Man, this is a big one tonight. I am starting to actually use the machine.
I have simply been chipping away at just about everything. Got the new control panel done, I keep trying different adjustments with the 3D printing (and getting good results!) and figuring out where everything needs to be set to mill PCBs properly. In short, it is going quite well.
In the coming days, I hope to finish the vacuum table and get it to a point where I can simply clamp it down and mill, without much (or any) set-up time or adjustment.
The milling (with the router) is 100% in size, squareness and dust control --everything is perfect.
The 3d printing is going much better now but I fear that I will never truly be able to produce proper prints with the CNC/ extruder controller set-up I have now. Beyond getting the two perfectly in sync, the fact that I cannot control either the temp or extruder speed on-the-fly (via the Gcode) is going to be a brick wall in terms of getting to the "next level". I have ordered a Sanguinolu board and the few parts I will need to finish the board. This will control BOTH the extruder and the machine itself. In essence, it will become a very large, heavy RepRap. 3D printing is hard and has a big learning curve --that learning curve just keeps getting harder as you stray away from the way everyone else is doing it. --I think the Sanguinolu will simplify a lot.
There is a super-bunch o' stuff in tonight's slideshow:
And you get a double post tonight!
I had discovered that I recorded a video last night but did not get it posted. So tonight, I give you 2 videos. Now, I could bore you with a lot of text right here, but instead... Why don't your bore yourself with almost 20 minutes of video?! Woo Hoo!
The machine is really big.
And of course, the Pictures:
You got time to build or time to blog, rarely both...
I have been pretty busy. This last week or more, I have been putting in 8-10 hours each day on the CNC and then finding myself in the office doing well, office stuff. I am a tired little boy. As it sits, I have been doing a lot more welding than blogging but I hope tonight's post makes up for it. The video is not too rambly and you get a pretty big slide show as well. Enjoy.
As you can see in the video, the bottom line is a lot of thick, heavy metal and a lot of weld. --That's the trick! 'Ting!
I think the video is clear enough and so I will refrain from repeating its contents in text format. Instead, I give you tonight's slide show.
I have started version 2 of the CNC!
To get to the point, we ain't screwin' around this time... I have gone from 1 x 1.5 x 1/16 box tube up to 2 x 3 x 3/16 tube. It is twice the size and 3 times the wall thickness. The assembly shown above has gotta be 100 lbs already. I am chamfering the begeebers out of the ends of the steel before welding, producing almost 100% penetration on the welds. Where the angle iron (which supports the glides) hits the box tube frame, we have almost 1/2" of metal thickness. Put simply, whatever is not supposed to move, won't. Period.
Where's the beef? Yeah, I got your beef RIGHT HERE!
I am printing things in 3D, yo!
It works! It works!
Now, I can't stress enough how much more I have to do on this one but, proof of concept, it is working. I am currently running the cnc via Mach3 to do the printing. An external stand-alone Arduino is taking care of the hotend temp and is running the extruder stepper motor via a Easy Driver from SparkFun. Here is the quick run-down:
- The .stl file comes from a Thingaverse download or exported from SketchUp. Sketchup requires a plugin to export to .stl but it is quite easily found and installed.
- The .stl file then goes to a Processing sketch called "SuperSkein" which converts to gcode.
- The gcode ends up in a simple text file.
- The text file is opened and, with a simple find-and-replace, a few "commands" are changed. I.e. M101's are changed to M8's and M103's are changed to M9's. M8/M9 are the codes to turn the coolant pump on and off (if this were a large milling machine) but instead, we take that "coolant pump output" and connect it to a pin on our Arduino. As those codes go by, the extruder is thus turned on and off.
- That now-converted text file can now be opened directly by Mach3.
The Arduino controller is quite simple. It does a read of the hotend temp every second and turns the heat on and off as needed. A "set temp" is inputed by a pot on an ADC input. A second pot allows the extruder's speed to be adjusted. Finally, a 20x4 LCD displays the values. There is a very tiny bit of very simple code involved and for now, it simply works. I think my next version will be via a Propeller chip and the code will have to be a bit more precise. I am finding out that very small adjustments make big differences when printing.
I have also discovered that very small adjustments in the mechanics can make a big difference as well... After my first print of a 2cm cube took (I swear) 9 hours, I decided to have a look at the extruder and figure out why it was having such a hard time advancing the filament. It seems that the "hobing" I had added to the drive bolt was not so much gripping the plastic as it was grinding it away and making tons of blue sawdust. I tried 4 or 6 different ways of preparing the bolt, and grinding different textures on it before I finally found the "tap trick". I cannot tell you how much it made a difference. (The video has more on the "tap trick"). All-in-all, I would say I am about 50% there to getting to proper prints.
Next on the list is to start some proper experiments. I need to download some test prints and grab my pen and paper. I have a lot of temp settings to read and record, extrude speeds, feedrates etc. Each one needs to be isolated and tested so I can start "tightening things up" on the ol' printing. I also need to write some clean, working, good code to control the hotend temp and extruder stepper. --Like I said, probably going to switch to a Prop for this one. And last but not least, I am getting very close to feeling like I have "figured out all the kinks" and can order steel and start work on the v2 machine.
(v2 has a working title but, it is not quite ready for prime-time... (It's a good one, though))
Here are your pictures for tonight, folks...
Before I start building the v2 machine, there is simply no reason not to try out the version one.
I came to the conclusion that in addition to the chatter/flex issues I am having, that there may be more problems and issues I have not found yet. I thought it best to go ahead and use the version one machine and see if I could "pre-solve" anything I could find before starting v2. In the end, my v1 machine actually works quite well and with some slowed-down feed rates, works a treat --even with a less-then-proper bit.
I will continue running stuff, learning the software and trying different things. Here are just a few of the items I was able to mill. Enjoy.
If you have not seen day 10 yet, please go back and watch that video --it will get you up to speed on the issues that led to this rebuild.
That's all you get tonight, the rebuild starts...
I'm going to be honest with you folks, I think my design for the Y axis just ain't gunna work.
The issue started during the install of the gantry to the base. There was a little crunchiness in the bearings and some less-than-smooth action when sliding forward and back. I had dismissed it at the time as simply needing to adjust the bearing tension. In hindsight, it was probably an omen. Here we are now, day 10, and I have had a couple days of running programs and observing the results. It seems the "chunkiness" of the bearings is gone but the herky-jerky action still lives on. Put simply, I am getting (call it what you will) chatter, bounce, loading/unloading etc. when the gantry is moved forward and back. No amount of adjustment of the bearings has fixed anything nor did physically widening the bearings themselves. Upon further inspection, it appears that this chatter is coming form a very small amount of flex from multiple places. The undermount ballscrew for the Y axis, and its very long cross-member wrapping around to attach to the gantry, is one culprit. The 6" spread on the bearings themselves may be a bit narrow. The fact that the main mass of the gantry is quite high and far away from the thing that is actually moving it is not good. And finally, the gantry itself may have a bit of twistablity to it. Add all this up and you end up with a router shaking +or- a few hundredths of an inch in all directions. This has lead to a full-stop, a rethink and moving toward the idea of a total rebuild.
In short, I am leaning heavily to doing a fixed gantry and a sliding table --A solution that solves quite a bit of problems, actually. As it stands now, there is no way I will be able to 3D print with the gantry as it sits. Trying to get 70 lbs to slam forward and back at the speeds associated with printing is just not gunna happen. Once again, this leads me to the sliding table idea. The gantry can be fixed if the table slides and would allow me to over-build it with no need to worry about weight. In fact, heavier would be an asset here, actually. There will be a few draw-backs I'm afraid, but all-in-all I am leaning this direction. I hope to have more solid thoughts on the matter soon. Until then, here is you video...
If only my pen were a router... ...and in a few days, it shall be...
I had a pretty good night last night and a pretty good day today. I tackled the set-up of Mach3, got all my pins set, my step count (hopefully) correct and everything moving at the correct speed and direction. It will even home when I ask it. I moved on to LazyCam and started working to export a design from InkScape as a .dxf file, import it and convert to G-code. It was less than successful but, with some tweaking, got some results. The first issue I had was the fact that my design came in greatly oversized. A 2" box, drawn in InkScape, turned into a 149" square in LazyCam. I simply scaled it to the proper dimensions but I look forward to figuring out the proper solution. Next, some of time when I imported a design, only some of it came through. Seemingly at random, parts and lines were just not there --again, looking forward to fixing this one as well.
In the end, I got what I wanted, which was for anything at all to come through that would allow me to simply test the machine. It did at that. I found myself quite happy with about 90% of the build with the remaining 10% being a bit of a let-down but ultimately very fixable. I think the bottom line here is that I have a LOT of adjusting, shimming, testing and more adjusting in my near future.
Here is a far quicker video of just the drawing part:
In conclusion, and in addition to the mechanical issues, it looks like my next issue is to start looking for some new design software. I have some suggestions from some friends, I just need to sit down and start playing with different stuff. Hopefully, I can find something better than InkScape that will play a bit nicer with my machine. Again, I feel another large learning curve coming on...
Today's blog post is brought to you by the letters X, Y and Z.
Oddly enough, I have had some of my biggest build days over yesterday and the day before and yet this post will probably be short and sweet. I gotta be honest with you folks, I am a tired little boy and my brain is a bit fried from starting to go through Mach3 Tutorials.
(I gotta be honest, I watched this back and its a bit rambley)
And some poorly-exposed photos:
Well, Ladies and Gents, I'll give it too you straight... I found tonight's blog post a bit lackluster and weak. I'd give it a 5, maybe 6. I apologize and acknowledge the issue. I will attempt to produce posts of greater quality in the future.
Thank you for your time.