How will IR thermometers do on sandblasted steel?

[Zeyneb]

Hi there,

I'm thinking about purchasing a high temperature range IR thermometer. The UNI-T UT305A+ has a range of  -50°C~1850°C. And the cost is also a high number btw

But how good of a reading will I get on sandblasted steel? This is because I would like to use it for welding and heat treatment applications.

Thanks in advance.

[bookaboo]

You need to consider the emissivity of the surface of the steel, which (if a quick google search is correct) will be ~0.5-0.6, though there are ways to measure your specific sample.
Ideally you would be measuring a "Black Body" with an emmisivity of 1. In reality nothing is perfect, but you want as close to 1 as possible.
It's not likely that you can do much about the emissivity property of the surface, so you will just have to take note of the background temperature and use some calculations. Your background temperature will presumably be lower than the surface temperature, so the relatively high emissivity will cause your readings to be low.

https://www.flir.co.uk/discover/professional-tools/how-does-emissivity-affect-thermal-imaging/

Most thermal cameras can calculate this out automatically if you adjust the settings correctly for each material.

[PartialDischarge]

Unless you calibrate the emissivity with some kind of contact temperature probe, it will measure badly.
The reason is that the emissivity varies wildly with material, oxide layer thickness, rugosity and even with temperature itself. 30% error is something normal.

[Zeyneb]

I do understand the property of emissivity. That's why I selected a model with the emissivity to be able to set in the device. Also the UT305A+ model has a Type K thermocouple port and a thermocouple sensor in order to do that calibration to have good readings in the IR mode.

You might want to look what the device is capable of:
https://meters.uni-trend.com/product/ut305plus-series/#Specifications

What I would like to learn is if beyond the property of emissivity does sandblasted steel have other nasty effects that can ruin the accuracy of my measurements.

[bookaboo]

There are better informed TI experts than me on the forum, but I'm not aware of anything aside from emissivity that you need to be aware of. Possibly the fact that emissvity will vary depending on how shiny your steel is, that variation could cause you to be way out (especially if your subject is much higher than background).

A touch of black spray spray or tape often solves. But I'd guess not feasible in this instance.
How hot is this steel when you measure? You might not need all the range that Uni-T has, perhaps that widens your range of options.

[Zeyneb]

One example is 4140 steel. A heat treatment will make the steel stronger but there are also some negative aspects to it. For most products it would need a process called tempering afterwards to make it less brittle.

During heat treatment it could get to a temperature of 870°C or 1600°F. When it is at that temperature it is suddenly submerged in some oil at room temperature. That sudden drop in temperature is making the steel strong.

[CatalinaWOW]

These thermometers use the ratio of two infrared bands to measure temperature.  Thus the absolute emissivity is less important than the constancy of emissivity across the two bands.  Sandblasting could help or hurt.  It will tend to make the surface more uniform than a surface which is a mix of rust, mill scale, oil char and bare metal.  But rough surfaces increase emissivity, and tend to a larger increase for shorter wavelengths. 

You need to know your material well to know if this matters.  For most materials and many applications high precision is not needed to get good results.  But I know of one case where to get the right heat treatment results the window is less than a couple dozen degrees C.  Even the pros with the best equipment have trouble measuring and controlling that.

[CatalinaWOW]

Some relatively easy experiments could give you some pretty good insight into this.  Set up a thermocouple and measure the iron temperature over a range of lower temperatures that you can keep constant (kitchen oven for example).   This is well outside your desired measuring range, but can give you some idea of how things are going.  Use the infrared thermometer to measure the temperature at a few spots on the iron untreated, and then try it after sandblasting.  If you have access to different grits for the blasting process this would give best insight into the impact of the roughness.  Once the surface roughness irregularities are large (several times) the measurement wavelengths the difference in emissivity change should be small.

I haven't done any really precision checking of results with these thermometers, but my casual tests have indicated fewer problems than many people seem to worry about.  Except for the obvious fault cases like highly reflective surfaces where you are really measuring the temperature of a reflected image.  Do remember that reflectivity is better at longer wavelengths.  A simple hack for this is mud.  Rust is also generally good so sometimes it is better not to clean the surface.

[Zeyneb]

Some relatively easy experiments could give you some pretty good insight into this.  Set up a thermocouple and measure the iron temperature over a range of lower temperatures that you can keep constant (kitchen oven for example).   This is well outside your desired measuring range, but can give you some idea of how things are going.  Use the infrared thermometer to measure the temperature at a few spots on the iron untreated, and then try it after sandblasting.  If you have access to different grits for the blasting process this would give best insight into the impact of the roughness.  Once the surface roughness irregularities are large (several times) the measurement wavelengths the difference in emissivity change should be small.

I haven't done any really precision checking of results with these thermometers, but my casual tests have indicated fewer problems than many people seem to worry about.  Except for the obvious fault cases like highly reflective surfaces where you are really measuring the temperature of a reflected image.  Do remember that reflectivity is better at longer wavelengths.  A simple hack for this is mud.  Rust is also generally good so sometimes it is better not to clean the surface.

Alright. There is a DIY sandblasting shop in my country. I can see what the effects are of various sandblasting grits.

There is a research article on this topic. For me it goes way over my head. But if you're interested into it. You can download the full article there. I couldn't get much information out of it that is relevant for what I'm looking for.

https://www.researchgate.net/publication/374379835_Prediction_of_Isotropic_Rough_Surface_Directional_Spectral_Emissivity_with_Surface-Morphology-Dependent_Modelling

But what do you think of the UNI-T UT305A+ for my application? Would there be more IR thermometers with a huge temperature range that I could consider?

[bookaboo]

Is a spot thermometer good enough for the application? How will you know you don't have variations over the target piece

This thread needs Fraser!

[Zeyneb]

Is a spot thermometer good enough for the application? How will you know you don't have variations over the target piece

That is surely a good point. I think I need to develop some skill to heat it up evenly with a torch. The object I have in mind is a rear axle beam of a car. That won't fit in an oven.

Maybe I will rent some thermal camara as well to see if I'm doing it evenly from room temp. to 200°C or so. Or maybe a Eevblog member from the Netherlands with a thermal camera would like to assist me with that. But I still have to buy plenty of supplies to get there.

[CatalinaWOW]

I can't speak to suitability for your application.  I don't know the material of your axle, or of any heat, forging, shot peening or other processes which give it it's current properties, nor do I know what loads your axle might sustain after your proposed modification.

The referenced paper gives no direct information of use.  It does suggest that for a thermal camera, good estimates of emissivity can be obtained, but you would need two apply the methods described in the paper to the two different bands used by the UniT unit and compare their differences.

Most of the unknowns in your question can only be answered by you.  What heats do you you need.  How do you accurately do you need to know them.  Is uniformity of temperature needed, or do you needed a controlled pattern. 

I can only give a general answer.  I have a UniT thermal viewer and it gives results I believe to be within about 10-20% for temperatures up to a couple hundred degrees.  And I have several spot thermometers similar to the UniT (possibly off the same assembly line) which provide results that are comparable.  I am not saying that these devices have errors of 10-20%, I am merely saying that I have no sound evidence that they are better than that.