discriminate brass from aluminum

Question

I am attempting to identify brass casings vs aluminum casings. I'm been trying to figure a reliable way to differentiate between the two. I've thought about

• Color: This may work, but I fear that dirty optics and targets may not work so well.
• Weight: I'm not really sure how to implement this in my context.
• Conductivity/Resistivity: Perhaps an inductive sensor? But aluminum is very close to the same value as brass.
• Ultrasonic properties: This may not work with varying levels of dirt. But I really don't know much about it.

I prefer non-contact, and a solution that is impervious to contamination, but I'll take any suggestions given.

It seems that an inductive sensor would be the way to go. I'm okay with rolling my own, perhaps a couple of tuned coils which the target passes through?

---

Okay, thanks to Charles for getting me on the right track. It looks like I'm needing an eddy current sensor.

These are shell casings, and when run through the machine will all be of the same caliber, so the uniformity should be good. For material identification, distance from the coil is important, and since the target is not always in the same position, I'll run the casings through the coil center, which should be immune from position errors.

It looks like some aluminum alloys overlap the brass alloys in resistivity, but I bet that the alloys used in cartridges are very consistent. I do have another bit of information, and that is the actual identification of the head stamp. Headstamp coupled with resistivity should get me a very accurate indication of what type of metal the cartridge is made from.

Answer 1

run them through rollers and down a chute so they're all moving the same speed then pass them horizontally through a strong magnetic field, the lighter and more conductive aluminium casings will be slowed more by eddy currents than the denser brass ones and fall into the near bucket. The brass will be slowed less and fall into the far bucket

Answer 2

Archimedes had just this problem over 2000 years ago. King Hiero II had supplied the pure gold for a votive crown to be made, and was suspicious that the metal had been adulterated with silver by a dishonest goldsmith while making the crown, and gave Archimedes the problem of proving or disproving his suspicion without damaging the crown. Archimedes noticed when taking a bath that his body, regardless of its irregular shape, displaced its own volume of water, and so took to the streets naked shouting "Eureka!"

Alli has a density of about 2.7 times that of water, brass is about 8.0. With such a huge difference, a density discrimination by noting the change of weight when immersed in water should be very easy.

Answer 3

Are the samples to be measured of uniform physical characteristics? Size, shape, weight, etc? I would imagine that magnetic properties and/or their reaction to ultrasound would be varied enough to do a reasonable differential identification.

Answer 4

Eddy current is probably best, but if you wanted to pursue the color method, color sensors have become very sophisticated. For example, the TCS34725. You can easily eliminate a lot of effects of dirt on the sensor or illumination. Very simple I2C interface.

Of course if they decide to start anodizing the aluminum brass color it may not work so well.

Answer 5

I'd expect brass and aluminum both to be variable in color, hardness, and both are diamagnetic. So, I'd moisten a swab with hydrochloric acid and wipe the metal. It won't do much to anodized aluminum, but it turns brass red (attacks the zinc, leaving a copper-rich surface). Unanodized aluminum will etch, but won't turn red.

Answer 6

If it's acceptable to place a probe in good physical contact with the object and to take a few seconds over the measurement, you could discriminate the two metals on the basis of their thermal conductivity (figures in W m^-1 K^-1 at 25 °C):

Aluminium        205 
Brass            109

You could do this by applying a fixed rate of heat input (i.e. power) and measuring the rate of temperature change. If you have examples of brass and aluminium samples to calibrate the technique with, you don't need to actually calculate what the thermal conductivity is, only to decide empirically what the threshold is between the two.

Note this method might be more difficult if the samples are hollow with thin walls and the wall thickness is not consistent - it can still be done but the measurement would need to be quicker and more sensitive and would then be more susceptible to variations in the quality of thermal contact between the probe and sample.