[loska]

I just solved a problem that has been plaguing me for nearly 2 years so I thought I would share!

My 6090 router from Jinan iTech has been slowly losing position in all axes, especially on complex jobs. This is most obvious in circles and arcs being non-concentric, and the Z height slowly dropping when doing engraving work.

The solution was to set a jumper in each of the M542 stepper drivers to change the pulse logic to negative. The jumper is marked 'J1' and is in a little bank of 6 header pins, visible only by unplugging the driver then removing the plastic cover.

Now it is fixed, I am pretty impressed by the accuracy and versatility of the machine - from ripping up slabs of plywood to engraving tiny details on PCBs.


Background -

The machine runs on a Weihong NK105G3 DSP, with 3 x Leadshine M542 stepper drivers. I use mainly ArtCam and also Eagle plus PCB-Gcode.

I first noticed this effect particularly when using 'area clear' toolpaths on holes.

I managed to contain the problem by switching to a finer micro-stepping resolution and avoiding area clear toolpaths, good enough for most work.

But - I have been running some fine PCB work recently and accumulated errors of near 1mm were terminal!

The error was repeatable - each time I ran the same job I would get the same result - but otherwise confounding as there was no obvious skipping, backlash or other failures I could see.

Returning the machine to relative zero, would end up in a different position to where it started, even though the readout showed 0.000, 0.000.


The technical bit -

The DSP controls the stepper drivers through 2 pairs of control wires sending 'step' and 'direction'.

There are a few different schemes but one variation is 'pulse logic' which can be positive, meaning that each step occurs as the voltage rises from 0 to 5V; or negative, meaning each step occurs as the voltage falls from 5V to 0.

If you look in the NK105 manual on page 83 there is a diagram showing this - the section relates to a servo driver but this is a clue as to what the NK105 is outputting (note the manual says to choose # 2, single pulse stream negative logic).

The M542s on the other hand are set to positive logic by default, at least mine came that way. Page 3 of the M542 manual sets this out.

The repercussion of this is that every time there is a direction change, one micro-step gets counted in the wrong direction. On a complex job, this adds up to a visible error.

Those area clears involved lots of little concentric circles and therefore lots of direction changes!

[lael]

Hi,

We are looking at getting a CNC with a Weihong NK105G3 DSP controller. Do you have a post processor for it that might work with Inventor HSM? do you know if there is one available?

[MarryU]

Hi,
Generally NK105 G2 or G3 are DSP offline terminal used to control each axis,not the software for making G code,so it has no post processor acturally. You need to look for post processor of HSM.

Mary

[lael]

Thanks Mary, but Inventor / HSM needs a post-processor to work with NK105 G2 or G3 right? Do you work with Weihong?

[MarryU]

You welcome Lael, yes I confirm with their technicion,they dont have.

[Archil]

Hi
I have exactly the same problem with Weihong NK105 and Leadshine MA860H
Can you please tell me how exactly should I change "J1" jumper pulse logic?
Thank you in advance

[bigz1]

I am having problems on the Y axis drifting(see attached). Can you show how you cured the problem?

Mine is Weihong NK105G2. I having a problem with the spindle gears. When turning on the spindle the the inverter reads the following when cycling through the gears:-

Gear Inverter
0 0
1 150
2 320
3 470
4 0
5 150
6 320
7 470Attachment 361974

[falacncrouter]

There is a little different between Weihong NK105 G2 and Weihong NK105 G3. I think you would better to ask CNC Router manufacturer from which you bought it.

[loska]

Hi Archil

Well it has been a year or so but I just found your message. Did you ever solve this problem?

Looking at the ma860h manual it looks like you need to add a jumper to J3 in your case to achieve the same setting change (active at falling edge)

See page 4 http://www.leadshine.com/UploadFile/Down/MA860Hm.pdf

Hi
I have exactly the same problem with Weihong NK105 and Leadshine MA860H
Can you please tell me how exactly should I change "J1" jumper pulse logic?
Thank you in advance

[loska]

Hi bigz

Well better a year later than never did you ever solve your problems? The drift you show here is a lot more severe than what I was experiencing so it might be a different issue,

The spindle gear problem however - I also had but fixed. The cause is that the 3 wires from the NK105 to the vfd for speed control were connected incorrectly. These are marked S_0, S_1 and S_2 and need to connect to corresponding lines on the VFD (markings will depend on your model of VFD).

I hope this gives you a clue!

I am having problems on the Y axis drifting(see attached). Can you show how you cured the problem?

Mine is Weihong NK105G2. I having a problem with the spindle gears. When turning on the spindle the the inverter reads the following when cycling through the gears:-

Gear Inverter
0 0
1 150
2 320
3 470
4 0
5 150
6 320
7 470Attachment 361974

[bigz1]

Hi Loska

Thanks I fixed the drift with your method. I will have a look at the inverter thanks.

Have you had any luck wih 4th axis?

Happy New Years