Machsupport Forum
Mach Discussion => General Mach Discussion => Topic started by: prototype on January 26, 2012, 01:28:16 AM
-
is there a formula to know what the accelleration and velocity need to be set at? when i set it to 20ipm feedrate, not sure if its right or not. and its not cutting correctly in arcs, circles, or engraving. seems to be gaining steps. also i have servos but not sure why encoders arent recognizing that it is out of position. im hoping it can be fixed by getting accel and vel correct. Thanks
-
With Servos you should be able to work out the velocity easily assuming you know the voltage you are supplying and the v/1000rpm of you motors and what the max RPM of the motor is. Acceleration is just a case of trying and seeing. If you are getting out of position then your servo tuning is bad and if it does not fault then the drive has the following error set too large.
Hood
-
I am going to get a voltage and ohms reading on my power supply tomorrow. I have been debating on getting a laptop to run my machine on. If it meets system requirements will I be able to use it?
-
If you are using the parallel port then a laptop may or may not work, if using an external controller such as the Smoothstepper then a laptop should be fine.
Hood
-
is there a formula to know what the accelleration and velocity need to be set at?
Yes and no, by that I am meaning that it is not straight forward calculation since acceleration, velocity, steps per, and other parameters need to be considered and they all
are relative to a specific machine and what you are going to do with it. For the typical user it is just easier to do some testing and use the values found. But in my opinion that
is not giving the user an understanding of what it all means, how it relates to Mach configuration, and thus the same questions will continue to be posted.
In light of the above, I started a writeup that eventualy will be put in Members Docs providing a discusion of the big picture.
RICH
-
Is there a certain factor that would make the laptop not work? Or is it just a gamble? Thanks
-
The newer the laptop the less chance it will work but some old ones wont either so its just a chance you would take. It is mainly the power saving features of laptops that cause issues with the drivers timing and also the onboard graphics. As said this should not be a problem if using an external controller with the laptop, in fact Artsoft only officially supports laptops if using an external controller.
Hood
-
I found someone that retrofitted the same mill I have to mach. So the hardware on the machine is identical. Only difference is the drives, Bob, and encoder count. Since the same servos, belt reductions, lead screws and all of that is the same, will my steps per unit, acceleration, and velocity be the same?
Sent from my DROIDX using Tapatalk
-
No, drives, power supply to them and encoder count etc can all make a difference.
Hood
-
The input input values in motor tuning change.
There are three parts to the total system, and to keep it simple, they are:
Mach - used loosely here, it includes the PC ( which includes how well the pc will function for using the controlling software, namely Mach ), controller ( which includes drive type, power supply ) , and all the rest of the items which touch the generated pulse stream
MACHINE - Its does the task and is made up of mechanical items (all those items which make up an axis) and electromechanical ( motors which must provide power to do work).......which in this case is identical....but adjustment of the machine can influence preformance.......so lets assume identical
MACHINING- is the task to be done which requires that the software and machine are capable of providing for it.
All the parts are interelated, changing one part affects the other parts of the total system.
RICH
-
Would I be able to list all of my data and you be able to get me in the ball park?
Sent from my DROIDX using Tapatalk
-
I don't don't use servos and the tuning is different than tuning steppers. That said, post if you wish and see
where it goes.
RICH
-
Accel is impossible to say, its machine/motor/drive /tuning dependant, Velocity should be possible to give a close value for but again servo tuning could affect that.
Hood.
-
HMMM....so we are right back to what the orignal question was and instead of echoing what has been posted i'll try again.
Yes, one can calculate, but one must be aware of what and how the calculations are made and what they are telling them based on
varying inputs.
In practical applications one defines an acceleration desired and finds what is required to achieve it …or … accepts the value based on some other parameters.
That said, if one has all of the information required they could use one of the computer programs to calculate / evaluate a system. The problem is that unless
the data is very close to what the program uses for information and comparison ( motor / drive system ) you will just be playing with numbers. The manufactures
programs provide a motor choice to fit the conditions, but in the end only actual testing ( motor tuning ) confirms the reliablility. The three M's all interelated.
Maybe try out one of the programs and see if you really have the data and understanding required.
Have fun,
RICH
-
Thanks for the input guys. I will just start playing with settings and making adjustments. Someone had mentioned to me ( in person ) that I need to be careful of "noise" in the lines and possibly getting some type of magnet to put at the pc connection and the servo connections. But he doesn't know what the magnets are called so I have yet been able to find them. I am going to find them and put them on. What other precautions can I take to reduce my risk of noise causing problems?
Sent from my DROIDX using Tapatalk
-
I would imagine the "magnets" he was talking about were in fact ferrites and not magnets.
Hood
-
What other precautions can I take to reduce my risk of noise causing problems?
Shielded cables. Eliminate or reduce the noise source if possible. Using chokes, those ferrite magnet things , can help but depends on what is actualy happening.
RICH
-
Im going to measure my wires and order Ferrites tomorrow. My system uses differentials ( I think its what they are called) to run the 4 wires from the encoder to an rj45 plug. Then cat 5 wires to the drives. Do they sell shielded wire for cat 5?
Sent from my DROIDX using Tapatalk
-
If your encoders are differential then shielding is not really required as the differential signalling is very noise immune.
Hood
-
That's what they are. Differential line drivers. Where else may a noise issue come from? All of my step and direction signals are ran through the flex that the servos were wired through with its original setup.
Sent from my DROIDX using Tapatalk
-
Step and Direction should be shielded and the shield properly grounded to a star point in your control cabinet.
Personally I use differential step/dir as that saves having to be extra careful and no shielding is required, just twisted pairs. Obviously you need to make up or buy a line driver to do that and also if your drives dont accept differential step/dir (mine do) then you will need to make or buy a receiver to go at the drives.
As said though if using single ended step/dir as long as its properly shielded and grounded it should be fine.
Hood
-
I have been playing with settings and its still overtravelling the same amount very consistently. My steps per unit for x and y is at 15000 and z is at 20000. With a dial indicator its correct if I run each axis individually. Just running a program (fanuc post) that runs multiaxis at a time that it acts up. I don't see anything in the code that looks out of place. Its in inch and absolute program.
Sent from my DROIDX using Tapatalk
-
Can you attach your xml and also what line encoders, any reduction between motor/screw and what pitch screw? Also what type of motors/drives do you have?
Hood
-
i attached xml and replied, it said that name existed so i renamed and sent again. there seems to be a delay so im not sure if it posted. i will check back in a bit. thanks
-
i have posted this twice but havent seen it on here, gonna try again. ;D i have 500 line count encoders, which would be 2000 counts per rev in quadrature. the encoders are mounted to the motors. thread pitch is .200 on x and y, .18 on z. there is a belt reduction of 2:1 on all axis' . im using dugong drives from cnc4pc. i have SEM motors ( 30M4-59 ) max rpm 2400, max volts 142, pulse amps 30. i checked and they have 110 volts being supplied to the dugongs. thanks again ;D
-
X and Y axis calcs from your figures
500 x 4 x 5 x 2 =20000
Z axis from your figures
500 x 4 x 5.5555555 x 2 = 22,222.22222
So seems like you have your steps per units set wrong on all axis.
For your velocity you should be able to get in the region of 1690 rpm so for X and Y that would equate to 169 IPM and Z 152 IPM
Hood
-
Thank you. Did the xml look fine?
Sent from my DROIDX using Tapatalk
-
Apart from steps per unit being wrong, velocity being lower than I would expect and also accel would seem slow for what I would expect from servos.
Sounds to me like your tuning in your drives may be needing tweaked.
Hood
-
I guess that my steps per unit being too low meant it was turning the axis' more times than needed thus creating the overtravels. Thanks a million. I will plug those numbers in and check servo tuning. Y and z are quiet but x is a little noisy so im sure it needs it. Hopefully I can test run it again tonight after my regular job. ;D
Sent from my DROIDX using Tapatalk
-
It would actually under travel with the settings you had. For example you tell Mach to move it 1 inch then Mach would send out 15,000 pulses (as thats what you had set in motor tuning) but to actually move 1 inch you would need 20,000 pulses so the axis would only move (1/20,000) x 15,000 = 0.75 inch
I am not sure but I think the Dugongs have electronic gearing in them, maybe you have that set to something other than 1:1?
Hood
-
I will check into the dugong gearing
Sent from my DROIDX using Tapatalk
-
Just curious. Are you guys self employed machinists?
Sent from my DROIDX using Tapatalk
-
I am self employed, mainly repairs to fishing boats, do welding (Stick, Mig, Tig) Milling, Turning and general spanner work. Used to be a fisherman myself until I got tired of being thrown about and soaked with freezing water 7 days a week ;D
Hood
-
Cool. I used to work with a diemaker who is now self employed in Michigan working on boats. Mostly sailboats.
Sent from my DROIDX using Tapatalk
-
The guy that sells the dugongs said they don't actually have adjustments. Just gonna start tuning servos again and cross my fingers.
Sent from my DROIDX using Tapatalk
-
If u pull hair out, does it grow back? I retuned my servos. I believe them to now be correct. But now im getting zero motion from mach3. ????????????
Sent from my DROIDX using Tapatalk
-
Afraid I dont know anything about the Dugongs setup but kind of sounds like you may be in tuning mode which only accepts commands from a tuning programme?
Hood
-
I was able to get movement before I came to work. The tuning looked good but was too weak to push servos I guess. I went in and raised all the numbers and got the graph looking ok again. Still not travelling correct distance at 20000 and 22222.222. Getting a little discouraged but going to fight with it again tomorrow
Sent from my DROIDX using Tapatalk
-
How much out is it?
Hood
-
Not 100% sure. Today I zeroed over a dot I have in the center of my material. I let it run an engraving program with z raised up. When it returned to zero at the end of the program ( which had many direction changes ) it was off to the left and down over. 500 each way. Going to use dial indicator tomorrow
Sent from my DROIDX using Tapatalk
-
Off subject again. 8) hood, at what point did you know it was time to invest in your equipment and be self employed? Thanks
Sent from my DROIDX using Tapatalk
-
When I got tired of getting soaked to the skin by freezing water and getting thrown about every day of life ;D
Hood
-
Yeah. But you already knew you could make enough money and get enough work to live on?
Sent from my DROIDX using Tapatalk
-
Who says I do make enough money ;)
I had been fixing the boats I worked on for years and also fixed some of the other boats if time allowed, ie bad weather when we couldnt get to sea, not many of those days though as we tended to go out in anything but the most severe of weathers :D
Anyway I got a wee boat of my own but it was too small to work in the winter so I had a lot of spare time then and ended up being asked to weld and repair other boats. Decided I liked that better so decided to pack up the fishing and do the welding/machining full time.
Dont make as much money, no where near but I like it better and have enough to pay my bills :)
Hood
-
Thanks. I've been asked to do lots of small parts I've been turning down. May have to break down and get a small lathe and put a dro on my machine until I can get the cnc accurate
Sent from my DROIDX using Tapatalk
-
I am going to order a digital readout for the mill. Hoping that will help me get steps per unit calculated
Sent from my DROIDX using Tapatalk
-
Nice to have a DRO but you should be able to calculate your steps per unit without it. IIRC, Hood likes to calculate it rather than use the calibration tool with Mach.
-
Yes I prefer to calculate as it means it will be as accurate as your hardware where using the utility will be as accurate as your measuring instruments and skills, most hardware is more accurate ;)
Hood
-
Hopefully with the scales I can tell it to travel 14 inches and know how far it actually moved within. 0002
Sent from my DROIDX using Tapatalk
-
Got readout installed last night. Hoping to check cnc accuracy and get it calibrated tonight
Sent from my DROIDX using Tapatalk
-
I have been trying to calibrate steps per unit using mach 3 settings and checking with digital readout. Not consistent. It will be off. 018 going in positive direction and. 786 in negative direction.
Sent from my DROIDX using Tapatalk
-
Those numbers are off using 10 inches of travel
Sent from my DROIDX using Tapatalk
-
Calculate the steps per unit from your hardware specs.
Hood
-
I will set x and y to 20000 and z to 22222.2222 . Then hopefully adjust settings until its accurate. This will fix the problem of traveling different in positive and negative directions?
Sent from my DROIDX using Tapatalk
-
It is set at 20000 steps per unit. It is consistently travelling .990 in one direction and 1.06 in the other.
Sent from my DROIDX using Tapatalk
-
As Hood stated your hardware specs determine the correct step/unit, other methods are approximations subject to the accuracy of your measurements. You might want to take a look at a stepper calculator http://www.machsupport.com/forum/index.php/topic,16315.0.html
to verify/confirm your calcs.
Also are you allowing for backlash in your measurements? Backlash could account for what you are seeing.
al
-
Im using a digital readout for the travel measurements
Sent from my DROIDX using Tapatalk
-
I got the steps per unit correct in the settings. I entered 3 inch moves and it was correct with the digital readout. But when I went to mdi and entered a motion it over traveled again. Feeling like its never going to be right to be honest.
Sent from my DROIDX using Tapatalk
-
Now its moving correct with straight cuts but not angles or arcs. I need help PLEASE. I have to get this fixed and im learning as I go.
Sent from my DROIDX using Tapatalk
-
You never answered in regards to backlash. After you got your distances traveled correct did you measure for backlash? Do you have it entered and enabled? What % is your arc slow down set to? Do you have CV or Exact Stop selected?
-
It does help to read the date :-)