Hello all

I have read a lot about my problem and even if not all topics fit exactly to my problem, I have tried all possible solutions and also implement all procedures, unfortunately so far without success. I would not say that I am an expert but I have some years of experience with CNC milling and have built all my machines myself. 1 small milling machine 450x550x200mm travel, 5.5KW ATC spindle for steel and aluminium, large milling machine described below, CNC lathe conversion Emco Compact 8, CNC plasma cutter with water basin. Also don't want to brag here by any means, just want to share that I'm not completely new to the area.

About my problem:
It happens 95% of the time after a tool change, but it has also happened once in the middle of milling that the Y-axis loses the zero point and then continues from a different position and of course destroys the work piece so. It's always just the Y axis. X and Z are always correct. What I can say is that it doesn't happen slowly, it happens all at once but as I said actually always right after the toolchange. The machine runs cleanly through the first part, then toolchange and then the Y-axis is shifted, where and when exactly it happens I cannot say. It is also not about 1-2mm, mostly 25-30mm so very much. I first thought that it might be Mach4, that I did something wrong there (not long Mach4 2-3 Monte but even after I tried Mach3, the same problem. Since I am slowly very frustrated, I hope somebody here can give me a hint, what I can still check or where I may have a mistake. Worked the last 3 years with Mach3 and now Mach4 with all my machines never had a problem, or rather never a problem that the cause was not clear. But now with my newly built milling machine it is. However, same control and same wiring as all other machines.

What I would like to note and also seems a bit strange (guess this has a connection to my problem) to me is that recently, since 2-3 weeks, I have to enter half of the pulse/unit per motor when setting the motors on the Y-axis (2 motors), so that the axis runs at the same speed as X and Z. Thought first that was perhaps a peculiarity of Mach4 but then tried it with Mach3 and on other machines and also there I suddenly had to halve the steps of the Y-axis even I never changed something at that machine. As an example, on the 2 milling machines, 1605/2005 ballscrews with 1/16 steps have a pulse/unit of 640 default on X and Y and 800/640 (1604/1605 ballsrews) on the Z. Now I have to enter 320 on the Y and 640/800 on the others so that the axes run at the same speed. I don't know if anyone knows this problem, but if so I would be interested to know where this comes from.

About the hardware I am using:

PC:
New HP Desktop 600pro, I5 8Gen, 3.3Ghz, Win 10pro (configured as a pure milling computer, purely Windows and everything that is not necessary and disturbs turned off), 24Gbit Ram not connected to the Internet. Is also used only for milling nothing else. Is used for all machines, can simply be snapped to the desired machine and connected by Lankabel. (Since all machines have identical control)

Software:
For construction Fusion 360 with Mach4 postprocessor and Mach4 as control software.
Postprocessor has not been changed except that for toolchange  it also takes the offset of the G43 H... tool, since I work with offsets.

Hardware machine:
4 JMC CL stepper motors 3NM Nema24 (Y axis 2 motors master/slave coupled with toothed belt).
KSU 2005 on X Y and 1605 on Z, X and Y 25mm rails and carriage, preload medium and hard. Z axis 20mm rails and carriage preload medium.
3 Power supplies 24V for stepper motors (but I will change to 36Volt the week) 1 for X and Z, 1 for Y, and one for the BoB and rest.
ESS Smoothstepper with BoB from cnc technics.
Inductive switches for limit and home
Spindle at the machine 2.2KW Chinaspindle G-Penny
Standard suction system with 0.75 KW motor switched via relay
Centrifugal compressor 1.5 KW for vacuum chambers.
2.2KW FU with line filter spindle
1.5KW FU with line filter centrifugal compressor
All cables shielded and grounded.
Machine grounded.
Control cabinet from Hager all grounded.
Supply lines 2x16A fused, 1x for spindle and suction, 1x for power supplies.

Machine structure:
Portal milling machine travel X 1300mm Y 1200mm Z 110mm.
Built primarily for wood, plastic and simple aluminum machining.

Standing mass:
6 chambers vacuum system, control cabinet, frame aluminum, Y axis with motors and rails, suction system, pipe and switching system vacuum system +- 250kg.

Moving mass:
X and Z gantry, spindle, suction head +- 40Kg.

Here is what I have already tried / tested.

• Mechanical test Y-axis, 2x disassembled and reassembled, gantry runs smoothly on the rails, can be pushed back and forth without KUS with the little finger without force over the entire length. With ball screws rotatable with 2 fingers, even if coupled with toothed belt and this is the same over the entire travel length.

• Backlash of the Y-axis x-times repetitively checked and "still none", because new ball srews C5 and 2 ball screw nuts braced against each other (slightly no pressure).
  

• Grounding of the machine, no matter where passage to PE.

• Problem with coupling Y-axis with toothed belt as well as without toothed belt, both motors identical.

• First PC 2 times new set up, then PC replaced by new with enough power and ram and also there with 2 operating systems tried, Windows 7 and Windows 10.

• Mechanical strength of couplings motor/shafts and ball screws portal cheek.

• G code of PP Mach3 and Mach4 nothing conspicuous. (Can't attach the code here where the error came yesterday unfortunately, but can post this tonight if needed).

• Machine settings, tried various settings, error always the same. (Also here can post the settings of the machine, machine.ini tonight if necessary).

• Also still to say, tried to readjust error, small test part 100x100mm, 5 times toolchange , since no problem does not lose the coordinates, only with larger machining operations that go longer.

• Ballscrews and carriage, all completely disassembled, cleaned with isopropanol and lubricated with lithium soap grease.

• Speed and acceleration motors tried different. However, I must say that would be more of a problem if there was a continuous slow displacement of the axis, please correct me if I am wrong. The speeds are always the same on the machines and so far no problem. Normal X Y max 4500mm/min 800m/s Velo. Z 4000mm/min velo 800m/s.

• Still small things tried but without success.

I apologize for the long text, just wanted to try to describe that as accurately as possible so that it can be understood.

Do you have any idea how I could proceed and what I could still try, I would be very happy about help, so that the problem is finally gone and I can work normally again.

Many greetings

Oli

do you have close loop steper? if not ,you have not any way to avoid it because no any way to get alarm when driver cant close the loop
so my first suggest is move to servo( now servo not so high price) if you so strict with the price move to steeper close loop

Hi,
I would guess that you have g0 move immediately after your m6, and your y axis is intermittently losing steps on that move.

What I would like to note and also seems a bit strange (guess this has a connection to my problem) to me is that recently, since 2-3 weeks, I have to enter half of the pulse/unit per motor when setting the motors on the Y-axis (2 motors), so that the axis runs at the same speed as X and Z. Thought first that was perhaps a peculiarity of Mach4 but then tried it with Mach3 and on other machines and also there I suddenly had to halve the steps of the Y-axis even I never changed something at that machine. As an example, on the 2 milling machines, 1605/2005 ballscrews with 1/16 steps have a pulse/unit of 640 default on X and Y and 800/640 (1604/1605 ballsrews) on the Z. Now I have to enter 320 on the Y and 640/800 on the others so that the axes run at the same speed. I don't know if anyone knows this problem, but if so I would be interested to know where this comes from.

This is very strange and is I suspect either the cause or at least a symptom of the fault you describe. Solve this and I'd bet you problem goes away.

You say you have two y axis motors? Are they identical? How have you slaved one to the other?

Can you disconnect each coupler so that you can move them independently. I would start by making sure that you can set the steps/unit in some logical fashion for each motor individually.
I'm wondering if you have a signalling clash or a signal timing issue that is causing onlt every second pulse to be read.

Craig

do you have close loop steper? if not ,you have not any way to avoid it because no any way to get alarm when driver cant close the loop
so my first suggest is move to servo( now servo not so high price) if you so strict with the price move to steeper close loop

Hi, yeah having closed loop stepper motors those I‘ve described in the overview.

Hi,
a 24V power supply for your steppers??? That's bloody hopeless, and 36V is not much better, you should be at least 60VDC if not 80VDC.

Craig

Hi,
I would guess that you have g0 move immediately after your m6, and your y axis is intermittently losing steps on that move.

Quote

What I would like to note and also seems a bit strange (guess this has a connection to my problem) to me is that recently, since 2-3 weeks, I have to enter half of the pulse/unit per motor when setting the motors on the Y-axis (2 motors), so that the axis runs at the same speed as X and Z. Thought first that was perhaps a peculiarity of Mach4 but then tried it with Mach3 and on other machines and also there I suddenly had to halve the steps of the Y-axis even I never changed something at that machine. As an example, on the 2 milling machines, 1605/2005 ballscrews with 1/16 steps have a pulse/unit of 640 default on X and Y and 800/640 (1604/1605 ballsrews) on the Z. Now I have to enter 320 on the Y and 640/800 on the others so that the axes run at the same speed. I don't know if anyone knows this problem, but if so I would be interested to know where this comes from.

This is very strange and is I suspect either the cause or at least a symptom of the fault you describe. Solve this and I'd bet you problem goes away.

You say you have two y axis motors? Are they identical? How have you slaved one to the other?

Can you disconnect each coupler so that you can move them independently. I would start by making sure that you can set the steps/unit in some logical fashion for each motor individually.
I'm wondering if you have a signaling clash or a signal timing issue that is causing onlt every second pulse to be read.

Craig

Hi Craig

Hmm about the G0, will have a look tomorrow morning and will give u a feedback. Will definitely try it, thanks.

About the different steps/unit yes this is strange, more strange is, I took my computer to my small mill and connect the ESS who is built in there and suddenly the same problem there too. I did no change anything at all and now the same with the Y axis, 320 pulses instead of 640😬 I‘m using one desktop to control both machines cause they standing next to each other and have the same setup but both one separat ESS built in.

Yes two motors for the Y axis, one setup as Y the other one as A (Slave of Y), had the option to choose in the ESS plugin to set it as slave (just get the same signal then Y but Mach will not see it as a other motor) or in Mach4 control in mapping. I‘ve use the last option, so in Mach4 under control, motor 1 master Y slave A. And yes all the same motors XYZA.

About the couplers, the idea of an signaling clash or a signal timing issue that is causing only every second pulse to be read sounds like a thing I need to check, but how can that be, if I switch the mill, suddenly the same strange thing happen at the small mill, which I have since 2 years and never had that before. Could it be that something is about using one computer for 2 ESS?
Cheers Oli

Hi,
a 24V power supply for your steppers??? That's bloody hopeless, and 36V is not much better, you should be at least 60VDC if not 80VDC.

Craig

Hi Craig

Got these steppers on this machine, max volt are 50.

https://www.upload.sorotec.de/doku/manuals/DS_iHSS57+iHSS60en_soro.pdf

Hi,
ALL steppers lose torque the faster they go, that's just plain physics. It applies to open loop AND closed loop stepper equally.
Closing the loop does not increase the power, the torque, the speed or anything else the manufacturer tried to BS you on.

There are two broad strategies for reducing the torque degradation with speed; is to minimise the inductance, and your steppers have 2.0mH, which
is quite respectable. The other strategy is to use the highest possible voltage you can, usually limited by the electronics of the driver, in your case 50V.
I would use 50V, or as you are likely to find 48V is more common.

Craig

Hi,
while I can well believe that at 24V your stepper would miss steps, maybe without the torque authority to catch up, but it should at least have faulted as 'Excess Following Error'.

How have you programmed the Following Error Window? How wide is it?

Craig

Hi,
ALL steppers lose torque the faster they go, that's just plain physics. It applies to open loop AND closed loop stepper equally.
Closing the loop does not increase the power, the torque, the speed or anything else the manufacturer tried to BS you on.

There are two broad strategies for reducing the torque degradation with speed; is to minimise the inductance, and your steppers have 2.0mH, which
is quite respectable. The other strategy is to use the highest possible voltage you can, usually limited by the electronics of the driver, in your case 50V.
I would use 50V, or as you are likely to find 48V is more common.

Craig

Hey Craig,

Ok will do it, will order some 48Volt power supply units.