[bluumax]

I'm getting accumulation in X+, Y- & Z+.
I put up a page on my site that might help explain better than words.
http://www.bluumax.com/Taig.html
I've had this problem since I got the mill a year ago, I've tried everything I can think of.
I've checked & re-checked the Mach settings, but still might have missed something.
Hope someone can help with this.


Taig 2019ER
Xylotex 3 axis drive & 269 oz/in motors
CNC4PC C11 Breakout board
Mach3 latest ver 3.037
Mill only PC: P4 266, 724MB

Shawn.

[GeneK]

I'm getting accumulation in X+, Y- & Z+.
I put up a page on my site that might help explain better than words.
http://www.bluumax.com/Taig.html
I've had this problem since I got the mill a year ago, I've tried everything I can think of.
I've checked & re-checked the Mach settings, but still might have missed something.
Hope someone can help with this.


Taig 2019ER
Xylotex 3 axis drive & 269 oz/in motors
CNC4PC C11 Breakout board
Mach3 latest ver 3.037
Mill only PC: P4 266, 724MB

Shawn.

Shawn, it's time to send the part to Q. A. and have them verify all the features. I took both files and concatinated them into one file and imported into Mach 3 and each hole is dead center in each square as best I can see. I'm sure if I cut it on my set up it will cut correctly.
It may help some to have the pads cut in the same pattern as the holes are drilled, i.e. vertical columns. It does seem to be related to back lash, and it is being distributed differently since the patterns are different.
Have you quantified how much back lash each axis has and tried backlash comp in Mach 3?

GeneK

PS
Shawn, I was just thinking some more. I would take my height gage and scribe the grid very precisely on the board and then machine the pattern. That should show what is skewed and an idea of how much.

GeneK

[Stepper Monkey]

I haven't had a chance to go through your website or the problem in detail yet, but I see where Genek is going and I think he is likely on the right track.

The exact same item cut with different toolpathing can show marked difference - no matter how little backlash a machine has, you can write code to unintentionally maximize it, and therefore accumulation error. By the same token, you can take a machine with very large backlash errors and write code to minimize it or almost effectively counter it right out.
I frequently cut low-tolerance items like this to consistently under .001" on a very old and abused Taig showing better than .002" backlash, without any compensation, just good toolpath strategy. By the same token, you can easily cut sloppy parts on an accurate machine with a bad choice of path.
Don't know if this is the answer, but it is a really good place to start.

[cyclestart]

Are you positive the part is clamped squarely on the machine? Also that the pads are equal distance from a datum edge? Judging by that website you do very careful work but the question should be asked.

What does a caliper (and square if needed) read for hole alignment. If the part is very small, maybe not much.

[bluumax]

GeneK: I've checked the & adjusted the BL so many times I can do it blindfolded.
If I tighten the nut anymore it starts to bind.
There's only a couple thousands BL in X & Y, Z is not adjustable.

Stepper Monkey:
The program, PCB Gcode decides the drill & route patterns.
Drill is always Y+ to Y- direction.

cyclestart:
I thought maybe the fixture was possibly moving, that was last year.
This is a problem that has plagued me for a year now.

I just didn't want to start posting until I sure I didn't just miss something stupid in setup, etc.
I don't think it's backlash.
Is there a some type of setting I could have set for lead screw compensation that's hidden?
The only thing I can think of is, somehow I got 3 lead screws that are not exactly 20TPI.
I do a ton of trouble shooting with electronics & mechanical, so I found that to be highly unlikely, but still possible.
Remember Apollo 13? It's possible someone assembled it with rejects or something. I've worked as a temp in industry enough to see some really scary stuff.

[GeneK]

GeneK: I've checked the & adjusted the BL so many times I can do it blindfolded.
If I tighten the nut anymore it starts to bind.
There's only a couple thousands BL in X & Y, Z is not adjustable.

Stepper Monkey:
The program, PCB Gcode decides the drill & route patterns.
Drill is always Y+ to Y- direction.

cyclestart:
I thought maybe the fixture was possibly moving, that was last year.
This is a problem that has plagued me for a year now.

I just didn't want to start posting until I sure I didn't just miss something stupid in setup, etc.
I don't think it's backlash.
Is there a some type of setting I could have set for lead screw compensation that's hidden?
The only thing I can think of is, somehow I got 3 lead screws that are not exactly 20TPI.
I do a ton of trouble shooting with electronics & mechanical, so I found that to be highly unlikely, but still possible.
Remember Apollo 13? It's possible someone assembled it with rejects or something. I've worked as a temp in industry enough to see some really scary stuff.

Shawn, you need to quantify what is happening. just tweaking the adjustments means very little if you don't know what you are trying to fix.
Have you used a dial indicator on one of the axises and programmed to move and see if it moves the *amount* you tell it to move, both directions? In other words, I have a 0 _2" dial indicator. I mount it to the spindle with the tip against a block on the table. If I move the table 1"and the dial winds 1", then reverse and see if it goes back to 0. Any diference is the back lash. Do the same for each axis. (also repeat several times and in each direction for each)

The other thing to do is as I mentioned before, use a height guage to scribe a 0.1" grid on the board and then cut and drill. Now you can evaluate how the squares are drifting in relation to the actual position and the drilled holes also.

If the thread pitch were wrong,( not 20 tpi) the pattern would be ok, but just scaled to a different size.

GeneK

[The Blight]

Bl does not acumulate. The only thing that could do this (as far as I can see) is that your X and Y axis is not exactly 90deg apart. A varying pitch in the lead screw can do so too, but I find that unlikely. Because you have had this problem for so long, I doublt that this is the problem. Check if the angle between the 2 axis is correct.

Backlash is added when the axis switch direction. It removes it when changing direction again. It doesn't add the backlash if you continue in the same direction. So the tolerance would then be +/- the amount of the backlash.

[bluumax]

I square the fixture on the table using 2 pins & one in the collet, running it back & forth on the X, (many times) I KNOW that it's square.
It's not measurable with an MDI by running back & forth a few times.
There is a couple thousands slop, but as The Blight said, It doesn't add the backlash.
I don't know of a way to square up the X & Y as they're not adjustable that I know of.
I don't have many good accurate tools to scribe a pattern.
This is just really weird.


Shawn.

[awerby]

First, although backlash itself doesn't accumulate (any slop will be taken up when the axis reverses), inaccurate backlash settings in Mach could very well do that. Essentially, you're telling the machine to add a certain amount to each move in a different direction, to compensate for the distance the screw has to travel before it engages. So if that number is incorrect, you will start to see the sort of drift you're complaining about. I'd start by turning off backlash compensation altogether, and see if the results change any. Later on, when you've eliminated this as a potential problem, you can carefully measure your backlash, put the correct number in the box, and try again.

The other thing that's likely to do this sort of thing is an inaccurate steps/unit setting in Mach. Are you using inches or millimeters? Inches are pretty straightforward, since the Taig is made in the USA, and uses 20 tpi screws (which are quite accurate, actually). 23-frame stepper motors are typically 200 steps per revolution, and (I think) the Xylotex uses 8x microstepping. So you multiply 200 x 8 x 20 to get the number of microsteps in an inch (32000), which should be entered in the appropriate place in Mach. If you're using millimeters, make sure you're converting correctly. You need to divide the figure above by 25.4 to get the number of microsteps in a millimeter; it comes to 1,259.8425. Is that the number you've got set as your steps per unit in Mach? Rounding it off would cause the sort of accumulating error you're experiencing.

Andrew Werby
www.computersculpture.com

[bluumax]

This is the way it's cut, NOT a goofed scan

[GeneK]

This is the way it's cut, NOT a goofed scan

Ok Shawn that scan shows me a lot. I think you need to check the gib on the Y axis. It is probably moving in and out with direction changes, and in one direction it is binding and causing the Y axis to lose steps. I don't know just how the gibs are set up on the Taig, but it sure looks like a comon problem with tapered gibs that have slop. On the Sherline the gibs is positioned by an L shaped wire and the end can be bent so there is a bias and takes out the backlash *in the gib*.
One thought that I have to test what is happening is to set your 0,0 and run the program several times 'in the air' and then return to 0,0 and see if there is any lost steps. your error should not excede your actual backlash. If there is a greater error after several runs vs only one run you are losing steps.

GeneK

[wongstein]

I had another thought. Mach3 doesn't differentiate whether you're using 8 pulses per step with a 20tpi leadscrew, or if you're using 4 pulses per step and a 40tpi leadscrew - it just knows to make 32000 steps for each inch of travel. So there's nothing to say that it's going to always send pulses in pairs (or groups of 8, which would be ideal).

So every time there's a change in direction, you could lose steps that way. This wouldn't matter much if your code sends the steppers forward and back the same number of steps each time, since you'd lose the same number going each way, but since you're drawing all those little squares in columns, the number and length of the movements that you're making on one axis is different in one direction from the other (because of the columns and rows of squares). This means you're losing more steps in one direction than the other during your router routine. Hence the drift along both axis.

Also,

On the other hand, your drill pattern comes out almost perfectly square - because there are fewer movements that aren't balanced by an equal opposite movement (and there are fewer movements altogether).

Basically, I think you're getting a rounding error by using 1/8 microstepping without explicitely telling the software about it. Why don't you try half stepping and see if it helps? I really don't see how Mach3 can send 1/8th microstepping effectively without a proper setting for it.

Good luck,
Anthony

[juzwuz]

I finally looked through my EMC2 settings for acceleration and I have my default set to 2 in/sec^2 with a max of 2.5 in/sec^2 and my max velocity at 30 in/min. I tried higher accelerations before and I always lost steps.

[neilw20]

I notice you are using an incrementing looping program and using G90.
So ever loop around it it adds 0.1" in a particular axis.
That's 2.54mm. Assuming internally Mach3 is done in metric (and is probably using floating point to make things even worse.) then at 32000 steps per inch that is 1259.(84252) steps per mm.
Seeing as how we can't send 0.84252 of a microstep to the axis, if we want to move 1mm we send out 1259 steps. If we want to move 0.1" we send out 319786 steps, which is 22 steps short. That is 0.0006875" error.
Not a noticeable problem if we are off positioned by that much on one pad.

This error could also introduced by the same sort of scaling problem in a PCB package if this is creating the code. If you use the earlier Protel PCB versions, and select metric, it can be out by 1.016% because the used 25000 binary steps internally so that there was no display rounding.
If you laid out a PCB in metric, then changed to inches, and plotted you PCB subtle errors would appear. Selecting inch or metric should only ever change what is displayed, not the absolute unit of measure. To support inch and metric to 0.001"=(2540 units) and 0.01mm=(100 units) you need some strange units so that both types of units can be represented with no errors.
Using incremental coding as you have shown in a fragment of program, every time we make a 0.1" step in the routing we are actually 22 steps short of the correct position. Due to the incremental G90 Mach3 re zeros itself (100 times for 100 pads). Now we have lost 2200 steps which is 0.06875" and scaling on your photo confirms almost exactly this amount of your error.

Try and hand code some linear code. DO NOT USE INCREMENTAL CODING over many steps.
Incremental code using G90 and Lxx for looping is very easy, but try the long way first.
It makes for a quick and easy program, but subtle scaling errors can occur internally in the program.

Your hole drilling does not appear to have incremental errors, but the pad pattern does, and it very closely matches the above numbers.

Forget the backlash and any other mechanical problems. You could put a digital counter on the stepper drive step signal to confirm the actual number of steps for 0.1" (should be 3200) did occur for drilling, and didn't occur for the pads (3198 steps). That's all it takes.

[bluumax]

I'm not using backlash in Mach, it does weird things, & I shouldn't need it for less than .005 anyway.
Steps are set correctly, 32000 (1/8th step).
It doesn't matter if I run 2 IPM or 18 either.

[GeneK]

I'm not using backlash in Mach, it does weird things, & I shouldn't need it for less than .005 anyway.
Steps are set correctly, 32000 (1/8th step).
It doesn't matter if I run 2 IPM or 18 either.

Shawn, Mach does weird things with backlash in
CV mode. Turn off CV mode to run with backlash comp.

GeneK

[bluumax]

I am losing steps then.
I set the step pulse to 2uS before I did this board, also, just incase.
I'll check the Y, but I'm pretty sure it's not binding as I've been @#$%ing with this thing so much looking for anything that could cause it.
I'll check it now.

[bluumax]

It feels smooth, no binding.

[bluumax]

X & Y are the same, Z max vel is 15

Soft limits were off anyway

[bluumax]

I've tried it with CV & exact stop, but no noticeable change