TSL1401 Linescan Sensor as an Optical Micrometer?
dbc1218
Posts: 33
I am working on a project and wanted to ask for some advice from some who has experience with the TSL1401 linescan sensor, hopefully Phil will see this. I am considering using the sensor to measure the diameter of the plastic filament commonly used in 3d printers. The material is abs or pla plastic and comes in rolls of round rod. The diameter of the rod is important to the consistency 3d printers and by continuously measuring the diameter the quality of printed objects can increase. My basic idea is to have the rod pass between the sensor and a back light and then use the diameter measurement techniques explained in the manual for the sensor to determine the diameter. This would be similar to what a laser micrometer can do, like this http://www.keyence.com/products/measure/micrometer/ls7000/ls7000.php I realize that some calibration will need to be done to convert pixels to millimeters but I'm pretty confident that this will be possible. My main question is about the possible resolution of the diameter in this situation. I would like to measure any diameter rod form 0.5mm to 4mm with a resolution of 0.01mm, which is 0.0004". I know form the data sheet that the distance between pixels will be 8micrometers or 0.008mm but don't know how this will translate to actual distance.
Does this like sound like something the TSL1401 might be able to do, or is there a better sensor for this application? What am I not considering that will have an effect on the resolution of the measurement?
Thanks,
Doug
Does this like sound like something the TSL1401 might be able to do, or is there a better sensor for this application? What am I not considering that will have an effect on the resolution of the measurement?
Thanks,
Doug
Comments
Yours is definitely a good application for the TSL1401. I'm not sure where you got the pixel pitch of 8 microns, but it's more like 62.5 microns. The entire length of the array is about 8mm, and there are 128 pixels. For that reason, you will need to use a lens to form a magnified image at the sensor's focal plane. The lens that comes with the TSL1401-DB might work if you can focus the filament close enough. You'll have to do some experimenting, though. You can always obtain a lens for it with a longer focal length. Longer focal-length lenses are better in another regard, too, in that the subject doesn't have to be quite as close to the lens as it would otherwise. This means that small differences in subject distance won't cause as much of an apparent change in filament diameter. You should still plan to confine the filament to a fixed position as it passes through the viewing area, though.
Anyway, if you can magnify the image of the largest filament such that it covers the field of view, your resolution will be about (4mm / 128) * 2 = .0625 mm. You could possibly do better than that by defocusing the lens slightly and analyzing the Gaussian curve of the grayscale image response to get sub-pixel resolution. But bear in mind that the TSL1401 has 128 pixels and that .01mm out of 4mm is one part in 400. So you won't get there just by counting dark pixels against the backlight.
-Phil
I looked at the wrong data sheet for the TSL1401 on the AMS website, thats why I thought 8microns. It also seems the TSL1401R is no longer avaiable.
Do you know where I could get other lenses and how will I know what will fit the cover?
As far as your resolution calculation, I don't know why you multiplied by 2. I would think if the line segment is 4mm long wouldn't the resolution be just be 4mm/128 =0.0312mm. What I'm I missing?
Doug
Try these suppliers for alternate lenses:
-Phil