How to tell a sensor to stop counting

Thread Starter

McNuggies

Joined Feb 26, 2019
11
Instead of counting the reel, a digital encoder could be used on the string , these come in various resolutions (pulses/rev) and also an option of a one/rev marker pulse to replace the hall sensors.
The encoder would monitor the string as it traveled over a couple of rollers that had the encoder attached, the resolution could be tailored by way of roller dia etc.
The (high resolution) counting pulses could be used to detect any stoppage of the encoder, as their resolution would be much high than the marker counter.
Without knowing exact details, just an option to work with.
Max.
This does sound like a good idea
 
Last edited:

Thread Starter

McNuggies

Joined Feb 26, 2019
11
Two big constraints of the project are a tiny budget that has to be kept to and no use of arduino is allowed so dedicated pre-built rotary encoders are a bit outwith the budget of 20 pounds for the whole project
 
Last edited:

Thread Starter

McNuggies

Joined Feb 26, 2019
11
Another idea I had was having a wheel on top of the spool that presses down on the spool and the string and turns as the motor turns the spool so the string doesn't wrap around but simply gets fed through while still counting pulses but without having to deal with wrapping around the spool
 

Thread Starter

McNuggies

Joined Feb 26, 2019
11
If the length of the sting is unknown then there is no choice but to have some mechanism to detect the end of the string. That could be as simple as an LED and phototransistor set up with a narrow beam that the string would block until it ended. There are many models of a device with that arrangement available from many suppliers, so the hard part would be holding the srting so that it blocks the light beam.

Another way, which is mechanically easier, is to first pass the string over a roller before being wound on the spool. Then the roller can have some scheme to generate a string of pulses as long as it is turning, and the pulses stop when the string runs out. There will need to be a second roller pressing the string against the first one so that it will always turn as the string passes.The pulses can keep resetting a short timer that enables counting the spool pulses until it times out.. Easy, simple, and cheap, AND it does not cause any wear on the string.
This is similar to what I'm been thinking about
 

iONic

Joined Nov 16, 2007
1,662
Not sure how you are measuring string at all given that it appears that you may be counting a rotation based on the outer diameter of the spool, therefore one rotation in the beginning does not equal the same length of string as it will when the spool is filling up with string.Is the end goal to simply(and I say that work loosely) to measure the length of a string? We could really use some pictures of the current setup! What is the uC doing aside from counting magnets? Where are the magnets exactly? You will need some kind of mathematical algorithm to do that.
 

iONic

Joined Nov 16, 2007
1,662
This is similar to what I'm been thinking about
Now your talkin.' The string passed over/around a single roller of fixed diameter. Or the string passed between two rollers being pulled by the take-up spool. Larger rollers will require more freq placement of magnets, smaller rollers will allow for greater resolution, but will be more difficult to add additional magnets....depends on the desired accuracy required. If there is a tiny bit of friction on the rollers, they will stop when there is no more string, which is your end game. If no additional counting, the length can be then calculated.
 

MaxHeadRoom

Joined Jul 18, 2013
28,619
Another idea I had was having a wheel on top of the spool that presses down on the spool and the string and turns as the motor turns the spool so the string doesn't wrap around but simply gets fed through while still counting pulses but without having to deal with wrapping around the spool
This was the encoder principle.
Use a measurement on the string, not the spool.
Same idea.
Max.
 

sparky 1

Joined Nov 3, 2018
756
I agreewith MrBill2; the roller is key. counter gear
Look at the mechanical parts of a line counter.
reference to cable footage counter diagram:
https://www.fish307.com/troll-master-seahorse-cable-footage-counter-parts/
The diameter of the spool and the gear train are the functional parts.
The designer examines what criteria gives accuracy. Using a digital
counter is simply a modification to replace the mechanical counter in this scenario.
 
Last edited:

Thread Starter

McNuggies

Joined Feb 26, 2019
11
Now your talkin.' The string passed over/around a single roller of fixed diameter. Or the string passed between two rollers being pulled by the take-up spool. Larger rollers will require more freq placement of magnets, smaller rollers will allow for greater resolution, but will be more difficult to add additional magnets....depends on the desired accuracy required. If there is a tiny bit of friction on the rollers, they will stop when there is no more string, which is your end game. If no additional counting, the length can be then calculated.
Yeah so a light sensor that stops the program and takes into account whatever length was still left between the sensor and the powered roller (with the string being between the powered and nonpowered roller that moves with the powered one) then a hall effect or reflective optical sensor with strips of reflective material on the roller or just whatever that counts pulses as the string is fed between two rollers instead of being wrapped between anything, basically something along these lines
 

WBahn

Joined Mar 31, 2012
29,979
Two big constraints of the project are a tiny budget that has to be kept to and no use of arduino is allowed so dedicated pre-built rotary encoders are a bit outwith the budget of 20 pounds for the whole project
So what is the budget for this part of the project? If the entire project is to make something that can measure the length of a piece of string, that's very different than having to measure the length of a piece of string as one small portion of a much larger project all of which has to be done for under 20 pounds.

It would be a great help to have a better description of what the actual problem is. The initial description implied that the problem was add a measurement capability to a system that winds string onto spools to determine the length of string that was actually wound onto a given spool. But then it becomes evident that winding the string onto the spool is not only not an objective, but is completely dispensable altogether.

So what problem is it that you are actually trying to solve? Not what problem are you having with how you are thinking of solving the actual problem, but what is the actual problem?

Are you merely trying to determine the length of a piece of string?

If so, then a number of factors can have a significant impact. Such as:

Is it always the same kind of string?
Is it essentially a loose piece of string (i.e., not connected to anything except what you choose to connect it to as part of making the measurement)?
What is the actual accuracy requirement? Saying "mm" is too vague. Big difference between ±0.5 mm and ±5.0 mm.
What is the total length of the string expected to be? Big difference between 1 mm out of 10 cm and 1 mm out of 100 m.
Is it the total length of string that is being measured, or just the length from a specific point on the string to the end?
How is the string going to get connected to the measuring system?
How is the initial length of the string (from the lead end to whatever point will trip the end-of-string event) going to get taken into account?

In your current system, how closely are the magnets spaced on the spool? Are they really close enough to be able to detect the amount of rotation that corresponds to 1 mm of string? If your spool has a diameter of just 1 cm, there would need to be more than 30 magnets embedded in it to get into the 1 mm resolution range.
 

djsfantasi

Joined Apr 11, 2010
9,156
If you use the same mechanism that is used on a mouse wheel, you could get amazingly accurate.
Do mouse wheels still exist? The principle remains. The “wheel” rotated a shaft, on which a slotted disk was mounted. This disk rotated between a photodetector. Software counted the pulses and distance was calculated by the count times what each slot represented in distance.

Let’s say there was 50 slots. And the circumference of the wheel was 100mm. So a count of 50 was a measurement of 100mm. A count of 4 was a measurement would be 8mm. If the wheel circumference was 150mm, the same count of 4 would be 12mm.

If you used the pinch roller design, one of the rollers can have a slotted disk to use in this scheme. And if the tension was adjusted so that in the absence of the “string”, the rollers would not contact each other, a sufficiently long absence of counts mean you’ve run out of string. I am sure that you can specify this detection time to account for variations in the drive motor sufficient for your purposes.
 

pmd34

Joined Feb 22, 2014
527
Using a roller sounds the most reliable idea, though I would put a couple of turns of the string around it, so its more like a "capstan" and you are less susceptible to slipping. For the sensor, the cheapest would be to use a "digital potentiometer rotary encoder". If encoder and toller were able to move slight with respect to the reel against a tensioned spring, when the string ended as you wound it up the encoder part would spring back away from the reel and you could detect this with a micro switch.
 

iONic

Joined Nov 16, 2007
1,662
Do mouse wheels still exist?
The "mouse wheel" I was referring to is actually the scroll wheel. ...but the principle remains.! Been thinking, could a bycicle wheel be the "wheel." The larger the wheel, the finer the resolution(counts/revolution). And if you can't use a uC, then some LED's representing a binary count could work.
 

djsfantasi

Joined Apr 11, 2010
9,156
The "mouse wheel" I was referring to is actually the scroll wheel. ...but the principle remains.! Been thinking, could a bycicle wheel be the "wheel." The larger the wheel, the finer the resolution(counts/revolution). And if you can't use a uC, then some LED's representing a binary count could work.
It’s sll good. My description actually referred to mouse balls.
 

AnalogKid

Joined Aug 1, 2013
10,987
Depending on how many pulses the roller system makes per revolution, you might get a more accurate measure of string length than with the magnets on the reel. One sensor solves both problems.

ak
 
Top