Stepper motor wires size, possible step loss?

[QinX]

Hi All,

I've bought myself a homemade CNC Router locally from someone that makes them for fun, but now that I have the machine in my basement I'm noticing there is plenty of things that could have been made better or aren't up to my expectations, which might have been too high from the beginning.

I'm in the process of making it work the best I can but there is something that is bothering me and I hope to be able to get a hand here.

The machine has got a belt driven X and Y axis with NEMA 23 motors.
I've tuned the motor so that it moves 150mm bang on when I tell it to.

However when I'm trying to run a simple circle or box I get a error of about 0.5mm and it doesn't seem to matter if it's a 10mm or a 50mm size it's always 0.5mm undersized.
measuring between 2 holes it does keep the exact distance. I think it might be some sort of deflection, but when I do a finish pass or even a spring pass it still stays on that 0.5mm undersize.

Before I start chasing ghosts, I do find the following odd.

All the cabling is done using Cat5 cable, even for the stepper motors. And it's about 3-4 meters of it. My basic electronic instinct tells me that can't be right but before I replace all that cabling with a thicker gauge I'd like to hear your opinions.

I'm really new to this so this post might be sparse on actual details.
Please let me know what information you think is necessary to give better advice.

Relevant information:
Nema23 Motors
X and Y are belt driven
Microstepping set to 1/2 (Option for 1,1/2,1/8,1/16)
http://i.imgur.com/nA5VA1P.jpg

Similar Threads:

• 9 wires stepper motor help identify.
• stepper motor with 8 wires
• Need Help!- stepper motor with 4 wires
• autopsy of a 8 wires nema 23 stepper motor

[CitizenOfDreams]

I think you are indeed chasing ghosts. From your description, you do not have a step loss problem, you have a backlash and/or deflection issue.
Cat5 feeding NEMA23 motors is pretty bad, but I don't think it's causing the problem you are seeing.

[QinX]

Thanks for the information, I've tried to check if I have backlash, but since I don't have a dial indicator I'm having to use my calipers and I'm either not able to measure backlash or it's not there.

I'm hesitant to believe it is deflection, if I repeat my finishing pass there doesn't seem to be any more chips coming from the workpiece.

I move my X or Y 150mm over the calipers and going back and forth between 0mm and 150mm is how I've calibrated the steps for the motors.
What I've read to measure backlash is to move the X to -10, then move it to 0mm, go to 150mm and back to 0mm. the amount you are short of reaching 0mm should be the backlash on that axis.
Is this approach correct? Or am I missing something?
If I do it like that I'm not measuring any backlash.

Another thing I've noticed and explored today, my Z seems to drop 0.5mm.
How I went about checking this:
I go all the way up with my Z and go down and touch the the workpiece, I zero my Z.
Then I go up and bounce the Z by jogging it up and down.
I go to my 0 position visually and the DRO says I'm around +0.5mm instead of 0mm on the Z.
If I turn on my router and go to 0 on my DRO I actually start router into the workpiece.

On another note, I will be replacing the wiring for the stepper motors with 1.5mm² wiring, that should be plenty thick.
Just out of curiosity what could happen if I keep using the CAT5 cable for the steppers, outside of the voltage drop and heating of the wires?


Since this issue is probably not stepper related, should I start a new topic in the correct subforum?

[James Newton]

Dropping Z is VERY common on underpowered systems. Add a spring, weight over a pulley, long bungee cord, etc... to counteract the effect of gravity on the Z axis.

If that problem happens only after the machine has been running for a while, check motor heat. Stepper motors CAN lose strength when they get hot enough (as in hot enough to burn) and a simple CPU cooler with some thermal paste can make a world of difference.

If the CAT5 isn't getting warm, it's probably not an issue for the motion, but the cable moving with the motor WILL be an issue. CAT5 isn't designed to move repeatedly, so I fully support your decision to replace it. Extension cords are a fine low cost source of good wire. Use two per motor, and just ground the ground wire... it helps to reduce electrical noise.

If none of that helps, buy better drivers.

As to the repeatable error... isn't that about half the diameter of your cutting bit?

[QinX]

My dad is coming over this weekend and he's dropping of a dial indicator so I can check for backlash better.

The 0.5mm undersized hole happens regardless of bit size, I've tried 4, 6 and 8mm size end mills.

I've looked up the part numbers on the stepper motors, currently I'm running at 1/8 microstepping.

X-Axis and Y-Axis
23HS30-2804S-SG5 (Gear Ratio 5:1 Spur Gearbox High Torque Nema 23 Stepper Motor 23HS30-2804S-SG5)
Gear Ratio 5:1 Spur Gearbox High Torque Nema 23 Stepper Motor 23HS30-2804S-SG5 - US$
Electrical Specification
Motor Type Bipolar Stepper
Step Angle 0.36°
Holding Torque 5Nm
Rated Current/phase 2.8A
Phase Resistance 1.13ohms
Recommended Voltage 24-48V
Inductance 5.4mH±20%(1KHz)

I'm also reading this:
Backlash at No-load <=3°
Might this be related to my problems?


Z-Axis
57BYGH56-401A
NEMA 23 Stepper Motor 12.6Kgcm 1.8Degre 4Leads 56mm 57BYGH56-401A $27.87 Free Shipping @GoodLuckBuy.com
Electrical Specification
Motor Type Bipolar Stepper
Step Angle 1.8°
Holding Torque 175 Oz-in/12.6Kgf.cm/1.24Nm
Rated Current/phase 2.8A
Phase Resistance 0.9ohms
Recommended Voltage 24-48V
Inductance 2.5mH


When you mention that the machine is underpowered, what exactly do you mean by that?

here's an album of the CNC Router.
CNC Router - Album on Imgur

[coherent]

Knowing the motor size/torque is only part of the equation. The size of the machine, weight of the gantry or spindle, quality of the linear rails and bearings, quality of the ball screws/nuts and or rack and pinion (all which affect friction and torque required to make things move) as well as what you plan on using the machine for and required cut speeds are all factors.
Cat 5 is not acceptable for stepper motor cabling if it's the kind used to run long lengths... even for very small steppers. Like James stated it's not made to continuously flex. 18-20 gauge is common on Nema 23 steppers. 22ga will work for lower amperage motors. 4 wire, multi strand (not solid wires) and shielded is best. If the cat 5 is like the flex jumper cable you would use to run between a router and a computer, it may be multi strand and ok for flexing, but is still very light wire for 2.8 amps/48 volts. If it contains solid single strand wires and one wire breaks while the machine is in operation the likelihood of it frying the driver for that axis is almost a given. My guess is if you feel the wire and it gets warm when the machine runs it's too small a gauge of wire.

[James Newton]

When you mention that the machine is underpowered, what exactly do you mean by that?

Z axis dropping. Very common, very clear sign that the machine is underpowered. It can't lift it's own head against the force of gravity. Just like a baby that hasn't developed it's neck muscles. Babies that can't hold there heads up after a month, or that don't have full control of their head position by 6 months should see the doctor. A CNC machine that loses height on the Z axis should be diagnosed and fixed. But unlike a baby, stringing the z-axis up with a bungee / spring / weight over a pulley is an acceptable work around. It makes gravity go away.

As coherent said, if a motor lead breaks during operation, it's VERY likely to kill the driver. I forgot to mention that in my post.

On the repeatable error... I have no idea... but if it's repeatable, you can compensate for it.

[QinX]

@James Newton
Ok so I tried to figure out what is exactly wrong with the Z axis, looking into the fact that you mention it might underpowered.
Well, if I hold the stepper at the coupler en try to jog it, it skips steps, there is very little force required to keep the stepper from moving. You are right, which I'm happy about but also bummed out now.
I can move it while it is holding and when I'm jogging the Z axis, there just seems to be not enough power there to overrule my power.

[A_Camera]

However when I'm trying to run a simple circle or box I get a error of about 0.5mm and it doesn't seem to matter if it's a 10mm or a 50mm size it's always 0.5mm undersized.
measuring between 2 holes it does keep the exact distance. I think it might be some sort of deflection, but when I do a finish pass or even a spring pass it still stays on that 0.5mm undersize.

I think we need more data here. What material is the machine made of? How does it look like? What is the pitch, steps per value? What is the speed per minute and the acceleration? Power supply voltage/currents? Honestly, I am pretty sure you are not losing steps. Losing steps means random size and shape change during the work. If you ALWAYS get squares and circles which are 0.5mm smaller than I simply think that your belts are out of tolerance and your machine should be fine tuned.

This is what I'd do:

Mill a straight track in X and then another in Y direction, say 100.00 mm long, measure them as exact as you can. From that value calculate the exact steps per value and use that in your setup. Some software, like Mach3 and UCCNC, has even a function for this kind of fine tuning, and in my opinion it is necessary to do, especially with a belt driven system. Measure also back lash as accurate as possible and eliminate it if you can. Backlash compensation is not working well in Mach3 and UCCNC in every situation, so it is best if it can be eliminated mechanically, which is not always possible in a belt driven system.

Check every screw and bolt/nut. Check also that the machine is squared. Even if it was squared when the seller sold it, it might get out of square after you installed it on it's final place, or even after some use.

No, Cat5 cable is not right for this type of use, but that is most probably not causing your problems, so don't do anything about that right now.

Set micro stepping to 8. That's a good value and it gives a better precision. You can never get better precision than the steps per value gives you and 1/2 micro stepping means that your stepper turns a full rotation circle after only 400 step pulses, which in my opinion is too rough. Increased microstepping means loss of torque, so it is a compromise between torque/speed/accuracy but I think 8, or even 4 is much better than 2, not only in terms of precision but also in terms of running smoothness.

I don't know if your machine is under powered or not, but again, that conclusion can not be made based on what you posted so far. I am using NEMA23 motors, milling aluminum without issues with the feeds and speeds given by FSWizard, have a heavy Z and a machine with a total weight of over 90kg, so NEMA23 does NOT automatically mean underpowered.

Don't start fixing what is not needed fixing. Identify the problems first. Use the included software to calculate your steps per calues, and experiment with different settings and changes. I found it pretty useful in calculating axis values.

[A_Camera]

@James Newton
Ok so I tried to figure out what is exactly wrong with the Z axis, looking into the fact that you mention it might underpowered.
Well, if I hold the stepper at the coupler en try to jog it, it skips steps, there is very little force required to keep the stepper from moving. You are right, which I'm happy about but also bummed out now.
I can move it while it is holding and when I'm jogging the Z axis, there just seems to be not enough power there to overrule my power.

Once again, that is not automatically a proof of an underpowered machine. It might be a problem with the power supply, or maybe even the driver or the motors or something else. If you can manually move or prevent it from moving my simply holding the motor then something else is wrong. I can't even hold a NEMA17 back if it is running and powered as it supposed to. Don't just change the motors without knowing why.

[A_Camera]

My dad is coming over this weekend and he's dropping of a dial indicator so I can check for backlash better.

The 0.5mm undersized hole happens regardless of bit size, I've tried 4, 6 and 8mm size end mills.

I've looked up the part numbers on the stepper motors, currently I'm running at 1/8 microstepping.

X-Axis and Y-Axis
23HS30-2804S-SG5 (Gear Ratio 5:1 Spur Gearbox High Torque Nema 23 Stepper Motor 23HS30-2804S-SG5)
Gear Ratio 5:1 Spur Gearbox High Torque Nema 23 Stepper Motor 23HS30-2804S-SG5 - US$
Electrical Specification
Motor Type Bipolar Stepper
Step Angle 0.36°
Holding Torque 5Nm
Rated Current/phase 2.8A
Phase Resistance 1.13ohms
Recommended Voltage 24-48V
Inductance 5.4mH±20%(1KHz)

I'm also reading this:
Backlash at No-load <=3°
Might this be related to my problems?

No, I don't think so. BUT... that is a weird motor since it has a gear box in it which is a 5:1 gear box, so it has a mechanical "micro stepping". I'd not set micro stepping in the software to less than 0.5 (1/2) in the software. Maybe it is an idea to try without micro stepping... 1/8 micro stepping on these motors is just too much, in reality it means 1/40 and you basically lose all torque with that setting. Never the less, it is not a motor I'd select for a CNC, especially not for a belt driven machine.

Of course, if that gear box is crap then you will get problems. Since you use this motor on X and Y, it may be so that the accuracy of the gear boxes is such that by not exactly tuning the steps per values, the error you get is always 0.5mm Try tuning to more precise values, as I have described in a previous post.

Z-Axis
57BYGH56-401A
NEMA 23 Stepper Motor 12.6Kgcm 1.8Degre 4Leads 56mm 57BYGH56-401A $27.87 Free Shipping @GoodLuckBuy.com
Electrical Specification
Motor Type Bipolar Stepper
Step Angle 1.8°
Holding Torque 175 Oz-in/12.6Kgf.cm/1.24Nm
Rated Current/phase 2.8A
Phase Resistance 0.9ohms
Recommended Voltage 24-48V
Inductance 2.5mH


When you mention that the machine is underpowered, what exactly do you mean by that?

here's an album of the CNC Router.
CNC Router - Album on Imgur

Your motors are NOT under powered, but should run with the right power supply and not set too high micro stepping value.

One alternative is to remove the gear boxes and use the motor as an ordinary 200 steps per revolution (1.8 degree per step) motor. In fact, that is what I'd do first. Costs nothing and easily done. Now that I had seen this post I am pretty sure that your problems comes from these strange motors with the crappy gear box.

[coherent]

The motor and power supply you list should be fine. The micro steps is normally set on the stepper drive itself if selectable, not via the software. Some drivers have fixed microsteps (like the Gecko G540) most others have DIP switches that allow you to select the microsteps. Also on many drivers also have DIP switches to select the motor amperage/power. Other stepper drivers require inline resistors on each motor cable/connector or have POT that is adjustable. Without knowing the model of stepper drivers you have it's impossible to tell you what to check. It sounds like you may have a power issue... either a bad power supply, incorrect setting or bad connection/wire.

The microstep setting will affect your steps per but has very little affect on motor torque. Your gear boxes are another issue and without seeing one or knowing the model, hard to judge their effect on your machine. Gear boxes change the gear or pulley ratio, they do not and can not change the micro steps of the motors. only the stepper driver can do that. Most CNC belt drives or gear boxes increase torque, although some decrease torque and the movement increase speed of the axis.
I would disconnect/remove any couplers or gear boxes from the stepper motors and test operation and torque. There is no way you should be able you stop a moving nema 23 stepper with your hand. If you can, it's a power issue/setting or motor tuning. Once the steppers are disconnected from each axis, you can also check to ensure all the axis movement is free and smooth and not binding etc.

[James Newton]

Please note: I said the MACHINE is underpowered. NOT that the motors, drivers, power supply or any other specific part was underpowered. If the machine can't lift it's head, it's underpowered. That might be because all the power is being eaten up in crappy mechanicals. It might be that the motors are great but the driver is crap. It might be that the wiring is causing the problem... who knows? But if it can't reliably lift the z axis.... it's underpowered. Ipso freaking facto.

Troubleshooting the problem is up to you. Or you can "cheat" and just counteract gravity on the head, as I suggested.

Also: Why do people buy cheap crap from China and then express disappointment when it turns out to be cheap crap? And yes, I do the same sometimes. I just bought a magnetic disconnect USB C adapter on ebay that isn't USB C, doesn't disconnect if hit from the right angle, and falls out of the USB micro connector because it isn't even made right for that. It comes from China and costs half what it does from a US store? There. Is. A. Reason.

We all need to stop acting surprised about this.

[ger21]

I'm hesitant to believe it is deflection, if I repeat my finishing pass there doesn't seem to be any more chips coming from the workpiece.

Depending on the direction of cut, the tool can be deflected into the woprkpiece, which would cause the results that your seeing.
Cut two identical parts, but with the toolpaths in opposite directions. One will likely be smaller than it's supposed to be, and one will likely be larger.

[A_Camera]

Gear boxes change the gear or pulley ratio, they do not and can not change the micro steps of the motors. only the stepper driver can do that. Most CNC belt drives or gear boxes increase torque, although some decrease torque and the movement increase speed of the axis.

If you have a gear box which gears down the rotation by 5:1 (0.36 degree per step) that is the same as having set the micro step to 1/5 on a motor without gear box, if that would be possible. That's what I meant by the comment: it has a mechanical "micro stepping". The results are the same regarding rotation angle. It changes the necessary micro steps of the motor to get a certain rotation angle per step pulse, and the number of pulses necessary for a full 360 degree turn. For example, if you have a 5:1 gear box and you set the micro stepping to 8 it means that you will need 8 x 5 x 200 = 8000 pulses just for one single rotation of one single axis. Since he is using the same motor on both X and Y that means 16,000 pulses to turn the motor around once. If he is using 5mm pitch then it means 1600 pulses for one single mm move. That's a lot for a stepper system.

So yes, in my opinion, the gear box, even if it is perfect with no errors, MUST be taken into the calculations of micro stepping and indeed changes the micro stepping in that the micro stepping must be reduced.

Like I said, the first I'd do is to remove the gear box and control the motor rotation only through micro stepping settings. There is no benefit in using a gear box, not here.

[QinX]

I'm sorry for not replying any more, I've been tied up with other project so I had to put the CNC on hold.
I'll be trying out all the suggestions mentioned until I've fixed it, but it'll be a slow process.

Depending on the direction of cut, the tool can be deflected into the woprkpiece, which would cause the results that your seeing.
Cut two identical parts, but with the toolpaths in opposite directions. One will likely be smaller than it's supposed to be, and one will likely be larger.

My dad actually thought the same thing so I had tested this, The holes are the same size.

Since I have to give back to dial indicator tomorrow I've had to make some time today to check for the backlash. And it seems both the X and Y have a 0.2mm backlash and having Mach3 compensate for that seems to solve resolve that.

I've also checkout the Z stepper and how it doesn't hold it's position when I try to turn in by hand.
shorting both the A and B pair at the same time results in me not being able to turn the stepper by hand.
Having only the A or B pair shorted I can turn the stepper the same way.
I've measured the wires all the way back into the stepper driver and the solder points on the PCB for the terminals.
Does this mean something is wrong with the driver? Where should I look?

I'm sorry again that I'm not able to respond to all other suggestions, but I will get back to them, I've got them written out in Notepad