PDA

View Full Version : Calibration issues, advice needed



Kernow Jim
18-01-2014, 12:52 AM
Hi everyone,

I have just got my stepper motors running and have done a test cut with the router. The calibration on the z axis seemed fine, but x and y, the project was smaller than expected but in proportion. I re-calibrated both x and y, then rechecked the amount of travel and had to calibrate again. After a couple more test cuts, the x axis is better but the y is still out.

I am using M542H controllers, which I set the dip switches to a recommended setting from this forum, but I can't remember which post it was, but it has made the motors run quieter than on the settings I had them.

I think the calibration problem is related to the the settings on the controllers. Could someone explain the steps, velocity, acceleration etc in Mach3 and the relationship to dip switch settings. I thought I understood some of this but I seem to be confusing myself even more.

I am using 16mm chineese ball screws on all axes and x and y seem very slow when jogging manualy compared to before I calibrated x,y and z.

Hope someone can help, thanks in advance

Jim

Kernow Jim
18-01-2014, 01:20 AM
The dip switch settings are Jazcnc's recommendations in reply to Liamo1991's post in this category.

Thanks again

Jim

Neale
18-01-2014, 11:57 AM
Are you talking about being wildly out (e.g. twice or half size) or just a little bit - a few per cent, say?

JAZZCNC
18-01-2014, 10:58 PM
Jim.

Your drives main dip switches let you set the Motor Amps and Micro stepping. Motor amps is easy just select the nearist lowest setting to the motors Amp rating.

Micro stepping (MS) is dependant on how you want to run the motors. In general setting the MS to a higher setting gives smoother running motors and to a lesser degree some resolution gain. The trade off is it works the PC and parallel port harder has it needs more pulses to move the same distance has a lower setting. There is also a limit where it's no longer benifical going higher, this around 2000 pulses, but won't get into this now.

What you want or need to know is the relationship between the MS value, ballscrew Pitch and Mach's Step Per setting. . . . It's quite easy really when you get your head around it.!!

Mach Steps per is asking for the number of pulses(steps) to move 1 unit, so if mm's then 1mm.
To calculate this you need to know the Pitch of the Ballscrew, NOT diameter, Pitch is the amount it travels for 1 revolution. Along with the MS the drives are set too.
Then divide the MS by the Pitch. Ie: MS=1600/5mm pitch=320 Pulses(Steps) Per/mm.
If your machine is moving the wrong amount it will be these settings causing it.

The velocity and Acceleration are then adjusted based on your preferances on speed and what the machine is capable of. There are far to many factors which affect this and really nobody can give you any accurate settings for these only guide lines has it's mostly trial and error which is based heavily on the individual machine.

Kernow Jim
19-01-2014, 12:55 AM
Hi Neale and Jazz

Thanks for your responses.

Thanks Jazz for your explanation, I think I understand better now.

The problem I had, was after I calibrated and then re-calibrated to check that it was moving the right amount, it would still be out, so I would do it again, and again and after 4 or 5 tries, still not knowing if it was calibrated properly. When I started out, both the x and y were nearly double on the measure. When I thought I was done I did a test cut, just a circle set at 100mm. I ended up with an elipse! The x axis was almost spot on but the y axis was only 38mm.

I have been out with the machine again, the MS drivers are still set to Jazz's settings and after calibration the 'steps per' on all 3 axes are between 1275 and 1286. It seems to cut at a reasonable rate according to the feed rates. I was expecting it to move quicker on rapids and jogging by hand.

Any way I am a lot happier with it now.

Thanks again guys

BTW did you see the picture of my machine on my intro post, I'd like your comments.

Jim

JAZZCNC
19-01-2014, 01:29 AM
I have been out with the machine again, the MS drivers are still set to Jazz's settings and after calibration the 'steps per' on all 3 axes are between 1275 and 1286. It seems to cut at a reasonable rate according to the feed rates. I was expecting it to move quicker on rapids and jogging by hand.

That's a very high number of steps per.!! . . . The figures make think you have the Micro stepping set far too high and working on the steps per figures you have given I'd guess you have drives set at 6400 micro steps if using 5mm pitch screws or 3200 if 2.5mm pitch screws. In either case your set too high and should reduce to 2000 at most. Anything above this is just wasting pulses and robbing you of speed. If I'm correct then This will be why you can't get very high rapids.

So Please give details about machine.

#1 Are you using mm's or inch units in Mach3
#2 Micro steps drives set at.
#3 Ballscrew or lead screw pitch.

Edit: I've just been to your Intro and can see you have either 5mm or twin start screws so 10mm pitch so this confims to me you have the Micro stepping set way way to high.
Drop the micro stepping down to 1600 and if 5mm pitch set the steps per to 320. or 160 if 10mm pitch.

Now if the Kernal speed is set to 25Khz you should get in the region of around 4000mm/min velocity with Accel of around 800-1000s/s.

Raising the Kernal speed will allow more pulse rate so a higher velocity can be reached but I wouldn't go any higher than 45khz and even then I don't recommend you change it has every change from 35khz makes the machine potentialy unstable. You'll need a fast PC with good parallel port to higher than 45Khz.

Neale
19-01-2014, 09:58 PM
Jim - from what you're saying, I'm not sure if you've gone the right way about setting up your machine. Unfortunately, copying someone else's DIP switch settings sounds like a great idea, but only if your machine is identical to theirs!

Briefly - a leadscrew/ballscrew pitch of 5mm means that the nut moves 5mm for every turn of the screw (or 10mm movement for a 10mm pitch screw, or whatever). A standard stepper motor will turn one revolution in 200 steps, so each step will turn the leadscrew by 1/200 of a turn, and this will move the nut by 5mm/200 = 0.025mm (if I've done my sums right, and assuming that the motor is geared to the screw 1:1). The job of Mach3 is to take a line of gcode that says, perhaps, "move the cutter along the X axis by 10mm" and translate that into the number of "step" pulses that it sends to the motor controller board. This board then turns those pulses into the rather more complex signals with lots more power that are used to actually turn the stepper motor. That's why you have to tell Mach3 about the stepper motor configuration and the pitch of the screw and it can do the sums it needs from that. If you tell Mach3 in the settings dialogue the right numbers, then the movement will be pretty close to exactly what you intend.

It gets slightly more complicated if you use micro-stepping. I said that a standard motor will turn exactly one revolution for every 200 steps sent to it. However, you might want it to move in smaller steps. The controller board can do this for you by modifying the signals that it sends to the motor. What you have to do is tell the controller board how many steps you want per turn. A typical number might be 800 steps per rev. If you look at the manual for the board, you can find the DIP switch settings that specify 800 steps/rev (might also be called 1/4 steps or some such). The motor will now turn one rev for every 800 steps, but you also have to tell Mach3 about this so it knows how many pulses it has to send to move the cutter by any specified amount. So, the settings page of Mach3 asks for the microstepping amount. Again, though, if you put in the right numbers you will get the right amount of movement and while you might want to double-check by measuring it that it's right, you should be very, very close without any need for "calibration". Something else that you need to tell Mach3 is if there is any gearing between the motor and the screw; if you have 1:2 gearing so that the screw turns at half the speed of the motor, then clearly you need twice as many steps sent to the motor to move the leadscrew nut by a given amount because the motor has to go round twice for every leadscrew turn. Again, you tell Mach3 about this gearing and it does the appropriate sums when it calculates numbers of step pulses to send.

Why don't you just use the highest available number of microsteps per rev, then? Won't this give better accuracy because there are more steps per rev so the smallest amount you can move is smaller, so you can position the carriage in smaller increments? Well, in principle, yes, but there are other factors. Stepper motors aren't actually that accurate as the number of microsteps is increased, certainly not when you get to very high numbers, and the actual torque available (in effect, the amount of "shove" that the motor can apply) reduces. Another very important point is that you need more step pulses out of Mach3 to move the carriage by a given amount and there is a limit to how fast Mach3 can generate pulses. As Jazz has said, this is related to the kernel speed setting, and in turn this is governed by the PC hardware. Some hardware is capable of higher speeds than others, but Jazz has given a reasonable value. Now, if you ask Mach3 to deliver more pulses more quickly than its kernel speed setting allows, it will just limit at some lower value. Quick example with 5mm pitch screw. You want to move 50mm in 1 sec. With 200 steps per rev (standard motor setting, no microsteps) you need 10 turns per second which means 10x200 = 2000steps/sec. With 16x microstepping (3200 steps/rev), you will need 16x2000 = 32000 steps/sec. If the kernel speed limits the pulse rate to, say, 20000 steps/sec, then you will not achieve 50mm/sec movement because you just can't send pulses that fast and your machine will move slower than you expect. OK, in practice, there are some other issues involved but this gives the basic picture. This might help explain why Jazz has suggested that your microstep settings are too high.

If you can give the basic settings as Jazz has asked, it gives a lot more understanding of the machine setup and is a good way to move forwards with some better values to try.

Sorry I've rambled on a bit, and this stuff is basic knowledge to a lot of people on this forum, but for people new to the area it might give a little bit more insight into what the numbers mean and how they relate together. Anyway, you only have to go through it all once, and then the machine is set to go!

JAZZCNC
19-01-2014, 10:43 PM
Something else that you need to tell Mach3 is if there is any gearing between the motor and the screw; if you have 1:2 gearing so that the screw turns at half the speed of the motor, then clearly you need twice as many steps sent to the motor to move the leadscrew nut by a given amount because the motor has to go round twice for every leadscrew turn. Again, you tell Mach3 about this gearing and it does the appropriate sums when it calculates numbers of step pulses to send.

Ok that's a good explination and correct but would just like to make this section more clear.?. . . .You don't actually tell Mach anything about gearing YOU need to factor in the gearing when you calculate the steps per. Mach has no place or setting for this gearing and infact knows nothing about any gearing.

Also after about 2000 Msteps your mostly gaining nothing other than a tiny bit smoother running motors which comes at a cost often not worth it due to the instabilty's that can come with it. To get any decent speeds for a router with high MS you'll need to raise the Kernal speed so this can and does make the system more unstable so for this reason alone your better keeping the MSteps down between 800-1600.

Kernow Jim
19-01-2014, 11:22 PM
Hi Jazz and Neale,

Thank you both for your explanations, I am understanding things a lot better now. You Guys are legends to spend all this time typing all that info for a greenhorn like me. Lets hope it helps others as well.

I managed to sneak off to the workshop/mancave this afternoon.
I am working in mm's in Mach, the MS drivers now set to 1600, the ballscrews are 5mm pitch. After calibrating, 'steps per' worked out to 637.8 on x, 631.1 on y and 642.6 on z.

After another test cut all axes were faster, I then had a light bulb moment and checked the velocity and acceleration settings they were set to 600 and 200, the settings recommended in the instruction mini disk, which I haven't altered since initial set up. I changed them to 2000 and 1000 respectively, now I have much better speed.

After coming back and seeing Neale's in depth explanation, everything seems much more clearer now.

Thanks again guys, you are awesome.

The trouble with us Cornish boys, we are strong in arm and weak in head.

Jim

JAZZCNC
20-01-2014, 02:04 AM
I managed to sneak off to the workshop/mancave this afternoon.
I am working in mm's in Mach, the MS drivers now set to 1600, the ballscrews are 5mm pitch. After calibrating, 'steps per' worked out to 637.8 on x, 631.1 on y and 642.6 on z.

Something is wrong here if your direct driving the ballscrews so this makes me think you have a 2:1 ratio so your affective pitch is 2.5mm.

If your direct driving the ballscrews then the steps per should be 320 exactly.

How are you coming up with the 637.8 figure .? Are you using the Axis calibration in Mach.?

Kernow Jim
20-01-2014, 11:26 PM
Hi Jazz

It is a direct drive to all ballscrews. The 'steps per' figures of 637.8, 631.1 and 642.6 on x,y and z respectively was by doing the axis calibration within mach. That is the figures that was coming on the dialogue box 'do you accept these settings'

Armed with my new knowledge from yourself and Neale, I was proposing to set the MS drivers to 800 steps and type into the 'steps per' box 320, and increase the velocity to 4000 and the acceleration to 1000. Unless you advise against it

My visit to the workshop yesterday, when I changed the velocity and acceleration settings certainly made a huge difference to the jogging and rapid speeds. Yesterday I also changed the MSteps from 800 to 1600 but this made the motors more whinier (is that a word?) more noisy. My machine is housed in my Fathers workshop at the moment, for the purpose of building it (more tools and equipment) which is about 14 miles away. It will be better when I have finished the build and bring home to my smaller less equipped workshop, so I can't exactly nip out to change some settings and report back.

Thanks again Jazz

Jim

JAZZCNC
21-01-2014, 12:02 AM
Armed with my new knowledge from yourself and Neale, I was proposing to set the MS drivers to 800 steps and type into the 'steps per' box 320, and increase the velocity to 4000 and the acceleration to 1000. Unless you advise against it

No remember it's a simple a formula to find the Steps per. Divide the Micro steps by the Pitch. . . .So 800/5=160

This is an absolute calculation to within fractions of the calculated figure so when doing the Axis calibration it should be very very close if not the same has the calculated figure. If it's different by a lot then something is wrong.

Things that can cause this differance are the drives micro stepping not being set to what they say they are.?? . . Doesn't happen often but it does happen. (Or you are setting them wrong.!. . It's Common to mix up ON / OFF on Dip switches.?? )

Regards Velocity and Accelleration then you'll just have to play about has every machine is different. 5000 and 800 for X & Y shouldn't be too taxing if you have decent amount of volts on the motors. Slightly less on the Z axis.

Kernow Jim
22-01-2014, 01:25 AM
No remember it's a simple a formula to find the Steps per. Divide the Micro steps by the Pitch. . . .So 800/5=160

This is an absolute calculation to within fractions of the calculated figure so when doing the Axis calibration it should be very very close if not the same has the calculated figure. If it's different by a lot then something is wrong.

Things that can cause this differance are the drives micro stepping not being set to what they say they are.?? . . Doesn't happen often but it does happen. (Or you are setting them wrong.!. . It's Common to mix up ON / OFF on Dip switches.?? )

Regards Velocity and Accelleration then you'll just have to play about has every machine is different. 5000 and 800 for X & Y shouldn't be too taxing if you have decent amount of volts on the motors. Slightly less on the Z axis.

Hi Jazz

You were absolutely right, something was wrong. As luck would have it I forgot my disk with the documentation regarding MS driver settings chart. I had to resort to removing a driver from my control box to look at the chart on there. Lo and behold, it seems I have the wrong docs for my drivers. The dip switch settings I have been putting in, eg, 1600 steps has been actually 3200, hence when I calibrated in Mach I was getting a figure of around 640 'steps per'.

I have now set the MS drivers to 1600, manually typed in the 'steps per' box 320. I played around with the velocity and acceleration settings and have ended up at 3000 and 800 respectfully on all three axes.

Everything seems to be running better now.

I don't know how long I would have gone on before making the discovery, good fortune has intervened by me forgetting the document disk, but it has been your insistence that something was wrong that has pushed me on to check and re-check settings.

Thank you again Jazz for your patience and knowledge for a greenhorn.

Jim

JAZZCNC
22-01-2014, 02:51 AM
Thank you again Jazz for your patience and knowledge for a greenhorn.

No problem just pleased you found it and sorry if I was blunt.!!

If your happy with those speeds then great and the 1600MS should give you nice smooth running motors but if you do want a little more speed then try at 800 and you should push it up around the 5000mm/min or higher.

What voltage are you running the drives on.?

Kernow Jim
22-01-2014, 08:18 PM
Hi Jazz

The drivers are running on 36v, but I have the dip switch set to half power. When I was testing the motors they seemed to get quite hot at standstill with the holding torque when the drivers were on full power. I haven't changed them back.

One other question regarding settings in Mach, in the motor tuning the settings for step pulse and direction pulse (1 to 5) I have set to 2 in both boxes from my documentation. A tutorial video on you tube suggested that these are normally set to 0, any suggestions/recommendations?

Thanks

Jim

JAZZCNC
23-01-2014, 05:54 PM
One other question regarding settings in Mach, in the motor tuning the settings for step pulse and direction pulse (1 to 5) I have set to 2 in both boxes from my documentation. A tutorial video on you tube suggested that these are normally set to 0, any suggestions/recommendations

Depends really on your drives so If your manual says 2 then leave it set at that. If you don't know what to set at then just set the Dir slightly higher than Step so 2 & 5 or 0-5.

Kernow Jim
23-01-2014, 07:18 PM
Thanks Jazz, I will leave them set as they are as things are running OK.

Thanks again for all your help and advice.

Jim