Source for Linear Optical Encoder?

Bill Chennault

All--

Does anyone know of a source for linear optical encoders with hobbyist pricing?

Thanks!

--Bill

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
You are what you write.

Bill Chennault

All--

I think I have answered my own question: There is no such thing as a "hobbyist" priced optical linear encoder. Further, there is no such thing as a hobbyist priced linear encoder of ANY kind.

--Bill

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
You are what you write.

John R.

What distance/accuracy are you looking for?

I've made "Goldbergian" encoders before by simply mechanically driving a potentiometer, either via rack and pinion on the pot shaft, or via pulleys and cables. Not necessarily very accurate, but functional within it's limitations.

You could also do the same thing with a rotary encoder.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
John R.

8 + 8 = 10

Bill Chennault

John--

I have come to the same conclusion: If I want linear encoders, I will have to build them.

Accuracy? I dunno. I was thinking maybe 32 counts per inch? Does that sound like too much for a home-built encoder? I want to measure the angle between two legs. Actually, in practice, a simple count will suffice. I really do not care about the angle. In practice, the count will tell me what position the legs are in relative to one another.

A pot might work on the axis of the joint itself. I am a newbie and have never tried something like that. (For that matter, I have never tried using a linear encoder of any kind.) If I use a pot, I imagine that I will need to do some AD conversion?

For a bit, I was very excited about the linear encoder technology available . . . until I saw the prices.

Thanks for any input you care to share!

--BIll



▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
You are what you write.

Larry

Bill Chennault said...
I think I have answered my own question: There is no such thing as a "hobbyist" priced optical linear encoder. Further, there is no such thing as a hobbyist priced linear encoder of ANY kind.

Bill, you can get a very accurate linear encoder for not a lot of money by converting inexpensive Harbor Freight digital calipers. They contain a serial interface that is pretty easy to tap into. 6" calipers can be had for $16, 12" ones for maybe twice that.

Take a look here:

www.pcmx.net/gauge/

or start googling variations on dro (digital readout)

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔

Bill Chennault

Larry--

What a neat idea! Building the thing is probably beyond me, but I am sure going to give it a hard look. Plus, HF is just a few miles from me. And, I think I can buy the same thing from Grizzly, online.

Thanks!

--Bill

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
You are what you write.

kelvin james

The processing ability has to be considered first before implementing the hardware.

kelvin

Bill Chennault

Kevin--

Of course, you are correct. When I thought about the data--using my limited knowledge of the Stamp--I thought, "Gee, will the Stamp know what to do with that floating point number?" Then I began to think of ways to just GET the output as an integer and "massage" it.

In any case, it is a neat idea and it got me to thinking out of the box into which I had slipped.

--Bill

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
You are what you write.

helluvawreck

Hi, I'm new to the forum and have only been working with stamps for a couple of weeks. I have used plc's and encoders in my work, however. Encoders are expensive. However, I believe that with a little ingenuity you could build an encoder with a plastic disc with holes in it and use an LED as a light source and the rotation of the disc would allow the light to pass through the holes as the disc turned. Slots around the circumference of the wheel would also work I suppose. You could also have a second LED and a single hole located on a second radius from center that would send a signal at each complete revolution. The sensitivity of the sensor on the opposite side could be adjusted to ellimnate false readings. I am new to electronics, as well, but I have seen relatively simple circuits that would accomplish all of this. It seems like just the other day while surfing I ran across some people using small encoders in a discarded mouse for this. I have never taken a mouse apart and looked inside. Unfortunately my main experience has been with working on large industrial machinery that is controlled by plc's and conventional relay logic. Up until recently electronics has been a black box to me but I intend to change all of that. These stamps are fascinatating things to me and also electronics. I hope I can learn a lot from all of you.

BTW, if you have access to a drill press you can take a discarded sprocket and build a simple jig to drill the holes. Just make a spring loaded arm to engage in the sprocket teeth. Rotate a simple drilling table from tooth to tooth and you can space your holes in a nice fashion. A thought just occurred to me. I saw a small sensor just the other day that (On Jameco, i believe) that was a combination sendor and reciver built into one piece that could be used for this purpose. I don't think it was very exspensive. It had a slot in it so a wheel or lever could pass through. There are more exspensive versions of this same type sensor for industrial machinery. I'll see if I can find that part and pass it on to you.

Good luck,

helluvawreck

helluvawreck

I'm interested in a cheap encoder as well. I just tried to look this up where I saw that sensor. Here are some links:

http://www.swiss.ai.mit.edu/~pmitros/encoder/
http://www.robotstorehk.com/sensors/sensor.html

I believe I will order some of these little sensors myself to see how well they work. I haven't read this article yet but it looks interesting.

helluvawreck

Speaker

I have used the NTE3101 recently, w/ a 'home made' wheel encoder. Run the output thru a schmidt trigger to clean it up. Digital count for under $5.

helluvawreck

Speaker, I'd be interested in the schematics for the circuit that you used if you have it handy or know where to direct me. I would like to do something similar. I'm just a newbie in both electronics and stamps but I'm trying to learn fast. I have a couple of robotics books and I believe they have some circuits for this and have seen some on the internet. I want to try my hand at a simple robot to get my feet wet. You seem to have a simple solution that you have actually completed and used. Are you a roboticist?

helluvawreck

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
I have found that:
We are never too old to learn new tricks.
We should never be afraid to ask questions.
There are no stupid questions.
We should stay humble and be thankful for our blessings.

Speaker

It occurs to me that the simple optocoupler idea doesn't indicate direction, just distance (number of pulses).· I really prefer NTE's designation of 'Photon Coupled Interupter Module.'· But a rose by any other name.· I don't remember part numbers, but there are optocuplers with 2 devices in one physical package.· I think the idea is that one pulse is detected slightly ahead of the other, thus indicating direction.· (But I'm not gonna swear to that)· The 555 is my 'Schmidt Trigger.'· Pin 3 on the 555 is the output.· I fed it to a couple of counters.· You would likely send it to a stamp.· I'm a rookie to microprocessors, but I wouldn't think the code would be very complicated.· 1.· Look at both inputs to determine direction.· 2.· One direction increments,· the other decrements.

Bill Chennault

helluvawreck--

I have built many rotary optical wheel encoders in the past . . . like about 30 years ago. Today, my need is for linear encoders. Rotary optical encoders are cheap. Linear encoders--using any technology--are not. I hope someone proves me wrong, or else I will have to build them!

--Bill
ps Thank you for doing all that research. Tonight, when I have a moment, I will look at those sites you posted.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
You are what you write.

bubblehead

Here's an article about a school robotics class that used optical mice·to measure wheel rotation:

http://delivery.acm.org/10.1145/1170000/1165895/9103.html?key1=1165895&key2=7552081611&coll=ACM&dl=ACM&CFID=15151515&CFTOKEN=6184618

Basically, they cut a mouse pad into a circle and mounted it on the side of the robots wheel.· A bracket held the optical mouse next to the mouse pad.· As the wheel turned, the optical mouse would track the movement and report back to·the robot controller (an old PC motherboard).

The article·got me to thinking about other applications for the sensors in an optical mouse.· So I·researched (a al google)·some of the chips used to implement optical mice.· They are basically an imager, microscope, and DSP in a single chip.· The imager/microscope·captures an image of the surface the mouse sits on, and the DSP analyzes the images to determine if the mouse has moved.· Depending on the chip, they can be configured to provide serial, PS/2, USB, or quadrature output and can have resolutions of up to 800 counts per inch.·

It should be possible to hack an optical mouse to provide a 2-D linear encoder.· A·microprocessor like a Stamp/SX/Propeller should be able to read a serial, PS/2, or or quadrature output.

I was thinking such a sensor·could be used·to directly measure the motion of an x/y table.· The advantage being that you don't have to worry about gear backlash, wheel/belt slippage, etc.· However, I have no idea how precise and/or repeatable the output of one of these sensors is.· For example, I suspect that if you moved the sensor in a circle so that it returned to its starting position, the reported sensor position/counts would not be the same as when you started.

Some of the sensor chips let you read out the actual image.· Perhaps with a suitable pattern on the working surface (think the dots on the ball in an optical trackball), the image data can be used to improve the repeatablility.

Here is a spec sheets on·an encoder chip:

www.st.com/stonline/products/literature/ds/11894/vt5363.htm

I just found·links to experiments using an Agilent mouse chip for robot navigation.· They use a Basic Stamp.

http://www.contrib.andrew.cmu.edu/%7ettrutna/16-264/Vision_Project/

http://www.contrib.andrew.cmu.edu/~ttrutna/16-264/Vision_Project2/

Larry

Another place to look for linear encoder hacks is totake apart an old HP printer. the sensor and linear encoding strip are in there--you just need to hook them up to the stamp.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔

Bill Chennault

Kevin--

It looks like all the FP math could easily be handled by the Floating Point Coprocessor V3. However, one wonders, especially me, about straying from the point of what I want to do with the microcontroller: I/O and some other low-level functions. Everything else, like pseudo-intelligence, floating point math, big memory, etc, should be up-loaded to a pc-type device.

BUT, if getting a good (cheap!) linear encoder demanded the use of the fp unit, then I might think about it.

--Bill

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
You are what you write.

Bill Chennault

Larry--

Hmmm. I have about all the old HP printers one could want. Thanks for the idea.

--Bill
ps Think hundreds.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
You are what you write.

helluvawreck

Bill, I was referring to industrial encoders. I have not found these encoders to be cheap unless they were purchased as surplus on perhaps Ebay for example or on the used market. The only linear encoder that I have ever used was on a panel saw that has a pusher that indexes plywood sheets as they are being cut to width and length. It has a twelve foot long magnetic tape and a reader that runs along with the carriage and rides about 2 mm from the magnetic scale. The cost of the whole rig was about $400. It was made by Siko. The pusher probably weighs 1500 lbs and is 12 feet long. Each end runs along with rack and pinion so it stays very square and parallel to the saw carriage. The saw cuts as many as 5 sheets of plywood at a time so the load being pushed is also heavy. The scale and reader are accurate to 0.002". The cut, of course, is not this accurate but it easily maintains an accuracy of well within 1/64" which is plenty good enough for a cabinet making shop. These encoders run for years on an industrial machine like this. The rotary encoders that would be used on something like this would perhaps have as many as 6000 pulses per revolution. Again the encoder, even though rotary, is not cheap and will cost at least $175. BTW, I'm not a trained technician or machinist. I've just worked in industrial plants for 35 years and have had to pick up what little I know the hard way - you either fix the machinery and keep it running or you go broke.

I'm totally new to hobby electronics but I find it fascinating - so much so that I just don't know where to start. I've been studying electronics on my own for about 2 months and got my stamp about 2 weeks ago at Radio Shack. I would have gotten it sooner but the store manager said that you couldn't do much with them and that they were just toys. The stamp is amazing to me. Now, since I got my stamp, it's hard for me to study my basic electronics because of the stamp. Bill, do you build robots?

helluvawreck

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
I have found that:
We are never too old to learn new tricks.
We should never be afraid to ask questions.
There are no stupid questions.
We should stay humble and be thankful for our blessings.

Bill Chennault

helluvawreck--

Your price figure of approximately $400 is amazingly accurate! I managed to actually get pricing from some of the industrial sales people and that pricing confirms everything you said. All I wanted to start with was a ball screw . . . too bad, can't afford it. So, I will settle for building my own. It will lack the precision, but I am confident I can build what I need.

Yes, I have built several robots . . . BUT, it was about THREE DECADES ago. Therefore, like you, I am a newbie and am excited and think that when I make an LED blink using a BS2p24 BOE--which is about 1/10 the size of my original Z80 microcontroller at something larger than 12" x 12"--that I have really achieved something! I am going through WAM and StampWorks, although I think I could skip a lot of it. However, there are nuggets of electronics information contained in both which I have missed out on in the last 30 years and which I cannot afford to skip.

I taught Robotics in 1979. In fact, I wrote the syllabus and had it approved here at the college. It was much like this forum; people helping people. We used a "thing" (kinda looked like a triangle with wheels with the Z80 microcontroller mounted on top) that I bought and built for the execution of student code written in ZBASIC. (I have forgotten the name of the microcontroller. It may have been "ZBASIC Microcontroller", or something.) The microcontroller was not reliable, as it would often reset with a slight flex of that huge board. But, it had reed relays, I/O pins, LEDs, and some other stuff. The main thing, as it is here, is that it was programmable in some version of BASIC.

The class I taught was the most successful class in robotics ever taught by the college. Even after we spent MILLIONS of dollars for equipment and REAL electronics/mechanical/hydraulic engineers to teach. However, I advised AGAINST all this expenditure for one simple reason: Our goal, formed at a higher level than me (back then) was to teach General Motors all about robotics. WHAT A JOKE! At the time, GM was NUMERO UNO in the robotics world. There was NOTHING we could teach them. In fact, I went to Albuquerque to the first (I think) annual Robotics World (or some such) just to meet with the high priests of GM robotics engineering. Although nice guys, they basically told me exactly what I had already figured out.

Today, I am the Dean and CIO of the college (in KC) and thinking about retirement. I not only want to learn modern stuff for something to do in retirement--as if I needed something else!--I am thinking about reviving that old robotics class that I designed and taught, which was one hell of a lot of fun for the students and me, and come back part-time just to teach it. And, thirty years after I taught the first one, I bet the microcontroller will be FAR more reliable! <g>

Obviously, the store manager with whom you spoke did not know beans, much less computers. I GUARANTEE you that the Stamp you bought at Radio Shack has FAR more capability than what I started with!

--Bill

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
You are what you write.

helluvawreck

Bill, I envy your electronics background. I went to UGA and studied every science class they taught in freshman/sophmore. I couldn't figure out what I wanted to do. Then I tranferred to GA Tech and majored in Mech Eng and went to tech for a couple and a half years but I had a family then and dropped out early in my senior year and went to machinist school for 6 months with the intention of going back to school and working my way through as a machinist. But I never did. After 6 months of machinist school it was getting hard financially for my family so I asked the machinist instructor if I could make some parts for a machine that I wanted to build and grade me on those parts instead of the ones I was making (they didn't make much sense to me). He said that he couldn't do that so I quit. I talked my father into buying a bridgeport mill, a drill press, a horizontal bandsaw, a welding machine, and a cutting torch and started building some stuff for our small furniture mfg business. I also began maintaining the equipment and made it a point to ask every electrical technician that came into the plant a bunch of stupid questions and watched everything they did. After a couple a years I was doing all the maintenance including electrical. Unfortunately we lost that company in a big law suit and I went to work for another mfg company for a couple of years as a maintenance mechanic. Then my brother borrowed a little money and started another mfg company and I went to work for him as a "slave laborer" to earn stock in it (I also got paid but it wasn't nearly what I was worth). After 10 years of hard work mostly in maintenance I acquired my stock. Our plant is now 160,000 sq ft and has a lot of machinery. I have a small machine shop inside the plant and still do most of the engineering and maintenance. I don't use much of anything that I learned in school except how to think. Most everthing I use I had to learn by doing. I would love to have an electrical engineering degree but it won't ever happen (I'm 56 now). So I'm just going to learn as much about electronics as I can on my on in the next few years because I have some stuff I want to build and it will take electronics to control it. Your right, these microcontrollers are something else.

helluvawreck

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
I have found that:
We are never too old to learn new tricks.
We should never be afraid to ask questions.
There are no stupid questions.
We should stay humble and be thankful for our blessings.

Bean

Guys,
I have a CNC mill (a small one) that I can make a linear encoder maybe 14" long or so.
What length do you need ?
Here is an example of a 3" wheel encoder I made.

Bean.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Cheap used 4-digit LED display with driver IC·www.hc4led.com

Low power SD Data Logger www.sddatalogger.com
SX-Video Display Modules www.sxvm.com

"People who are willing to trade their freedom for·security deserve neither and will lose both." Benjamin Franklin

Tommy Bot

Thinking out loud here:

How about taking Bean up on making you an encoder wheel, but have him cut 1/32" teeth around the perimeter.

Then buy an "Incra Jig" which is a linear woodworking jig that works as a precision marking gauge, a freestanding depth-of-cut gauge, a fence alignment gauge and a center finder and other uses.

http://www.incra.biz/Products/IncraGauge.html

It employs two pieces with 1/32" teeth cut along·one face and placed together teeth side in. When you loosen the knob and slide it to length and re-tighten it, they claim a 0.004" repeatability. I have had one for woodworking for years now and swear by it.

Take the two racks apart, customize one and place it on your linear piece, and mount the wheel on the teeth.

Sort of a linear to rotary converter.

Or look at the fence replacement racks (same 1/32" system)

http://www.incra.biz/· (Incra homepage, look for replacement parts)



Just a thought



TT





▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
(Frequently heard from other's)

Tommy, I know it wasn't designed to·x, but can you make it·do x·anyway?

·

Bean

Tommy,
I was offering to make a linear encoder. I'll have to measure to see what the maximum length I can cut is.
I not sure I can make 1/32" teeth. I've never tried.

Bean.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Cheap used 4-digit LED display with driver IC·www.hc4led.com

Low power SD Data Logger www.sddatalogger.com
SX-Video Display Modules www.sxvm.com

"People who are willing to trade their freedom for·security deserve neither and will lose both." Benjamin Franklin
·

kelvin james

Bill

You can get a linear encoder module and strip from usdigital for around $50. They are quadrature and have index. I think this is a reasonable price for a fairly high resolution linear encoder.


kelvin

www.usdigital.com/products/em1-heds/

helluvawreck

kelvin, I surely do appreciate this tip about usdigital. I just went to that site and they definitely look very interesting and are certainly affordable even for many hobbiest purposes. thanks

helluvawreck

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
I have found that:
We are never too old to learn new tricks.
We should never be afraid to ask questions.
There are no stupid questions.
We should stay humble and be thankful for our blessings.

Tommy Bot

Bean (Hitt Consulting) said...
Tommy,
I was offering to make a linear encoder. I'll have to measure to see what the maximum length I can cut is.
I not sure I can make 1/32" teeth. I've never tried.

Bean.

Bean, Sorry about that.

Maybe a plastic gear with the proper teeth already on it, then cut the encoder holes and viola~.

But that being said, Kelvins link looks very promising. I think it can be made for less, but $50 without the fabrication hassel isn't bad at all.

TT

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
(Frequently heard from other's)

Tommy, I know it wasn't designed to·x, but can you make it·do x·anyway?

·

Bill Chennault

Tommy and Bean and All--

Gee, and I thought this thread was dead! Thanks for all the ideas and offers. I need to do more design work and quite frankly, learn more about the Stamp and how to use it before I actually build something that will involve the linear encoders.

But, I am going to bookmark this thread! [noparse]:)[/noparse]

--Bill

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
You are what you write.

Bill Chennault

kelvin--

Thank you! Those linear strips REALLY interest me!

--Bill



▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
You are what you write.

Bill Chennault

kelvin and All--

And, at USDigital, this looks good, too!

http://www.usdigital.com/products/ma2/

--Bill

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
You are what you write.

crgwbr

I have experiance with the US Digital MA-2p encoders in industrial settings.· I highly recomend them.· To convert them from rotery to linier I, you could glue/epoxy a foam hobby wheel onto the shaft, and have a flat surface that it rides against.· If there is enough pressure between the wheel and plate, it should not slip.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
NerdMaster
For
Life