Sensing winch cable position

[David Lowy]

I need to sense the position of a cable attached to a 2000lb winch, similar to the one pictured below.  

The project is to supplement the winch's  manual controller with an automatic one. 

The winch is mounted to the aluminum frame of a tent-like aircraft hangar and is used to open and close two doors through a set of pulleys, an action I would like to perform without having to hold the button for the 90 seconds or it takes.  The sensor needs to tell the circuit  when to shut the motor off  when either of two cable positions is passed (open and closed). 

Currently the 5/32"  cable  is not going through a guide, but one can be built and sensor mounted to it.  

 

Would a hall effect sensor from a bike computer work?  Would a couple of small magnets crimped to the cable be sufficient?  

Anything clamped  to the beginning of the cable would have to withstand the cable being  wound on top of it without interfering with it  too much. 

Any ideas?

 

[Dan Royer]

I understand you don't want to put something on the head of the cable (because it will be wrapped into the coil and possibly damaged).

How about if you put an encoder on the side of the winch and a hall effect for the tail end of the cable?  Then you'd know how much was unwould by counting the steps and you'd know when you are close to the tail (eg stop winding!).

I would also consider putting two hall effect sensors, one after the other.  That way you can automatically determine the winding on the cable.  (It acts as a backup emergency stop, too.)

Dan

---

From: da...@lowy.com
Subject: Sensing winch cable position
Date: Wed, 20 Mar 2013 09:07:37 -0700
To: vhs-g...@lists.hackspace.ca

[Dave Hylands]

On Wed, Mar 20, 2013 at 9:15 AM, Dan Royer <aggr...@hotmail.com> wrote:
>
> I understand you don't want to put something on the head of the cable (because it will be wrapped into the coil and possibly damaged).
>
> How about if you put an encoder on the side of the winch and a hall effect for the tail end of the cable?  Then you'd know how much was unwould by counting the steps and you'd know when you are close to the tail (eg stop winding!).
>
> I would also consider putting two hall effect sensors, one after the other.  That way you can automatically determine the winding on the cable.  (It acts as a backup emergency stop, too.)

You may want to consider putting the sensors on the doors rather than on the winch.

I think you'll find that the winch my wind up in different positions depending on how the cable bunches up during winding. (i.e. wind it several times in a row and the door may wind up in different positions). Maybe the differences are too small to care about, but its something to consider.

The advantage of using the doors is also that you have absolute position, rather than relative position, so you don't need to reset everything if you get a power failure during door opening/closing.

--
Dave Hylands
Shuswap, BC, Canada
http://www.davehylands.com

[Tom Keddie]

I feel like some of these systems work by sensing the current drawn by the motor, this is good for safety too.  If something obstructs the doors the motor thinks they are open/closed and stops.

Cheers,
Tom

[Miststlkr]

Just for another option than the initial one, maybe a sensor in the doors to know when they are closed (magnet sliding door alarms come to mind) and a mechanical endstop the door will trip when opened.

Alternatively, what are the chances of Assn optical sensor picking up a brightly clotted strip of (paint/tape/wire) on the cable to use as a positional refrence?  Place the sensor before the spool and a contrast sensor would say (silver cable: go, silver cable: go, silver cable: go.... Black stripe means stop!)  (or opposite if your cable is dark..)

[David Lowy]

Thanks all for your thoughts so far:  

Dan:  I like your thinking of putting a sensor on the winch. If I understand correctly you are suggesting counting spindle revolutions. Yes, it will be easy to put one or two sensors on the tail end because it doesn't get wound, but unfortunately when the doors are open those sensors will high up in the air, attached to the original hangar door ropes. 

 

Dave:  Good idea, but it may not be practical to sense the doors, which are canvas strapped to an aluminium frame.  The door bars would swing open in a fairly predictable manner, but the fabric folds can bunch differently and there is no good place to put the stationary end.  A solution on the cable or winch seems more practical. 

Tom: Yes, and I would like to look into this all-electronic solution of sensing current, if it is not too difficult, for the safety aspect.   Thanks for reminding me of this.   In the open position there is an increasing amount of load on motor. However, towards the end of the closed position, unfortunately there is no load on the motor.  Dozens of spindle turns are required to release enough slack before the doors can be latched shut.   I suppose a timer might work, releasing x seconds of cable after the current stabilizes. 

Miststlkr:  Door magnet sensor problem addressed above (doors are not shut until I pull them closed after the cable has slackened enough) but I will keep in mind using optics. 

Thanks!

 

David Lowy

778-889-7558

[Miststlkr]

Bah.  I missed the "tent-like" comment. I had big ol' sliding hangar doors in my head.  Sorry!

[David Lowy]

Excuse the reply to myself, but a couple thoughts and clarifications: 

On 2013-03-20, at 9:57 AM, David Lowy wrote:

Dan:  I like your thinking of putting a sensor on the winch. If I understand correctly you are suggesting counting spindle revolutions. Yes, it will be easy to put one or two sensors on the tail end because it doesn't get wound, but unfortunately when the doors are open those sensors will high up in the air, attached to the original hangar door ropes. 

.... so of course they will only be useful for sensing when the door is open.    

For a number of reasons sensing the closed position seems to be the challenge. 

Dan, can you elaborate on your "encoder" idea?

 Tom: Yes, and I would like to look into this all-electronic solution of sensing current, if it is not too difficult, for the safety aspect.   Thanks for reminding me of this.   In the open position there is an increasing amount of load on motor. However, towards the end of the closed position, unfortunately there is no load on the motor.  Dozens of spindle turns are required to release enough slack before the doors can be latched shut.   I suppose a timer might work, releasing x seconds of cable after the current stabilizes. 

"Dozens" is an exaggeration.  However, many zero-load turns are required before the doors have enough slack to manually latch together.  Until this time, they are hanging supported only by the canvas which has to be pulled tight by hand, and at the end a levered latch. 

[Dan Royer]

I'm only using the sensor as a limit switch.  That means it does two jobs:
- prevent the winch from winding too far, regardless of which way it is wound
- provide a reference point from which to move

If you know it takes N steps from the automated winch to unwind (read: open the door) then you can tell the door position at any moment by tracking the number of steps that have been taken.

But what if the machine loses steps?  Yeah, well, that's why it uses an encoder wheel.  it's a disc of black and white marks that moves with the spinning of the spool and a sensor picks up on the marks to tell how much has actually moved.

If you prefer, wrap the cable once around a wheel and put the encoder on that.  eg,

winch -> encoder -> hall effect -> door

this way even if the winch gets tangled or something the encoder should maintain a pretty good count of how much cable has actually moved.

Dan

---

From: Da...@Lowy.com
Subject: Re: Sensing winch cable position
Date: Wed, 20 Mar 2013 10:18:58 -0700
To: vhs-g...@lists.hackspace.ca

[David Lowy]

Dan, I think you may have made a typo or are imagining  the system backwards. 

To be clear, the cable winds onto the spool to open the door, and unwinds to close it.   

A magnet on the cable tail end (and/or a current sensor) can tell when to stop the winch  from winding  (opening) too much.

As far as unwinding (closing), yes a dark/white pattern applied to the spool will become visible as the cable unwinds. Which end of the spool unwinds first should be the same as which end wound first. 

Da...@Lowy.com

Sent from my iPhone

[David Lowy]

By "opening" and "closing" I meant the door. 

To prevent further confusion perhaps  we should just talk about the winch, which "winds" the spool up to become full of cable and "unwinds" to become empty of it. 

Da...@Lowy.com

Sent from my iPhone

[Dan Royer]

I stand by my previous statements.  an encoder wheel turned by the cable and positioned between the winch and the hall effect sensor would give you good distance-moved measurement and a limit switch, everything you need for repeatable motion control in your use case.

Dan Royer
Marginallyclever.com - 3D Printing & Robots
+1 604 916 2281

---

Subject: Re: Sensing winch cable position

From: Da...@Lowy.com
Date: Wed, 20 Mar 2013 12:52:58 -0700
To: vhs-g...@lists.hackspace.ca

[Kevin]

Counting winch rotations would be inaccurate depending on how the cable winds on the winch because the diameter of the wrap per rotation is not constant as it accumulates on its self increasing the diameter or unwinds, decreasing the wrap diameter.

Your best bet is to feed it through three rollers.  You can mount them to a plate, the rollers should deflect the cable path enough to use the cables existing tension to turn the rollers.  The center roller can use a magnet, or cog, or a light dark pattern to send on-off pulses to a micro controller of some kind, an arduino mini perhaps.  Use a digital pin for the button you use to activate it, count the pulses to remember location watch the button for user input, trip a relay on or off to control winch.

Write the position to flash memory so that it will still know the winch position when or if the arduino loses power.

-Kevin

[Kevin]

Using the non-volitile memory on an arduino.

http://www.arduino.cc/en/Reference/EEPROM

[David Lowy]

Dan, 

I didn't understand until now that you meant a separate encoder wheel turned by the cable.  I pictured markings on the rotating winch spool.

I will think on this.  It sounds good. 

Dave

[Simon Lyons]

I know you said a door switch was out but I just saw in the latest flyer from http://www.goldmine-elec.com/ there is a "Wide Gap Door Alarm Switch" (p/n G19262). FWIW.

// Simon

[Keenan Tims]

Sensing the position of the doors with a pair of microswitches not viable?

--
Sent from my mobile. Please excuse my brevity.

[Keenan Tims]

Oops sorry folks, this got stuck in the mail queue on my phone. I *have* read the thread...

K