Testing the status of a hinge

lukesmith

I want to build a device that can test the 'closedness' or status of a hinge. So the hinge would be fully open and would be able to close 90 degrees (like a door hinge for example) So one extreme it would be completely flat and the other extreme it would be a right angle. Does anybody have any ideas how I could test that status of the hinge and return the angle of the hinge?

lardom

If you are looking for the angle my first thought would be to put a sensor on a servo horn and rotate it until a threshold is reached. Perhaps an ir beam would be blocked after a servo rotates (x)°. Welcome to the forums.

Ttailspin

If you need digital,(fully open, fully closed.) then limit switches would work.
If you need analog,(anywhere between full closed and full open,) then a variable resistor seems like it might work.
As for how to attach any of the above, will need a bit more information about what it is going to be attached...

-Tommy

Erlend

You could glue on a tiny magnet on one part and a tiny reed switch on the other. It's reliable, but takes some experimentation to find the swich point. This is for Open or Close. If you want continous 90° why not use the same magnet, but a Hall effect sensor to measure the changing magnet field. Would be very reliable, but again, takes some experimentation - to determine the x° vs B (magnetic field) curve. Actually, I'd like to make one myself.

Erlend

lukesmith

Thanks for the replies guys. The reason I want to do this is to measure fingers on a glove. So I would measure each joint on the finger to see how far closed it is. The hall effect sensor sounds interesting but I'm guessing it wouldn't work for this function as there would likely be too many magnets too close together being detected by the wrong sensor.

dre

Hi,

It would appear that the sensor you might consider is a deflection or bending sensor.
A quick 'Google' brought up this:

http://www.sensorwiki.org/doku.php/sensors/flexion

Is this approach of any value?

cheers, David

Chris Savage

Testing for open/closed status is relatively easy. Getting angle involves a lot more mechanical and electronic intervention since in the closed position versus open it would be difficult to have a mechanical interface span the range and still be "hidden". If you just want open/closed status you can use a microswitch in the door jam.

Tracy Allen

Am I correct to think that there will be no metal hinge involved, only a person's fingers, perhaps in a glove to which the sensing harness is applied?

I'd go first to the hall effect sensor, although what you say about cross talk between neighboring fingers could be a problem. Maybe smaller magnets and detectors close to the joint.

Another thing that comes to mind is the inductive variometer principle. Have one coil excited by a high frequency and another coil is a pickup. The signal picked up is a function of the position of the two coils. You might arrange it as rings, one on each joint of the finger (insert grain of uncertainty here). The signals could be separated by selective drive to the coils, or by different frequencies, or by something like PhiPi's recent Walsh functions thread. Would be a nice Prop application.

Beau Schwabe

Tracy Allen,

you are correct in your assumption with relation to two parallel coils (hinge closed) versus two perpendicular coils (hinge open).

If he goes with two open coils, the general rule of thumb is that the average diameter of the coil equates the the outward sensing distance for open air coils.
So the distance from center to center of the open hinge needs to be determined. i.e. on my office door that's about 2 inches. So my average coil diameter should be at least that. So to fit in the door jam I made a coil that's 2 inches by 1 inch (see attached image) ... average diameter about 1.5 inches and it's just barely enough to sense a reading... I would go larger to make the full 2 inch average at least.

Note: the "red" coil that you see is literally a single loop of wire. This would prevent the door (the moving part of the hinge) to require any power.

A simple LC oscillator should work. Attached is an example of one that I am using...

Note: the output frequency is high in the MHz usually, so you will need to divide it down with something like a 4040B as shown in this e-bay item.
http://www.ebay.com/itm/251191079787?ssPageName=STRK:MESELX:IT&_trksid=p3984.m1555.l2649

ElectricAye

Run cables from each joint back to the arm where (bulky?) sensors can measure the amount of springy stretch over each joint? Might that approach make things less bulky around the fingers?

Duane Degn

lukesmith wrote: »

Thanks for the replies guys. The reason I want to do this is to measure fingers on a glove. So I would measure each joint on the finger to see how far closed it is. The hall effect sensor sounds interesting but I'm guessing it wouldn't work for this function as there would likely be too many magnets too close together being detected by the wrong sensor.

I don't suppose you could tell us what you want to accomplish with this? It would keep the guessing down.

You original question used a door hinge example. You'd likely use very different methods to measure the angle of a door hinge than you would to measure the position of a finger.

What kind of gloves? Do you assume the glove is needed to attach the sensors or do you have a specific glove in mind?

Do you want to measure the angle of each joint in each finger or do you care more where the tip of the finger is?

There are so many different ways to do this and a lot will depend on the specific requirements you have. Is this an input device for a vehicle? Game?

More information on your end goal would be helpful.

Erlend

Fit a bracelet to the glove, which can be clamped firmly to the wrist. Run tape/ribbon from tip of finger, along the ridge, and into the bracelet. In the bracelet convert longitudinal movement into a signal - lots of options here - by connecting to a linear potentiometer, and then to a spring. At startup, hold a book,and calibrate the signals. This is close to how the body&brain does the same thing, so it's a proven method.

Alternatively find a semi conducting elastomer and measure the stretch directly as changing resistance of the tape. The spongy stuff that comes with IC's have this property - go ahead and measure resistance of one while pressing down on it.

Erlend

lardom

I'm wondering if you're trying to design a 'hand-like' gripper...

Chris Savage

I missed the glove reference earlier. I saw the first message about door hinges and was on the wrong track. Most of the VR gloves I have seen measure finger bends using a flexiforce style sensor. Essentially a resistive strip that changes resistance as it is bent.

Beau Schwabe

I missed the glove reference also ... an induction sensor could still apply. A coffee straw wound with wire and a flexible steel cable moving in and out of the straw would work.

Example using a BIC pen and a piece of coat hanger, but a coffee straw and steel cable would work just as well...
http://www.youtube.com/watch?v=9sBm1zO7G0w

The basic oscillator schematic can be located in post #9 of this thread, then simply divide the frequency down with a 4040B (12-stage binary ripple counter) and use a Basic Stamp or Propeller to read the pulse width or count the number pulses over a fixed time period.

erco

This old Mattel Power Glove is just begging to be hacked... it's cheap (currenntly $20) since it's without the ultrasonic sensors that go on the TV set. AFAIK it uses the flexible resistors in the fingers. Plenty of buttons to hack, too!

http://www.ebay.com/itm/Nintendo-power-glove-with-manual-/171020515697?pt=Video_Games_Accessories&hash=item27d19df971#ht_101wt_917

TV commercial at http://www.youtube.com/watch?v=FMkTq3NW-jE

lukesmith

Awesome replies guys! Sorry about the hinge analogy I made things pretty confusing. Yeah my aim is to measure the angle of fingers joints on a glove and then use then feed the data to a pc to animate a 3D mesh. I managed to get hold of a flex sensor so I will see what that's like - though I have read that it is only accurate to around 10 degree increments. I will try out these other methods and compare the results. Anybody got any ideas what kind of thing might give the most accurate results?

ElectricAye

lukesmith wrote: »

...Anybody got any ideas what kind of thing might give the most accurate results?

Have you considered doing a patent search on this sort of thing? People have been doing virtual reality gloves for decades, so I'm guessing many of the problems you are likely to wrestle with have already been solved. Just a thought.

lukesmith

I have tried looking on the USPTO but I didn't have much luck. Pretty much every data glove I have seen uses a flex bend sensor but I have never been hugely impressed with the result. However testing this flex sensor that I picked up, the accuracy is actually pretty good and with some smoothing I think it would be a good approach - so perhaps its more down to the implementation to get a decent result. Testing and comparing the accuracy of some of the ideas suggested here vs the more commonly used flex sensor sounds like a decent project.