum (and that is assuming that VW have used the best options with their turbo manufacturer - 790celcius is a practical maximum for most housings)

As the temperature still appears to be rising on your high-gear acceleration, then you would have to assume one of the following;
2) A valve-timing fault exists on this car
3) A fundamental calibration error exists in this VW

I look forward to seeing the result with a new sensor!!

Fat Freddy
TwoWaves
Posts: 318
Joined: Fri May 04, 2012 4:52 am

### Re: How to measure a Bosch LSU 4.9 wideband lambda sensor

I'm still mulling this all over. But I do agree with the temps being a bit high.

The literature I have found suggests temps lower than 700. Although I have only found a few claims on this.
I do wonder if the resistance of the leads may effect the value. I measure 0.3 ohms total. IIRC 0.003 approx = 1*C. There fore if my calculations are correct thats100*C.

Cheers
FF

Fat Freddy
TwoWaves
Posts: 318
Joined: Fri May 04, 2012 4:52 am

### Re: How to measure a Bosch LSU 4.9 wideband lambda sensor

I did a measured run last week. 300TDi, 15.5psi boost, pre turbo TC, 32*C ambient, full gas up a steep gradient (speed limited to 80Kph) measured with a multimeter with a K type TC input.

The temps under load were 598*C (had to back off) and 632*C (still creeping up before backing off). During extended overrun temps would drop to about 180*C, creeping up to 210*C during idle.

Kim.
Are you able to log live data during a run? Your testing defiantly is giving some interesting results which I expect will differ from what I have seen on my old clunker. It's also given me some idea's to work with poss try to put a RTD alongside the TC.

Cheers
FF

KimAndersen
TwoWaves
Posts: 213
Joined: Fri Nov 30, 2012 2:53 pm
Location: Denmark

### Re: How to measure a Bosch LSU 4.9 wideband lambda sensor

Hi FF

Yes - I did log live data from the test run. The Test Leads from picoscope were long enough to reach inside the car.

There's nothing as a live signal to look at during real test where against the calculated values presented by the OBD
software sometimes can be misleading to look at.

What there is amazing about this exhaust gas temperature sensor (G235), is to watch how sensitive it is to small changes in the engine load.

I´m still puzzled about the high reading i was getting - they are definitely wrong and I might come one step closer to correct this issue.

I have found some interesting data/test values from volkswagen whereby they claim that with the sensor (G235)
disconnected, there should be a supply voltage at 3.35 V (±.5 V) from terminal pin 1.

When I measured at terminal pin 1 was i getting 5.00 volt !!!!!!!. What this means to my lookup table is that, its totally wrong and i must make a new one.

A difference at 1.65 volt at the supply voltage correspond to correction factor of 550 degrees celsius - roughly said.

I´m really confused about this supply voltage to this sensor !!!! - is there anyone that can confirm this as being correct in respect to the 3.35 volt from the ECU

Regards
Kim

Fat Freddy
TwoWaves
Posts: 318
Joined: Fri May 04, 2012 4:52 am

### Re: How to measure a Bosch LSU 4.9 wideband lambda sensor

Hi Kim

Finally got a run in with the same conditions - ambient ect. I'm happy with the results. First run I found an irregularity in the look up table around the 600*C mark. So I can't say the exact peak temp, but very close to the expected 630* peak. Run two was a right off. And the third I had to back off - although the temp came very similar to what I had seen on a similar run when backing off.
I haven't made a look up table for the AD595 output yet but measuring straight off the TC is given temps very close to actual measured with a meter. So the LUT must be close.
I had to add cold junction temps into a maths channel for final measurement.
The K type is reasonably fast. Roughly speaking, the voltage output will increase by about 85% of total required of a immediate temp change within 1 second.

Any hoo's. I thought I'd post up the results.