Are Open-Loop Steppers Worthless for Accuracy?

This forum is dedicated to those Hobbyists Interested in CNC machining & 3D Printing in their home shops. (Digital Read Outs are also topical, as is CAD/CAM as it relates to CNC)

Moderator: Harold_V

User avatar
SteveHGraham
Posts: 7788
Joined: Sat Jan 17, 2009 7:55 pm
Location: Florida

Re: Are Open-Loop Steppers Worthless for Accuracy?

Post by SteveHGraham »

I am Googling around trying to see why linear is harder to control than rotary. I found out something interesting. Heidenhain uses linear encoders.

Nothing is simple.

This has no practical impact on my efforts, but I'm reading up on it anyway.

Every hard-fried egg began life sunny-side up.
User avatar
SteveHGraham
Posts: 7788
Joined: Sat Jan 17, 2009 7:55 pm
Location: Florida

Re: Are Open-Loop Steppers Worthless for Accuracy?

Post by SteveHGraham »

The interesting thing about this is that a linear encoder tells you where the tool IS, while a rotary encoder just tells you where it SHOULD be.
Every hard-fried egg began life sunny-side up.
User avatar
ctwo
Posts: 2996
Joined: Tue Mar 27, 2012 12:37 pm
Location: Silly Cone Valley

Re: Are Open-Loop Steppers Worthless for Accuracy?

Post by ctwo »

I think rotary encoders still tell you where it is. Like the odometer in your car. Open loop steppers are programmed to go where they SHOULD be, but you don't know if it made it. Adding a rotary encoder tells you where the stepper motor has turned, and will tell you if steps are lost, and therefore compensate, so it ends up where it's supposed to be.

Either feedback mechanism is far better than open loop, but I think it only matters when you are running at the edge, either speed or tolerance. Most open loop systems still seem to work just fine as long as you do not over drive them.
Standards are so important that everyone must have their own...
To measure is to know - Lord Kelvin
Disclaimer: I'm just a guy with a few machines...
John Hasler
Posts: 1852
Joined: Tue Dec 06, 2016 4:05 pm
Location: Elmwood, Wisconsin

Re: Are Open-Loop Steppers Worthless for Accuracy?

Post by John Hasler »

If you use only a linear encoder all the backlash of the leadscrew mechanism is inside the entire loop. It's quite hard for the control loop to deal with this and remain stable. A partial solution is to use inner and outer control loops where the inner loop uses a rotary encoder to control the speed and position of the motor under direction of the outer loop which issues commands to it based on the linear endoder data. This makes motor errors and leadscrew errors seperable. The outer loop can then treat the inner loop as an ideal actuator.

Linear encoders are expensive, though, so you don't use one if you don't need it.
User avatar
Dave_C
Posts: 960
Joined: Mon Apr 21, 2008 10:34 am
Location: Springfield. MO.

Re: Are Open-Loop Steppers Worthless for Accuracy?

Post by Dave_C »

CTWO,

IF (and I say a big if) you have absolutely no backlash in the machine you would have a decent chance of making it work.

So how many of our home machines have "0" backlash in the whole setup?

We are a "home machinist" forum and not the industrial tool makers site so I was speaking in terms of what we might expect from our home machines.

Dave C.
I learn something new every day! Problem is I forget two.
User avatar
ctwo
Posts: 2996
Joined: Tue Mar 27, 2012 12:37 pm
Location: Silly Cone Valley

Re: Are Open-Loop Steppers Worthless for Accuracy?

Post by ctwo »

Steve will find that out very quick, but I was talking about home use. Folks have hacked DRO's into their controls, not likely in industry though. I do not see inexpensive off the shelf systems, however. It seems the whole amateur controller market is a bit of a mess. You have a few choices.

I was just hoping to clarify for Steve that linear encoders do exist, even if they are not practical (cheap/easy) for home use, and that rotary encoders work just fine.

I suspect I have zero backlash ball screws and did consider linear encoders. The machine works fine open loop so I am not going to worry about it. I mean, I expect accuracy ±0.001 and I seem to get it. I just need to go slow and stop breaking cutters.
Standards are so important that everyone must have their own...
To measure is to know - Lord Kelvin
Disclaimer: I'm just a guy with a few machines...
User avatar
Dave_C
Posts: 960
Joined: Mon Apr 21, 2008 10:34 am
Location: Springfield. MO.

Re: Are Open-Loop Steppers Worthless for Accuracy?

Post by Dave_C »

CTWO,

I see where you are coming from and agree completely. I use Mach 3 which can read an encoder but does nothing with the data it reads. Wow, that was helpful...Not! (it only displays it in one of the small DRO readouts)

My machine is pretty tight but I struggle to get a good setting on my tools so that what I program is what I cut. Of course the same issues I had with manual lathe work carries over to CNC. Things like spring for example.

So CNC is not a substitute for good machinist practices. You still need good feeds and speeds, depth of cut and so on.

My solution was to buy motors with encoders already on them. Mach 3 has nothing to do with the encoders, just the motor drives. If the motor drive says go three steps and the motor does not go three steps it faults out and stops. No missed steps to worry about.

Another thing it does is vary the current to the motor. It supplies enough current up to the stall point to try and move the motor the required amount. SO under low load, the current stays fairly low, under heavy load it can go up to the max. I also used a 70 volt power supply so I can max out the motors if needed. So far, no issues and not stalls!

There are some tricks to settings these motors for accel/decel so they don't error out.

Dave C.
I learn something new every day! Problem is I forget two.
User avatar
ctwo
Posts: 2996
Joined: Tue Mar 27, 2012 12:37 pm
Location: Silly Cone Valley

Re: Are Open-Loop Steppers Worthless for Accuracy?

Post by ctwo »

I would move that CNC is not only not a substitute for good machining practices, it is far less forgiving of poor machining practices.
Standards are so important that everyone must have their own...
To measure is to know - Lord Kelvin
Disclaimer: I'm just a guy with a few machines...
User avatar
SteveHGraham
Posts: 7788
Joined: Sat Jan 17, 2009 7:55 pm
Location: Florida

Re: Are Open-Loop Steppers Worthless for Accuracy?

Post by SteveHGraham »

I do hope people have figured out that I am not looking to put linear encoders on my lathe.

The option of motors with encoders doesn't sound bad. I am looking around to see what kind of prices we are talking about. I think my motors are NEMA 34, but they may be NEMA 23.
Every hard-fried egg began life sunny-side up.
Magicniner
Posts: 532
Joined: Thu May 30, 2013 4:40 am

Re: Are Open-Loop Steppers Worthless for Accuracy?

Post by Magicniner »

ctwo wrote:I would move that CNC is not only not a substitute for good machining practices, it is far less forgiving of poor machining practices.
That's not only very true but also not well understood by those who think CNC is an "easy way" to make things with less "value" than manual machining ;-)
User avatar
Bill Shields
Posts: 10459
Joined: Fri Dec 21, 2007 4:57 am
Location: 39.367, -75.765
Contact:

Re: Are Open-Loop Steppers Worthless for Accuracy?

Post by Bill Shields »

THERE IS NO REASON that you cannot use linear encoders for feedback....assuming the control will ACCEPT the feedback.

A position is a position and in fact, most CNC machine manufacturers use linear encoders to determine axis position since they inherently tell you exactly where the table is, not where the screw 'thinks' it is.

Keep in mind that there is a little difference between a DRO and a linear encoder, so please do not get them confused. Most (not all) DROs are just that READ OUTS and do not have 'outputs' that can be fed back to the control.

As for how all of this will work on 'open loop steppers' is another kettle of fish since by definition, an OPEN LOOP STEPPER goes so many steps to attain a position and that is it.

A CLOSED LOOP control looks to where it is going (as communicated by some sort of position encoder), and adjusts accordingly. You do not need to 'map' anything.
Too many things going on to bother listing them.
Post Reply