I'm guessing you won't be assemling/customizing MakerBot's for friends and family???
LOL! That's an understatement. They couldn't pay me enough!
Today has been a bit frustrating. Last night I tried building a simple golf tee for the UPS driver and noticed after I removed it from the platen that the cross-section was not round but oblong. This is a sign of a loose belt. But, because I didn't notice it until after I removed it, I didn't know for sure which was the offending axis. After running another circular object today, I determined it was the X-axis. Drat! That's the hardest one to adjust. To get to the screws that hold the motor, I had to remove the top cap at the left end of the X-axis stage, and remove one X-axis support rod, since it blocks two of the screws. Because I had used Loc-Tite, this was more of a struggle than it might otherwise have been. After getting to the screws, I loosened them and used a couple small C-clamps to pull the motor back against the belt using the heads of the two far screws, while tightening the two near ones, then tightening the two far ones.
Got it put back together and started a build on something else round to check my results, but the X-axis would not home. The X-axis idler screw had begun to rotate out, blocking the carriage. To get to the nut on the bottom, I had to remove the X/Y stage altogether, which meant loosening the Y-axis motor, slipping the belt off the pulley, and sliding out both Y-axis support rods. That's when I discovered that the nut holding the idler support screw was gone entirely. So I decided to replace it with a nylock nut, which I did.
Got everything buttoned up and the Y-axis belt retensioned; but, of course, in the midst of all this sturm und drang, a nut had slid down the rabbit hole into the electronics section. That simply would not do, so I had to remove the bottom -- again -- to fish it out.
Screwed the bottom back in and turned the unit on to run the cylinder. It came out round this time, at least, but it's still not what I wanted, which was a solid ring, as shown in the ReplicatorG model window:
What I got instead was two concentric, thin-walled cylinders:
'Dunno what's up with that. The quest for perfection adequacy continues!
I'm using the big stuff (3mm). I did the design in Rhino3D and exported it as an STL file. Something may have gone awry there, but I would think that ReplicatorG would have displayed it the way it would be built. The other possibility is that the extrusion rate was too slow for the X/Y rate, resulting in too thin a deposit. (I used the "Print-O-Matic" setting for this build.)
Phil, try setting the number of shells to zero in the Print-O-Matic settings. Zero really means one. For a thin walled object there is not enough room for another shell so it just leaves it empty.
Roy, the settings take into account the filament diameter, layer height, nozzle diameter and wheel diameter (on the extruder) to calculate how much filament to extrude.
Cluso99, to make a shell you would set the infill to zero. To make solid, set it to 100. Generally I use around 25% infill. It generates a honeycomb pattern inside the object. Other fill patterns are available too but I like the honeycomb.
Pleasant3D, a Mac program, will display the toolpaths of the objects after RepG has created the Gcode. Very useful.
This is the Gcode to make Phil's extruder. You can see that there are two shells, which means a value of "1" was entered in the settings. I don't remember what the infill ratio was but you can see the zig zags which make up part of the honeycomb pattern. The next layer runs perpendicular and is a bit different to make the rest of the hex patterns.
'Started work on a simple golf tee for the UPS guy. This is to be printed upside down, with the point facing upward. The tricky part is the concavity where the ball sits. Initially, I was hoping that I could pull off a hemispherical recess, IOW a dome with the tee upside down. But that didn't work very well. Once the interior wall of the dome reached a critical angle, the extrudate failed to adhere and collapsed into the space underneath. Obviously, a more gentle angle, reaching into the tee's shaft was called for. And that seemed to work.
The other problem I encountered was when the tee got to a certain height near the pointy end, it was being dragged around by the extrusion tip. It turns out that this was not because the plastic was bending, but because the belt on the automatic build platform was flexing. So I decided to add a much wider base to the build, which I could clip off after the build was done:
I had a heck of a time getting the "wagon wheel" to print, though. The attempted print was moving very fast, and there didn't seem to be anything coming out of the extruder -- at least nothing that was sticking to the build platform. And I ended up with a mess of fine filaments strewn about. So I slowed down the build from 30mm/sec to 15mm/sec. That seems to have cured things, but now it looks like a have to shim the build platform after all. The extrudate sticks better to the rear of the platform than to the front.
Okay, the tee has just now finished. Here's a photo:
Here's another from the top with the "wagon wheel" clipped off and after some hand-finishing to remove burrs (inserted into a nylon spacer to keep it upright):
Rich: I'm curious about the front flange that holds the clamp on the extruder. That represents a major undercut with the extruder lying on its back. How was it supported during printing?
Thanks, Rich. Now I understand where the tooling marks between the two flanges came from.
Hmm. Looking at those cable chain pieces makes me wonder why the top of my much smaller dome collapsed. It may have something to do with extruding in a circle vs. stretching the extrudate in a straight line over a gap.
Here's an example of the kind of thing I was hoping the MakerBot would be useful for. When I was designing the ColorPAL, it was necessary to come up with a custom shroud to separate the LED from the sensor. I designed one in Rhino3D and mocked it up on the CNC mill by milling layers from 1/16" G10 and gluing them together. (This was before I got the laser cutter, which would have made the job even easier.) Today I used the same Rhino model to produce an STL file for the MakerBot. Here's a photo comparing the G10 mockup (left), the resulting injection-molded part (center), and the MakerBot prototype (right):
The MakerBot prototype has been cleaned up a bit by reaming excess "spaghetti" from the circular orifice, and drilling out the mounting holes. This part not only had issues with bridging the circular gap at the half-way level, but the height of the lower chamber is too small -- owing, I think, to the splooging of the bottom laminate onto the build platform. I would not consider the MakerBot prototype to be useful for evaluating this particular part.
BTW, I'm seriously considering removing the belt from the ABP and covering the build platform with Kapton tape or film. That belt seems to be more trouble than it's worth. It won't take much effort, and the belt could be put back quite easily if I change my mind.
The coil of plastic "wire" that feeds into the extruder was turning into a twisted, snarly mess. So I had to come up with a way to tame it. I got a 5-gallon plastic bucket, an empty Pepperidge Farm cookie can, one of those cheap ball-bearing lazy Susan gizmos, and four rubber feet. This is the result:
The whole assembly rotates on the ball bearings, so the wire does not get twisted. But I'm not entirely sure it will stay contained in the bucket. I may have to add an internal rim around the top of the bucket part, if that turns out to be a problem.
The large diameter of the spool means that the filament unwinds with very little pull. The guides ensure that the filament can't get caught up in anything. I routinely leave the TOM printing unattended for hours on end and have not had a single issue with the filament since making this upgrade.
My Thing-O-Matic is in my shop's computer work area, which I keep somewhat dark on purpose. As a consequence, viewing builds in progress was somewhat of a strain. So I decided to add a worklight. I built it from a Parallax extrusion section in the standard 2.75" length, some white LEDs I had left over from a previous TCS230 build, and a PCB that I designed about 13 years ago for red LEDs, and which I had to modify by cutting the series strings in half and adding load-balancing resistors. The plastic ends and other pieces were custom-cut today on my laser cutter. Here's the result:
It's powered from the Thing-O-Matic's PSU (12V rail). Here's what it looks like installed:
My biggest concern right now is stress cracks in the acrylic.
Phil,
Just an FYI, cracking the acrylic can easily be repaired with a drop of acetone (100% from the drug store.) The cracks were simply becuase it was too tight.
Also, in another post, you mentioned that you used loc-tite on the bolts. That won't help that much, because the wood will collapse due to use (vibrations, etc.) and your nuts/bolts will still loosen up a bit... however, the loc-tite will prevent you from tightening it up when it does loosen up.
BTW... I have a couple of CupCakes and am currently building a ToM.
The volume of plastic used is 3.5cc, according to the ReplicatorG program. With a filament diameter of 3mm, that amounts to about half a meter of filament.
I've removed the troublesome ABP conveyor belt and replaced it with a 0.090-inch-thick aluminum build platform that I cut out on my CNC mill:
I got some silicone heatsink compound from RadioShack to use as the interface between the HBP and the aluminum. I've never used this stuff before, and I must say: it's a lot more pleasant to spread than that old white greasy stuff they used to sell.
The long skinny black things are threaded replacements for the individual nuts that I cut from acetal on the laser cutter. I got tired of chasing the nuts around, and now I've got two loose in the MakerBot that I cannot account for. They're trapped somewhere, because they don't rattle or slide around when I shake and tilt the machine.
I didn't have any of the extra-wide Kapton tape, and trying to cover the aluminum with strips of 3/4" tape didn't work out very well. So I used some self-adhesive polyester film that I had in stock that I could cut out on the vinyl cutter. That's the white stuff in the photo.
Here's a photo of the new build platform installed:
I was not entirely sure the molten plastic would even stick to the polyester, but it turns out that the bond is excellent -- much better than with the belt. One thing that is much different now is the amount of time it takes the build platform to come up to temperature. The additional thermal mass of the aluminum entails a much longer wait time for a build to start. Nonetheless, I'm quite happy with the results; and I know I can enjoy more reliable builds, now that I've taken the time to make this modification.
__________________________________
This pretty much wraps up the presentation part of this thread. It remains to be seen how valuable an addition to my fabrication tool repertoire this thing will be. The Epilog laser cutter was a slam dunk. It "just works", and I never have to mess with it, except to clean the mirrors and lenses once in awhile. The Thing-O-Matic, by contrast, does not have the same mechanical integrity and remains more of a fussy machine to keep running. I'm sure I'll have fun with it, but I 'm not yet sure I can depend on it for any serious production or prototyping work. Perhaps that will change as I gain more experience with it.
Even though the machine is now built, and I'm ending this part of this thread, I'll still be happy to answer any questions y'all might have.
I'll still be happy to answer any questions y'all might have.
"Y'all"? You must be from SOUTHERN Port Townsend! I'm from SC originally, I slip back into "yawl drawl" when I visit there.
Great build description all the way, PhiPi. You made many ingenious mods throughout. I know several people with these, most have reached the same conclusion as you. Fun, but not ready for prime time just yet.
erco
'
Its " you'll " not " yawl "
'
If you need some help here. I can learn you the right way to spell some things. LOL!!!
'
Its the 3D files I worry about that might get out of wack from the translation.
I think I'm on my way to justifying the Thing-O-Matic purchase after all. I had to test and assemble a small batch of TSL1401 daughterboards for Parallax today. This is the first batch to use the new lenses I got from China. The lens barrel on these is much longer than that of the older Korean lens, which means that the time spent twisting them in by hand can really add up quickly. So I designed a "lens twister" in Rhino3D that I could press onto a 1/4" drive bit and use to rotate the lens in with a portable electric drill. (The electric screwdriver is too slow, it turns out.) Here's a photo:
The part that grips the lens is toothed, which shows in the perspective view from Rhino but not very well in the photo. Also shown are two lenses: one barely screwed into the holder; the other, fully screwed in. That's a 1 cm distance. Divide that by the 0.5mm thread pitch, and you get 20 full rotations; so the time savings are significant.
I'm happy with the way the MakerBot handled the job. It's really the only tool in my shop that could have done it so easily.
My pcbs for the stepper motor controllers have just arrived back from mfr. Unfortunately I am away so cannot post a photo. The pcb is 45x45mm (1.76"x1.76") and plugs into my prop BaseBlade1 pcb (or any prop pcb with 5V, GND, and 10-12 Prop I/O pins. It has provision for 4 stepper motor daughter pcbs - either my stepper boards (0.6"x0.8") based on the Allegro A4982/A4984 or the Pololu stepper boards based on A4983/A4988. BTW this pcb "could" be driven by an Arduino as an interim step to getting to a prop driven reprap.
Comments
LOL! That's an understatement. They couldn't pay me enough!
Today has been a bit frustrating. Last night I tried building a simple golf tee for the UPS driver and noticed after I removed it from the platen that the cross-section was not round but oblong. This is a sign of a loose belt. But, because I didn't notice it until after I removed it, I didn't know for sure which was the offending axis. After running another circular object today, I determined it was the X-axis. Drat! That's the hardest one to adjust. To get to the screws that hold the motor, I had to remove the top cap at the left end of the X-axis stage, and remove one X-axis support rod, since it blocks two of the screws. Because I had used Loc-Tite, this was more of a struggle than it might otherwise have been. After getting to the screws, I loosened them and used a couple small C-clamps to pull the motor back against the belt using the heads of the two far screws, while tightening the two near ones, then tightening the two far ones.
Got it put back together and started a build on something else round to check my results, but the X-axis would not home. The X-axis idler screw had begun to rotate out, blocking the carriage. To get to the nut on the bottom, I had to remove the X/Y stage altogether, which meant loosening the Y-axis motor, slipping the belt off the pulley, and sliding out both Y-axis support rods. That's when I discovered that the nut holding the idler support screw was gone entirely. So I decided to replace it with a nylock nut, which I did.
Got everything buttoned up and the Y-axis belt retensioned; but, of course, in the midst of all this sturm und drang, a nut had slid down the rabbit hole into the electronics section. That simply would not do, so I had to remove the bottom -- again -- to fish it out.
Screwed the bottom back in and turned the unit on to run the cylinder. It came out round this time, at least, but it's still not what I wanted, which was a solid ring, as shown in the ReplicatorG model window:
What I got instead was two concentric, thin-walled cylinders:
'Dunno what's up with that. The quest for perfection adequacy continues!
-Phil
Anyway, what you produced looks like it is actually working nicely
Have you started to learn Blender yet? There are a few video tutorials from a guy in France that I found helpful.
I'm using the big stuff (3mm). I did the design in Rhino3D and exported it as an STL file. Something may have gone awry there, but I would think that ReplicatorG would have displayed it the way it would be built. The other possibility is that the extrusion rate was too slow for the X/Y rate, resulting in too thin a deposit. (I used the "Print-O-Matic" setting for this build.)
-Phil
This is some really good stuff. Thanks for taking the time out of you busy day to update us.
Does Browser think that you just made a nice new box to call home?
Jim
Roy, the settings take into account the filament diameter, layer height, nozzle diameter and wheel diameter (on the extruder) to calculate how much filament to extrude.
Cluso99, to make a shell you would set the infill to zero. To make solid, set it to 100. Generally I use around 25% infill. It generates a honeycomb pattern inside the object. Other fill patterns are available too but I like the honeycomb.
This is the Gcode to make Phil's extruder. You can see that there are two shells, which means a value of "1" was entered in the settings. I don't remember what the infill ratio was but you can see the zig zags which make up part of the honeycomb pattern. The next layer runs perpendicular and is a bit different to make the rest of the hex patterns.
I had no idea that the extruder was not completely solid. That honeycomb infill is amazingly strong!
-Phil
The other problem I encountered was when the tee got to a certain height near the pointy end, it was being dragged around by the extrusion tip. It turns out that this was not because the plastic was bending, but because the belt on the automatic build platform was flexing. So I decided to add a much wider base to the build, which I could clip off after the build was done:
I had a heck of a time getting the "wagon wheel" to print, though. The attempted print was moving very fast, and there didn't seem to be anything coming out of the extruder -- at least nothing that was sticking to the build platform. And I ended up with a mess of fine filaments strewn about. So I slowed down the build from 30mm/sec to 15mm/sec. That seems to have cured things, but now it looks like a have to shim the build platform after all. The extrudate sticks better to the rear of the platform than to the front.
Okay, the tee has just now finished. Here's a photo:
Here's another from the top with the "wagon wheel" clipped off and after some hand-finishing to remove burrs (inserted into a nylon spacer to keep it upright):
Rich: I'm curious about the front flange that holds the clamp on the extruder. That represents a major undercut with the extruder lying on its back. How was it supported during printing?
-Phil
'
really cool
Support isn't always needed. These cable chain pieces have significant bridges but do not droop at all.
Hmm. Looking at those cable chain pieces makes me wonder why the top of my much smaller dome collapsed. It may have something to do with extruding in a circle vs. stretching the extrudate in a straight line over a gap.
-Phil
The MakerBot prototype has been cleaned up a bit by reaming excess "spaghetti" from the circular orifice, and drilling out the mounting holes. This part not only had issues with bridging the circular gap at the half-way level, but the height of the lower chamber is too small -- owing, I think, to the splooging of the bottom laminate onto the build platform. I would not consider the MakerBot prototype to be useful for evaluating this particular part.
-Phil
I'm sure you're right.
BTW, I'm seriously considering removing the belt from the ABP and covering the build platform with Kapton tape or film. That belt seems to be more trouble than it's worth. It won't take much effort, and the belt could be put back quite easily if I change my mind.
-Phil
The whole assembly rotates on the ball bearings, so the wire does not get twisted. But I'm not entirely sure it will stay contained in the bucket. I may have to add an internal rim around the top of the bucket part, if that turns out to be a problem.
-Phil
The large diameter of the spool means that the filament unwinds with very little pull. The guides ensure that the filament can't get caught up in anything. I routinely leave the TOM printing unattended for hours on end and have not had a single issue with the filament since making this upgrade.
It's powered from the Thing-O-Matic's PSU (12V rail). Here's what it looks like installed:
-Phil
Phil,
Just an FYI, cracking the acrylic can easily be repaired with a drop of acetone (100% from the drug store.) The cracks were simply becuase it was too tight.
Also, in another post, you mentioned that you used loc-tite on the bolts. That won't help that much, because the wood will collapse due to use (vibrations, etc.) and your nuts/bolts will still loosen up a bit... however, the loc-tite will prevent you from tightening it up when it does loosen up.
BTW... I have a couple of CupCakes and am currently building a ToM.
Bill
This is such cool CNC machine and parts look nice. But, I am wondering how much plastic is used to make one part, like the ones you are showing?
Here's a whistle that I built from a model on http://www.thingiverse.com:
The volume of plastic used is 3.5cc, according to the ReplicatorG program. With a filament diameter of 3mm, that amounts to about half a meter of filament.
-Phil
I got some silicone heatsink compound from RadioShack to use as the interface between the HBP and the aluminum. I've never used this stuff before, and I must say: it's a lot more pleasant to spread than that old white greasy stuff they used to sell.
The long skinny black things are threaded replacements for the individual nuts that I cut from acetal on the laser cutter. I got tired of chasing the nuts around, and now I've got two loose in the MakerBot that I cannot account for. They're trapped somewhere, because they don't rattle or slide around when I shake and tilt the machine.
I didn't have any of the extra-wide Kapton tape, and trying to cover the aluminum with strips of 3/4" tape didn't work out very well. So I used some self-adhesive polyester film that I had in stock that I could cut out on the vinyl cutter. That's the white stuff in the photo.
Here's a photo of the new build platform installed:
I was not entirely sure the molten plastic would even stick to the polyester, but it turns out that the bond is excellent -- much better than with the belt. One thing that is much different now is the amount of time it takes the build platform to come up to temperature. The additional thermal mass of the aluminum entails a much longer wait time for a build to start. Nonetheless, I'm quite happy with the results; and I know I can enjoy more reliable builds, now that I've taken the time to make this modification.
__________________________________
This pretty much wraps up the presentation part of this thread. It remains to be seen how valuable an addition to my fabrication tool repertoire this thing will be. The Epilog laser cutter was a slam dunk. It "just works", and I never have to mess with it, except to clean the mirrors and lenses once in awhile. The Thing-O-Matic, by contrast, does not have the same mechanical integrity and remains more of a fussy machine to keep running. I'm sure I'll have fun with it, but I 'm not yet sure I can depend on it for any serious production or prototyping work. Perhaps that will change as I gain more experience with it.
Even though the machine is now built, and I'm ending this part of this thread, I'll still be happy to answer any questions y'all might have.
-Phil
"Y'all"? You must be from SOUTHERN Port Townsend!
Great build description all the way, PhiPi. You made many ingenious mods throughout. I know several people with these, most have reached the same conclusion as you. Fun, but not ready for prime time just yet.
'
Its " you'll " not " yawl "
'
If you need some help here. I can learn you the right way to spell some things. LOL!!!
'
Its the 3D files I worry about that might get out of wack from the translation.
-Phil
The part that grips the lens is toothed, which shows in the perspective view from Rhino but not very well in the photo. Also shown are two lenses: one barely screwed into the holder; the other, fully screwed in. That's a 1 cm distance. Divide that by the 0.5mm thread pitch, and you get 20 full rotations; so the time savings are significant.
I'm happy with the way the MakerBot handled the job. It's really the only tool in my shop that could have done it so easily.
-Phil
My pcbs for the stepper motor controllers have just arrived back from mfr. Unfortunately I am away so cannot post a photo. The pcb is 45x45mm (1.76"x1.76") and plugs into my prop BaseBlade1 pcb (or any prop pcb with 5V, GND, and 10-12 Prop I/O pins. It has provision for 4 stepper motor daughter pcbs - either my stepper boards (0.6"x0.8") based on the Allegro A4982/A4984 or the Pololu stepper boards based on A4983/A4988. BTW this pcb "could" be driven by an Arduino as an interim step to getting to a prop driven reprap.