Changing a servo's angular speed to give a constant lateral spooling speed.

[david]

Hi All,
Not sure if that's a good description.  I made a servo lash-up to spool thread on to a reel with the servo arm (~70mm) moving +/-60 degrees about centre and taking about 10 seconds to travel side to side.  It worked ok but of course the linear angular rotation means the arm spends more time at the sides of the reel and hence the thread builds up.  To correct this I believe I need a Cosine function such that when at the 60 degree extremes it is traveling twice as fast.
I can do this either by increasing the servo step size or by reducing the time delay between steps (probably easier).  Now this widget doesn't need to be precise and I have pondered calculating the reduced time steps but the alternative is a look up table (about 140x Word size variables).  I should mention that I'm currently using an old 12F675 but this is not fixed.
Is there a better way to achieve this function of should I keep going with the above approaches and see how it works out?

Cheers,
David

[Stephen Moss]

If I am understanding correctly the issue is that the servo arm movement is circular and not liner.
If so, then rather than some complex calculation could you not either use a linear actuator or if too expensive just attach a gear instead of an arm to your current actuator spindle which drives a gear rack to covert the circular motion a linear one yourself.   

[RGV250]

I was thinking the same thing as Stephen, this is cheap enough to give it a try.
https://cpc.farnell.com/kitronik/2595/linear-actuator-kit/dp/MK00992

Bob

[david]

The servo is typically constant angular velocity whereas what's needed is constant linear velocity.  A linear drive would be ideal but they do tend to be expensive and pitched mainly at industrial applications. 
A rack and pinion drive is the ideal option but you need to find a gear with a correct spline count to fit the servo, then a rack (and bearing to run in) to match the gear.   The one linked to would likely have considerable gear backlash but something similar with perhaps finer teeth would be a starter.
I'll have a go at the variable time/fixed step size approach and see how it looks.  0 to 60 degrees should only require a 2:1 speed change so not too bad.

Cheers,
David

[See_Mos]

I think a linear drive could be put put together quite cheaply from a few simple parts starting with a stepper motor, some screw thread or long screws and some nuts or threaded spacers. Depending on the span an old CD drive might provide a ready made item.

[JonW]

Might seem daft but what about a windscreen wiper setup?

[david]

I think a linear drive could be put put together quite cheaply from a few simple parts starting with a stepper motor, some screw thread or long screws and some nuts or threaded spacers. Depending on the span an old CD drive might provide a ready made item.

I've used stepper motors before on projects and they sound like an easy option but if the situation arises where the motor can stall it loses it's location so you then need to put limit switches and encoders on it to give it a sense of location. The servo has the position feedback built in along with a hefty gear train making it very convenient but not particularly well suited to this application.
I'm about to try the Cosine variable timing lookup table with a fixed step size so will be tuning for maximum smoke shortly.

Cheers,
David

[david]

Might seem daft but what about a windscreen wiper setup?

Well the magnitude is about right.  The application is for a mate's motorised kontiki fishing reel and ideally the thread needs to be spooled across the drum when lifting about 10kgs.  Speed may be the issue - it requires about a 10 second traverse time.
I may have to revisit these options if the Cosine table bombs out.

Cheers,
David

[RGV250]

Hi,
I might be wrong but surely a windscreen wiper motor will have the same issue as they are motor with a disc turning into linear motion so the speed will not be constant. That is assuming they have not improved in the last 20 years.

Bob

[JonW]

What about a motor with a cam lobe pushing a spring tensioned arm.  Change the shape of the cam to alter the speed and deflection.
Or use the same mechanism that's in a fixed spool fishing reel as the principle is the same and they spool perfectly.  Maybe not the same width but could be modified or enlarged with 3d printed cogs

[david]

Well after much fluffing around I finally have a Cosine timing table running.  It's barely detectable except for about the last 15 degrees either side where it does speed up due to the reduced timing interval.  The code is ugly and would not lend itself to an easy reel speed change which would be a logical requirement.  We will wait for the more powerful servo to arrive in about 10 days and do some field tests (literally) with a bag of bricks.
I appreciate the other ideas submitted but at this early stage we are looking for a simple addition with minimal mechanical changes and want to avoid the need for sand and sea protection.  A waterproof servo ticks a lot of the boxes but not all.  I may be able to get a speed take off from the main reel motor and index the spooling arm to it which would be a major improvement provided the new servo has the required torque.   I'm not yet convinced this is the right approach but it's an easy one however I can see we may need to fall back on more directly driven mechanical systems.

Cheers,
David

[joesaliba]

David,

Saw this late.

If you would like to use something similar to the rack and pinion but using a servo: -

1. Make a slot in the servo arm
2. A pin will pass through the servo arm to the rack
3. When the servo is moving to the ends of the rack, the pin will slide outside the servo arm thus increase speed. When near the middle the pin will slide in, towards the center of the servo arm and decrease the speed the rack is moving

Mechanical, but should work. Adjusting servo speed will automatically change the speed across all rack movement.

I drew something as an example.

Joe

[david]

Great drawing Joe!
I did initially consider a slotted arm but was talked out of it by my mate. Some of the reasons were-
1. Big change of torque between centre and min/max positions.
2. Need to mount the rack and provide bearing.
3. Difficulty in avoiding the rack when free spooling on outward leg.

I think with some minor changes items 2 and 3 could be easily overcome so I think it's worth revisiting.
Many thanks for the reply.

Cheers,
David