Machsupport Forum
Mach Discussion => General Mach Discussion => Topic started by: russh on November 14, 2015, 07:39:40 AM
-
Hello All,
Can some one offer some help on an issue I seem to have
Basically, when set my zero positions, I then come to run a program and now and then (more times that not!) it seem to loose its zero positions. I've had a few cases where Z or Y have lost Zero and the tool has crashed in to the work piece.
Initially I though I might be cancelling or calling up an incorrect fixture offset in the G-code with G54, but even when I remove this from the safe line, it still happens
Any help would be greatly appreciated
Thanks
Rusty
-
Hi Rusty,
Sounds like you are loosing steps.
To resolve the problem just one step at a time ( no punn intended)...
Try reducing your Velocity and Acceleration settings ,for each axis, by 50% (in motor tuning) and save the changes - and then see if the problem persists.
Tweakie.
-
Tweakie,
Huge thanks for the quick reply
Would the "losing sets" issue occur in manual mode as well, I.E If I jog the table around with the MPG wheel, I don't seem to loose zero (well not that I have spotted, but I will recheck this statement as well), it only seems to be during code running
By the way, The problem occurs even if I manually over ride the feed rate right down to a snail pace during program run as well
Cheers
Rusty
-
Hi Rusty,
Reducing the feed-rate is not quite the same thing so try reducing the Velocity and Acceleration by 50% as suggested and report back on what happens regarding loosing position.
Tweakie.
-
Hi Tweakie,
Right, I ran a few tests and record a bit of position information first. I turn the machine on and zero'd the tool to each edge of the vice I had in, then ran a simple movement program with G00 and G01 commands. The machine lost position from the zero points by X: 3-4mm, Y: 3-4mm, and Z: 16mm.
I then changed the Velocity and Acceleration values as you said by 50%.
Original Values
X: Velocity=6000, Acceleration=300
Y: Velocity=6000, Acceleration=300
Z: Velocity=5100, Acceleration=300
New Values
X: Velocity=3000, Acceleration=150
Y: Velocity=3000, Acceleration=150
Z: Velocity=2550, Acceleration=150
I then ran the same simple program (after re-zeroing), and repeated the program a few times. It still lost position, but not as much this time. X: 0.5mm, Y: 0.5mm and Z: 2.4mm
Should I reduce by half again? or do we need to look at something else now?
-
Reduce it by half again I would say.
I run mine at 1000 max for velocity, and even that I find is too fast when the machine is cold!
-
So 6000 velocity was WAAAAAAY to high then!
That was what was in the manufacturers config file!
I will give it a go at half again, and if it still runs out will knock it down a little more
-
Hi Rusty,
Wow those Velocity figures you have been using are high.
Perhaps try; Velocity 2000 and Acceleration 300 and see how that goes.
Tweakie.
-
Right,
I'm now down to 1000 mm/m velocity and just had a more stable run in X and Z, but I still lost about 0.8mm on the Y.
I'm now trying 800 mm/m but is this beginning to look like another issue, or is sub 1000 common?
-
How are the motors sounding? Smooth and humming or jerky and noisy?
-
The movement in the X and Y seems smooth, cant feel any grinding or jerking on the table. There is a touch of noise in the Z, I'm tempted to say like a grinding noise, but I wouldn't say its actually grinding.
The motor sound nice, like any electrical hum as the work
-
I've just ran at 800 mm/m and it seem stable, I measure delta from start Zero of X: 0.0069 Y: 0.01 and Z: 0.09, but this may just be down to measure error. I'm going to keep running this movement programme through a few more times and remeasure
-
Seems you are getting to the motors sweet spot with the velocity settings. It took me a while as I thought faster was best, how wrong was I, you may find that 600 will suit your Z axis, no harm in trying?
-
Right, so this low velocity is pretty normal then, I guess it depends on the motor spec?
What I dont understand, if the machine manufacturer states 5m/min for X and Y, and 3m/min for Z, which im way off
I will give 600 a try
-
Seems I still have a small problem
Now down to 600mm/m
After running a 25 minute program of a repeat set of movement
Y and Z seem stable
X lost roughly 0.3mm
-
Hi russh,
what is the pitch of your lead screws? I use 5 phase Vexta steppers, I wanted the better resolution (500 step per rev verses 200 step per rev for more
common 2 phase steppers) and smoothness at full step. When researching I found the Vexta website handy because they have a graph of motor torque
vs speed. As it turns out a motor rated at say 200 oz.in typically measured at around 100 rpm is likely to be only 20 oz.in at 1000 rpm.
More research has shown that all steppers are broadly similar in that the torque diminishes rapidly at speed. Microstepping makes it worse.
The bottom line is that stepper motors have no grunt at speed, and you will lose steps or stall if you try.
My steppers are fitted with 10:1 low backlash planetary gear reducers, with the reduction even modest steppers produce great torque but very slow. My 'rapids'
I have settled at 600 mm/min with 5mm pitch screws ie the screws at 120 rpm and the steppers therefore at 1200 rpm. When tuning/experimenting I tried a lot faster,
one stage I had rapids of 2100 mm/min ie screws at 420 rpm at steppers at 4200rpm. They ran hot and very unreliably.
As I progressed I found that at 600 mm/min I could back the drivers off a wee bit and still have reliable operation even with 30 kg on the table. At cutting speeds
around 150 mm/min in steel I have 80kgf of thrust so plenty of grunt...
My advice is don't try to run steppers fast, they don't like it and you will break tools like they're going out of fashion.
Craig
-
Hey Craig,
Really appreciate that post. I dont know the answers to your questions of pitch and stepper motors, but leave it with me and I will strip the machine back slightly to take a look.
Its interesting cause its mostly the Z axis loosing position the most, and that has the large cast head, spindle and motor all attached, so the motor will need alot more torque than X and Y
-
Hi russh,
sounds like the weight of the z axis will be a challenge. Maybe you could consider a gear reducer, they really improve torque/thrust.
Most of my toolpaths are 2d in the sense that the z axis lowers to to fresh cut of say 2mm and then x-y take over until a fresh cut is to be made.
If my z axis is slow it doesn't matter to much.
Low backlash planetaries are not cheap even second hand but are available for most steppers are pretty easy to fit. The ones I use are 23 size
so a 23 stepper screws directly to the input and the output mounting is 23 size as well so you just stick it in between as it were.
The only time the reduced axis speed is really going to hurt is when you do co-ordinated moves like helical cutting or 3d contouring. Even then you
are likely to be at federate which is so much slower anyway.
I amongst other things engrave circuit boards and particularly when drilling all those little holes with tons of small z moves do I miss fast rapids.
Craig
-
What I dont get is this is a purchased machine, not a DIY job, and there are your tube videos of it in operation working pretty fast. I'm wondering if I've been ripped off with cheap stepper alternatives by the manufacturer. I will strip it a little and get some more details.
This is the machine
http://youtu.be/sop3aiFxaws
-
Hi russh,
watched the vid and I'm jealous! Looks the business. I noted that the note that introduced the clip said 2200 W BLDC spindle and
SERVO control. Is that how your machine is fitted out?
Might be worth an email to the manufacturer to clarify what 'servos' were actually demonstrated in the vid and if your machine is the same
spec.
Craig
-
Hi russh,
followed up and searched for your machine and it lists servos as an option, microstepping stepper motor as standard.
I imagine you would know if you had paid extra for servos.
Craig
-
Russ, it may sound daft, but have you checked the electrics/electronics. It does begin to sound as though something in that direction is not set correctly, voltage, amps, steps, loose wire maybe. Just a thought because the machine itself does look a good one.
Stewart
-
Hi Russ, I think Stewarts suggestion is very good. I rather think the manufacturer would be reluctant to release a machine that did not
perform as advertised given that the controller is Mach3 and any problems will be aired on the forum, the very same customers the
manufacturer is trying to attract.
There are articles on the net about stepper torque and microstepping, I have lost the links tho. The upshot is that deep microstepping
causes torque to decrease even worse than running them at speed.
May I suggest that you disable microstepping, most likely switches or jumpers on your drives, and rerun your test. Given that the z axis seems to be
the problem child just the z axis will do. The articles I have read suggest that half stepping actually improves torque, 120% of full step torque. Set your
drives to half step.
It may well mean your resolution is less than you would like but probably a good experiment which may help you to home in on the fault without
having to spend anything or pull things to bits.
Craig
-
Hi Guys,
Please don’t get dragged into this micro-stepping / torque myth.
There is so much garbage posted on the net about this - go to the stepper driver chip manufacturers sites for the true information regarding the micro-stepping torque figures.
Tweakie.
-
Guys,
Huge apologies for not getting back with some motor info sooner, I've not been too well in health lately so had to down tool for a while.
I managed to get at the Machine yesterday to try and find a data label on the axis driver motors, unfortunately there is not a single label on them, so other than the following picture, I don't have much to offer and not sure what I can do.
The manufacturer is uncontactable unfortunately!
(http://s27.postimg.org/fzqjomf9f/20160116_194118.jpg)
(http://s24.postimg.org/cnhuxto39/20160116_194130.jpg)
-
Hi Russ, hope your feeling better?
Now these look like steppers and not servos, can't tell from the pics what size they are so do a quick measure of the front mounting plate and see what size you get, measure in imperial.............2.3" = NEMA 23, 3.4" = NEMA 34 and 4.2" = NEMA 42. We can then take it from there.
-
Hi Stew,
The size of them is just shy of 3.4", and have a length of around 4.5"
-
OK Russ, these are NEMA 34 motors, unfortunately without any details on the ratings we can't check if you have the electronics set up accordingly.
Did you buy the machine brand new or second hand?
My suggestion is to contact the manafacturers directly and see if they can give you direct support, we can only guess and estimate with the details you give us and it can be frustrating for both sides.
Hope you get it sorted Russ as it does look a neat machine.
Stewart
-
Hi Stewart,
I've tried to contact the manufacturer a number of time with no response, and to be honest I'm down right annoyed with them as they have cost me thousands of dollars in orders that this machine was bought for.
If there is anything you can advise on checking as starting points I would be grateful
Thanks
Russ
-
Hello Russ,
Please check your PM's and email.
Thanks!
JB
-
Sorry for asking the obvious but you bought this as a ready to run unit? If so has it run ok from the start and then the problem arose or has it done this from day one? Have you done any alterations to it? Any details of the electronics (stepper drivers,PSU, etc) how is it connected to PC or is a built in unit? The more you can show and tell the better chance someone may find the problem. Another check the forum guys can do is look at your XML file, copy it from the Mach folder and post it on here.
Stewart
-
Hi Stewart,
It has done this from day one. I have made no modifications to it and ran the machine from the XML file the manufacture provided with the machine.
Let me gather all that information together with photos and list it out on the forum. I should be able to start taking photos later today.
Thanks again for you help
-
Hi Rusty,
If you could also post your .xml file - not the one sent with the machine but the one from your Mach3 folder - copy and rename it to say rusty.xml then post it using the Additional Options tab we can check your settings and see if we can spot a problem with your set-up.
Tweakie.
-
Hi Tweakie
Will do, although the one I'm using in the Mach3 folder is indeed the one that came with the machine
-
Hi Tweakie
Will do, although the one I'm using in the Mach3 folder is indeed the one that came with the machine
Yes indeed sir, but the one in your Mach3 folder reflects your actual set-up (which may contain installation, accidental or other errors) and these (if any) are what we are trying to find.
Tweakie.
-
I had a very similar issue with my 1st lathe.
iirc, inverting the step signal and/or lengthening the duration solved it for me.
Russ