This is the learning curve............
Disappointing...........
1.) UCCNC will only make your workpiece as smooth as you setup the tolerances if you work in constand velocity mode (G64), the CV parameters in the general config or the parameters set with the G64. As far as I remember the default is something like 0.3 units tolerance which is fine for this type of work what you doing if you working with millimeters, but is a large tolerance if working with inches. 0.3 inches of tolerance is rough.
2.) Your VFD might not be able to read so short pulses. ( 1/400000 = 2.5usec )
3.) ???
4.) G53 Z7.5 is at Z7.5 in machine coords. Where it really is depends on your home coordinate. Did you home the machine prior to the job? Is your Z home coordinate proper?
This is the learning curve............
Disappointing...........
Been doing this too long
Too bad about your bowl. I remember driving the bit through more than one work piece when I was first getting started too. Now when I want to test run a critical project I'll get a piece of styrofoam and try it on that. Its hard to ruin a tool or damage something cutting styrofoam.
FYI, I run my kernel speed at 400Khz and my spindle step/direction settings are: 41 step per rotation and; 16000 acceleration steps s/^2.
DDgitfiddle
Remove the G53, you don't normally use a G53 in a move like this
The jerk factor, this really shows up when your servo motors have a low pole count,and Encoders have a low resolution, there is no way to change it, without using a higher resolution Encoder, and better quality servo motors, the lowest I always recommend is no less than 12 Bit 16 Bit= 64384 PPR, is a better starting point, anything lower than 12 Bit, you will see and have the jerk factor, there is no substitute for high resolution Encoders, if you want smooth machining
Mactec54
David,
I have not played with Smoothing in Fusion, but I will try. What you describe for air cutting is exactly what I was trying to do...it's just that first move that messed me up. After reading some of the other replies, it looks like I need to rethink my post-processor a little.
Thanks,
Robert
David
Romans 3:23
CurlyWoodShop - www.etsy.com/shop/CurlyWoodShop
David Falkner - www.youtube.com/user/difalkner
difalkner - www.instagram.com/difalkner
I'm not a Fusion 360 user, but . It may help with your F 360 setup.
Work has kept me away alot lately, so I haven't had time to post any updates or really try many of the ideas above in a few weeks. I did change my setup in UCCNC a little to set my Z-home position to 0 instead of 11.4. I also re-edited the Fusion360 post for UCCNC so that the G53 command sent the axis to 0 instead of 7.5. This has eliminated the confusion that I had that caused the crash. Originally, I thought using positive Z values would be less confusing, but now I realize that I don't really use machine coordinates that much, so I'm not really working with negative numbers that much.
So technically, my system is operating like a stepper, not a servo because I am using the ClearPath SDSK series. A agree that a servo system would be smoother and more accurate, but that was a little more than I wanted to bite off for my first machine. The resolution on the motors is rated at 0.057 degrees with a repeatability of 0.03 degrees. Teknic doesn't indicate what the encoder resolution is, but it would be effectively somewhere in the 12-14 bit range. It might be higher since it would make sense to have slightly better encoder resolution than output resolution. Again, I'm not sure that is relevant to the performance since the motors pretend to be steppers.
I will try to post some pictures of the ridges I am seeing and also try out some different settings in UCCNC. Unfortunately, I'm now under the gun to finish some Christmas presents, so that may have to wait. None of these issues is a show-stopper at the moment.
-Robert
If you run at slower feed rates it should help to smooth it out, you will have to find that spot where spindle RPM and Feed rate will work for your machine
No they don't even come close the 12-14 Bit, they are way too expensive for what you are getting, people have been buying these on hype, and no real time testing, like what you have found, it is good that you do this because it can help others make better choices of what they buy, most don't know how important an Encoder resolution is, until they see what you are seeing, I found this on there web site,
The number of possible positions per revolution that you can set is less than the number of encoder counts per revolution. For example, certain models have a maximum “input resolution” of 800 positions per revolution. For these models, you can command the motor to any one of 800 shaft angles (i.e., every 0.45 degrees), and you will achieve any of these commanded positions with a repeatability of 1/12,800 (0.028 degrees) or 1/64,000 for an integral horsepower model.
“Enhanced” resolution models (those with an “E” immediately after the third hyphen in the part number), have finer input resolution and can be set to as many as 6,400 positions per revolution (16,000 for integral horsepower models). As an aside, there are a number of technical advantages to using an encoder with a higher resolution than the “positional setability” of the motor, but that discussion is beyond the scope of this FAQ.
Mactec54
So after a frustrating second failure on milling out the bowl (entirely my fault, I programmed the retract heigh too low and the bit milled a slot right through the bowl), I decided to try a different piece. With a little more experience and attention to detail, I managed to separate this one from the stock in one piece. It even looks like the CAD. After two screw-ups, I was jumpy every time the machine made a rapid move towards the stock.
I'm pretty happy with the outcome. I am still seeing some of the same jittery motion I saw on the bowl, but the machine was predictable and did a nice job. I haven't tried alot of the suggestions made here, yet...I just needed to get this finished. I'm starting to come up with some basic work practices that are making sense to me. Now I need to get that bowl done!!
-Robert
That looks great, Robert! Is that African Mahogany? Do you have any photos of this on the machine? I'm curious how you held it in place.
David
David
Romans 3:23
CurlyWoodShop - www.etsy.com/shop/CurlyWoodShop
David Falkner - www.youtube.com/user/difalkner
difalkner - www.instagram.com/difalkner
David,
That is African Mahogany. I found a piece that was 48" x 18" in the shorts bin at my local supplier. I picked it up on a whim because it was calling to me.
Here's the piece on the machine.
I flattened the piece on the router (I'm loving my 2 3/8" facing bit) by shimming it and clamping it similar to the picture. I got the bit as close as I dared to the clamps while flattening, then flipped the piece over and clamped it to the spoil board with another scrap piece of MDF underneath. The scrap was slightly smaller, so it only contacted the flattened area. I flattened the top side then trimmed off the un-machined ends. I squared up the stock on three sides on the table saw and jointer. I line up one of the two parallel edges with my T-Track and clamped the board down. I then put a stop block against the board in the other direction. This allows me to flip the board for machining the second side. The only critical part is that the stock size needs to match the CAD or the front and back would be out of alignment. The biggest downside here is that there is alot of waste.
The Incra hold-down clamps are nice, but the screws and knobs are pretty tall and get in the way of the dust shoe. I need to acquire shorter T-bolts and a different knob...maybe just a nut and a washer.
-Robert
Because
Thanks, Robert. I have some African Mahogany but it's reserved for guitars. But I have plenty of Walnut and have been talking about making some bowls and dishes.
David
David
Romans 3:23
CurlyWoodShop - www.etsy.com/shop/CurlyWoodShop
David Falkner - www.youtube.com/user/difalkner
difalkner - www.instagram.com/difalkner
If only I had about a 20" band saw, the rest of this piece or wood might make a decent top and back for an acoustic. I have been slowly picking up a few pieces of mahogany and curly maple for my first full guitar build (hollow body electric). That was the reason for buying the CNC to begin with. I've built one solid body electric, but the neck was purchased (Fender-style bolt on). I'm looking forward to getting started on that project, but I need to build some jigs first. (And get the computer off my work bench).
-Robert
The main reason I built the CNC in the first place was to make very accurate jigs, templates, and fixtures for building acoustic guitars. In the last 14 months or so since the CNC has been up and running I have only used it about 2% for that purpose but I plan to get back on track this next year. I may even see if I can draw a neck in Fusion 360 and cut blanks on the CNC. I might even get to thinking 'I have arrived!' if I can accomplish that.
David
David
Romans 3:23
CurlyWoodShop - www.etsy.com/shop/CurlyWoodShop
David Falkner - www.youtube.com/user/difalkner
difalkner - www.instagram.com/difalkner
I can do a little of both but not certain I excel at any of it. I'm probably better than average but likely not stellar at anything except patience and detail and taking longer on projects than I really should. That counts for something, right?
This is what I have so far, and that's not the bridge I'll use. That's a Walnut practice bridge. The actual bridge is EIR and is about 20% smaller. No CNC on this, btw, except for the bridge.
David
PS - not trying to hijack your thread
David
Romans 3:23
CurlyWoodShop - www.etsy.com/shop/CurlyWoodShop
David Falkner - www.youtube.com/user/difalkner
difalkner - www.instagram.com/difalkner
difalkner, Did you wish to get into this guitar tooling stuff?
i'll send the zipped 3Dmodel. This is for my friendly local luthier.
Been doing this too long
Absolutely! Whatcha' got for me??
David
David
Romans 3:23
CurlyWoodShop - www.etsy.com/shop/CurlyWoodShop
David Falkner - www.youtube.com/user/difalkner
difalkner - www.instagram.com/difalkner