Can a smooth rods only towers printer be stable enough?

Hi guys,

This is a very quick question: I've seen quite a few Rostocks online that have only the smooth rods as a vertical tower and was wondering how stable can those be and what would be the speed limit before the assembly gets too wobbly with just the rods.

One example I can quickly find online is geetech's rostock mini G2.

I see the rods are about 12mm but I've seen others with 8mm (I think i remember right).

---

RepRap Lander concept on Concept Forge
RepRap Lander concept on RepRap Forums
My Things, mostly experimental stuff

This is probably going to cause an argument, but I'm going with no. With great materials and large enough rods yes it could be stable, but if it's not adding bracing is almost impossible (welded brackets would work). That said I think the basic kossel extrusion build is weak at the top corners especially. I'm trying to figure out how the make my large kossel more stable. Basic engineering and martial arts #1, triangles are strong. With only vertical rods diagonal braces interfere with the carriages.

I bolted some Home Depot aluminum angles diagonally from top extrusions to bottom extrusions on my expanded Kossel Mini. I guess it reduces volume a bit, but it's quite minor. And I can sit on the thing while it prints. Highly recommend.

Well my delta has 16mm rods and it's very stable (200x200mm print size)

I see no reason in principle why smooth rods alone cannot provide sufficient stability and rigidity. The key is securing them to the triangle base and top in a stable and rigid manner. I doubt whether this can be done adequately in a large printer by the use of printed ABS parts.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Quote
pkm
Well my delta has 16mm rods and it's very stable (200x200mm print size)

16mm rods makes indeed a hefty printer.

Quote
epicepee
I bolted some Home Depot aluminum angles diagonally from top extrusions to bottom extrusions on my expanded Kossel Mini. I guess it reduces volume a bit, but it's quite minor. And I can sit on the thing while it prints. Highly recommend.

I have seen threaded rods and also steel cable used for this purpose (using cable tensioners) and I like these solutions because they are adjustable. If you were to rivet or bolt aluminum angles to a frame you have to be 100% sure that your Delta is 100% True and that it will not change in time.

Can you post some pictures of your reinforced Delta printer?

I do believe that adding some more columns/braces in between at least two sets of towers is easy enough for a delta that only uses the rods as towers. Aluminum angles, extrusions or even some acrylic boards bolted to the top and bottom plates would offer the extra rigidity with minimal modifications of the original frame.

---

RepRap Lander concept on Concept Forge
RepRap Lander concept on RepRap Forums
My Things, mostly experimental stuff

Quote
dc42
I see no reason in principle why smooth rods alone cannot provide sufficient stability and rigidity. The key is securing them to the triangle base and top in a stable and rigid manner. I doubt whether this can be done adequately in a large printer by the use of printed ABS parts.

The problem is that if your rigidity is low then the printer will have lower resonant frequency. If this frequency is comparable with the print head oscillations when printing small infills then you will have a problem. The printer head can easily generate frequency at about 200 Hz at rather common accelerations of 7g. Single 70 cm long Ø 8 mm smooth rod will have resonant frequency the area of low tens of Hz. Of course this is a very rough computations but it does not sound promising for Ø 8 mm rods. But Ø 8 mm smooth rods have also a problem of dynamic bending in the range of about 0.1 - 0.2 mm so they are just not suitable even inside a rigid frame.
But one can get much thicker smooth rods easily and aluminum extrusions should do much better too.

Quote
hercek

Quote
dc42
I see no reason in principle why smooth rods alone cannot provide sufficient stability and rigidity. The key is securing them to the triangle base and top in a stable and rigid manner. I doubt whether this can be done adequately in a large printer by the use of printed ABS parts.

The problem is that if your rigidity is low then the printer will have lower resonant frequency. If this frequency is comparable with the print head oscillations when printing small infills then you will have a problem. The printer head can easily generate frequency at about 200 Hz at rather common accelerations of 7g. Single 70 cm long Ø 8 mm smooth rod will have resonant frequency the area of low tens of Hz. Of course this is a very rough computations but it does not sound promising for Ø 8 mm rods. But Ø 8 mm smooth rods have also a problem of dynamic bending in the range of about 0.1 - 0.2 mm so they are just not suitable even inside a rigid frame.
But one can get much thicker smooth rods easily and aluminum extrusions should do much better too.

8mm does seem rather thin for this application. I was thinking of 12mm smooth rods, like the ones on my Ormerod. The RepRapPro Fischer kit uses 8mm rods, but they are only 426mm long.

Edited 1 time(s). Last edit at 12/15/2015 07:14AM by dc42.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Quote
dc42

Quote
hercek

Quote
dc42
I see no reason in principle why smooth rods alone cannot provide sufficient stability and rigidity. The key is securing them to the triangle base and top in a stable and rigid manner. I doubt whether this can be done adequately in a large printer by the use of printed ABS parts.

The problem is that if your rigidity is low then the printer will have lower resonant frequency. If this frequency is comparable with the print head oscillations when printing small infills then you will have a problem. The printer head can easily generate frequency at about 200 Hz at rather common accelerations of 7g. Single 70 cm long Ø 8 mm smooth rod will have resonant frequency the area of low tens of Hz. Of course this is a very rough computations but it does not sound promising for Ø 8 mm rods. But Ø 8 mm smooth rods have also a problem of dynamic bending in the range of about 0.1 - 0.2 mm so they are just not suitable even inside a rigid frame.
But one can get much thicker smooth rods easily and aluminum extrusions should do much better too.

8mm does seem rather thin for this application. I was thinking of 12mm smooth rods, like the ones on my Ormerod. The RepRapPro Fischer kit uses 8mm rods, but they are only 426mm long.

Yeah, should I have known that before buying an off-market DreamMaker OverlordPro... It has 8mm rods that are ~510mm of unsupported length with another ~50mm on both sides that goes into the injection molded case. Now the working length for this particular printer is about 270mm so the rest 240mm of the unsupported rod (the length below the carriages in the previous link) is something that design-wise is a mistake. The Overlord has 2mm aluminum extrusions on the outside (please see the linked pic) which can be bolted to at least a 3D printed part that would clamp the rods just below of the carriage when the nozzle is at Z=0.

Would this make a good 8mm rods Delta configuration? Of course I am also looking into encasing the whole printer somehow but so far I have no idea of an easy way. That enclosure would provide even more stiffness.

Edited 2 time(s). Last edit at 12/15/2015 08:37AM by realthor.

---

RepRap Lander concept on Concept Forge
RepRap Lander concept on RepRap Forums
My Things, mostly experimental stuff

Quote
hercek
The problem is that if your rigidity is low then the printer will have lower resonant frequency. If this frequency is comparable with the print head oscillations when printing small infills then you will have a problem. The printer head can easily generate frequency at about 200 Hz at rather common accelerations of 7g. Single 70 cm long Ø 8 mm smooth rod will have resonant frequency the area of low tens of Hz. Of course this is a very rough computations but it does not sound promising for Ø 8 mm rods. But Ø 8 mm smooth rods have also a problem of dynamic bending in the range of about 0.1 - 0.2 mm so they are just not suitable even inside a rigid frame.

I am trying to understand so please help me to. From my non-scientific pov, the only direct hit on the rods is the effector inertia when it moves away or towards the rod. I believe all moves in XY direction act upon all rods to some degree but mostly upon the rod that the movement is closest to a straight line passing through its center.

Now, that dynamic load would induce a certain amount of deflection, which I might be able to calculate eventually if I knew the max acceleration and the effector's weight (F=m*a).

There is rotational momentum also when the effector moves from side to side relative to one rod which induces torsional stress in the rod but I don't have a clue about that calculation for the purpose of determining the safest acceleration for a certain weight of the effector OR even if that is of any importance in our case.

Besides these two forces I can't think of much, nor do I understand what the resonant frequency means in this case. I assume the dynamic bending you are referring to is the deflection I have been talking about.

In my case, the smooth rods that are unsupported are ~510mm and because they are inserted into the supporting structure about 50mm on each side should decrease their flexibility by quite a bit compared to unsupported beams. I also plan, as I have suggested in the previous post, to reduce the unsupported length to a bit more than the working Z height that the carriages are capable of covering between the end stops and the nozzle-touches-the-glass point. I am pretty sure that at that resulting rigidity the weak point will be the 16bit controller that will not be capable of calculating greater accelerations than a certain limit.

So I want to take this backwards and first calculate what accelerations is an Arduino mega capable of knowing the end effector's weight and then calculating the deflection in the rods for the resulted force at that maximum acceleration. I will then get the length of unsupported 8mm rod that is acceptable before reaching critical deflection.

Thanks for helping me figure this out.

Edited 3 time(s). Last edit at 12/15/2015 06:20PM by realthor.

---

RepRap Lander concept on Concept Forge
RepRap Lander concept on RepRap Forums
My Things, mostly experimental stuff

Quote
realthor
...Now the working length for this particular printer is about 270mm so the rest 240mm of the unsupported rod (the length below the carriages in the previous link) is something that design-wise is a mistake....

I think you will find that when the effector is at the edge of the bed opposite one of the towers, the unsupported length below the carriage on that tower is somewhat less than 240mm. But IMO your idea is sound, if there is still significant unsupported length below it.

When you flick one of the smooth rods, what sort of twanging sound does it make?

The movement speed when running delta firmware on an Arduino Mega is limited by the speed at which it can calculate and generate stepper motor pulses. The acceleration that is possible is limited not by firmware, but by mechanics, effector mass. motor inertia, and supply voltage to the drivers.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Quote
dc42
The movement speed when running delta firmware on an Arduino Mega is limited by the speed at which it can calculate and generate stepper motor pulses. The acceleration that is possible is limited not by firmware, but by mechanics, effector mass. motor inertia, and supply voltage to the drivers.

Ok, so is it something that can be calculated? I Want to know how much of the still unsupported length should I reduce or how a light of an effector should I aim for to reach printing accelerations that would still make good prints at a faster peace. I guess I should work at the before mentioned items before dealing with supplying more voltage to the drivers. Ans what is motor inertia? Is it something stated on the motors spec sheet?

Thanks.

Edited 1 time(s). Last edit at 12/15/2015 06:52PM by realthor.

---

RepRap Lander concept on Concept Forge
RepRap Lander concept on RepRap Forums
My Things, mostly experimental stuff

CPU power of your controller will determine the maximum speed you can run. The other important limiting factors for speed are stepper inductance and the driving voltage and how quickly you can cycle your stepper drivers together with microstepping used.

Maximum acceleration depends on stepper motor torque and weight of your platform, diagonal rods, and carriages.
Notice also that the stepper motor torque depends on the stepper rpm (see torque charts of your steppers).

Yes, dynamic hotend position errors are the ones which happen because the hotend is accelerating (i.e. changing its speed).
Getting the estimates of dynamic behavior of a rostock like printer is rather easy. The equations for forces, accelerations, beam bending are simple. You can find them all even in wikipedia. Ignoring oscillations, you can consider only the maximum forces to find the top limit for the dynamic errors. And the maximum forces are limited by the torque of your steppers. I have a maxima notebook for this but it is not ready for publishing and probably never will be since it is a lot of work to get it into the publishable shape (as I learned on my delta calibration notebook).

I thought this thread is only about frame rigidity. Moreover the frames which are supported only by the smooth rods(*). From that point of view I do not believe that only six Ø 8 mm rods will be rigid enough (for a printer which has at least 50 cm height). If you are concerned of the maximum dynamic hotend position errors then your biggest problem is not Ø 8 mm smooth rods but the belts. E.g. glass core GT2 belts can have dynamic error of about 2 mm! That is nothing compared to about 0.2 mm of rod bending. Most belts used on delta printers are too springy. If you care than search the history of my posts. I already posted a lot about belts.

All this is a problem only if you intend to run the printer at high accelerations.

(*) your picture indicates something else

Yes, in my pictures you can see that the printer is not supported only by the rods but I fail to see how the extrusions help with deflection over such a length of unsupported rods (500+mm). It might help with keeping the top section concentric with the base yes, but many errors, like you also suggest, happen just between the rods and the effector, with little help from a more rigid frame.

But I am a bit overwhelmed by the plethora of things to consider. I can understand deflection (bending of the rods while the effector is accelerating). I can also understand the relation between the voltage, motor torque and the maximum acceleration that they can output.

Quote
hercek
CPU power of your controller will determine the maximum speed you can run. The other important limiting factors for speed are stepper inductance and the driving voltage and how quickly you can cycle your stepper drivers together with microstepping used.

I assume that for a Mega the top speed I can run a Delta with is a known thing by now right? There are countless mega/ramps Delta owners that for sure have tried to max out the speed.

Regarding these "dynamic hotend position errors" I believe you are referring to a sort of backlash (the effector accelerates and when decelerating to a halt it overshoots by those errors?). I can see these errors increase with acceleration. I see the hotend/effector position errors a compound metric made of rods deflection, mechanical overshooting due to acceleration and belt-induced overshooting due to belt elasticity/backlash in the pulley.

All in all I assume that there will be a weak link among all the above, which will limit everything else so I can focus maybe on that one.

So far we have:
1) Maximum speed dependent on CPU throttling => Top speed limited by CPU power;
2) Maximum acceleration dependent on motor specs, voltage in the steppers
3) Maximum acceleration dependent on effector mass, due to the compound effect of rods deflection, mechanical positional overshooting (backlash?) and belt elasticity (=> backlash, overshooting)

Am I correct with these conclusions?

PS: this thread is about frame rigidity allright. I am trying to find out the maximum height of unsupported rods for which the errors at maximum speed/acceleration are acceptable. Now I have 500mm of unsupported rods and I am planing to reduce that by clamping the rods somewhere without affecting the Z-range of the carriages.

---

RepRap Lander concept on Concept Forge
RepRap Lander concept on RepRap Forums
My Things, mostly experimental stuff

Quote
realthor
...
So far we have:
1) Maximum speed dependent on CPU throttling => Top speed limited by CPU power;
2) Maximum acceleration dependent on motor specs, voltage in the steppers
3) Maximum acceleration dependent on effector mass, due to the compound effect of rods deflection, mechanical positional overshooting (backlash?) and belt elasticity (=> backlash, overshooting)

Am I correct with these conclusions?

As hercek already pointed out, maximum speed can also be limited by the supply voltage being too low for the combination of motor inductance and steps/mm used. I had to increase my supply voltage to 24V when I changed to 0.9deg/step motors, in order to get speeds greater than about 150mm/sec.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Quote
dc42

Quote
realthor
...Now the working length for this particular printer is about 270mm so the rest 240mm of the unsupported rod (the length below the carriages in the previous link) is something that design-wise is a mistake....

I think you will find that when the effector is at the edge of the bed opposite one of the towers, the unsupported length below the carriage on that tower is somewhat less than 240mm. But IMO your idea is sound, if there is still significant unsupported length below it.

Just measured the length of rod under the lowest positions of the two carriages when the third carriage is up as the effector is touching it and you are right, it came down to less than 100mm, 93.5 more precisely. I still believe that rather than having 500mm unsupported rods, 400 is 20% better right?

Overlord's motors are pretty weak, 1.1A 200steps. I set the voltage according to the polulu formula for A4988 stepper drivers (Current Limit = VREF × 2.5). For 1.1A motors I got VREF=1.1/2.5=0.44
I have 2.5A motors prepared for another printer I want to build, a corexy, but I wait for the DRV8825 steppers to arrive. But I don't think I am going to have those motors on the delta... they kind of make more sense on the corexy.

Edited 1 time(s). Last edit at 12/16/2015 08:54AM by realthor.

---

RepRap Lander concept on Concept Forge
RepRap Lander concept on RepRap Forums
My Things, mostly experimental stuff

Quote
realthor

So far we have:
1) Maximum speed dependent on CPU throttling => Top speed limited by CPU power;
2) Maximum acceleration dependent on motor specs, voltage in the steppers
3) Maximum acceleration dependent on effector mass, due to the compound effect of rods deflection, mechanical positional overshooting (backlash?) and belt elasticity (=> backlash, overshooting)

4) Maximum speed you can melt plastic, determined by the heat capacity of your head, the heating element, the length of the melt zone, the surface area of your filament, the amount of viscosity of the plastic, and the torque of your extruder..... and probably some other factors.

The nice thing about deltas, is that most of the errors in backlash and so forth should cancel each other out, and the belts are pretty good at staying where they're supposed to be. I don't even mind a bit of flex in the arms, because it acts as a damper for some of the vibrations-- but it *does* limit the maximum print speed. You'll probably find with a reasonably solid build, because of acceleration, steppers, etc., you're going to be limited to 60-80mm/s max print speed anyway. Going beyond that will require some serious engineering.

You want a solid frame, built "square" as possible (uprights perpendicular to the horizontals), 120 spacing on the towers, smooth action on the carriages, and arms that are all identical in length.

I'd rather have a printer that's consistent-- last night, in the middle of a 6 hour print job, I had to go out for an emergency. Not willing to leave my printer running unattended, I paused the job, turned off the heaters, and killed the power supply (leaving the printer running off USB power). Came back an hour later, turned the power on, restarted the heaters, and when they heated up, restarted the print. You can't tell where I paused it.

For me, that's more important than outright speed.

Also-- If you want to speed up your printer, turn off honeycomb infill. It's pretty, it's elegant, it shakes the $@% out of the effector.

All of this is largely my personal opinion, and your mileage may vary, contents may settle during shipping, etc, etc, etc.

Mega can print at least 25 cm/s if your delta_segements_per_second is about 90. But your print speed will be limited by your extruder (how quickly can it extrude melted plastics).

Dynamic errors - the errors which happen because of interaction between inertial forces (F = m*a) and elastic forces (e.g. F = k*x for a belt .. the x in this equation is elongation ... it is the value of dynamic error).
Backslash is a different kind of error it is always about the same and happens at each direction change).
Then you can have static errors which are always the same at given location (regardless of acceleration).

To learn how to compute dynamic error due to smooth rod bending check out this: [en.wikipedia.org]
The point to use it easily is that bending moment (in a simple case like our 3d printer smooth rod) is changing linearly with the beam length. Well probably not exact since the smooth rod is continuous at the carriage location but it should be a good enough approximation.

But again, if you care about the dynamic errors check your belts first.

I have calculated in a rather less painful way the deflection for smooth rods at static loads (only by weight of the supported load at the middle of the length) in the "Is there any best practices for smooth rod diameter per hanging length?" thread I have opened some time back.

I am not trying to get lost in a notebook of calculations, but rather to learn a bit about how the whole thing behaves and to have an idea of where should I look for improvements and where to look for downsizing my expectations in order to have good prints. I don't want stellar speeds and crazy accelerations, I'd rather have accurate prints at slower speeds but I want those slower speeds to be the highest for that particular precision

About CPU throttling via complex calculations I can't do much about but changing the borad (will not do for the time being).
About the motors, again nothing, I have them already, will use what I have. I will change the stepper drivers though (A4498 to DRV8825) although 32 microstepping will not increase the quality but might lower the noise at least.
Abut the structure rigidity maybe I can do something like clamping the smooth rods for the exact length the carriages can travel on them anyways.
About the extruder maybe a geared extruder can be used instead of the direct drive extruder.
About the belts I will switch in my case from MXL to GT2.
About the effector there are ways to make it lighter, outsourcing the extruder (to a flying extruder for example if the current bowden setup is not satisfactory) and even the fans to maybe bowden-style silicon tubes that can also come from the flying extruder.

But for the purpose of this thread I will just conclude that clamping the smooth rods as close as possible to the maximum range of the carriages to the external frame is a good measure to improve quality and I will just do that. Then, for the remaining of unsupported rods, I will run some calculations to decide on the maximum accelerations that they can take before flexing too much, but for that I will first need to lighten my effector.

So, in the order of priority:
1) clamp the rods to shorten the unsupported length of 8mm rods;
2) make the effector lighter (Overlord's effector is 300g+effector plate+rods+the cable that has all wires and bowden tube, I am guessing at 350-400g at this time)
3) calculate max deflection that will still sport good results => update settings like acceleration,speed,etc
4) ? (tbd)

Because all these are mostly printer-specific I will not update this thread with the results, maybe I will open another one. At this point the conclusion would be that enough thick rods can be used for a no-frame delta but in my case the answer is not, which would satisfy my initial question.

Edited 2 time(s). Last edit at 12/16/2015 05:37PM by realthor.

---

RepRap Lander concept on Concept Forge
RepRap Lander concept on RepRap Forums
My Things, mostly experimental stuff

The "calculate deflection" discussion is highly speculative until somebody put some straingauges or others means of measuring the deflection, as to verify which approximations holds and what is important and not.

Obviously the longer or the thinner is bad, but when is it too bad?

You are now going to make a lot of modifications to a printer that you have not printed a single piece with? How are you going to know if it has improved or not?

It is cool that you want to hack your printer rather then print, but if you just read forums and dont stop and measure then it ends up being hacks for the fun of making them

Oh. This is a hobby. I also make a lot of stuff just because it looks interesting/ fun to do. Just think careful about why your doing it and don't forget to have fun hope that negates the somewhat negative wibe of before

Quote
LarsK
Oh. This is a hobby. I also make a lot of stuff just because it looks interesting/ fun to do. Just think careful about why your doing it and don't forget to have fun hope that negates the somewhat negative wibe of before

Haha, yes it does . I don't have too much time to work on it but between two other computer-related jobs I can read a post or write a post. I have printed with it calibration prints and wasn't impressed but that's because I haven't calibrated it properly. My try square seems to be not enough square so I have ordered another one and I am waiting for it. In the meantime I can't really calibrate the Delta and i want it calibrated physically first and then from software.

I didn't buy it to fill my life with plastic junk but to learn and use it whenever I have a need for something. And so far there is nothing that puts pressure on using it. It's like when you buy a new vacuum cleaner it's not like you will be so eager to use it every day

---

RepRap Lander concept on Concept Forge
RepRap Lander concept on RepRap Forums
My Things, mostly experimental stuff

Quote
realthor
About CPU throttling via complex calculations I can't do much about but changing the borad (will not do for the time being).
About the motors, again nothing, I have them already, will use what I have. I will change the stepper drivers though (A4498 to DRV8825) although 32 microstepping will not increase the quality but might lower the noise at least.

Changing to 1/32 microstepping also increases the load on the CPU, so it may reduce the maximum speed. Also the DRV8825 drivers don't work well with some types of motor. A better option is to stick with A4982 or A4988 drivers and change to 0.9deg/step motors. But I agree with LarsK, calibrate and use your printer before you change anything. Then you will find out what its weaknesses are and you can focus on those areas.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

realthor:

I looked at your link and the method you used seems to get the same results than my method. E.g. you got 0.06 mm deflection for 8x360 and I got about 0.15 mm for 8x700. So it is good enough. I also made a measurement (only an approximate one since I was lazy to build a precise rig for it) and the measurement is in agreement in the theory.

If your printer does not use steel core belts then the first thing you should do is to change the belts for some which use steel core.

Then you can clamp the smooth rods more.

To lighten the platform you can also use flex3drive or your own variation of it. Its error due to torsional deflection should be small:
[forums.reprap.org]

If your frame vibrates much when printing then you can mount something heavy on the top of the printer.

But the very first thing to do is a proper calibration of what you already have. Because you indicated you did not do this yet?!

Edited 1 time(s). Last edit at 12/17/2015 11:29AM by hercek.

Yes hercek, i started on the wrong foot with this printer... first a motor was making a terrible noise while traveling at high speeds, took the printer apart, troubleshooted, took the motor apart, etc etc ...
While doing this I was discussing here, so there wasn't much I could do anyway.

Now I found that one motor was wrongly wired, wired the connector the right way and put almost everything together. So i started calibration... and the numbers weren't matching up. I soon realized that my try square was not very square afterall ...so ordered a new one, now waiting for it.

I could try to make the frame true by using some other seemingly straight item I find around (like a bathroom scale square glass) but if the square wasn't square what are the chances the glass is square (even if made in a factory with industrial cutters).

Anyway this is where i am. I I don't want to do the calibration further without having the bed-to-towers squareing. I don't like having to sit here close the printer and not doing much but there's little I can do right now.

Ok, enough with the heartbraking story

Now, my calculations were made for static load. There is difference while the effector is moving there is a dynamic load that is pressing on the rods but for Deltas that shouldn't be too much trouble. Still I'll try to crunch some numbers while I still have time.

Edited 1 time(s). Last edit at 12/17/2015 01:09PM by realthor.

---

RepRap Lander concept on Concept Forge
RepRap Lander concept on RepRap Forums
My Things, mostly experimental stuff

This topic tried to answer my current question, and I'm glad I found it. I'm really wondering what the latest tests showed since the last post.

I plan to build my own delta, and I too want to use 6 smooth rods as rails AND frame pillars.
The difference is I want to make it 1000-1100mm tall, and plan to use 16mm rods.
Actually, I was roaming the forums to see if anyone used this method with a slimmer rod, like 12mm. The price difference is negligible. I just don't want to call a friend when I want to move this printer around, but I will if that's what it takes. Plus, I don't think I'll move it much from the designated place.

The rods will have threaded holes on both sides, along the rod axis, and will be bolted to both plates with M8 or M10 bolts.
I thought about threading the outside of the rods, and screw them in a threaded flange, like this one:
[www.nero.co.uk]

...and than bolt the flanges to the plates, but I'm concern that the rod threads do not offer enough tension to reach the desired rigidity.
The price of the rods made from good steel is acceptable from what I've seen, but I didn't decide what type of steel I'll actually use.
Maybe... I'll add 2 slim rods diagonally, opposed to each other (in a V shape), to eliminate CW and CCW motion of the frame, but than I'll definitely use 12mm main rods.

For bearings I plan to use these:
[sc01.alicdn.com]
or these:
[cdn3.volusion.com]
I'm waiting for a response regarding the actual difference between them in relation with what I plan to use them for.

I was surprised to see nobody included the base and roof into discussion.
I plan to use a 5mm thick base and roof, with precisely machined holes for the rod's bolts. A stage for the heated bed will be mounted at a certain height from the base plate.
Of course, underneath the stage I will mount the motors, PSU and what not.

I plan to make it unintentionally over kill, instead of discovering that it's a flimsy design.
Before you say anything about aluminum 2040 and so on, please don't. I like it, but this design brings expenses to a new level.
I wanted a very good frame, with prices that won't get me divorced. I'm not sure what's happening, but some companies mock people with their prices.

I also know about Robotdigg, and their entire linear rail design parts, but the shipping fee sucks big time. A simple order from the dc42's BOM costs me 98$ just for shipping to eastern Europe.
Add the actual products, and than to all that add 24% customs fee, and you start to wonder if it's worth it.
BTW, I tried to give you a second replay Mr. David, but it didn't worked, so I posted here.

I emailed Robotdigg regarding a more friendly shipping fee, and this is what they replayed:
"Please place your order online our webstore for more, thanks."
Short, funny and frustrating at the same time.
Sorry for the long post, and the english mistakes that I made.

The Fisher Delta printer uses smooth rods for its towers. However, it has acrylic panels (with large cutouts in them) bolted to the sides to make the structure more rigid. There are some pictures in the build instructions at [emaker.io]. I guess you could do a similar thing using plywood panels.

Edited 1 time(s). Last edit at 04/28/2018 01:56PM by dc42.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Of course smooth rods can be used for a frame, they are just like profiles except the hole is filled and they are round
Quality smooth rod is hardened steel and you can't thread . Anyway, better secure it in a hole, a slit and a bolt to squeeze it, rather standard method.

Edited 1 time(s). Last edit at 04/28/2018 02:55AM by MKSA.

---

"A comical prototype doesn't mean a dumb idea is possible" (Thunderf00t)

@Mr. David
That Fisher tip was gold, platinum even. Some of my plans changed a bit thanks to Fisher. I'm almost embarrassed to use his design in helping mine. So much work, detailed explained pictures, .stl , cad, and so on.
It shorten my work A LOT.

Quote
MKSA
Quality smooth rod is hardened steel and you can't thread . Anyway, better secure it in a hole, a slit and a bolt to squeeze it, rather standard method.

You mean secure it with this through the upper and lower plate hole?
[cdn2.bigcommerce.com]

I imagine it would be very... I'm not sure at all...
Are you sure you can't thread the rods like this? :
[flamcogroup.com]

Quote
Gheorghe
@Mr. David
That Fisher tip was gold, platinum even. Some of my plans changed a bit thanks to Fisher. I'm almost embarrassed to use his design in helping mine. So much work, detailed explained pictures, .stl , cad, and so on.
It shorten my work A LOT.

No need to be embarrassed, the whole point of open source development is that you can take other people's designs and adapt or improve them.

There is a forum section here for the Fischer, so you might want to ask the folks there whether they have encountered any issues that can be blamed on using that type of frame.

Edited 1 time(s). Last edit at 04/29/2018 09:14AM by dc42.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].