[mattbesquare]

I am correcting the problem with pushing my steppers out of their "comfort range" and losing steps.

But I have research funds coming up again and am wondering about encoders (the reason I got dual shaft steppers). I am increasing the moment of inertia on my axises to help decrease movement per cut , and smoothing gibs, considering a linear bearing set up, but that introduces a new set of problems. But I am having personal issues trusting 6 hour run times. The goal is extreme precision, not speed persay.

So an intermediary controller between mach3 communicating with encoders and sending step information would be a simple task with my previous experience and I have all the hardware on hand except the encoders. Anyone see a good deal out there?

To make the coding easier and price lower 200 ppr absolutes would be perfect, or any factor of 200 please (would really make my algorithms easier).

Thanks everyone, sorry been so quiet, when I lack funds and am not machining, reading of others awesome is hard.

Going with incremental encoders I could run a gcode preprocessor to insert code modules that zero out the head position at given command increments.. that should be a very good solution for extended run accuracies...


Koyo Series E2 decent? 50.00 each I am still confused how I will handle a stop right at a ttl channel hig/low change, hoping the 1.8 degree movement will keep my encoders well withing their high/low settings wave

[Iron-Man]

I am currently researching the AMT203. Digi-Key does not have stock at this time, but shows $ 63.18 for part 102-2050-ND.
{ AMT203-V modular encoder kit, outstanding tool for experimentation and familiarization with the AMT203 features. }
http://www.amtencoder.com/Product

Hope this helps,

Iron-Man

[HawkJET]

I am correcting the problem with pushing my steppers out of their "comfort range" and losing steps.
[snip]
The goal is extreme precision, not speed persay.

Then the obvious is to slow things down and get the steppers back into their "comfort range" as you put it.

If you use steppers, then you are already confined to this. If you want some redundancy, you could use the step verification within Mach3.

Extreme precision is not achieved by what you suggest. If you want true closed loop, get true closed loop! Don't use a band-aid.

There is nothing inherently wrong with a stepper motor system when designed correctly. You just need to stay within the capabilities of that system. If you are loosing steps, then by definition, you are exceeding the capabilities of the system.

[The Blight]

http://www.evarobotics.com/_product_...evelopment_Kit

There are Stepper Servo systems out there with high precision positioning. Noticed it over in the Linear and Rotary motion forum. Mariss from Gecko is also developing something similar. I have seen a few other too, but don't have the links. A search on google will turn something up, but they cost quite a bit.

Why settle for the limitations of a stepper motor in open loop, when you can close the loop and get better performance? If you are just going to correct or check for missed steps, then it doesn't give you much advantage, except maybe some peace knowing your machine wont ruin itself if it loses its position (many reasons this could happen. Collision is one of them).

[HawkJET]

There are Stepper Servo systems out there

A stepper motor can be used as a brushless DC motor (Back in the '80s, I designed a large stepper motor with 150 ft-lbs of torque and it turned out that controlling it as a brushless DC motor worked better for the application). In that case it is a stepper in name only. It is no longer controlled in a fashion that has any resemblance to a stepper and therefore it no longer has the advantages (or disadvantages).

[The Blight]

So a stepper motor, when used for something else is not a stepper motor any more? Whats your point?

A stepper motor can be used in a closed loop as a stepper motor too.

[HawkJET]

So a stepper motor, when used for something else is not a stepper motor any more? Whats your point?

An ordinary brushless DC motor can be controlled in the same way as a stepper motor and it will step like a stepper motor. So what is a stepper motor?????

An very many ways, a motor is a motor is a motor. A stepper motor can be commutated just like a BLDC motor so you could say it is a BLDC motor. You could even make a mechanical commutator for a stepper motor and turn it into a brushed DC motor. So YES, when a stepper motor is used for something else it is something else - what is it that makes something a stepper motor anyway?

A stepper motor can be used in a closed loop as a stepper motor too.

Yes, but there is a difference between merely closing the loop and actually having a servo system. A closed loop stepper system has limitations - it is essentially a step verification system. When you make it a servo system, those limitations are gone and it is no longer a stepper system.

I suspect we are caught up in semantics and definitions. I have designed both stepper motors and brushless DC motors. There are features that one could say defines them as one or the other - but not necessarily....

So what is it that makes a stepper motor a stepper motor anyway????

[Al_The_Man]

A stepper motor can be used in a closed loop as a stepper motor too.

But has it been used in a true PID loop as a servo is?
I would think if you are considering trying to go closed loop via an encoder, there is not much more to make it true closed servo loop, IMO.
Al.

[HawkJET]

there is not much more to make it true closed servo loop, IMO.

Al, I disagree here. The nature of the way a stepper motor is controlled does not lend itself to a servo system. To get there, you essentially throw away the stepper controller and make it a BLDC controller. Once you have that, now you can add a servo controller.

Basically the only thing in common would be the motor and encoder. All the other electronics is different.

[The Blight]

You are right about that JET.

But I still see it as a big plus, knowing where your machine is, even if it does something stupid.

[Al_The_Man]

Al, I disagree here. The nature of the way a stepper motor is controlled does not lend itself to a servo system.

Actually that was the point I was trying to make.
I should have said servo motor loop, and not just servo.
Al.

[HawkJET]

I still see it as a big plus, knowing where your machine is, even if it does something stupid.

I agree! My apologies if I have implied that step verification is a bad thing. My comments are about what the OP offered as a problem and his solution.

If he wants extreme precision (his words) rather than speed, he is much better off to make sure he does NOT loose steps in the first place. Loosing steps and discovering it sometime later will allow for an error for that period of time. I would not consider that extreme precision.

With that said, it is certainly of value to KNOW that you have lost steps so you can take corrective action. Mach3 allows you to do this. I have been considering doing it myself on my Boss 3.

The mechanical DRO on the X & Y axes of the Boss are really nice to have. I basically use them as kind of a step verification now. I wish I had something for the Z axis

[The Blight]

I had no idea Mach3 had this feature. Never used Mach3, so I can't test it. Anyone tried it?

[Andre' B]

So what is it that makes a stepper motor a stepper motor anyway????

A quality motor, brushed or brushless is designed with as little DC resistance as possible. So that when you feed the motor a voltage of X it will run at a known speed and pull as much current as needed to do it. The controller needs to be smart enough to protect the motor from melting down.

The way I see it is a stepper motor/system is intended to provide an inexpensive position/motion control.
To do that something needs to give, (the old fast, cheap, or reliable, pick 2).

So they build in a little DC resistance which lets the motor be stepped and a position can be held with a simple pulse without the current getting to high when there is no back EMF being generated.
It also means for a fixed voltage the torque drops as the speed goes up, and it runs hot, etc..

A more complicated stepper controllers chop the power up to control the current so that the DC resistance can be smaller and you can go faster. At this point you are on your way to a brushless DC motor.

[HawkJET]

Motor design is beyond the scope of this thread. There is a lot missing in your apparent understanding of motor design. What you have posted contains many inaccuracies and assumptions.

[Iron-Man]

I am currently building my own servo system, completely from scratch but many still use stepper driven systems. I am thinking about designing a system around the TMC457 from Trinimac. The TMC475 has built in PID and encoder interface. From what I can tell from the spec sheets, this device would make an excellent stepper motor motion control system. It even has step-dir inputs. The only thing needed would be the proper power (driver) section for cnc. Here is a link to the site http://www.trinamic.com/tmc/render.php?sess_pid=444
An evaluation board is also available.

Hope this helps,

Iron-Man