5
\$\begingroup\$

I am an automotive engineer with very very little or almost zero electrical/electronics experience. I am working on logging real time temperature values on an engine exhaust valve and mounting a thermocouple and routing the cable through the valvetrain and valve cover. Instead of that, someone asked if the thermocouple voltage values can be transferred wirelessly (a sender unit mounted on valve tip, i.e. top most on the pink thingy and a receiver on the spring seat)

A valve's picture

I would like to know if this is even possible and how difficult will this project be for an automotive engineer?

\$\endgroup\$
3
  • \$\begingroup\$ Nicely interesting problem. :) Questions: 1. What are the valves made from and what's their thermal coefficient of expansion? 2. Have they (the valves) been worked so their temperature coefficient of expansion isn't isotropic? 3. If so, what's the ratio and direction of the anisotropicity?. \$\endgroup\$ – EM Fields Apr 19 '16 at 17:10
  • \$\begingroup\$ Can the thermistor be placed on the wall in the exhaust port to avoid movement? You'd have to sneak the thin(?) wires out of the header gasket (might leak?) This may be a better reading of the heat the valve is experiencing. There is probably a loss of temperature along the valve stem. \$\endgroup\$ – user107405 Apr 19 '16 at 18:22
  • \$\begingroup\$ My biggest concern would be that there's not exactly a thermal equilibrium. The hot exhaust gases are frequently replaced by cool air, into which fuel is injected which then evaporates cooling the interior even further. \$\endgroup\$ – MSalters Apr 19 '16 at 21:56
9
\$\begingroup\$

A good way might be to use a pyrometer. It is an optical device that receives the heat/light emitted by the object whose temperature needs to be measured and then uses part of that "spectra" to calculate the temperature.

It relies on understanding the "black body" radiation from the object and its emissivity: -

enter image description here

I've used it for temperatures as low as 600 degC on aero-engine turbine blades but, in theory, can be used much lower. You can pipe the radiation off in a suitable optical fibre and it's a non-contacting method.

It does rely on filtering a small fraction of the bandwidth of the emitted radiation and there are optical filters that can do this for you. I've used an optical filter and the photo-diode's natural spectral response to select the 1.7 \$\mu\$m part of the emission. If you can find a device and filter that is good for band-pass filtering about 3 \$\mu\$m then you should be in business.

The only area I'm unsure of is the sub 600 degC region but if you are looking to measure temperatures at or above this point it can certainly be done.

\$\endgroup\$
2
  • \$\begingroup\$ A quick google search suggests 650 C peak. \$\endgroup\$ – MSalters Apr 19 '16 at 21:57
  • \$\begingroup\$ And a bit more searching turned up this - 200 to 1600 degrees, 0.6 mm spot size @ 50 mm lens distance, actual sensor approximately one meter away. \$\endgroup\$ – MSalters Apr 19 '16 at 22:19
6
\$\begingroup\$

I think this could be done with a ribbon or thin wire thermocouple attached to the top of the valve.

I would suggest looking at making a blind hole (say via EDM plunge) in the valve and put a polyimide-insulated (check the expected temperatures) sensor deep within the valve itself. A type K or N thermocouple (those are nickel alloy types, as opposed to, say, J or T) might have the required characteristics, but you'd have to check the fatigue life against the expected run time of the test.

Such a thermocouple can fit into a very small diameter hole and will give an accurate and reliable reading of the temperature at the junction as long as the connections last (and a break is easily detected within milliseconds).

I don't think electronics attached to a valve stem is going to be very practical, at least not at first look, however there are some RFID sensor chips that might be promising if the environment was not excessively harsh- if it's running much over 120°C probably not practical.

enter image description here

\$\endgroup\$
1
  • \$\begingroup\$ So I am moving forward slowly on this project. Is EDM the only way for machining the cut for the valve? Are there any other ways? \$\endgroup\$ – Vish_evo Sep 9 '16 at 2:59
6
\$\begingroup\$

This is tricky, so the first thing you need to do is get someone that knows what they are doing with electronics. The high temperatures and vibration make this difficult, regardless of how you eventually get the data from a sensor to the outside world.

I'd want to minimize whatever has to go onto the moving valve. That probably means just a high temperature thermistor with flexible wires that will work long enough to give you a few useful readings. Then put all the rest of the electronics where it's cooler, not moving, and not covered in dirty oil.

Some sort of wireless transmitter and then a wireless receiver, even if only a transformer seems like a lot more trouble than its worth at first glance. Again, I'd want to minimize the electronics in this harsh environment. Keep it to just a thermistor and wires.

The reason I'm saying thermistor (varying-resistance temperature sensor) and not thermocouple is because with a thermocouple you are stuck with certain materials for the wires. You want to choose the best material to withstand lots of flexing, which will be multi-strand copper. You can't have both wires to a thermocouple be the same material. High temperature "platinum RTD" type of thermistors should work here.

I'm expecting there is oil splashing around, else I'd suggest IR temperature sensing.

\$\endgroup\$
1
  • \$\begingroup\$ @EMFields Fulfilling Godwin's Law so quickly? I'll take that as a failure on your part to come up with anything constructive to say. \$\endgroup\$ – W5VO Apr 20 '16 at 2:15
1
\$\begingroup\$

Yes, this absolutely possible.

Below I am going to summarize the proposed solutions with their associated problems and propose one idea of my own.

Solution 1: pyrometer has been proposed by Andy aka and mentioned by Olin Lathrop. The associated problem is sticky dirt blocking optical tract.

Solution 2: copper wires have been proposed by Olin Lathrop. The associated problem is much wear and tear on the wires.

Solution 3: transformer/coil has been mentioned by Olin Lathrop. This can be used with a thermistor, or you can see if your valve has a useful slope of its magnetic characteristics in the region you intend to measure. I would prefer this approach, especially if it works with valve as the core and no other stuff. Maybe Olin will explain associated problems.

Solution 4: something like an RFID sensor chip. The associated problem is the need of custom high-temperature sensor chip made of gallium nitride or some other temperature-resistant semiconductor material. A major engine manufacturer can very well order such a chip, but you will certainly need professional consulting. Also, I'll mention here that there is a lot of alternative options to power the RFID chip in the walve, including possibly piesoelectric or thermoelectric generation, so the power does not need to be transferred wirelessly, only the signal.

Note that first two solutions attempt to solve the original problem, not building of an LC circuit.

\$\endgroup\$
1
\$\begingroup\$

Your question is about the implementation of a particular solution to the problem of determining the temperature of a moving engine valve. And you have received various solutions. However, a simpler solution would be to install one (or more) thermocouple(s) on the valve guide (not being identified).
Although the temperature of the valve is not being measured directly, a correlation between the temperature of the valve and its guide can be determined, and readings "adjusted" to reflect the temperature of the valve.
Whether the thermocoule wires are to be routed, or its voltage values wireless transmitted outside the valve cover, either method can be implemented.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.