[nambass]

I am having a problem that my DIY CNC is not cutting perfect circles. Do anyone experienced this or can give advice why this happens?

Attachment 220488

[ssutton]

If you could shoot a video while your machine is running and cutting this type of material that might be helpful to understand what may be going on.

[nambass]

Thanks, will do!

[LeeWay]

Right. It could be a number of things or combinations of things.

I also think you can help pinpoint what is going on, by machining a softer material. Say foam, plastic or soft wood.

[FineLineAuto]

Right. It could be a number of things or combinations of things.

I also think you can help pinpoint what is going on, by machining a softer material. Say foam, plastic or soft wood.

I second Leeway's suggestion. Using foam is at least going to eliminate the potential problems of the machine not being too rigid enough or using the wrong feeds and speeds.

Are you using Mach 3? Are you using the parallel port to drive your electronic setup? What electronic setup to you have?


Nate
Fine Line Automation
www.finelineautomation.com

[nambass]

Sorry, for being incomplete:

Yes, I use Mach3 on windows XP PC with the following configuration:
- Table size is 2m x 3m
- Mach 3 Break Out Board (parallel)
- 4 x Gecko G203V Stepper drive
- 1 x 42VAC 300VA Transformer
- 1 x HP PSU Board
- 1 x Heatsink
- 3 x 34HSX-104D Stepper motor (2 for X-Axis, 1 for Y-Axis)
- 1 x 23HSX-102D Stepper motor (1 for Z-Axis)

I then installed 2010 Screenset
The Router is a Bosh CE1600

Also attached my mach3 xml file (note I changed the name to *.txt because xml not allowed to be uploaded.

[ger21]

If you're machine isn't flexing, you either have a lot of backlash, or possibly a loose coupling.

[fixtureman1]

Are you sure that both the x steppers are working I have had one of mine drop out and get that.

[louieatienza]

It's either backlash or flexing. If you're using helical couplers their flex can cause position errors as well, especially if they're the dual-helical type (which should only be used for axial misalignment only.) Or, if you buy Lovejoy couplers on eBay, they may have spiders that don't fit well or too soft which can cause this as well; a couple brass shims can snug them up. Or, the setscrews for your couplers are moving or loose; you should have a flat or flats on the stepper shaft, and use LockTite on the setscrews.

If you're using rolled ballscrws with a single nut you might have backlash there as well (despite what the manufacturer may claim.)

Finally, for the most ridgid setup use the shortest tool that gets the job done, with the shank as far in the collet as possible (before the flute fadeout), that will clear any fixtures, screws, or clamps.

[nambass]

I am also worried about that, The machine is very firm. I am using 2 x 90mm x 90mm aluminium extrusions bolted onto the 30mm Aluminium side plates of the gantry. it is heavy and very firm.

I am using 20mm Linear rails with 20mm Linear blocks (2 x blocks per side) For my X-axis I am using a rack and pinion system with a 15 tooth Pinion. The Pinion is connected to a 60T T5 Timing pulley driven with the T5 420 Timing belt. through the 15T T5 Timing pulley which is connected to my stepper motor. The tension of the pinion is set with a screw and tension spring. It could be the tension is not set tight enough as well?

But what I did noticed, when I home the machine - it touch the limit switch (x-axis) and move off the switch. In that process it seems the left motor/side moves little more than the right motor/side. Remember the y-axis is 2m long.

I am concerned about this and think this might be the cause of the problem - maybe I am wrong. To test and see if my eyes is telling the truth, I tried moving the gantry by pushing only on the one gantry side plate to see if it moves more that the other side, and is true - it does move slightly more. I have tighten the bolts holding the gantry together as much as possible and is very tight.

I attached the image of how the extrusions are fixed to the gantry sides. The linear blocks seen on the image is for the Y-Axis.

[LeeWay]

For a slaved axis system, you really want easily adjustable home switches. I'll include an image of how I installed then on my plasma.
Basically the switches are mounted on a plate and then that plate bolted to the machine. In my case, the plate simply has holes and slides in the 80/20. If using a different structure, you can machine slots in the plate instead.
It is critical to get both side of the gantry to stop square with the table when homed after they come off the switch. If they start out of alignment, they will cut that way throughout the job.
This is rather easy when you have adjustable switches. I don't think that is the issue you are seeing though.

Like to see a picture o the whole machine as well as a close up of the Z axis.

[nambass]

Attached is some pictures on my machine.
I tried to show the X-Axis driving mechanism, The design of the Gantry and then also the limit switch. The limit switches sit on the gantry, at the bottom of the side bars which is fixed to the actual gantry side plate. These bars is for additional stiffness.

It seems there is about 1mm movement between the to gantry sides. this means that when I move the one side of the gantry, it moves about 1mm before the other side also start to move.

[ger21]

To test and see if my eyes is telling the truth, I tried moving the gantry by pushing only on the one gantry side plate to see if it moves more that the other side, and is true - it does move slightly more.

If the machine moves when you push on it, then you need to fix that. If it's moving, it's not going to cut well.

It seems there is about 1mm movement between the to gantry sides. this means that when I move the one side of the gantry, it moves about 1mm before the other side also start to move.

Under it's own power? Are the pinion gears held tightly to the rack?

[LeeWay]

If the axes are slaved, they should move synchronously. There should be no delay. A delay suggest a loose drive train part or backlash of some sort. Both of which will cause your issues.
To sync motors correctly, they need to be the same type really and have duplicate settings. Both of the Gecko's need to be tuned as well for good results.

[Bob La Londe]

What are you using to generate the code? Look at the toolpaths carefully on the screen to see if they are wonky. I had that happen once with a Mach 3 Wizard. It did it consistently bad. I usually just use my CAM software now, but I have found the Wizards in the New Fangled group are ok. It was one of the single Wizards that did that.

[nambass]

I am using cambam as well as Aspire 4

[nambass]

I have ensured the settings for the 2 drives is identical. Apart from this, how should I tune the 2 drives?

[nambass]

Greg, when I push the gantry I need to push harder so that the motors also moves. As you can see in the image, I have a tension spring that holds the gear tight to the rack, so yes the gear is held firm to the rack. I can tighten the tension screw some more.

[LeeWay]

I have ensured the settings for the 2 drives is identical. Apart from this, how should I tune the 2 drives?

I'm not really sure how the best way to go about this would be now. The drives are tied together. They should be tuned separately. The instructions are in the Gecko manual on how to tune the pot setting. I suppose it would still work if you temporarily disengaged the slaving and then crept each motor in turn as slowly as you could get. Maybe 1" per minute. That would give you a few seconds to adjust the pot without travelling far or disconnecting anything.

I doubt this is the issue either, but could contribute if the drives are wildly out of tune.

Oh and nice looking machine BTW. Looks like you used some nice hardware and electronics.

[louieatienza]

Looking at the pics: If the gantry leadscrew coupler is similar to the one on the Z axis, and has the same spider... I would change the spider to one of a harder material. Are they Ruland (McMaster)? If I am not mistaken black is probably the softest spider material, and while they're great for some systems where shock-absorbing is necessary, it can cause problems with CNC's. Just another place to check....