I have moved from the test part phase to the final part phase. I spent many hours making test pieces to know the printer limitations. I highly recommend before you start a project like this make some simple test pieces to know how the limits of your printer or how it handles things like holes or detail reproduction. An example is rivets, I made a small piece that had rivets of all sizes I planned to use (and then some), which I printed and examined to see what did not show, and what was the actual printed size of what did show. Do the same with gaps (like spaces between boards or seams) and holes. Making these test pieces will save you a ton of time and frustration because you can model the parts to the size they need to be for them to print to the size you want. An example is a wooden boxcar, you may model the boards to their scale size with scale spacing only to have it print as a smooth wall because your printer (or the resin you are using) will not reproduce that fine of a gap. What you will have to do is over exaggerate some features and knowing by how much will allow you to model your part with those exaggerations to start and not have to go back and fix it later, which can be an issue with some 3D modeling software (especially the free versions).
And speaking of software, if 3D printing is something you want to do I would recommend buying software rather than using free versions. I use Autodesk Inventor, which is very expensive, but I am learning Fusion 360, which is very similar but much cheaper. I will still use Inventor but by learning Fusion I can help friends that use it. I have seen many people use Sketchup but recently I have noticed that it dose not make round parts round, the are a series of flat surfaces which will show in the final print, Fusion and Inventor make round parts round.
Anyways, back to the project. I had many issues printing the frame. I use an Epax X1 printer and the print bed is not large enough to print the frame in one piece. Even if I print the frame vertical it would still be too long. Also, printing vertical would take forever because with resin printers it is the height of the print that determines the time, not the amount of parts on the build plate. Currently I have the final boiler print being printed and it is 83% done and 20 hours of printing so far.
The issue I had with the frame is that I modeled it all to scale, and when I printed the parts they were just too flimsy. I increased some thicknesses that I felt would not be noticeable which helped but then the second issue came up, the warped. There are ways to fix warped parts but that would not guarantee them to be in tolerance for what I wanted. Since the frame is the foundation of the entire model I needed the frame to be straight and in specifications so everything would screw together, and all seams would be tight and square.
When I looked at my test frames, which were printed on an FDM printer, there were very straight and square. I came up with a way to print a subframe on my FDM printer using PLA, which I laminated resin printed parts over. My FDM printer has a 12” print bed so I was able to print a skeleton frame, which include bolsters, boiler mounts, (what I am calling) sub-pilots, and sub-running boards. Then I changed all my resin frame parts to be halves with half flanges on the back sides to cover and hide the PLA. When the resin parts were glued over the PLA the back flanges met face to face and basically encapsulated the PLA. I used plenty of CA gel to glue the parts all together, and one thing for sure is that resin parts and PLA love CA glue, almost too much so.
My FDM printer has a special print bed material that I added, which produced an almost perfect top surface of the running boards, but it was “almost perfect” and not perfect so I laminated 0.010” black styrene over it, this time with water thin CA, then I trimmed it out and sanded the edge smooth to match the PLA.
With the frame starting to look like something I added the wood end beams on the pilots. I used basswood for these beams which had sockets cut into the backs to accept the sub-pilots that were part of the PLA subframe. Also on the back of the PLA subframe where small tabs that I included that were used to hook the back of the coupler pockets, which basically connected the subframe to the couplers, this takes the pulling force of the couplers and spreads it to the subframe rails and not relies on the glue joint between the wood beams and the frame rails.
Speaking of coupler pockets, these were designed to accept Kadee On3 couplers unmodified (almost). If you do not do On3 then you can skip this paragraph if you want. One issue with modeling in On3 is that you only really have one coupler choice, Kadee. However, unlike HO scale Kadee couplers which have many varieties for different situations, in On3 they make only one type (offered in two colors). The shank on the On3 Kadee is long and uses a spring and a slot to center the coupler and give it the slack action effect. I find this method of centering annoying because light cars just make the train do this annoying inch worm thing as it moves along on level track, where the end of the train springs out until all the centering springs are compressed, then the entire train shoots forward until all the springs are uncompressed. What I have been doing is cutting off the tabs on the bottom of the coupler and blocking the spring with styrene so the coupler can only pivot. I modify Kadee HO scale #5 centering springs to make the coupler center without the slack action. My coupler pocket design is HO #5 width and does not have the On3 slot to accept the tab I sliced off.
I wanted my coupler pocket to look in scale yet allow for the Kadee coupler to be inserted full size and held in with a lid that uses a screw to retain all of it. I first modeled the part to be to scale, then I modeled the coupler pocket on the back side and started making modifications to the “scale” features to get the coupler to fit. I slightly re-proportioned the outer parts so that it is very hard to tell that is has had some “adjusting” done to it. I am very happy with the final part and I am very happy to know that I will have a functioning coupler on my shay. Those that have experience with brass steam locomotives, especially shays, will know that it is very frustrating that they are not made to have functioning couplers.
I will post some photos of the progress so far. The next big hurdle I will have to solve will be the crankshaft. I have to mount a gear on the crankshaft because I am driving my shay from a center mounted motor. On my test shays I just had a brass rod to simulate the crankshaft, now I have to make the real deal. I have a brass PSC crankshaft that I may cut in half to mount the gear, but my first try will be something different, more on that later.
I have seen many people use Sketchup but recently I have noticed that it dose not make round parts round, the are a series of flat surfaces which will show in the final print, Fusion and Inventor make round parts round.
That's an adjustable setting in Sketchup. You can define the number of facets a "round" object gets. If you raise the number significantly (say, from 8 to 50) that will cure that issue.
Yes, I have experienced similar things - in my case I wanted to print a girder with some fine structures, right at the limit of what my printer can do. Yes, it is a "squeeze plastic from a spool through a hot nozzle" printer, but many things are similar...
One of the things I've found is that for fine details the orientation on the print bed matters - I can print a flat girder lying down, but I cannot print it standing up. At least not nicely. Other parts the orientation is more about the supports and how they can be removed... Or in what direction I want the layers to be (either for strength or looks - layers can be used creatively ;-)
As for the issue of roundness - I'm using OpenSCAD, which allows a parameter to be set (either globally or per part) to control the number of facets to use. Obviously, a cylinder with 6 facets will print as a hexagonal structure (which can be useful). However, if I want a perfect circle, I just need to increase the number of facets. How many I need depends on the size of the object. However, there's the thing with "resolution" and in this case it's working for me: the smallest structure my printer can print accurately is about a tenth of a millimeter (few thousands of an inch?). Once the facets represent an "error" of less than the resolution, it will print as a perfect circle.
Yes, OpenSCAD has another mechanism to control roundness, which is a fancy way to specify facets.
Yes, the Slicer software actually looses some accuracy between the model and the file that is sent to the printer. So what, I can still over-spec the 3D model (e.g. for 5/100 millimeter) and happily loose some of the accuracy which will still be better than the resolution of the printer. With a resin printer, there is a minimum resolution, below which there simply will not be enough pixels on the "screen" to create finer structures. While it may be "better" than my filament printer, you'll still get there eventually.
Some things might be different on a resin printer, however you do need to design the part for the printer you're going to use. At least once you get to a point where you're pushing the capabilities of the printer.
In a way, 3D printing is similar to metal casting. You'll get a nice part that needs to be cleaned up in places. I've found that I have to put extra material in places where I have mechanical interfaces, so I can sand it to fit.
With all the small, mechanical parts on the Shay, did I interpret the pictures correctly in that some of the mechanical parts are made from brass? Did you have any issues with the finer details being sensitive to printer parameters and orientation?
yes, all mechanical parts are brass or steel. Any place there is a rotating part (line shafts) the printed parts are sleeved with brass tubing.
printing in resin is not as much of an issue with orientation, but it depends. An example would be a diesel cab, if you print the cab normal to the build plate (bottom of cab parallel to build plate) the vertical side walls would print fantastic (ultra smooth with no visible lines and all details with print close to the as designed sides). The roof would be a different story. Let's say the cab was for an SD40-2 and it had a hole in the top center for a horn. The angled sides may have very slight lines that are visible to the eye but may not show when painted. on the flat roof the hole may be drastically reduced in diameter or may even not be present, which is caused by resin not draining out of the hole between exposure cycles. Small holes in horizontal surfaces (due to surface tension of the resin) will retain resin in them and the exposure of each layout will cure that resin that is stuck in the hole. The same diameter hole on the side of the cab will print fine because the hole is being created as the layer build. Actually what happens to small holes in the vertical plane is that they tend to look a little oval because as the hole starts to form it is wide open and all the resin drains away, but as more and more of the hole forms it retains a little more resin. I will often chase a hole with a drill bit to "round it out".
the biggest issue with orientation is supporting the resin until enough of the model builds that is self supports. Too little supports and sides can sag or parts look warped.
the benefits of resin printing is very very fine detail. as you can see in my parts there is no post process sanding on the visible surfaces, there is some sanding on the backside to remove support material bumps. Resin is also solid, of course you can do FDM prints solid. resin prints paint fantastic, the image of the painted firebox shows a print that I took right off the table, into the cure chamber, and then painted and weathered it, no sanding or primer. The disadvantage of resin is strength, sure a thick part is fine but for model railroad models the details typically are very thin. Resin can be a sticky smelly mess as well.
FDM printers have advantages also, depending on the material and how thick your sidewall / infill is the parts can be really strong. There is no post process curing but there may be lots of post process sanding if you want a smooth finish without seeing the print layers. Your typical FDM printer will print as low as 100 microns, but resin printers currently can do 25 microns, and think that is going to change very soon. The other limitations of FDM printers is wall thickness are limited to your nozzle width, I use .4mm, they make smaller but it is really hard to print without plugging. resin printers can print much thinner, but beware because the parts can get really fragile.
sanding resin prints are really easy, like sanding balsa wood. FDM printers can print in lots of materials and some can be sanded and some don't sand so well. I do not have a heated chamber so I have lots of issues printing in ABS on my FDM printer, but ABS really sands well. I typically print using PLA, it is cheap and prints with very little issues, but sanding can be a bit of a lost cause, go to fast and it melts, and it is hard as rock (to a point). I have just seen an interesting method for smoothing PLA prints, you brush the resin that you use in a resin printer over the PLA model, shine a UV light (or bring it out in the sun) for about a minute and the resin will fill in all the layer lines, and it can be sanded to a perfectly smooth finish with little effort.
In the end I believe both printers are worth having and both shine in certain areas, resin for its high quality detail reproduction, FDM for its cost, strength and speed (with little mess). I would say both are worth having.
No matter how many facets you put on a so called circle it will never be smooth or round. It have have so many small facets that it is irrelevant but it won’t be curved.
This is probably the largest of SketchUPs issues. But the program was not intended to be a commercial/professional program like AutoCAD even though it evolved to be. As such it has some issue.
The down side of cranking up the facets is that it increases the size of the file. I have seen round Dior knobs that were created with really high settings that were absolutely huge (file size). The way that SketchUP uses defines and stores these facets is why the program can become such a memory and resource pig. But this is the price you pay for the ease and relatively inexpensiveness of the original program.
I translated a 3D Sketchup chair mad or curved tubes and such into a flat 2D AutoCad File and it had more line segments then the rest of the 20,000 sq foot 3 story building combined. I had to replace the chair (all 12 places it occurred) to be able to work with the file and cut the file size to less the 1/12 the original file size. So SketchUp is a strange program and hard to predict sometimes.
But it is relatively cheep compared to ,any 3D options.
fusion 360 would be the better way to go over sketchup for model railroading stuff. The personal version (meaning you can't sell your designs or products made from your designs) is free, last I checked. When you download Fusion 360 is says it is a 30 day trial, but that is assuming that you are using it for commercial use, personal use stays free. Fusion 360 will not have file size issues like Sketchup does.
thanks for both of your replies. I wrote an elaborated and way too long reply and it seems to have gotten lost - probably for the better ;-)
Circles being round: between the resolution of my printer and my learning how to use it I have not have any issues with that yet. What I get does meet my expectations. But then I've had a 9-needle printer in the past and was happy with the output - better printers were well out of my reach then...
Brass parts: yes, that does seem smart.
Resin vs. PLA printers: you are right, both have their place. Actually I'm quite happy with my Mini-Delta PLA printer and have yet to "push its limits." Sanding works reasonably well, especially while watching a movie ;-)
Comments
progress report
I have moved from the test part phase to the final part phase. I spent many hours making test pieces to know the printer limitations. I highly recommend before you start a project like this make some simple test pieces to know how the limits of your printer or how it handles things like holes or detail reproduction. An example is rivets, I made a small piece that had rivets of all sizes I planned to use (and then some), which I printed and examined to see what did not show, and what was the actual printed size of what did show. Do the same with gaps (like spaces between boards or seams) and holes. Making these test pieces will save you a ton of time and frustration because you can model the parts to the size they need to be for them to print to the size you want. An example is a wooden boxcar, you may model the boards to their scale size with scale spacing only to have it print as a smooth wall because your printer (or the resin you are using) will not reproduce that fine of a gap. What you will have to do is over exaggerate some features and knowing by how much will allow you to model your part with those exaggerations to start and not have to go back and fix it later, which can be an issue with some 3D modeling software (especially the free versions).
And speaking of software, if 3D printing is something you want to do I would recommend buying software rather than using free versions. I use Autodesk Inventor, which is very expensive, but I am learning Fusion 360, which is very similar but much cheaper. I will still use Inventor but by learning Fusion I can help friends that use it. I have seen many people use Sketchup but recently I have noticed that it dose not make round parts round, the are a series of flat surfaces which will show in the final print, Fusion and Inventor make round parts round.
Anyways, back to the project. I had many issues printing the frame. I use an Epax X1 printer and the print bed is not large enough to print the frame in one piece. Even if I print the frame vertical it would still be too long. Also, printing vertical would take forever because with resin printers it is the height of the print that determines the time, not the amount of parts on the build plate. Currently I have the final boiler print being printed and it is 83% done and 20 hours of printing so far.
The issue I had with the frame is that I modeled it all to scale, and when I printed the parts they were just too flimsy. I increased some thicknesses that I felt would not be noticeable which helped but then the second issue came up, the warped. There are ways to fix warped parts but that would not guarantee them to be in tolerance for what I wanted. Since the frame is the foundation of the entire model I needed the frame to be straight and in specifications so everything would screw together, and all seams would be tight and square.
When I looked at my test frames, which were printed on an FDM printer, there were very straight and square. I came up with a way to print a subframe on my FDM printer using PLA, which I laminated resin printed parts over. My FDM printer has a 12” print bed so I was able to print a skeleton frame, which include bolsters, boiler mounts, (what I am calling) sub-pilots, and sub-running boards. Then I changed all my resin frame parts to be halves with half flanges on the back sides to cover and hide the PLA. When the resin parts were glued over the PLA the back flanges met face to face and basically encapsulated the PLA. I used plenty of CA gel to glue the parts all together, and one thing for sure is that resin parts and PLA love CA glue, almost too much so.
My FDM printer has a special print bed material that I added, which produced an almost perfect top surface of the running boards, but it was “almost perfect” and not perfect so I laminated 0.010” black styrene over it, this time with water thin CA, then I trimmed it out and sanded the edge smooth to match the PLA.
With the frame starting to look like something I added the wood end beams on the pilots. I used basswood for these beams which had sockets cut into the backs to accept the sub-pilots that were part of the PLA subframe. Also on the back of the PLA subframe where small tabs that I included that were used to hook the back of the coupler pockets, which basically connected the subframe to the couplers, this takes the pulling force of the couplers and spreads it to the subframe rails and not relies on the glue joint between the wood beams and the frame rails.
Speaking of coupler pockets, these were designed to accept Kadee On3 couplers unmodified (almost). If you do not do On3 then you can skip this paragraph if you want. One issue with modeling in On3 is that you only really have one coupler choice, Kadee. However, unlike HO scale Kadee couplers which have many varieties for different situations, in On3 they make only one type (offered in two colors). The shank on the On3 Kadee is long and uses a spring and a slot to center the coupler and give it the slack action effect. I find this method of centering annoying because light cars just make the train do this annoying inch worm thing as it moves along on level track, where the end of the train springs out until all the centering springs are compressed, then the entire train shoots forward until all the springs are uncompressed. What I have been doing is cutting off the tabs on the bottom of the coupler and blocking the spring with styrene so the coupler can only pivot. I modify Kadee HO scale #5 centering springs to make the coupler center without the slack action. My coupler pocket design is HO #5 width and does not have the On3 slot to accept the tab I sliced off.
I wanted my coupler pocket to look in scale yet allow for the Kadee coupler to be inserted full size and held in with a lid that uses a screw to retain all of it. I first modeled the part to be to scale, then I modeled the coupler pocket on the back side and started making modifications to the “scale” features to get the coupler to fit. I slightly re-proportioned the outer parts so that it is very hard to tell that is has had some “adjusting” done to it. I am very happy with the final part and I am very happy to know that I will have a functioning coupler on my shay. Those that have experience with brass steam locomotives, especially shays, will know that it is very frustrating that they are not made to have functioning couplers.
I will post some photos of the progress so far. The next big hurdle I will have to solve will be the crankshaft. I have to mount a gear on the crankshaft because I am driving my shay from a center mounted motor. On my test shays I just had a brass rod to simulate the crankshaft, now I have to make the real deal. I have a brass PSC crankshaft that I may cut in half to mount the gear, but my first try will be something different, more on that later.
Jeff Kraker
Read My Blog
(No subject)
Jeff Kraker
Read My Blog
I have seen many people use
That's an adjustable setting in Sketchup. You can define the number of facets a "round" object gets. If you raise the number significantly (say, from 8 to 50) that will cure that issue.
-----
Read my Journal / Blog...
Digitrax & JMRI NRail T-Trak Standards
T-Trak Wiki My T-Trak Wiki Pages
Round parts being round
Hi,
that is a very impressive model!
Yes, I have experienced similar things - in my case I wanted to print a girder with some fine structures, right at the limit of what my printer can do. Yes, it is a "squeeze plastic from a spool through a hot nozzle" printer, but many things are similar...
One of the things I've found is that for fine details the orientation on the print bed matters - I can print a flat girder lying down, but I cannot print it standing up. At least not nicely. Other parts the orientation is more about the supports and how they can be removed... Or in what direction I want the layers to be (either for strength or looks - layers can be used creatively ;-)
As for the issue of roundness - I'm using OpenSCAD, which allows a parameter to be set (either globally or per part) to control the number of facets to use. Obviously, a cylinder with 6 facets will print as a hexagonal structure (which can be useful). However, if I want a perfect circle, I just need to increase the number of facets. How many I need depends on the size of the object. However, there's the thing with "resolution" and in this case it's working for me: the smallest structure my printer can print accurately is about a tenth of a millimeter (few thousands of an inch?). Once the facets represent an "error" of less than the resolution, it will print as a perfect circle.
Yes, OpenSCAD has another mechanism to control roundness, which is a fancy way to specify facets.
Yes, the Slicer software actually looses some accuracy between the model and the file that is sent to the printer. So what, I can still over-spec the 3D model (e.g. for 5/100 millimeter) and happily loose some of the accuracy which will still be better than the resolution of the printer. With a resin printer, there is a minimum resolution, below which there simply will not be enough pixels on the "screen" to create finer structures. While it may be "better" than my filament printer, you'll still get there eventually.
Some things might be different on a resin printer, however you do need to design the part for the printer you're going to use. At least once you get to a point where you're pushing the capabilities of the printer.
In a way, 3D printing is similar to metal casting. You'll get a nice part that needs to be cleaned up in places. I've found that I have to put extra material in places where I have mechanical interfaces, so I can sand it to fit.
With all the small, mechanical parts on the Shay, did I interpret the pictures correctly in that some of the mechanical parts are made from brass? Did you have any issues with the finer details being sensitive to printer parameters and orientation?
Have fun and stay healthy
brass parts
yes, all mechanical parts are brass or steel. Any place there is a rotating part (line shafts) the printed parts are sleeved with brass tubing.
printing in resin is not as much of an issue with orientation, but it depends. An example would be a diesel cab, if you print the cab normal to the build plate (bottom of cab parallel to build plate) the vertical side walls would print fantastic (ultra smooth with no visible lines and all details with print close to the as designed sides). The roof would be a different story. Let's say the cab was for an SD40-2 and it had a hole in the top center for a horn. The angled sides may have very slight lines that are visible to the eye but may not show when painted. on the flat roof the hole may be drastically reduced in diameter or may even not be present, which is caused by resin not draining out of the hole between exposure cycles. Small holes in horizontal surfaces (due to surface tension of the resin) will retain resin in them and the exposure of each layout will cure that resin that is stuck in the hole. The same diameter hole on the side of the cab will print fine because the hole is being created as the layer build. Actually what happens to small holes in the vertical plane is that they tend to look a little oval because as the hole starts to form it is wide open and all the resin drains away, but as more and more of the hole forms it retains a little more resin. I will often chase a hole with a drill bit to "round it out".
the biggest issue with orientation is supporting the resin until enough of the model builds that is self supports. Too little supports and sides can sag or parts look warped.
the benefits of resin printing is very very fine detail. as you can see in my parts there is no post process sanding on the visible surfaces, there is some sanding on the backside to remove support material bumps. Resin is also solid, of course you can do FDM prints solid. resin prints paint fantastic, the image of the painted firebox shows a print that I took right off the table, into the cure chamber, and then painted and weathered it, no sanding or primer. The disadvantage of resin is strength, sure a thick part is fine but for model railroad models the details typically are very thin. Resin can be a sticky smelly mess as well.
FDM printers have advantages also, depending on the material and how thick your sidewall / infill is the parts can be really strong. There is no post process curing but there may be lots of post process sanding if you want a smooth finish without seeing the print layers. Your typical FDM printer will print as low as 100 microns, but resin printers currently can do 25 microns, and think that is going to change very soon. The other limitations of FDM printers is wall thickness are limited to your nozzle width, I use .4mm, they make smaller but it is really hard to print without plugging. resin printers can print much thinner, but beware because the parts can get really fragile.
sanding resin prints are really easy, like sanding balsa wood. FDM printers can print in lots of materials and some can be sanded and some don't sand so well. I do not have a heated chamber so I have lots of issues printing in ABS on my FDM printer, but ABS really sands well. I typically print using PLA, it is cheap and prints with very little issues, but sanding can be a bit of a lost cause, go to fast and it melts, and it is hard as rock (to a point). I have just seen an interesting method for smoothing PLA prints, you brush the resin that you use in a resin printer over the PLA model, shine a UV light (or bring it out in the sun) for about a minute and the resin will fill in all the layer lines, and it can be sanded to a perfectly smooth finish with little effort.
In the end I believe both printers are worth having and both shine in certain areas, resin for its high quality detail reproduction, FDM for its cost, strength and speed (with little mess). I would say both are worth having.
Jeff Kraker
Read My Blog
No matter how many facets you
No matter how many facets you put on a so called circle it will never be smooth or round. It have have so many small facets that it is irrelevant but it won’t be curved.
This is probably the largest of SketchUPs issues. But the program was not intended to be a commercial/professional program like AutoCAD even though it evolved to be. As such it has some issue.
The down side of cranking up the facets is that it increases the size of the file. I have seen round Dior knobs that were created with really high settings that were absolutely huge (file size). The way that SketchUP uses defines and stores these facets is why the program can become such a memory and resource pig. But this is the price you pay for the ease and relatively inexpensiveness of the original program.
I translated a 3D Sketchup chair mad or curved tubes and such into a flat 2D AutoCad File and it had more line segments then the rest of the 20,000 sq foot 3 story building combined. I had to replace the chair (all 12 places it occurred) to be able to work with the file and cut the file size to less the 1/12 the original file size. So SketchUp is a strange program and hard to predict sometimes.
But it is relatively cheep compared to ,any 3D options.
-Doug
Fusion 360
fusion 360 would be the better way to go over sketchup for model railroading stuff. The personal version (meaning you can't sell your designs or products made from your designs) is free, last I checked. When you download Fusion 360 is says it is a 30 day trial, but that is assuming that you are using it for commercial use, personal use stays free. Fusion 360 will not have file size issues like Sketchup does.
Jeff Kraker
Read My Blog
Lost reply
Hi,
thanks for both of your replies. I wrote an elaborated and way too long reply and it seems to have gotten lost - probably for the better ;-)
Circles being round: between the resolution of my printer and my learning how to use it I have not have any issues with that yet. What I get does meet my expectations. But then I've had a 9-needle printer in the past and was happy with the output - better printers were well out of my reach then...
Brass parts: yes, that does seem smart.
Resin vs. PLA printers: you are right, both have their place. Actually I'm quite happy with my Mini-Delta PLA printer and have yet to "push its limits." Sanding works reasonably well, especially while watching a movie ;-)
Have fun and stay healthy.