Messages

111

General Mach Discussion / Re: need mach motor tuning explained better

« on: March 06, 2009, 05:39:43 AM »

John - Looking at the Gecko, these are 10 microstep devices, so you have 10 micro steps, then 200 per rev for the motor - this equals 2000. And you said previously you have a 4.8 to one reduction to the final gear wheel.

The problem seems to be calculating how far this will move your table down the rack.

Here we are going to have to impinge on measuring (which I have just pooh poohed in the last post) although I am sure you could tell me the pitch. However, although I don't know where you are, this could be metric or imperial.

What I would do is paint one of the teeth on the rack white, and a slot on your pinion white, or mark with a felt tip pen will do - just so you have no doubt where you started. Enter 9600 in the steps per unit on the Mach motor tuning page, and keep the speed down fairly low, say at 10 units per minute. Keep acceleration at 1 upsps. Then I would try the command on the MDI line (with everything zeroed) of G0 X10 (you seem to have a fairly big table).

The cog should have turned 10 times, and the marked tooth should be back on the bottom. Accurately mark which tooth on the rack the cog has reached, and now measure the distance from start to stop. You can do this with a ruler, since we are looking at 10 times the unit size, and as I said before this should be a fnite number (of some sort). I you lay a tape along the rack, then, as Hood said, the carriage should have moved 160 teeth up the rack (if it is a 16 tooth cog) - what is that distance - 10 inches - fine. If it does not line up with inches - try millimeters - 100, 500.

If it is 10 inches, then your pulses per inch is 9600/10 = 960 (although this seems low, since it gives an accuracy of less than a thou) - but you get the general idea.
Calculate your pulses to give one turn of the final cog, then if you are saying there are 16 teeth on the final cog, then what is the distance between 16 teeth- because that is your unit - if you get that far, converting into inches or mm's is just maths.





 

112

General Mach Discussion / Re: need mach motor tuning explained better

« on: March 06, 2009, 04:58:02 AM »

Bill -

I don't know - my problem is that I am pedantic (that means stubborn Hood). I see most things as black and white, and in the case of the number of pulses it takes to move something 1 unit, then this is one of them. I don't care what the unit is, or indeed any of the other figures, but the final figure must be finite. So - if I have the same components as you, I need the same number of steps as you do, to move one inch - not four or five more or less, the answer is the answer.

Since everything is made to a finite tolerance these days (after Whitworth), this means that all our equipment can be accurate. I could understand where, say, you made your own leadscrew, and were unsure of the final pitch, but that is not so with modern items. The only time I saw this was on the American market recently, when some enterprising sole was selling ball screws said to be 5 turns per inch, when in fact they were imported metric screws at 5mm pitch - and totally ignoring the 0.4 (16 thou) difference.

The facility for measuring, on Mach 3, is one which I have never used, but is an ideal way of checking your calculations BUT you can never be entirly accurate when measuring, even with some pretty good equipment.By this I mean if you have calculated 60,000 pulses per inch and you check by measuring and get the same answer + or - 60 pulses, then you can be sure your 60,000 is accurate - if you only get 45,000 then you have made a serious error somewhere and need to check your figures.

I repeat that this facility cannot be for "tuning" the number of steps per inch, becasue - as I have already said, if I have the same equipment as you, I must get the same answer.

I'll have to post this now, so I can read the rest of the posts

113

General Mach Discussion / Re: Newbie needs help

« on: March 05, 2009, 05:56:52 AM »

If you are saying that Mach 3 says you are cutting at 0.0625 and you are infact cutting at 0.375 then you would appear to have an offset working - either a tool offset, or a machine offset.

You do not say whether you have fitted home switches - but I will assume you haven't. I do not know Cut2d but you must make sure that you do not put anything in the program that is a tool offset, or length.

Jog your axis to where you want to start your program - i.e.the X0 Y0 Z0 position of the program. This is usually, but not alwyas, the bottom left hand corner of the table, with the cutting tool resting on the workpiece. Check that this is correct for your program

Press the Machine Code button on Mach 3, the surround will light - "RefAllHome" and the Machine Co-ordinates will go to zero. Press the button again, and the surround will go out. The display is now showing Program Co-ordinates - these should be zero - if they are not, zero each one by pressing the zero button at the side of the DRO.

You now have Machine and Program co-ordinates all at zero - i.e. there are no offsets on your machine, and if you check Config/Fixtures the table should be clear. Check, while in there, the tool table - this should be clear as well.

If you jog your cutter clear of the work, and press cycle start, it should now cut at the correct depth. If it does not, Cut2d must be putting in some offsets. Stop the machine and check the offsets table and see if anything has altered.







 




 

114

General Mach Discussion / Re: need mach motor tuning explained better

« on: March 05, 2009, 05:15:36 AM »

If you are saying that your machine does not move the correct distance - say 1 inch, when you tell it to move 1 inch, then this is down to the correct number of steps per inch in your motor tuning page.

HOWEVER - you are NOT tuning it.

Steps per inch is a finite number, there is no discussion, no alteration, no adjustment etc, - just a number. It is made up as follows - all numbers multiply the previous answer.

I assume you are using stepper motors. Check, but these are normally now 200 steps per revolution = 200
Your drives will have a microstep facility - e.g. Gheckos at 10, others 4 or 8 or 16 - Ghecko =10   = 200 x 10 =2000
If you have any gear reduction between the motor and leadscrew - I have a 3 to 1 belt reduction = 2000 x 3 = 6000
The pitch of your leadscrew or whatever - mine was 10 turns per inch  10 x 6000 = 60000

For your system - assuming steppers and microsteps - start at 200 for the motors, I will leave out the microsteps for the minute because you don't say what you are using.

If the gear ratio between the stepper and the drive wheel is 4.8 to 1 then  200 x 4.8 = 960
Irrespective of the number of teeth, and the pitch etc. How many turns does your final gear wheel have to make to move down the rack by 1 inch. I don't understand what you mean by 1.0 pitch - is this one mm,cm,inch - my catalogue shows so many variations, I dont see one which matches yours.

In any case - the thing you need to know is - how many turns does the wheel do on the rack to move one inch - multiply 960 by say 1 - if the rack pitch is 16 and the number of teeth on the wheel is 16 = 960

And finally the number of microsteps your drives use - say 8 or 10.= 7680 0r 9600 etc

The thing I must impress on you is that this number is not negotiable. You cannot now try it and  measure the distance the rack moves and adjust the number of pulses, either you have it right, or you have it wrong. There are several reasons why your axis do not move the correct distance, but steps per unit is not one of them.
By all means try and measure to check you are reasonable accurate. - move the carriage to the right - set up your measuring equipment - I use digital calipers - move the carriage by an MD! move again to the right eg. G0 X1 - so x will move exactly an inch, and then measure. On my steel lathe, I expect to be to 1 thou - but I can't measure any more accurately than that. If you always move in the same directiom eg move right, measure, move right, measure you take out all the backlash from the system, i.e. the gears, belt etc have taken up all the spare movement in that direction.

If you make the final measurement to the right, then (again using the MDI) reverse the movement and measure the answer should be 0, but will not be, becasue the gear, belts etc all have to settle in to push the opposite way. This is called backlash. All systems have it (althoug on some it is so small you cannot measure it) and Mach 3 can compensate for it, but that is another story.

To make you testing more reliable, decrease the speed of your traverse on the motor tuning page - I don't know what you have it set at, and decrease the acceleration. This will reduce the chances of the motors missing steps. Once you are satisfied that the machine is accurate, you can up the speed again.

Sorry to be so pedantic - this is one of the Bees in my bonnet - see previous posts on the subject.

115

General Mach Discussion / Re: Missing identical steps on each axis

« on: March 04, 2009, 04:46:37 AM »

As Hood says, a little more precise information would help - but it depends on what stage you are at in your build.

Missing steps could be exactly that - the computer is putting out pulses which your motor/drive cards are missing.

If on the other hand you are running out say 25mm and then running back 25mm and finding the axis short of the start - then that could be backlash

116

General Mach Discussion / Re: home switches gone crazy

« on: March 03, 2009, 04:28:00 AM »

I think you have answered your own question.

The four wire stepper motor drive wires are handling a heavy current for a split second and so produces masses of what is in effect magnatism. If you then pass a copper wire through a magnetic field, you get an electric current, so, in effect laying the wires neatly side by side, you are exposing your switching circuits to all the interference that your stepper drivers can muster.

My stepper driver wires are flat, heavy duty copper loudspeaker cable, which comes in a four wire strip. These are looped to the three motors, and becasue it is heavy duty, designed for rough usage, it just lays along the bottom of my machine tray.

The switches all route through shielded multicore cable, but again, I do no mix those with quick pulsing detectors (index wheel etc) with my limits and home, which are essentially slow on/off.

You might find, however, that if you increase your debounce timings you will get rid of a lot of interference, if you still want everything nice and neat.

117

General Mach Discussion / Re: Geckos Faulting

« on: March 03, 2009, 04:14:41 AM »

Ditto - and what exactly is the fault - do they keep stopping??

What are your temperatures like after  6 hours continuous work, or does the temperature at that particular time of day rise. Dont know where you are or what it is you are doing - but I see Iron Works  - phew!

Have you tried ringing Ghecko, I think they are pretty helpful.

118

General Mach Discussion / Re: Constant Surface Speed.

« on: March 03, 2009, 04:07:15 AM »

Hood - Sorry, I was out all day yesterday.

The CSS worked very well, and it seemed to me, very smoothly. The wheels are 6 inches in diameter and I was cutting into the back by about 0.25 to lighten them - they are 1 inch thick. I wanted to start at the centre about 0.5 from the axle hole, and come out to a point about 1 inch from the rim.

I wrote the script that angled into the work, then ran parrallel to the  finish.  Botched the first one a bit (actually I needed a new tip on the tool) but after that it was a matter of mounting the hweel and pressing cycle start. I did all eight wheels in no time.

I can't really comment, I suppose, becasue the speed is constantly changing as the cut moves, but it seemed alright, and gave a nice finish.

As you know, I don't have a Bob, my index (single pulse) is optical reflector, straight to the computer.

119

General Mach Discussion / Re: Mach3 question: the ratio between mach3 and cnc machine

« on: March 02, 2009, 02:36:46 PM »

In Config/Motor Tuning, you need to set up each axis with the proper number of steps per unit. I assume, from putting in a command like X1000, then you must be in mm's.

This should be set first under Config/Native Units.

For each motor then you must tell the computer home many pulses it has to put out to move your table 1mm. This is not difficult. If you are using stepper motors, then they will probably need 200 pulses to turn one rev. Your drive units will be able to introduce micro steps, and for smoother use, you should choose 8 or 10, usually by adjusting the DIP switches on the card. Lets say is is 10. This now means that you need 2,000 pulses per unit.

If you have any gearing between your motor and your leadscrew then this must now be calculated - on my machine I have a step down belt drive of 3 to 1 - so in my case this would now be 6,000 pulses. If you have no gearing, then no multiplier.

The final thing is the pitch of your leadscrew - mine is 5 mm pitch therefore I DIVIDE by 5 to get the result - 1,200 pulses per mm.

Each axis might be different, so make sure you calculate each one.

Start your speed setting at a reasonable speed say 100 mm per min, (acceleration about 10mmpsps) and gradulayy increase this until the motor clearly start missing steps, then come back down until the run smoothly again.

Do the same with the acceleration.

120

General Mach Discussion / Re: CUT QUALITY REALLY RUFF

« on: March 02, 2009, 02:17:38 PM »

It is doubtful that your code will be in arcs, and is probably made up of many short, straight lines, since the only arc available in CNC is a circle.

To mitigate that, you can set Mach 3 to continuous velocity, which is what Hood said. This smooths out the short lines, by accelerating into the next line, before the previous lines has finished.G61 in your code means absolute stop after each line. G64 means continuous velocity.

I don't know what you are using to generate your code, or what system you are using to turn the plasma on and off, but check the code to make sure it is not turning the plasma off after every line.