Y axis stalling

[richie00boy]

I've got a 3020T and it seems fine apart from a problem with the y axis which seems to come on after the mill has been running for a while. It stalls like if you have set the feed rate too high and looses track of where it is, i.e. end up being short of where it should have got to. I've taken a video of it, this was the worst it has ever been at this point.

http://www.readresearch.co.uk/personal/mycncuk/mill_stalling.avi

I've got the motor and coupling close and centralised OK, the y axis is tighter than the others, but this is to be expected as it's got more weight to carry and a longer leadscrew? Also has a tiny bit of eccentricity.

Any ideas? I'm wondering if it's the control box.

[blackburn mark]

I've got a 3020T and it seems fine apart from a problem with the y axis which seems to come on after the mill has been running for a while. It stalls like if you have set the feed rate too high and looses track of where it is, i.e. end up being short of where it should have got to. I've taken a video of it, this was the worst it has ever been at this point.

http://www.readresearch.co.uk/personal/mycncuk/mill_stalling.avi

I've got the motor and coupling close and centralised OK, the y axis is tighter than the others, but this is to be expected as it's got more weight to carry and a longer leadscrew? Also has a tiny bit of eccentricity.

Any ideas? I'm wondering if it's the control box.

errrm... is that not your X axis?
first id go into motor tuning and knock your rapid speed down a bit on the offending axis and if that stops it we can assume that the drag on that axis has gone up a bit for some reason (has it been stood for a while?)

[JAZZCNC]

If it's sticking enough to stall going that slow then this should be easy to find.! It sounds very much like very dry or knackerd lead nut to me with it coming on after time.?
Could be a tight bearing
Do this in a methodical order and you'll track it down quick.
Start by disconnecting the lead nut and running gantry up and down feeling for binding.? If it does then loosen bearing mounts and again run up and down seeing if it lessens.? It could just be an alignment issue.? tighten up and if gone then happy days.!! . . If still sticking then chances are it's a bad or dying bearing.

Check the lead nut for wear or crap embedded in it. Run it up and down checking for tight spots. Unfortunately if it's worn or tight spots then nothing much can be done.!

One thing I've experienced with these machines is the wiring is rubbish. Recently I repaired one where the Z axis and Y would change direction for no reason which turned out to be stepper wires. New shielded wires and machine ran a lot nicer and stable.

[JAZZCNC]

errrm... is that not your X axis?

On this machine the run them other way round a mark.? Drove me crazy crashing kept crashing the bloody thing. . :emmersed:

[blackburn mark]

On this machine the run them other way round a mark.? Drove me crazy crashing kept crashing the bloody thing. . :emmersed:

ahhh! something new every day :)

[richie00boy]

Thanks Jazz will try what you said. I should point out that this is a new machine only used about 3 evenings so there shouldn't be anything worn. I have tried to search for problems with this machine and found some Youtube footage of a similar problem and they said it was the wires. I have tried jiggling the wires to reproduce the fault but can't. Also the Y axis wires never move.

I'm a total novice to CNC, is the lead nut the nyloc that is on the end of the y axis leadscrew? I contacted the seller (ebay) and they said not to touch this nut. Actually they seemed really helpful and quick to respond, but since I sent them the video they are ignoring me. I wonder if I have gone outside of some Paypal claim window or something, and they were trying to fob me off until that time.

[JAZZCNC]

Also the Y axis wires never move.

One of the issues I seen was that the wires are too tightly crammed into the cable guide and rubbing badly. They where also unshielded and I'm sure some of the trouble was electrical noise was interfering with the steppers because when changed for shielded cable the machine was much smoother and sounded loads better.?


I'm a total novice to CNC, is the lead nut the nyloc that is on the end of the y axis leadscrew?

No it's the bit that's fastened too the gantry cross brace.

How long running before this starts happening.? Also is the noise I'm hearing there all the time or does that come on after time has well.?

[richie00boy]

Not sure what you mean there, the noise is the stalling. If the machine has been on about an hour it seems more prone to it. If I was to rewire it any ideas where I can get replacement connectors to fit the blue control box?

Is the nut the part in the housing that the leadscrew tuns to move the gentry up and down? I have tried moving the stepper motor by hand to feel for any tight spots and there didn't seem to be anything that bad.

The stalling can happen anywhere on the y axis, doesn't seem to be related to x or z axis position.

[camhguh]

Not sure what you mean there, the noise is the stalling. If the machine has been on about an hour it seems more prone to it. If I was to rewire it any ideas where I can get replacement connectors to fit the blue control box?

Is the nut the part in the housing that the leadscrew tuns to move the gentry up and down? I have tried moving the stepper motor by hand to feel for any tight spots and there didn't seem to be anything that bad.

The stalling can happen anywhere on the y axis, doesn't seem to be related to x or z axis position.

Maplins sell the connectors :)

I had to replace the x,y and z cables on our machine to keep the machine going, although i used standard connecting cable because it was all i had to hand at the time. Machine has worked ok since.

Swap the cable over and see if the fault tranfers to that axis which should confirm a cable problem.

[richie00boy]

Thanks for that, looks like they are Maplin code FK24B.

When it starts doing it again I'll swap X and Y cables over and see if it changes over. Is there any great danger in plugging and unplugging connectors on the control box when it is on?

[JAZZCNC]

Thanks for that, looks like they are Maplin code FK24B.

When it starts doing it again I'll swap X and Y cables over and see if it changes over. Is there any great danger in plugging and unplugging connectors on the control box when it is on?

DANGER DANGER . . . . . ABSOLUTELY DO NOT DO THAT.!!!

You will destroy the drives which is an all-in-one board so affectively killing the control box.

[JAZZCNC]

Not sure what you mean there, the noise is the stalling. If the machine has been on about an hour it seems more prone to it. If I was to rewire it any ideas where I can get replacement connectors to fit the blue control box?

That noise doesn't sound like stalling motors it sounds more like a binding or sticking bearing or tight leadnut. The leadnut is the plastic part that sits in the housing that runs up and down the threaded shaft and is located under the bed in the centre of the gantry bolted to the cross brace.

You can re-use the connectors just de-solder them.!

[irving2008]

Thanks for that, looks like they are Maplin code FK24B.

When it starts doing it again I'll swap X and Y cables over and see if it changes over. Is there any great danger in plugging and unplugging connectors on the control box when it is on?

As long as the steppers are not powered at the time... if you can turn them by hand then its safe to unplug them... I wouldnt recommend trying it when they are powered up. Normally the control box can be powered but the system is halted in standby, which is not the same as stationary but enabled (when the motors are locked). But as Jazz says, if in doubt, don't.

[JAZZCNC]

As long as the steppers are not powered at the time... if you can turn them by hand then its safe to unplug them... I wouldnt recommend trying it when they are powered up. Normally the control box can be powered but the system is halted in standby, which is not the same as stationary but enabled (when the motors are locked). But as Jazz says, if in doubt, don't.

Irving this is a TB based pile of rubbish board and it will die if so much has looked at in a funny manner so ya can bet ye last dollar it will DIE even in halted state with the slightest spike. . . . Rich it will 100% DIE if steppers powered up and you unplug. . . Just turn it off it's honestly not worth the risk for the 20 seconds you'll save.!!

[blackburn mark]

Irving this is a TB based pile of rubbish board and it will die if so much has looked at in a funny manner so ya can bet ye last dollar it will DIE even in halted state with the slightest spike. . . . Rich it will 100% DIE if steppers powered up and you unplug. . . Just turn it off it's honestly not worth the risk for the 20 seconds you'll save.!!

I run a 5 axis TB china job and plug/unplug the 4th axis stepper all the time (as irving describes,,, in standby... nothing going to motors)

iv earned a quid or two on that "TB based pile of rubbish board" so don't be to disheartened richie00boy

I must admit though, I am tempted from time to time to bin it and splash out on something slightly more robust.... on a 5+hour run its lost enough steps to have a measurable error from the origin (0.2mm ish)

[John S]

No you will probably fry the card. Power down first to be safe.

[i2i]

if you can get this to me in Cardiff, i'll do my best to sort it for you.

[JAZZCNC]

I run a 5 axis TB china job and plug/unplug the 4th axis stepper all the time (as irving describes,,, in standby... nothing going to motors)

iv earned a quid or two on that "TB based pile of rubbish board" so don't be to disheartened richie00boy

Didn't actually say couldn't do what Irving said just not a good Idea with a cheap TB based board.!
I wasn't actually meaning to dishearten anybody so If I did Rich then sorry but at same time not going to say these boards are good because they are not. (0.2mm over 5 days would upset me.!!)
They are very flaky and while mark gets away with it doesn't mean you will. This board, has will most other drives, will 100% fry if you accidentally unplug while steppers are powered up and given your new to CNC then for safety and your wallets sake that's why I say turn it off.

[richie00boy]

I was just concerned that if I turned off the control box it might reset the problem. Will power down first and if it does then will just keep going till the problem arises again then make sure the motors are free to turn by hand before I do anything.

The noise Jazz, well it's the exact same noise that I got when my friend was setting the unit up and had the feed rate really high and it was missing steps. Maybe the camera doesn't make it come out right. I'll not rule out the sticking bearing though.

[JAZZCNC]

Well if it does reset then you know it's electrical so you've eliminated mechanical.!!

RE noise: Ok lets take a look at the motor tuning settings.? In Mach goto Config/motortuning/ then at right hand side click on each Axis individually and tell us what it says in the "Steps per" "velocity" & " acceleration" boxs.
You must click on each individual Axis button to see it's individual setting and if you change any settings always push enter and save before moving to next Axis.

[richie00boy]

Been running the machine for a while tonight trying to get it to do the stalling. Typical - when I want it to happen it won't! These are my settings. Sure if it was a setting it would be doing all the time?

http://www.readresearch.co.uk/personal/mycncuk/x-axis.png

http://www.readresearch.co.uk/personal/mycncuk/y-axis.png

http://www.readresearch.co.uk/personal/mycncuk/z-axis.png

[John S]

Change your step pulse to 5 or 6 and try that. ignore the direction pulse.

[richie00boy]

Not used the machine for some time until the other week when I did a couple of PCBs. Noticed that it was losing position a bit when drilling the holes as they were not quite central on the pads - I put up a thread about this.

After trying to reface my spoilboard it was stalling like mad on the Y axis again and I think actually this is what caused the PCB holes to be misaligned - the machine is just moving so slow that I can't hear or notice it then.

I turned off the controller and tried turning the stepper by hand using the knob on it, it was pretty stiff. So I decided to take it apart a bit and try Jazz's suggestions.

Well it moves freely on the rails when I unscrew the leadscrew mount. With the leadscrew mount loose I did rapid Y movement (shift and up/down arrow on Mach3) and it runs nice until the mount hits the lip of the gantry and then starts dragging the gantry along. I can see the mount bobbing up and down slightly sometimes. So I'm wondering if the leadscrew is a bit bent? I guess if it is this could be the cause of the problem?

If I do need a new leadscrew where can I get one and is there an upgraded type worth fitting?

http://www.readresearch.co.uk/personal/mycncuk/workings.jpg

[JAZZCNC]

If screw is bent enough to cause stalling then you should see it easy enough.? Bent screw will cause a vibration first then if really bad will cause stalling.
It it is indeed a Lead screw and not ballscrew(Hard to tell from pic) I would look at the lead nut it's self to see if it's tight or badly worn and it's binding under load.! What material is the lead nut made from.? looking at the colour of the lead screw I suspect Delrin.? Can you post a picture of it.

If replacing the screw/nut then really you want a ballscrew. They are far superior to lead screws and if bought from China then will probably cost only slightly more than a decent lead screw bought in this country. They will also machine the ends to your spec if the ends are custom machined, which I suspect they will be on this machine.!

[richie00boy]

I'll have to strip the machine more to get at the leadnut - there is grease on it so I can't see what it's made of. I woudn't have thought it was excessively worn as it's only run about 10 hours.

It does vibrate and stall.

Is the current leadscrew threaded rod that I can buy easily and get the machinist at work to do the ends for me? I don't want to spend more than 50 quid really, but the primary use of the machine is PCBs so I need to get low backlash.

[JAZZCNC]

I'll have to strip the machine more to get at the leadnut - there is grease on it so I can't see what it's made of. I woudn't have thought it was excessively worn as it's only run about 10 hours.

It does vibrate and stall.

Suggest you take it off completely and check screw against flat surface to see how much it's bent. It may be possible to straighten if not bad.

You also need to know why it's bent.? Has it bent because the alignment is off and one end is higher or offset to the other. If so what happens is the leadnut which is fastened to the gantry bends the screw when it approaches the end which is higher or lower than other. The screw fixed ends and leadNut need to be perfectly aligned on the same plane. This is a common cause of bent screws.!!


Is the current leadscrew threaded rod that I can buy easily and get the machinist at work to do the ends for me? I don't want to spend more than 50 quid really, but the primary use of the machine is PCBs so I need to get low backlash.

Like I say hard to see from picture.!! . . . BUT it won't be thread rod like you buy at local hardware store.
It could be what's called Acme Lead screw which is often used on cheaper machines, if this one of the early Chinese machines then chances are it is.? If so they used plastic lead nuts which wore out very quickly and could have done so even with 10hrs use that's how crap they where.!!!!

The later slightly better Chinese machine are coming with 5mm pitch ballscrews, Looking closer at the pics I'm tipping it's lead screw not ballscrew.

You will buy a ballscrew with nut from china complete with end machining for around £50 delivered.
TR14 lead screw will cost about £15-20, lead nut will be £10-15 then you'll need to machine the ends. You could be up and running for £20-£30 but you'll still have crappy screw with some backlash in the nut unless you buy better anti-backlash nut which will cost much more.!!. . . So you can probably tell were I'm going.? Not worth IMO so just Bite the bullet and buy a ballnut you'll get much smoother machine with better accuracy and probably more speed.

[richie00boy]

Thanks for the help. How can I tell if the alignment is off to bend the leadscrew? And is that fixable?

I do remember noticing that the leadscrew was not totally concentric when I first got it. I assumed it was just due to the leadscrew being made from really long thread which got slightly bent at some point in the Chinese factory.

[JAZZCNC]

Thanks for the help. How can I tell if the alignment is off to bend the leadscrew? And is that fixable?

With careful measurements and visual checks.!! Aligning ballscrew and nut is tedious job but very important to both prevent bending and getting the best performance.
If alignment is off enough to bend the screw it should easily be visible by manually running the gantry along it's length carefully watching the screw to see if it's bending. Also if you turn the screw by hand with motor disconnected and you feel it stiffen has it nears ends this is good indication some things off.!!

Can't really comment on what would be needed to fix without seeing how ends are fixed and how much off it is.?

[FatFreddie]

Thanks for the help. How can I tell if the alignment is off to bend the leadscrew? And is that fixable?

If you take one of the bearings off and turn the screw by hand, as the nut approaches the other bearing it should be fairly obvious if it's misaligned, particularly if you can get the nut quite close to the bearing.

[richie00boy]

OK, before I take things apart a bit more and get lost, here is what I have found.

1. The motor knob becomes stiff when the leadnut under the gantry can be at almost any point along the length of the leadscrew, i.e. not just at the ends.
2. I have unscrewed the leadnut from the gantry and can see the leadnut housing bobbing about relative to the gantry again at almost any point along the length of the leadscrew.

I'm thinking that there could be two ways for bending the leadscrew. One if the leadnut was not mounted centrally, and the other if the chassis is out of square or the leadscrew not exactly in the middle of each end piece. How can I check for all of these? Because there is also the possibility that a bent leadscrew has been fitted at the factory - which I'm hoping is the case and my machine is fundamentally sound. Also I would not be that bothered if the leadnut was not mounted centrally because if I convert to ballscrews I would be fitting a new mount.

[JAZZCNC]

OK, before I take things apart a bit more and get lost, here is what I have found.

1. The motor knob becomes stiff when the leadnut under the gantry can be at almost any point along the length of the leadscrew, i.e. not just at the ends.

Ok you can't feel for stiffening properly with the motor connected because of the Detent of the motors to do it properly you MUST remove motor.


I'm thinking that there could be two ways for bending the leadscrew. One if the leadnut was not mounted centrally, and the other if the chassis is out of square or the leadscrew not exactly in the middle of each end piece. How can I check for all of these? Because there is also the possibility that a bent leadscrew has been fitted at the factory - which I'm hoping is the case and my machine is fundamentally sound. Also I would not be that bothered if the leadnut was not mounted centrally because if I convert to ballscrews I would be fitting a new mount.

Without seeing the exactly how the ends are fixed and the leadnut mounts to gantry then it's hard to say. But if it's how I think then to get a rough idea you'll need to remove the bearing at one end and the run the gantry to that end with leadnut attached and see if the screw is central in the hole.? This should give you some indication if it's misaligned.
The most common reason for bending is because the leadnut is either lower or higher than the ends in middle of it's travel where it's the most bendy it then lifts or pulls down on screw when approaching ends bending screw in process.

To check if screw ends are equal is a simple case of measuring from one reference edge to screw centre(or edge) for both vertical and horizontal. IE: Machine side to screw for horizontal and End plate edge to screw for vertical.

Then if ok loosen the leadnut mount from the gantry so there's no pressure on it and run the gantry to one end. Now check for any gap between Mount and gantry.? Now Run the gantry to other end and see if the gap is the same. Do the same at several points along the screw.
If all is ok the gap should be same or very very close to none, ideally you want light interference fit.! Now it's just a case of shimming the gap so when tightened there's no pressure or pulling on the screw.
If there's no gap and there's tight fit between leadnut mount and gantry then you'll need to remove some material from leadnut mount.
If your leadnut mount doesn't bolt onto flat surface and instead has drop bracket that which fastens to gantry cross brace then you'll need to loosen the bracket to slip fit and put witness marks on bracket and brace. Then run gantry up and down it's length and watch if marks move.? If they do then try loosening lead nut from the bracket and find balance using both bracket and leadnut mounts, failing that you'll have adjust the end fixings.!!

If the machine is out of square then you'll feel it going tight when pushing gantry up and down it's length with leadnut disconnected. Being honest thou with the sloppy linear bearings on these machines I doubt you'll feel it. That said for it to be bad enough to bend leadscrew you'd most certainly be able to visually see it.!!

Obviously I'm giving adjustment suggestions blind has I can't see your machine and how things are fixed but hopefully should give you an idea of what i mean.!!! . . .. . . OR I could have just blown your mind. .:hysterical:

[richie00boy]

The stiffening is a lot worse than other axis though and I can see the leadscrew is bent so I think I'm satisfied with that as the fault diagnosis. The gantry runs nice and free up and down the rails with the leadnut disconnected, so I'm happy that things are square. I can't feel any play.

I don't think I can do these other tests being mentioned because the leadscrew is bent, therefore it's going to make the leadnut not align with the bearing hole when looking through, etc. If it was the leadnut alignment that bent it, well I'm ditching that anyhow.

So what I need now is to spec and purchase a suitable replacement ballscrew system. Then once I have that I can do these trueness checks. I'm going to have to do them anyway to work out the mounting for the ballnut aren't I?

I'll start a new thread for the ballscrew.

[richie00boy]

Just as a last check, here is a video of the leadscrew showing the runout. The leadnut housing has been unbolted from the gantry to show how it rocks due to the bent leadscrew. Hopefully you will agree that this is the cause of the problem.

http://www.readresearch.co.uk/personal/mycncuk/leadscrew.mp4

[JAZZCNC]

Just as a last check, here is a video of the leadscrew showing the runout. The leadnut housing has been unbolted from the gantry to show how it rocks due to the bent leadscrew. Hopefully you will agree that this is the cause of the problem.

http://www.readresearch.co.uk/personal/mycncuk/leadscrew.mp4

Take it you have the mounting bracket loose.? . . . But No not at all I've seen worse than that an still work fine.!!. . . . . For this to cause stalling it should only possibly be at very high speeds and even then with that amount of wobble over such short distance I doubt it would.!!

If this is causing your stalling then I think you may have the motors tuned too high.? What do you have set for Acceleration & Velocity in motor tuning for each axis.?

[richie00boy]

Tuning settings are further back in this thread and the comments then said they were fine.

I've now unplugged the motor and been turning the knob to move the Y axis up and down, it's not that stiff really. This is with the leadnut mounting loose, so it's just pushing the gantry on the lip and is free to rock about a bit.

I also tried putting the Y axis drive on the X axis motor and running that and there is no stalling. I tried by hand to give some resistance to the head moving and it just powered through it.

Could I have a dodgy motor?

[JAZZCNC]

Tuning settings are further back in this thread and the comments then said they were fine.

Ok I've been back to start to familiarise my self with original issue.!! . . . . . Did you change the wires.?


I've now unplugged the motor and been turning the knob to move the Y axis up and down, it's not that stiff really. This is with the leadnut mounting loose, so it's just pushing the gantry on the lip and is free to rock about a bit.

Does it go tight when you tighten the mounting.? If so then you may need to shim the space between mounting and gantry surface. Some times just a piece of paper can make difference between binding and not.!!


I also tried putting the Y axis drive on the X axis motor and running that and there is no stalling. I tried by hand to give some resistance to the head moving and it just powered through it.

No wouldn't expect it to stall either.!! The Y axis is tuned for half the speed of X axis and X axis is half the weight of Y axis. So in simple terms you have roughly twice the torque pushing half the weight.!!



Could I have a dodgy motor?

Very unlikely Steppers are very robust things they either work or they don't.

Basicly if you haven't got bad mechanical binding or some things badly bent or misaligned like it seems to be turning out you haven't, then it's either incorrect motor tuning or you have electrical/electronic problems, either from wires or failing drive or possible PC.(thou don't think so here)

knowing there's no mechanical binding then my next place to turn, if you haven't already, is to change the wires and connectors.
I've seen them cause this sort of thing on these TB based drives before, strange but it's happened.

In mean time It may be worth playing with motor tuning to find the point were it doesn't stall. Can be tedious but some times the difference between stalling and not stalling can be less than 50 on either vel or Accl.
Get pen & paper and write down settings for each axis before starting and be methodical, only change 1 setting at time and write it down then try.
Start with Y axis and with Velocity first then Accel straight after, do one then the other until you find the balance.
My approach for getting into ballpark is to half the current setting then try, if Ok, then increase by half the difference. . . . IE 1000 original try 500=Ok so try 750=Not Ok try 625=Ok try 560 etc
This work for both Vel & Accl but best if only change one setting then do the other, All you have to remember is you can't have high Velocity and high Acceleration. When you are near usually raising one means lowering the other.

One thing to try is lowering the other Axis settings slightly and see if it affects the Y axis.? Don't think so because it's not happening while moving other axis.? Or is it.??

All that said your tuning for the Y axis is painfully slow to me so going lower to stop stalling would drive me crazy.!!

After changing wires and lowering tuning if it's still doing it then I suggest you dump the Electrics and buy some real drives.!!

[richie00boy]

Bit of an update. I think it's now sorted. There were a couple of issues, first was that it turned out the steps setting was 65kHz in Mach3, set that down to 25kHz. This became apparent trying a friends PC which was a little lower spec than mine.

The other issue was/is some binding on the leadscrew which I have improved by loosening the motor mount screws and running the axis up to the motor end and tightening the screws back up. I had to do this twice, the second time trying to do it on the tightest position, as the first time there was still some binding up at the top end.

Will see how it goes, maybe the binding will get worse again once I put the heavier 800W spindle on. But for now it's running the best it's ever run. I was able to speed the axis up to match the x axis.

[Boscoe]

richie00boy I didn't know you were interested in CNC as well as diyaudio!

[richie00boy]

Well it's a means to an end rather than interest. All I need now is a big CNC for speaker cabinets lol.