Machsupport Forum
Mach Discussion => General Mach Discussion => Topic started by: Fishbone on September 30, 2019, 03:34:46 PM

I am sure this has been dealt with many times in the past but I have read so many post on this subject and so many have led down dead ends and were slightly different than my issue. So please forgive me if this redundant.
Here is my hardware;
Old Meistergram engraving machine built like a tank.
Old Dell running WinXP
Vexta PH26821, 5V, 1.5A, 1.8*, 125 Oz torque. Used on X and Y.
Sain Smart TB6560 BOB, ST330 Motor driver. (I know, but the motors run fairly well using these).
Setup in Mach3. Driver test passed, excellent.
Motors tuned. In both 8 step and 16 step. Running smooth and reasonably fast. No weird noise. I have set torque and decay on all the different setting and found the best for each.
The current has been set at max of 1.5A and 1.0A with no change.
My figures came out 1600 steps for 8 step micro and 3200 steps for 16 step micro. (200 x 8 =1600 and 200 X 16=3200) I have also tried putting a guesstimate of 2000 steps and tuned accordingly.
Units are set in inches. All of this testing and changes have taken more hours than i can recall.
So here is the problem.
When i go into setting and use the Wizard for setting steps per unit it has not come close to the unit in inches. I have used 1", 4", and 12". the travel is like a 1/4" to 3/8. When i accept the recommended steps from the Wizard
and run the setup again I get a very similar distance. maybe 1/32 longer. The Wizard has adjusted the steps up to over 300K and no change and not close to inch units.
So where have i made the error in setup because i know 99.9% of the problems are operator error. Are my step figures correct.
The only thing i haven't done is reload Mach3. The reason i haven't reloaded is i have set up a Dyna 2400 milling machine with Gecko drives and it runs fine with this version of Mach. Oh and the machine is not connected to the internet so it is not polluted, but makes it more difficult to update.
As always any help will be greatly appreciated and i know some of you human Wizards will be able to guide me through

My figures came out 1600 steps for 8 step micro and 3200 steps for 16 step micro. (200 x 8 =1600 and 200 X 16=3200)
These would be you steps per revolution of the motor shaft.
You then need to factor in any reduction through gearing, belt/pulleys and screw pitch to end up with the actual steps per unit (inch).

So the TPI is 8 which I believe is a pitch of .125. Do I then multiply the revolution by that number?

3200 x 8 turns per inch = 25600 steps per inch, if direct coupled to the screw.
The manual explains the procedure in great detail that is easy to understand. Do you have it ?

Well there it is. Right in front of me in plain sight, I do have the manual and I followed along for awhile and then read the wizard would take care of it. The short answer is I tried to short cut. The problem with that is it blinds you from what's right in front of you.
Thank you for the info and the patience. I needed it!

The short answer is I tried to short cut. The problem with that is it blinds you from what's right in front of you.
Very true. The steps per wizard is a perfect example of something that shouldn't even exist IMO. Folks that know what they are doing don't need it and folks that need it don't know enough to use it. But if they do it gives them a false sense of security (they assume their problem has to be something else). The only correct way to set steps per in motor tuning is to do the math (plain and simple, no 2 ways about it). There are too many other variables that can make one think the steps per are spot on when they could be off significantly. To add insult to injury any error in the steps per is cumulative. So, if you set you steps per incorrectly because you relied on a measurement that was flawed due to skipped steps, backlash, flex, etc. However much it is off in the distance used is multiplied by the times that distance will go into a move.
The best most of us have that is even capable of measuring close enough to do a good job checking steps per is a dial indicator with 12 inches of travel. Take the best indicator in the world and take measurements and adjust until you get the steps per perfect and you can still be way off if some of the related variables are not accounted for. Lets just say its off .010 in one inch of travel because backlash while measuring wasn't avoided. At 8 feet it would be off almost a complete inch (.980). Steps per can get out of hand real quick and the only right way to do it is doing the math.
Tons of material available to teach folks how to properly tune motors. Tons on this forum, in our docs and other places online. Anyone attempting to tune motors need to spend the time to understand what it is they are doing and how things not thought of can give false readings. There really is no shortcut to do it right. Sorry you learned this the hard way but glad you learned.

Hi,
there is one occasion where the steps calculator can have an advantage.....where you have a rack and pinon or a belt/chain drive.
You can do the calculation but it requires an exact 'pitch' diameter of the gear/sprocket/drive wheel. The pitch diameter will be
close but no identical to then outer diameter. My suggestion in those circumstances is do the calculation and get to within a few percent
and then use the wizard to refine the calculation to the required accuracy.
Craig

Echoing Brett and Graigs replies,
1. One should always calculate the steps per unit and understand that the number of steps
calculated is theoretical.
 Learn how to calc the steps per, no excuse's!
A link to a steps per unit caculator:
https://www.machsupport.com/forum/index.php?topic=16315.0
2. You must confirm the calculated steps per unit. Understand, that when you test, quality of axis components and their installation come into play. The testing is only as good as ones ability to accurately setup and measure a distance based on a physical standard and the method they use. For example:
Theoretical resolution for a steps per unit of: 20,000 ( 1/20000) = 0.00005"
" " : 253,000 = 0.000004"
? is your axis components linear travel that accurate and can you measure to actualy know.
? How much total backlash do you have and is it consistant? ALL axes have backlash it's just a
matter of degree. Total backlash is the sum of each individual axis component non
movement( ie; screw with nut, belt, coupling, gears, etc.) and each can vary depending on
operating conditions.
Practical setting of the steps per unit will be a value that satisfies their machines overall machining accuracy. The value can be as calculated, as physicaly tested, or a value close but different than both of those.
3. Think of the wizard as a software assist to check / adjust the calculated to the tested.
RICH