Light Spectrometer help :?:

[Tom McCurdy]

I am currenlty doing a mentorship project with a prof at WMU a local college, (I am in high school) and we are working on automating a huge light spectrometer... anyway I need some way of knowing which way a gear is turning.... the gears we need to keep track of are the two gears that are the same size to the right of the motor in the top view picture... What would be a way I could instantly know which way the gears are moving?

Each time the gears move around once the display moves 5 numbers.

Here are pics of what I am working with
www.quantumninja.com/hw/elec/mentorship

there are 3 pictures there

 

[TheOne]

**broken link removed**

You most probably will need 2 offset in such a way when one is over a tooth the other one is over a space. Then it is a simple matter to sense which one give a pulse first to determine the direction of rotation. You may need to work on rising and falling edges as well to get reliable results.

 

[williB]

hmmm that left gear shaft would be perfect to put an encoder on , but the white box there is in the way..
can you take the gears off the unit..??
if so you could drill two holes in the left gear and use two opto switches.
such as EE–SX498..
do the gears allways turn at the same rate ??

 

[williB]

http://www.cherrycorp.com/english/sensors/op_speed.htm

You most probably will need 2 offset in such a way when one is over a tooth the other one is over a space. Then it is a simple matter to sense which one give a pulse first to determine the direction of rotation. You may need to work on rising and falling edges as well to get reliable results.

nice idea but the gears look like they are brass, and the geartooth sensors need ferrus metal., plus the pitch of his gears is too small to work with those geartooth sensors..
i liked the ring magnet example though..he could attach a ring magnet to the gear with two hall sensors, correctly placed they could inform a Pic on the direction..

 

[TheOne]

The ring magnet is worth a try.

Another option is to use a reflective light sensor with a disc

**broken link removed**

http://www.convict.lu/Jeunes/Divers/Poor_rot.htm

 

[williB]

The ring magnet is worth a try.

Another option is to use a reflective light sensor with a disc

**broken link removed**

http://www.convict.lu/Jeunes/Divers/Poor_rot.htm

yeah that would work he could paste the encoder onto a gear and use two light sensors...

how about this you know those micro switches with the long flat tab on them..?
IF the gears can withstand some resistance(physical) he could attach one in such a way that when the gear was moving in one direction the switch would be turning on and off..
and when the gear was turning in the other direction the switch would allways be off...

 

[TheOne]

It may work, but a non-contact solution will be more elegant.

 

[williB]

It may work, but a non-contact solution will be more elegant.

true , when tom gets back some things we need to know are
1) how soon after changing directions do you need to know that it has changed.??
2) how fast do the gears rotate?

 

[bmcculla]

You can buy a 2000 count optical encoder for ~$50. I've seen ones that will mount directly onto an existing shaft. If you have a bit of money to spend this is the best solution.

You could use your gear as an encoder. A pair of cheap laser pointers and phototransistors could be set up to give you a quadrature output using your gear teeth.

Depending on how quickly the gears are spinning you could use a small DC brush motor that would be spun by the gear or shaft. The motor will generate a voltage depending on the speed and direction of rotation.

 

[Tom McCurdy]

hmmm that left gear shaft would be perfect to put an encoder on , but the white box there is in the way..
can you take the gears off the unit..??
if so you could drill two holes in the left gear and use two opto switches.
such as EE–SX498..
do the gears allways turn at the same rate ??

no the two gears will always be going the same speed as one another, but that speed can vary.

 

[Tom McCurdy]

It may work, but a non-contact solution will be more elegant.

true , when tom gets back some things we need to know are
1) how soon after changing directions do you need to know that it has changed.??
2) how fast do the gears rotate?

1) Immediatly---losing a single gear tooth would screw up the calculation of the spectrum
2) The gears vary in speed from very very slowly to somewhat quickly

 

[Tom McCurdy]

I will try to gather more information today....

I am visiting WMU again to do some more work on the spectrum.

 

[williB]

i got it

 

[panic mode]

how fast is the rotation of gear(s) you want to monitor?
what is the range (revolutons)? what accuracy do you need?
if counting gear tooths will do the job, that's te way tou go.
you don't need two offset gears - just offset the sensors
looking at the same gear (for example sensor #1 looks at the
middle of the tooth X while sensor #2 looks at the
edge of the tooth Y).

 

[williB]

a better question is why .?
i assume the the gears turning gets somehow transferred to the display ..
but why do they change direction??
Are you trying to make the mechanical display electronic ??

 

[williB]

Posted: Wed Dec 08, 2004 11:46 pm Post subject: Light Spectrometer help :?:

--------------------------------------------------------------------------------

I am currenlty doing a mentorship project with a prof at WMU a local college, (I am in high school) and we are working on automating a huge light spectrometer... anyway I need some way of knowing which way a gear is turning.... the gears we need to keep track of are the two gears that are the same size to the right of the motor in the top view picture... What would be a way I could instantly know which way the gears are moving?

Each time the gears move around once the display moves 5 numbers.

Here are pics of what I am working with
www.quantumninja.com/hw/elec/mentorship

there are 3 pictures there

why not use the signals that turn the motors that turn the gears .
i mean there must be a direction voltage signal ..why not just use that??

 

[Tom McCurdy]

the gears can also be turned without the motor manually

 

[Tom McCurdy]

You can buy a 2000 count optical encoder for ~$50. I've seen ones that will mount directly onto an existing shaft. If you have a bit of money to spend this is the best solution.

You could use your gear as an encoder. A pair of cheap laser pointers and phototransistors could be set up to give you a quadrature output using your gear teeth.

Depending on how quickly the gears are spinning you could use a small DC brush motor that would be spun by the gear or shaft. The motor will generate a voltage depending on the speed and direction of rotation.

We are thinking that we are going to try to use the 2000 count optical encoder... that way we can know position as well as direction. We need to know position to at least 1/256th of a revolution. We also need a optical encoder with a long life.... hopefully a million or more rotations. Could you give me a link to somewhere we could purchase one with the specs... 50 dollars seems like a good deal.

 

[williB]

this one looks nice ..and is available in 3/16 to 3/8 shaft diameters..

**broken link removed**


HUB SIZE: (shown in inches)

Specify Hub Bore +.0004/-.0000 [+0,010/-0,000mm]
_______ ______________________

3/16 .1873 [4,757mm]
1/4 .2498 [6,345mm]
1/4+ .2501 [6,353mm]
5/16 .3123 [7,932mm]
3/8 .3748 [9,520mm]
3/8+ .3751 [9,528mm]
4MM .1575 [4,00mm]
5MM .1969 [5,00mm]
6MM .2362 [6,00mm]
8MM .3150 [8,00mm]

 

[Tom McCurdy]

I didn't see a price on there... do you know what that would cost?