[Knut]

Do any of the stepper drivers sold handle encoder feedback? Gecko doesn't mention it. Most pages describing the process seem to think you might as well go with servos and do it right. The new steppers with encoders on them are spendy, and the used servos with the proper gear boxes for my needs (Plasma table with a drilling head and plotter pen) are spendy and not easily found.

JerryFlyGuy mentioned he was finding size 23 gear boxes cheap, not sure if I could match that, or couple up with the right servos either. Would like to buy
direct from a surplus jockey and get a spare right up front.

copper is so expensive, ruining one sheet would pay for some closed loop stuff. With all the talk of tuning and backlash problems with servos, I have retreated to using steppers but would still like to know if it were keeping count. Is there any fix to my concern? Or should I just get some big steppers and forget about it? The basic Vicon table does just that.

Thanks

[mcpltd]

Generally dumb drivers do not have encoder feedback capabilities.
You can get other boards to handle the encoder feeback though.
The only driver i know of that handle encoder feedback are the drivers with the controller built in.
As an example the IMS IM483IE, panther, miclolynx and the Mdrive can handle encoder feedback.

[InspirationTool]

Mariss has said that Gecko is working on such a driver, but there has been no sighting in the wild yet.

The large stepper system seems to work for alot of people.

-Jeff

[bradodarb]

I am building a gantry mill and have decided that I cannot afford a servo system just yet. I found a company that sells a closed loop stepper system for under $1,200
at [URL="http://www.maxnc.com"] I'm not sure if their double stack motors will be enough to power my machine, but I'm certain that it will handle the loads applied by a plasma cutter just fine. Seems to be that the controller, drivers and motors are a complete package. The only downside I can see is that the controller interfaces through the parallel port of your CPU. Probably not the best setup when interpolating 4-5 axis but again would be more than adequate for a 2.5 axis setup.

[ger21]

If you get steppers with encoders, and use Mach3, you can get a board from www.rogersmachine.net that will monitor the position of your steppers, and stop or pause the machine if any errors occur.

As for the maxNC system. For not much more than half the price, you can get 3 Geckos and some 400-600 oz in motors and get 3 times the performance. You lose the closed loop, but should have plenty of power to not have to worry about it.

[Knut]

Thanks for all the tips. Looks like sticking to proven stepper techniques and readily available boards for my first table will make things less problematic.

That Rogers board looks like my only option, not sure how to back up and correct a mistake or miscount in Mach3, but perhaps that will become known once I do some study thereon.

Thanks again.l

[bradodarb]

Gerry, does your proposed nearly half the price solution include a controller/breakout? I have'nt used mach3 or whatever software maxnc is using, but from the screenshots mach3 looks quite a bit more sophisticated.

[Torchhead]

Properly designed stepper systems don't lose steps. The event of a loss of position comes when something gets beyond the ability of the motors. At that point even with encoder feedback the motor can't do any more. The best use of encoders on steppers would be to prevent ruining a part if lost steps do occur (stop the machine) or with more sophisticated electronics to actually slow the motors down to prevent lost steps (rather than adding in more power to try and correct). Slowing the motors down too much with plasma can have really bad cut results.

[kreutz]

Even properly designed stepper systems could lose steps. While the designer will leave a torque reserve on the motor as a safety margin nothing will stop the user from adding more load or speed reaching the limit. Stepper motor drivers are generally open loop, so lost steps will accumulate.
By adding the encoder you close the loop, so the controller is able to detect and correct the position error similar to what a servo does.

You can use the encoder as verification tool, detecting and flagging an error when you lose steps and stopping the system, or you can use a real closed loop controller that corrects the problem and saves your part.

Steppers don't work as servo motors in the sense that you can accelerate the servo to catch up when the positioning error increases, but you can't do the same with the stepper because it will probably stall.

The stepper motor's torque decreases with rotational speed, so the faster it goes the less torque you have available. Using PID in that case is frustrating due to the non linearity of the stepper response. What many closed loop stepper controller designers do instead is accumulate the error and provide those extra steps at the end of the movement.

See http://digital.ni.com/public.nsf/all...0?OpenDocument

[Torchhead]

Perhaps I should have said "properly designed and used stepper systems". I have an open loop stepper system for engraving and light routing that has never lost position unless I did something stupid like ramming into a holddown. Pushing a design past it's limits will result in some form of failure...even in servo's.

Adding in steps at the end of a move is fine for "positional" applications where the load is moving to a given spot. I see this a lot in industrial postioning systems where something has to move to a very precise point but the path it takes is immaterial. In the real world of CNC cutting, mis-position along the toolpath (cutpath) should not be made up somewhere later and the forces that caused the mistake ususally cannot be corrected.

Stepper closed loop operation has little value in toolpath operations unless it has a way to actually lower the speed (all the way to zero if necessary) if the position is outside the error range of the control. At lower speeds steppers have more torque and could possibly power through the issue (perhaps while ruining the cutting tool!). It's the inverse of the typical closed loop.

The analogy is in cruise control of an auto. It works fine if your running straight on the highway but functions poorly if you have a blow out. At that point it needs to do something else besides adding in more gas pedal!

[kreutz]

There is a renewal interest on closed loop stepper control, there are several advantages including low cost and higher torque at low rpm than equivalent size servos, the problem so far has been controller complexity.

Steppers show a tendency to resonate, lose torque and stall in what is called midband instabilities. There are some known techniques to compensate but so far it has been proven complex to design a rugged closed loop control system on similar basis as the servo controllers. Sometimes reducing speed will take the stepper over one of those resonance prone zones. The new interest is in using the called flux-vector control method with or without external encoder feedback in order to overcome those limitations.

New published results are promising, as shown in some published research papers, flux-vector control method allows magnetic field optimization control to compensate for the voltage loss due to BEMF and extend the motor’s speed range almost 100% eliminating midband resonance effects. Hopefully DIY controllers based on this technologies will appear in the next few years.

Google: "CLOSED LOOP CONTROL OF STEPPER MOTORS WITHOUT POSITION SENSOR" and "NEW BENEFITS OF VECTOR CONTROL OF HYBRID STEPPER MOTORS WITH ENCODER" research papers for more information.

[Korellibopper]

Have you checked out how you can daisy chain the PMDX-131 to make a closed loop stepper system. You add a another PMDX board and it has the encoders hooked up to it then you connect that to the PMDX-131's grnd power supply and a couple inputs. Its in the intructions manual you can download them from PMDX.

[automationtechinc]

We have the closed-loop stepper system, pls check here

High-Torque Stepper Motor, Stepper Motor, Driver, Stepper Motor kit, DC Servo Motor, DC Servo Motor kit, Stepper Motor Power Supply, CNC Router, Spindle, and other Components. Automation Technology Inc

High-Torque Stepper Motor, Stepper Motor, Driver, Stepper Motor kit, DC Servo Motor, DC Servo Motor kit, Stepper Motor Power Supply, CNC Router, Spindle, and other Components. Automation Technology Inc

[awerby]

That looks interesting, but it doesn't seem to be available for another month or so. When it is, how will it deal with a servo error? Will it compensate for it by adding steps, or simply shut down the system? What CNC control software does it use - Mach3, EMC or something else?

Andrew Werby
ComputerSculpture.com — Home Page for Discount Hardware & Software

[bradodarb]

Hello,

I have been using a kflop (Dynomotion | Motion Control Boards) for quite some time with excellent results over a broad range of hardware.

It supports true closed loop stepper systems(as well as many other motor configurations). Rather than just stopping/pausing on an error it will adjust the pulse train(adding steps or reversing dir and adding steps) to maintain a closed loop position.

In the forums I have seen people use anything from resolvers, economy rotary encoders as well as glass scales for feedback.

Check out the CNCzone forum here::

Dynomotion/Kflop/Kanalog - CNCzone.com-The Largest Machinist Community on the net!

-Brad Murry

[Torchhead]

How does ANY stepper based "closed loop" do any meaningful position correction without increasing torque? Steppers run with fixed torque because they are current limited per step; just adding or subtracting steps will not help a stepper that has reached the edge of stall (which causes most lost steps and lost position) It already is getting more steps than it can process. The motor simply cannot move at the rate at the given load. Physics dictates that to get a stepper out of stall, you must:
a: Decrease the load (amount of torque needed)
b: Decrease the RPM to roll the motor back inside it;s RPM/torque curve.

a: is typically not possible in most motion control situations and b: requires a CNC application that has coordinated multiaxis moves and where ALL axis be rolled back at the same time and rate.

I can see if you are moving a single motor from point A to Point B and the only important thing was it arrives at point B exactly, than adding in steps would make sense even if it comes at the end. I have seen that used in industrial positioning machines where only the final position is important Trying to make real time position corrections along a toolpath that are coordinated with the other axis becomes much more difficult.

So if you have a controller that handles all of the axis, and does it's own trajectory planning you could (in theory ) buffer the pulses or modify the toolpath velocity on the fly but you cannot do it on an individual axis.

The motors in question appear NOT to be steppers since the signals are 3 phase and the encoder is internal and appears to be used to give shaft position and commutation information back to the controller, It sounds more like a DC brushless servo (BLDC) than a stepper (which steps the coils and has motor detents of typically 200 positions per rev. The torque curves are interesting and look more like steppers on the low end and more like a DC motor on the top end.


TOMcaudle
www.CandCNC.com

[BobWarfield]

Tom, loads do vary considerably throughout the course of running a program. Corners, for example, have far more load that peaks up almost instantaneously and then goes down again very quickly once the cutter has "made" the corner.

Another example where this is the case is overcoming "stick-slip" with conventional ways. Getting the axis moving (especially the Z in the "up" direction on a mill with a heavy spindle like an RF45) is painful, but once it is in motion the coefficient of friction goes way down.

The question is whether a slight slowdown could result in the stepper getting past the stall and catching up to where it should be again without exceeding the max following error for the closed loop. I don't have a hard time believing it could in many circumstances. If it couldn't, you wouldn't lose steps so much as coming to a dead stop at the first sign of trouble.

Cheers,

BW

[Torchhead]

Bob
I understand your scenario and I do agree loads do change a lot over the course of a cut but it's hard to get past the physics of how a stepper operates. First let me state that a properly engineered stepper system when operated inside the torque curve of the motor does not lose steps. You can present the system with a condition that asks for too much speed at too high a torque either from acceleration or back forces from the load, and cause the motors to stall. Usually that is not a real subtle condition because when the motor starts losing steps even continuing to send the normal steps cause the motor to collapse into a full stall.


The best analogy I can use is an airplane. All pilots know that most conventional airplanes will stall if you increase their angle of attack at a given power setting. Try to rise the nose too high and climb too quickly and the air over the wings ceases to flow fast enough to provide lift. The recovery from a stall is to either increase power or to drop the angle of attack.

The stepper is a "fixed power" device. Each pulse is limited in current so it simply cannot increase torque past it's point on the curve. So no more power left in the throttle. Only thing left to do is drop the RPM

I find that lost steps in a CNC machine are not just a few here or there and are not something that are correctable in the time frame of staying on the toolpath. So we are left with the consolation that the system will stop (if designed right....not just the one axis but all of them), and save the job.

I cannot address the value of a system that corrects for "minor" (occasional ) lost steps for the condition where the loss is only momentary and the motor self-recovers. I can address the value of a system that will stop motion and save an expensive part and the material and it has merit on some machines.

When writing about table physics and proper design and engineering I stress that you design around running steppers in the first 50% of their torque curve for normal cutting (max rates) and not push the machine into marginal areas where a stall can occur.

It all changes for servos because the motor torque curve is pretty flat and linear and they can (and Do) increase their torque (up to the peak or the limit of the drive)

My comment on the DYnomotion was in the context of a typical build using control software like MACH3 or LinuxCNC (EMC2). The Dynomotion is a complete motion control that uses their software and their motor drives and can sense the loss of position and make dynamic corrections and IF the software does it right it will slow down the feedrate before it starts adding in steps. It is not something that would work with the subject of this thread being the new "stepper".

We provide both stepper and servo controllers so I have no agenda to push one technology over another. We listen to what the customer is trying to build and
help them make valid decisions based on there goals.

We have been working on a true stepper closed loop solution but I am not ready to announce anything just yet. Things on paper have a way of changing when they go to proto and real world stages!.

TOMcaudle
www.candcnc.com

[BobWarfield]

Tom, the airplane stall analogy is an apt one for the corner example. Just as the plane's nose drops after a stall and eventually it recovers, so too the corner. The stepper or servo won't drive it fast enough in the corner, but it advances a bit, and the cutter clears the material at the slower speed which lets it move ahead again and repeat. Eventually we're through the corner, the question is whether the path that was described is of acceptable accuracy to avoid a fault, and whether a follow-on finish pass can clean it up to even better accuracy.

I hear the statement a lot that if we operate the stepper within its design parameters it won't lose steps. It always comes out sounding like, "If we don't place the stepper in a situation where it could lose steps, it won't lose steps." LOL, I don't get a lot of warm fuzzies from hearing it.

I do know one thing: my servo-based CNC's fault relatively often if I have them tuned aggressively for acceleration or speeds. By often I mean once or twice in a day of shop use at the aggressive tunings versus maybe once a month with more conservative tunings. That once a month has more to do with me forgetting to pump the one shot than anything.

I also frequently find spring passes and finish passes to be critical to accuracy, which tells me there is a finite following error fairly often that just doesn't happen to trip the servo fault.

Meanwhile, my stepper-based systems just lose the steps and keep on going. I have heard this many times from other machinists. The typical comment is that by the end of the day they're scrapping a lot of stepper cut parts if they don't frequently rehome their machines whereas the servo machines just keep on humming.

Every time I see my servo system fault, I always stop and think about how often the exact same thing is happening on my steppers and I never even knew it until I went to measure the part and wondered why it was so far off. It's just not that obvious when you lose a step or two here or there, and it does happen more often than you'd think.

There's at least two other active threads where guys are running the Z-axis up and down and discovering they're losing steps, mostly from the "up" direction. It's no big surprise--the head is heavy and fighting gravity going up. Losing that step is a nasty one. It means your next cut is that many steps lower than expected. Therefore, it is taking more force to perform the cut, so we may lose yet more steps. It can be a vicious cycle.

Between the two, I'd recommend the steppers for beginners. They're just way easier to setup and a bit cheaper. You can do awesome work with steppers. I'd be sure to rehome the machine frequently to make sure any lost steps are cleaned up. As you notice you're getting some during the homing process, you'll know more about what to avoid in future. Eventually, you'll run inside those design parameters and will seldom lose the steps. It does take a little experience though.

If you want ultimate performance, love the servos.

Cheers,

BW

PS Not sure any of that is applicable to a plasma table. Perhaps if there is reason to believe some parts of the travels have extreme friction relative to others? But the "cut" doesn't have the force feedback issues a mill would.

[hoikay]

There are some new products called hybrid servo motors ,they become more and more popular in stepper motor market, it combines the advantages of both servo system and open loop stepper, and it's cheaper than servo system,one of the brand name is Leadshine.