IDEX H-Bot
-
Yes, this is a terrible idea for many reasons. I simply want to know if this is possible to configure with the stock rep-rap firmware.
As the title suggests, the independent dual extruder machine would have two layers of the H-Bot motion system. Each extruder would have it's own X and Y motor, but the firmware would need to keep both extruders at the same Y value.
Thank you,
Will -
I don't think this is a terrible idea, and I think at least one user already has such a setup. It's quite interesting, really.
I think the stock firmware can indeed support this kinematic, but I'm not entirely positive. It may require creating and hiding axes in tool change macros.
*not actually a robot
-
Yes it's possible. See also CoreXYU and CoreXYUV kinematics, which are related to H-bot.
Duet WiFi hardware designer and firmware engineer
Please do not ask me for Duet support via PM or email, use the forum
http://www.escher3d.com, https://miscsolutions.wordpress.com -
@bot The reason I say it is terrible is complicated. While functional and useful, I am unsure of the marketability of the product. Other similar designs offer different advantages that may be preferable to most other people.
Honestly it is quite alarming how little I know about getting this to work at a firmware level. Any tips on where I could get better guidance?
If you are interested I wrote a little about it in response to a strictly theoretical build. While I have not received all the pieces to my machine from the machine shops, I regard my design as functional given that myself and two senior level engineers, one with a doctorates in mechanical engineering, have reviewed the design. Furthermore I have done extensive FEA on the assembly with a safety factor of 5 for a total deflection of 0.2mm on the Z, and 0.02mm on the X/Y axis. The critiques on my final design were in fine alignment and assembly, not mechanical functionality. It has been completely re-engineered 5 times.
"I have engineered an H-bot motion system. The reason my design works and this design will not is due to FEA. If you ran FEA on this design it would not be stiff at all.
Not to mention you would never actually be able to assemble this due to all of the excess rails. Getting them all aligned would be a nightmare. Not to mention you have all your pulleys and motors cantilevered outside the frame. You have all of the extrusions running the thin way where all of the twisting moment is applied to the y carriages.
Look at how the makerbot replicator H bot is designed. These are engineers from stratasys if I am not mistaken.
My H-Bot design is infinitely more stiff and requires no more space than a coreXY because of the type of pre-loaded linear rails I used and in what orientation I used them. Christ, I have them bolted to a half inch of blanchard ground steel almost a foot wide stretching the length of the machine reinforced by two 50x100mm aluminum extrusions and the rest of the machine is built with 40x120mm. Bear in mind, this is a large machine.
Having an IDEX H-bot is a mistake from a complexity standpoint, this did not stop me from trying. I mounted the pullies with roller bearings on 12mm steel rod with polymer bushings in between. Yes you can get it to work, but to what end? Two printers for the price of 1 but at 4x the cost of a simpler design? There would need to be strong motivation, like high precision and speed milling operations to warrant such a design.
CoreXY is not as accurate as H-bot simply because when you have the rigidity required for high accuracy, H-bot is simpler to design. Unfortunately, for using H-bot in 3d printing specifically, 3D printers are not required to be high accuracy. It is a very messy process, there is no benefit from having high precision.
As for speed, H-bot wins again. When you have enough stiffness, you can use huge motors. To re-iterate, H-bot is easier to engineer when you have enough stiffness. You need to take advantage of that stiffness if you do not want to lose speed.
Speaking of cost the original makerbot cartesean system is far superior to all the other 3d printing specific designs. You can make 10 adequate sheet metal cartesean machines for the price of one high-precision H-bot machine not including the labor intensive process of FEA engineering the H-bot. It took me a full year to design a great H-bot when I could have designed an adequate sheet metal machine of the same size in a few months. Granted it would need to have the Z axis bolted to a cement wall, or have multiple lead-screws ( eww ).
The motivation for my exercise in futility, encase you had not guessed, was space and time. I have space and electricity for one painfully large 500kg, 35A of 220V machine, not two. Furthermore, I want highly dimensionally accurate parts fast in PEEK, PEI, and PC. I cannot waste time and money playing with taking a part off the printer onto a $30,000 CNC I dont have. Ontop of that, I want to be able to have the option of having two half sized machines when I do not need the full build envelope. Hell, maybe other people feel the same way. Maybe they will pay top dollar for my work."
-
@dc42 Thank you so much! My next question, has anyone actually used this configuration? If so, how difficult was it to configure?
I did not get much after looking up "CoreXYUV kinematics" any links you know off hand?
-
I added both CoreXYUV and CoreXYU support in response to user needs, so both are in use. When someone requested MarkForged kinematics, I got fed up with writing new kinematics modules, so I generalised the kinematics to support any kinematics for which the movement of each axis is a linear combination of the movements of the motors.
-
@Sphyloid said in IDEX H-Bot:
................. I regard my design as functional given that myself and two senior level engineers, one with a doctorates in mechanical engineering, have reviewed the design. Furthermore I have done extensive FEA on the assembly with a safety factor of 5 for a total deflection of 0.2mm on the Z, and 0.02mm on the X/Y axis.
Given that layer heights tend to be in the order of 0.2 or 0.3mm, then designing a printer which has a predicted deflection equal to that layer height is not a good thing (in my humble opinion). I'm surprised any senior engineers (with or without doctorates) would say that is acceptable.
As for speed, H-bot wins again. When you have enough stiffness, you can use huge motors. To re-iterate, H-bot is easier to engineer when you have enough stiffness. You need to take advantage of that stiffness if you do not want to lose speed.
For FDM printing, the speed is largely governed by how fast you can melt and extrude filament. So while huge motors might allow one to use high accelerations in order to attain high speed, there is little point in doing so if the hot end is incapable of supplying molten filament at the required rate. The best I've been able to do is about 300mm/sec using multiple extruders feeding into a mixing hot end.
Honestly it is quite alarming how little I know about getting this to work at a firmware level. Any tips on where I could get better guidance?
@dc42 has written the firmware so there is little to know. All you have to do is configure it. For info, I'm a 67 year old (retired) mechanical engineer with not much in the way of paper qualifications and I have been able to configure and run a CoreXYUVAB with 6 extruders. So I'm sure that someone who is a clever as you will be just fine. The Duet Wiki contains all the information you need and if you get stuck, this forum is a good resource to use.
