Abgastemperaturfühler im AFN Krümmer... | Posts 16+

Hi,

@vanilla229: I now know several people who are using the VNT20 turbo on their AFN engines, and there are no problems installing it, and the turbo lag is relatively minimal.

What exactly do you mean by 'the only option would be to add fat?'
I don't understand?


Regards,
DJ

The perceived size of the turbo lag is certainly subjective. Only by looking at your stationary power and torque curves can one see that the turbo lag is kept 'within limits.' It is absolutely certain that no manufacturer would bring such an engine to market. There's almost nothing happening at the lower end of the RPM range.

The simplest way to increase turbine power/boost pressure is to increase the exhaust gas temperature. This is achieved by reducing the air/fuel ratio, which results in a richer mixture. However, this always leads to higher soot emissions.

simplified, more is injected at the lower end, which causes some soot, but it also makes progress...

Hey DJ!

It's great that you've decided to investigate the issue of exhaust gas temperature using measurement techniques, rather than just guessing or relying on hearsay.

I wish you the best of luck in advance. With this initiative, you can only gain, no matter the outcome.

But now, let's get to your questions, as the answers provided so far seem a bit incomplete to me.

The position of the sensor in the manifold is not arbitrary, as there can be differences of up to 50°C in the manifold, depending on the efficiency of each cylinder and the compression ratios, as mentioned several times – but that's not your problem, I'll come back to that later. The chosen position is fine, just make sure the sensor tip isn't too close to the manifold wall due to thermal radiation.
The problem isn't the temperature spikes or the dynamics in the exhaust, which you can't measure with either 3mm or 1.5mm thermocouples. To measure temperature quickly, you need extremely expensive equipment. However, it's not really relevant anyway, because the gas temperature spikes are only partially relevant due to the inertia of the materials – so it's not your problem. You want to see how things look at the intake manifold in a quasi-stationary state, not whether the turbocharger melts during a 10-minute drive.
It's important that your sensor is suitable for use in the exhaust system of a car, because the chemical and environmental factors are not to be underestimated (vibrations, thermal shock, etc.). It's best to ask the manufacturer for more specific information about this.
Your real problem, unfortunately, is that you don't have a reference point. If the sensor later displays 600 or 900°C, that doesn't necessarily mean that this value is even remotely accurate – I'm speaking from experience. Therefore, you will only be able to evaluate relative changes, but that's something in itself.
For this reason, the integrated shelf is rather insignificant.

Within the VAG group, there are several engines that already come standard with exhaust temperature sensors. These sensors are secured in housings and already offer the necessary durability with a moderate, but perfectly adequate, level of accuracy. Why not use these parts? Then you can also consider the placement more carefully. The key is to position them as close as possible and as centrally as possible in front of the turbine wheel.

Hi Bavarian,

You seem to be quite knowledgeable.
I'm glad you were able to find some answers to my questions!
I still have a few questions about your answers.

The position of the sensor in the manifold is, of course, not arbitrary, because, depending on the efficiency of the individual cylinders and the compression ratios, differences of up to 50°C can occur in the manifold, as mentioned several times - but that is not your problem.

Well, the 50°C was just an estimate and came from me.
I don't know how much of a temperature difference there would be if you move the sensor 3cm further to the left on the manifold, or if you position it in the center in front of the turbo. It could be 50°C, but it could also be more or less.

You want to see what it looks like in a quasistationary state at the VL, but not during the 10-minute AB run, because the charger melts.

Okay! The display should give me an approximate value so I have a general idea - it's better than nothing.

It is important that your sensor is suitable for use in the exhaust system of a car, because the chemical and environmental factors should not be underestimated.

Well, the sensor is only designed for passenger cars, or originally for gasoline turbo engines.

Your actual problem, however, will be that you unfortunately don't have a reference. If the sensor later shows 600 or 900°C, that doesn't necessarily mean that this value is even remotely correct in a physical sense - I'm speaking from experience.

I don't quite understand. What do you mean by 'reference'?
If the sensor displays 600°C, it's likely to be closer to 600-650°C in reality!?
At least, that's where it measures.

Within the VAG group, there are several engines that already come standard with exhaust temperature sensors. These are secured in housings and already offer the necessary long-term durability with a moderate, but completely sufficient, level of accuracy. Why not use these parts? Then you can also think more about the positioning. The principle here is: as close and as centered as possible in front of the turbine wheel.

Ah, I didn't know that. But I've already bought one (and it wasn't cheap), and I'm going to install it.

Regards,
DJ.

DJ1,

The temperature differences in the manifold are caused by potential differences in the exhaust temperatures of individual cylinders. Due to manufacturing tolerances, some cylinders may burn hotter or colder. In practice, these differences are typically around +/- 25°C at the exhaust port for standard hardware. Therefore, you might measure differences of up to 50°C in the manifold if you measure sequentially in front of a cold cylinder and then a hot cylinder. Since the individual cylinder exhaust temperatures are not known, it is impossible to estimate how representative your chosen measurement position is, as the individual cylinder outlets are not flow-separated.

Regarding the reference, I mean exactly what I said. There's no way to calibrate the sensor or the measurement chain within its operating range – and neither is anyone doing that with NiCrNi or PT100 sensors. The only possibility would be to use a second sensor to validate the measurement results, or to use the sensor in an engine where the exhaust temperatures are already known for a specific operating point, even if that point is faulty.
...or you might be lucky and your sensor is working 100%. However, you'll never know for sure.

Keep me updated, I'm very interested in your results.

Do you have the ability to record exhaust temperature data, or is it just a visual display? How much did the measurement technology cost, and do you have a link to the supplier?

Greetings.

Regarding the reference, I mean exactly what I said. There's no way to calibrate the sensor or the measurement chain within its operating range – and neither is anyone doing that with NiCrNi or PT100 sensors. The only possibility would be to use a second sensor to validate the measurement results, or to use the sensor in a motor where the exhaust temperatures are already known for a specific operating point, even if that point is faulty.
...or you might be lucky and your sensor is working 100%. However, you'll never know for sure.


Oh, but don't let any manufacturer of measurement technology or an ISO-certified developer hear that. Otherwise, every quality management system would go haywire.

In our case, with our measurement technology, it's not the same as what you wrote.

Our decentralized and centralized temperature measurement technology is calibrated within the range of -50°C to 1250°C. The error is less than 0.02K per °C across the entire range.
Calibration is performed using a reference voltage source that simulates the measurement voltage of a NiCr-Ni element.
NiCr-Ni sensors are classified into different error classes. Some of these sensors are calibrated, and their error levels correspond to their class. Most sensors are calibrated once a year, as stipulated by most testing regulations.

Therefore, you can confidently trust the temperatures displayed, which are based on our measurement technology, as they reflect the typical error margins for the class of equipment used.
Our customers, including engine developers and validators, do so at a minimum.

Regarding the peak exhaust gas temperature, you are correct that even a 1.5 mm thermocouple is only an acceptable compromise here. The problem is that these peaks are so fast that they can only be detected by extremely thin sensors (e.g., 0.02 mm spot-welded NiCr-Ni wire). Most measurement techniques are not fast enough to record these peaks. Our current product achieves 100 Hz, but there are few sensors that can show a temperature delta at a 100 Hz frequency, or they are so thin that they would immediately melt in the exhaust gas. If they existed, we could have become faster .

I'd be interested to know what kind of measurement equipment you used to gain your experience. (You can also send me the names as a private message if you don't want to share them publicly).
It is true that with an appropriate measurement chain, the readings can be off by up to 20°C in extreme temperatures (1200°C), but if you are getting a reading of 600°C when the actual temperature is 900°C, then you are using an 'estimation technique' and not a measurement technique. Our system is definitely better in that regard.

Best regards, Jochen.

Hi Jochen,

What is the purpose of this post? It has nothing to do with the content of these threads, nor does it relate to my post. I didn't say anything against the apparently 100% error-free exhaust temperature measurement technology at your employer.
That's not the point at all.
Based on your post, it seems you are advising DJ1, who is about to implement exhaust temperature measurement in a vehicle for the first time, using only one sensor and a measurement chain that neither you nor I are familiar with. This measurement chain is intended to be integrated with the engine protection system, and the accuracy of its readings will determine the life or death of the ATL/engine. Are you really advising him to trust the displayed measurement value without any verification? I find that questionable.

As a justification, you state:

Jochen_145 wrote:

The NiCr-Ni sensors are divided into different error classes. Here, too, there are calibrated sensors, whose errors correspond to the class. And most sensors are calibrated once a year. Most testing regulations stipulate this. Therefore, you can certainly trust the displayed temperatures, which our measurement technology provides, and compare them with the typical errors associated with the class of equipment being used.

The majority of sensors? How is a NiCrNi thermocouple calibrated within its measurement range? I'm not referring to the evaluation unit or circuit for the thermoelectric voltage, which you explained well. And why the majority? Are there thermocouples that are not calibrated?
The majority of testing regulations, which testing regulations?, are they from the thermocouple manufacturer, the engine developer, or your company? Are these regulations universally applicable?

Furthermore, if NiCrNi is used in the exhaust stream for more than a year in your company, or even if it's calibrated, I would personally be concerned. Engine developers typically only use NiCrNi for a single measurement campaign, after which it's discarded.

You also mention centralized and decentralized temperature measurement. What does that mean? I understand the term "engine developers," but what are "validators"?

Jochen_145 wrote:

I'd be interested to know what measurement techniques you used to gain your experience. (You can also send me a private message if you don't want to write the names openly). It is correct that with an appropriate measurement chain, the readings can be off by up to 20°C in extreme temperatures (1200°C), but if you are getting a reading of 600°C when the actual temperature was 900°C, then you are using an "estimation technique" and not a measurement technique. Ours is definitely better in that regard.

Over the past 10 years, I've worked with a wide range of measurement technology for vehicle applications of all types, but one thing they all had in common: out of the approximately 100 possible standard measurement points on the engine, the temperature sensor before the turbine is one of the most critical, and the failure and defect rate is very high. This is where the application in a test stand (PST) differs significantly from the application in a vehicle.
Maybe I've just always been unlucky, or maybe I've never used your thermocouples.

What I wanted to convey to DJ1 is the fact that exhaust gas temperature measurements in a vehicle, due to the lack of a reference point and regardless of the measurement technique used, cannot automatically be considered accurate, and that errors derived from this can lead to serious damage to the engine tuning. In my opinion, any further discussion is unnecessary.

Greetings.

The sensor you're using is interesting because it tapers towards the tip. This means the tip, which is the measurement area, is relatively small and responsive. Who is the manufacturer, and is it calibrated?


I'm not sure if it's allowed to post links from eBay in this forum.
Just type the term 'exhaust temperature' into eBay – you'll find my exhaust temperature sensor there.

Regards,
DJ.

@bafische,

What is the purpose of this post?
Simply put, you're introducing something into the world that can't stay like that. And that's exactly what I've depicted.
Regarding the measurement of exhaust gas temperature using a NICr-Ni sensor, I already mentioned in my previous post that I have abandoned the same plans (for the reasons stated above).

Secondly, you state that you cannot actually measure temperatures with NiCr-Ni elements.
'If the sensor displays a value of 600 or 900° later, that doesn't necessarily mean that this value is even remotely accurate from a physical standpoint.'
We can't simply disregard this, because measuring temperature using NiCr-Ni thermocouples is likely one of the most common methods used in development.
You can trust a measurement taken with a calibrated sensor and calibrated measurement equipment, and that's why calibration is done.



Yes, it is possible to measure exhaust gas temperature over a longer period. However, for this, 'stable' sensors (e.g., 5mm in size) are used. These sensors can be calibrated, but they are also relatively slow, so they do not capture exhaust gas spikes (as described above).

It is true that there are calibrated and uncalibrated sensors. The difference lies partly in the price, and partly in the fact that the measurement error of a calibrated sensor can be determined through a reproducible measurement, thus proving its accuracy. (For lower temperatures, this is often done in heated oil, while for higher temperatures, it is done in an oven, where the metal is heated and measured with a reference thermocouple.)
When measuring with a NiCr-Ni thermocouple, the thermoelectric voltage is measured, which changes depending on the temperature.

The testing regulations I mentioned are not generally applicable, but are usually defined by the user or their quality management system. (They must demonstrate that what they are measuring is accurate).

Decentralized measurement technology is located very close to the measurement point, which minimizes the error introduced by the NiCr-Ni compensation cable. Furthermore, the wiring effort is reduced because you only need to run one communication cable (e.g., CAN) to the logger. With centralized measurement technology, all measurement cables are routed to the logger, which is located in the trunk.




Almost everyone has confirmed that the location of the thermocouple near the turbocharger presents a very risky installation scenario.
Regarding durability, you've said exactly the same thing as I did. So, no problem!
It is absolutely correct that neither of us knows the measurement chain from DJ, and that the temperatures displayed on HIS system may not be accurate.

However, with the right measurement setup, it is possible to measure a temperature using only one sensor and still obtain a reliable result.
You are unfortunately not entirely correct with that statement.

If you had written it more generally, instead of as in your second post on this topic, I wouldn't have had to correct anything.


Best regards, Jochen.

P.S. We only supply sensors; we do not manufacture thermocouples. However, our measurement technology can be found, among other places, at a Stuttgart-based automobile manufacturer, for example, in continuous testing (vehicle validation).

Hello Jochen,

I generally think it's great when content is corrected here, and if you feel compelled to do so based on your knowledge, that's even better – that's what the forum is all about.
I'm not just here in the forum as a know-it-all, but also as a big fan of reliable statements, solutions, answers, and helpful suggestions that should be the end result of a discussion in a thread.
Therefore, in light of your second response, and even risking repetition, I am still wondering about the context of your statements.

In this thread, I/we have not discussed general temperature measurement techniques, but rather the now-known measurement setup by DJ1, who intends to install an uncalibrated thermocouple into a very hostile environment characterized by turbulent, pulsating flow, high temperature gradients, aggressive chemical substances, and high-frequency vibrations.
In the vast world, there are many other, much more precise and faster ways of doing things, and while your employer may be working admirably here, that's ultimately secondary and doesn't lead to a reliable answer regarding the original question.

It seems to me that deriving this conclusion using specific technical terms and internal guidelines, which are certainly familiar within your work environment but are incomprehensible to outsiders without explanation, only serves to further confuse the reader.

We can gladly discuss general questions about temperature measurement and data recording, as well as current possibilities for calculation/simulation in engine process calculations to represent exhaust gas temperature and the temperature levels of particulate filters during vehicle operation, over a beer or in a separate thread, as some of your statements require clarification.

Please consider the above in future discussions, especially when interacting with me.

...and finally, a very important comment from me:

Jochen_145 wrote:

You can trust a measurement taken with a calibrated sensor and calibrated measurement equipment, that's why it's done that way. However, with a suitable measuring chain, it is possible to measure a temperature using only one sensor, and one can also trust the resulting value.

I never believe or trust a measurement without further verification – that's something I've learned in my profession.

Hello Jochen,

We can gladly discuss general questions about temperature measurement and data recording, as well as current possibilities for calculation/simulation in engine process calculations to represent exhaust gas temperature and the temperature level of particulate filters during vehicle operation, over a beer or in a separate thread, as some of your statements require clarification.



@bafische,

Sure,
I'm curious to hear about your experiences.
My information comes from the underhood thermal management testing conducted in Rüsselsheim, where the DPF (diesel particulate filter) was 'burned clean' at the temperatures I described.

Best regards, Jochen.