Capabilities of a Hall-Effect sensor

[Michael Hill]

I am wondering if a Hall-Effect sensor could be use to sense how close a material is to it (using it much like a proximity probe). I'm talking about a range from 0-30 mils. The material being used is close to Inconel. Price is not an issue, so if you can find one, even if it's a couple thousand, please tell me. The only reason I'm asking here is because I know there are quite a few applications in robotics (such as gear tooth counting).

[Richard Wallace]

Do you need a sensor that will respond with a linear (analog) voltage that is proportional to the distance to a ferromagnetic target, in the range of zero to 0.030 inch?

Or will a switched (digital) voltage that indicates the passing of a ferromagnetic target, such as a gear tooth, within about 0.030 inch from the sensor suit your needs?

[Michael Hill]

Well, the application would be to judge tip clearance on a turbine in an aircraft. I'm not exactly sure of the capabilities of a hall-effect sensor, especially at these high frequencies. Blades will pass at around 336 kHz. A voltage relation between distance is what I will need. I'm not counting them, we already know how many will pass by since we are measuring RPM with a quadrature encoder on the shaft.

[Alan Anderson]

Originally Posted by Michael Hill

I'm not exactly sure of the capabilities of a hall-effect sensor, especially at these high frequencies.

At its most basic, a Hall effect sensor responds to a magnetic field. Current through the sensor is affected by the field to produce a voltage output. If you could perform tight enough calibration, you might be able to measure in the range of distances you're looking at, but it seems to me that it would be difficult to maintain accuracy. Any stray magnetization of the turbine blade or housing could throw off the results completely.

I wonder whether something like Doppler rangefinding can tell you what you want to know. The shape of a plot of the beat frequency vs time ought to be related to the minimum distance as the turbine tip goes by.

There's also the possibility of using optical interference of laser light as a gauge, though the maximum distance is rather large for that to work well.

[Michael Hill]

I'm not exactly sure how much magnetic interference would be caused by the casing (it is actually aluminum).

The main reason I'm asking is because we have a tip clearance probe, that well...was made ~20 years ago and it's malfunctioning. We are having some rotordynamics issues (i.e., our rotor isn't balanced). The worst thing that can happen is a blade rubs against the casing.

[IKE]

Originally Posted by Michael Hill

Well, the application would be to judge tip clearance on a turbine in an aircraft. I'm not exactly sure of the capabilities of a hall-effect sensor, especially at these high frequencies. Blades will pass at around 336 kHz. A voltage relation between distance is what I will need. I'm not counting them, we already know how many will pass by since we are measuring RPM with a quadrature encoder on the shaft.

I have used a special type of hall effect sensor for measureing deflection of the flywheel on an engine. These were good for measuring deflection during idle. I beleive the linear voltage vs. distance range was around 1mm (.4 to 1.4 mm = 0.5 to 4.5V). Less than 0.4mm the linearity gets really bad. I think we were able to sample cleanly at 10kHz, but not sure if you could get reasonable data from a step (blade pass) at that speed.

The company that produced the sensor we used was Microstrain, Inc.
www.microstrain.com

We used a Rotec Front-end to acquire from this transducer.

[Michael Hill]

wow, I had a major miscalculation, blades are only going to be passing at 4.6 khz

[mluckham]

Hall-Effect sensors require a ferromagnetic target - your blades are aluminum, and there would appear to not be a way to attach a magnet, or steel object to them to take advantage of eddy currents as the aluminum blades pass by. Assuming you are wanting to measure at full RPM and not on the bench.

However you might be able to use an inductive sensor, as described here? http://www.compad.com.au/cms/autosys/articles/129 http://www.ia.omron.com/product/fami...index_fea.html

Even if blade distance(s) were known, it would not necessarily be helpful when balancing the rotor?

Rather than measuring the distance you might only need to detect the closest approach of any blade to the outer casing, or whether any blade approached closer than a given clearance?


If the inductive approach won't work, it seems like ultrasound or optical would be the way to go. I hope these links are helpful:

http://www.iop.org/EJ/article/1742-6...7-9ae896a46c3b

http://www.laserfocusworld.com/articles/257213

http://gltrs.grc.nasa.gov/reports/19...999-209183.pdf

[IKE]

I thought the housing was AL, and that the blades were something else.
4.6 kHz is still pretty fast for moving blades. Could you do a vibration technique with accelerometers on the axle (not sure if one exists on a turbine), and then at various locations on the housing. If you phase reference the motion of 1 relative to the motion of the other you might be able to deduce distance deltas.

[mluckham]

Oops - I should have known better to skip looking up "Inconel" (a nickel-type alloy).

Some of Allego Micro's linear hall effects have a 20 KHz frequency response http://www.allegromicro.com/en/Produ...ors/linear.asp which should let him detect the leading and trailing edges of the blades.

[Michael Hill]

Like I said, I'm not counting the blades, I want to know how close the blades are to the casing. (i.e. the sensor), so I'm not totally sure 20 khz is going to cut it. I mean, we can sample with our DAQ system at 1 MHz, but I don't know if the reaction of the sensor is fast enough

[Al Skierkiewicz]

Mike,
I have been thinking about this since I saw your post yesterday. If I am reading this right, you need to know the projection of each blade as it passes a fixed point correct? BTW, I think your first estimate of sample rate was accurate. i.e. 20 turbine blades mounted on a shaft spinning at 15,000 RPM passing a sensor seems like it's in the right neighborhood for 336kHz. At that sample rate, it seems that ultrasonics would not work, Hall effect or other metallic sensors would likely not work either. That leaves optical or some form of radar I think. It seems I read somewhere about an optical interferometer being used in just such an application.

[IKE]

20X15,000rpm/60(sec/min)=5,000Hz. (I forget the 60 all the time).

I checked the response on the non contact sensors I was talking about and they are also 20kHz. With some conditioning, you would be able to get a statistical distribution and a minimum clearance, but there is a good chance it wouldn't work well.
You may want to contact a company called LMS International. I know they do a lot of NVH work with the Aerospace industry. Maybe they have encountered a need for measuring something like this before.

[Al Skierkiewicz]

Yep, I forgot the 60, too.

[mluckham]

He said 4.6 KHz, so by Nyquist theory double the sample rate to 10 KHz to detect a blade. To ensure a measurement at both the leading and trailing edges of the blades (and points between) double it again to 20 KHz. To measure points across the width of the blade, double it again?

Keyence http://www.keyence.com has a line of laser sensors "LK-G/LJ-G Series" with sampling rates up to 50 KHz and a number of measurement ranges. They also have laptop-based software ... a standalone solution would include a LK-G Sensor head, LK-G controller, and LK-Navigator software.

Why not contact them at 1-888-KEYENCE and discuss your application with a sales engineer?