help my 6040 yoocnc is losing steps

[messy modeler]

Hi.

i am a new newbie despite trying for over a year to work out my cnc 6040 machine. i started with motor issues later progressing onto software issues, thankfully i think have most of them licked but still im having problems.

is there anyone out there can can help me?

i have tuned (i think) my machine and the stepper motors are set to different values for x and y. not the 320 in the manual. i was pushing the machine too fast, as recommended again in the manual but that just made the axis stall. A lot!
finally after lotsa research and head scratching i slowed the feed rate down.
that cured the stalling.

its like the machine keeps on losing steps and not returning to x0,y0
i was getting oval circles and the 'tuning' helped but i dont think its enough.

im not sure what else to say so if anyone has any ideas please ask me the questions that may get to the bottom of what is going wrong.

i have tried many things right down to taking the machine apart, reading every scrap i can about step loss, pouring over the manual (i found different ones for the same machine that say different things (hows that possible?))
but to no avail.

i would really appreciate some help, thanks

[KOC62]

Well, I don't know how much help I can be since I don't have a running CNC. Secondly, I don't know your system.

First I would make sure all my axis can move smoothly without binding when turning the lead screw by hand. (One must eliminate the obvious.)

I would ensure that my motors are correctly connected to the drivers as per driver manual. Also ensure that you have a good working power supply that can handle your motor loads. Also check for any lose or poor connections to your motors and power supply.

Ensure that your PC-driver interface is properly working. Know what the pulse requirements are for the STEP/DIR signals. Then ensure that your software is correctly set according to whether the pulse is inverted or non-inverting and of the correct pulse width.

In Air Test:
Then I would jog one axis a given distance slowly and measure whether it did, preferrably close to the axis limits. e.g. if your axis can travel 25" then try for 20". If this axis seems normal then do the same test of the same axis faster and faster, until you reach your designed jogging speed. Does it lose steps? When? Maybe its an acceleration setting problem? Motor resonance?

Repeat a similar test for each axis one at a time. We want to know if each axis misses steps or only one axis. But don't change too many things at once since we are troubleshooting.

If no steps are missing during jogging then try a sample cut. We are trying to determine under what conditions you lose steps or if another problem exists that looks like lost steps.

[messy modeler]

Thanks.
I have tried most all of what you recommend but I feel my machine has different issues.
The x and y stepper motors are the same but have different step setting values.
X = 320
Y = 331.857 (still not perfect though)

This is the best setting I have found.
Still draws slightly small in y axis.

Machine never returns to 0,0 always slightly off.

Pc is in epp mode.
Kernel is 25k
Drive is 10
Pulse is 5

The more moves the machine does the more "out" it is.

I cannot understand why the step values would be different.
The ball screws are the same. 5 teeth per inch.

I'm not sure about micro stepping or how aNd where to chnange it.
I've read that the dip switches on the driver board does his but I'm not prepared to change it without advice.

I found the vaue for the y axis by doing a calibration test from in mach3 asking the machine to move a certain amount and telling Mach the actual amount.

If I rehome the spindle after a cut the machine will reliably recut the same path with the same discrepancy in size.

I have a simpleish rose as a test path and can redraw repeatable paths only if I reset the 0,0 after each run.

Like I said I'm kinda at a loss as to what to check next.

Thanks again.

[KOC62]

To me it sounds like your calibration is still off. This could explain the accumulative error of many small moves.

Can you assess whether the error is precisely the same (i.e. a fixed percentage) or whether it is random? You didn't indicate how you did the calibration and what error(s) you got.

If I ask the machine to the move the Y-axis exactly 1 inch and then repeat that command 9 more times I would expect that the Y-axis moved a total of 10 inches. But if it moved only .9 inches per command due to a calibration error, then after the 9 more moves it will have moved 9 inches, with a total error of 1 inch. In this scenario the error is a fixed percentage. A random error could be due to backlash.

How is 0,0 determined on your machine? I could see a "home" position procedure being set by limit switches, meaning it doesn't require absolute co-ordinate numbers. But I'm assuming that your 0,0 is some distance from the last position used. Hence a calibration error will cause your machine not to return to the actual 0,0 position relative to the last position used.

I'm not sure how people do their calibration. I will be into that procedure sometime this spring when the weather warms up. I would most likely do the biggest rectangle my table can do, without hitting the limit switches and one axis at a time, and measure as precisely as possible. Then I would do a diagonal on this rectangle to see if the axes are perpendicular. Also, this diagonal is the longest line my table could do, while still being a straight line, and thus easier to measure.

[messy modeler]

Thanks

I'm sure the calibration us still off too....

There are no limit switches on my machine.
You set 0,0 manually, ie find somewhere on the bed and reset mach3 values to 0,0

I will try the test you recommend and move in one direction in incramental steps.

The test I did is from within the mach3 screens for settings.
Labeled axis calibration.
Mach3 asks you how far you would like to move.
Ie 100 mm
The machine moves what it thinks is that amount.
You manually measure the distance travelled.
Then you input that distance into the waiting screen and mach3 recalculats the step value.

I think that's right.
It certainly made the movement better but it's still not good.

You mentioned backlash but I was lead to believe that my type of screw and bearings don't allow for backlash. I could be wrong. I'm in no way an expert.

How could I test for it?

I have a dial test indicator.
I have made a fairly good guess on how to use it but after lots of searching I cannot find anything that actually tells me how to use one.

I place the dial against my spindle mount.
I then use the MDI to move the axis in one direction.
The dial shows a little short of the inputted amount.
Ie I move 10mm
The machine / dial shows a little short...like about 9.8
I move the machine back 10mm
The machine sometimes returns to 0 sometimes returns short.
If I repeatedly move from 0 to 10 and back the results are increasingly off in one direction.
Ie the fault or loss only goes up.
Eventually the machine could be 1 or 2 mm off 0

From what I have tried I still think the machine could be tuned better.
I am totally confused by why the two identical stepper motors are set to different steps.

I will try to substitute the over and see if the step setting would need to change.
If the stepper is bust then when I change it to a new axis the fault would go with it.
That's my guess anyway.

[nikolatesla]

Yep I had this exact problem.

You have a faulty axis driver board

Swap x and y boards and you will see it changes .
What you have is one microstep not functioning in the driver chip

You also have a lot of other problems

wires are all crap, replace them now before you have the problem of axis heading south

Fit home switches because you will be losing steps depending on the job

Slide rail bearings need adjusting (tightening)

It might be worth trying to set your boards to full steps(switches 1 and 2 off) for repeatability on fast complex moves, slow your z axis way down

What you have is some good parts and some bad parts that look like a working machine .. Good luck and be patient .. eventually you will have a good machine

Lindsay

[messy modeler]

Hi Lindsey.

Thanks for your input.
I will try as soon as I'm able.

Do you know all the settings for the little red dip switch ?
I cannot find them anywhere.

I have heard that the cable is not very good b4, is it possible to buy an upgraded loom?

The side rail bearings were very loose but when I tightened them the machine became very very noisy so I backed them off again. Do you think I should tighten them up and if so how much? Really tight or just nipped up?

Thanks again for your help.
I will post any developments as soon as I can.

[messy modeler]

Update.

I tried Lindsey's dip switch settings.
Not good.
Axis moved very quickly and stalled repeatedly when it did.
Horrible screechy noises.

I'm guessing that if the dip switches are changed then something in mach3 has to change too.
1&2 were switched to off. 3 is on & 4 is off

The original settings had 1&2 on.

I am very interested in the dip switch settings.
If anyone knows what they correspond to, please enlighten me.

[nikolatesla]

Update.

I tried Lindsey's dip switch settings.
Not good.
Axis moved very quickly and stalled repeatedly when it did.
Horrible screechy noises.

I'm guessing that if the dip switches are changed then something in mach3 has to change too.
1&2 were switched to off. 3 is on & 4 is off

The original settings had 1&2 on.

I am very interested in the dip switch settings.
If anyone knows what they correspond to, please enlighten me.

Sorry it has been a few weeks since i set them but you will have to change mach3 as well ..you need to get your head around the step vs acceleration vs movement .

Just swap you boards and confirm the fault . Fix it and start from there

I would order a driver chip the big one on th board now because I dont know where to buy the boards ..or better still just get a geko540 . The controllers are not well put together .

Check your slide rail bearings for play mine had .5 mm play ..there is an adjustment key ,have a dig. you need to strip and rebuild any way as you will find out .

The wire I used was 240volt twin flex light cord Just make sure it is flexible multistrand and NOT in a rigid shell like the stupid chinese ones.

The chinese ones lasted just long enough to do a few carning tests .. in time to stuff up the first JOB..

There machines are a challenge to get right

[messy modeler]

thanks

do you know what the settings are for the dip switches?

i am cutting out some good parts but now im noticing that my z is getting lower and lower as the machine makes more cuts.
i am cutting patterns out using multipass and finding that each successive pattern gets deeper.
to the point that the bit cuts the scrap board below my material.

the outline sizes look better but im only cutting relativly small parts 90mm x 90mm so any incrimantal outs are minimal and hard to measure.

when i check the path screen on mach it looks good.

i think its still the driver boards but have little money to change them so have to work with what i have i guess.

the dip switches could be the way, ie full steps
but without knowing what to change on mach3 im a bit lost.

thanks for the help

[messy modeler]

UPDATE

new worries...

i have been a little more successful lately with cutting out flat patterns but my new problem that ive just noticed is that as the cuts go on the z axis now gets progressively deeper into the material.
im only talking about .1mm but when yr cutting 1mm after ten or so ups and downs the depth starts cutting the bed and gets deeper as the run continues.
i have been able to turn the z motor up ie twist on the thumb wheel when its resting.

checking the steps per move calibration sets the steps at 327ish instead of the 320 stated in the manual. for what thats worth.

i am at a loss to find where my machine is losing steps.

arrrrggggggg.......

[yoocnc]

Although I'm not sure you buy the 6040 What is the version,
Because different dealers have different requirements.
But I think your problem comes from the interference, rather than stepping drive.
Even the cheapest drive YOOCNC production is to meet the design requirements.
In my opinion you do not need to step drive of a few hundred dollars to replace the drive of a few dollars,
That only a handful of their own ideas.
Engraving machine is just a simple processing equipment,
Its performance by the mechanical parts of the limit
Change expensive drive only to let the machine run faster.
But in fact there is no obvious advantage
Because you are carving can not run too fast.
The spindle motor power limitation of its processing speed.

Try
Installed in front of the drive filter (250V10A),
Ring installed in the inverter output section.

If you buy a logo machine YOOCNC use have any questions,
You can contact directly with yoocnc@foxmail.com access to after-sales service.

My English is poor， ridiculous.
Maybe the GOOGLE translation tool allows you to see to understand.