CoreXZ

My z travel is 325 and my x travel is 350. This works to 23.43 wraps. I just took an empty spool and was able to get 26 wraps on it, so the spools I have should work.

I'm going to try a different approach. I'm going to put all 24 wraps on one side, and string that to the x carriage in the closest position - which I believe is at the bottom on the opposite side. Then the other end, which is totally paid out will need to go the max distance.

Springs are good if you don't have a perfect system and want to maintain tension.

Tuners are good for retensioning when the string starts to relax.

Both solve different problems. I personally rely only on the compliance of the string to maintain tension. It has been many months since I tensioned it last and when plucked makes a high pitched note.

Edited 2 time(s). Last edit at 10/04/2014 12:41PM by nicholas.seward.

---

ConceptFORGE
Wally, GUS Simpson, LISA Simpson, THOR Simpson, Sextupteron, CoreXZ

I had a little better luck. I used my wrapping strategy as discussed above, and it worked fine. I was able to reach all 4 corners of the printer frame. However, after a few minutes, I noticed a sound I didn't like, and as I looked closer, I saw that the line coming into the spool was fouling the line going out again. I think I am just too close to the limit of what the spool can handle. Before I go the printed spool route, I am going to try lighter line. I'll start with the 60lb test you mention .

I haven't purchased my lighter string yet, but as an experiment I printed my own spools, just a little bit longer and with a slightly larger diameter than the aluminum ones. I restrung everything, and was happy with the way everything worked. The string paid out of the spools nicely, but even better it was taken up by the spool very neatly.

The only difficulty I am having is knowing just how much of the string to wrap onto the two sides of the spool. The calculation given earlier gives the total length of string needed, but how that gets wrapped onto the spool - how much is put on the left versus right side - depends on where the gantry is at the time of stringing. I'm slowly learning that allowances need to be made for the fact that the string needs to be retensioned a couple of times after the initial stringing.

Another option is to move away from spools entirely, and instead use a traction drive-- I pursued the idea specifically to address the same problems you are facing. My current version has six windings instead of the five shown in that post, as much for the resulting cable path as for the additional traction, and I have not experienced any slippage when the line is tensioned.

I'm having good success with braided fishing line - no stretch, and a traction drive was pretty successful with just five winds around a rubber bushing on my drive shaft.

I had a old prusa i2 frame lying around as i used up most of the parts for my mendel max. I thought that a great way to bring it back to life would be to convert it to a corexz. That way it would get rid of all the z screw issues.

I get a printable area of 190 x 200 with this setup.

I have used 625vv bearings and 18mm aluminum spool.





files

Nick
I'm a newwwwby so this may be irrelavent but I put together a 2-up and added a gantry pulley system.
After trying a couple of "strings" I picked up a roll of jewelry bead stringing wire at Michael's Craft.
It works great - virtually no stretch since it is wire.
It comes in a lot of variations but what I got was:
brand: Beadalon
Strands: 49 -- this is VERY flexable
size: 0.024" - 0.61mm
length: 10' - 3.05m
Have u tried that?

@ruggb: I have found the fishing line in no way to be deficient. I have been operating mine for 6+ months at high tension and have not had to retension or restring.

The wire you have sited seems to be braided(?) stainless steel with a nylon coating. I suspect that to be a great choice for an application such as this. It the wire isn't braided then I would expect it to suffer from fatigue in short order. I would be interested to know if it is in fact braided and if so how long you can run it under tension until you fail. (I am guessing on the 2up that it is under very low tension and experiences low cycles so pretty much anything would work great.)

---

ConceptFORGE
Wally, GUS Simpson, LISA Simpson, THOR Simpson, Sextupteron, CoreXZ

Hello

First of all thank you for a great tread.

I've been following this post for a while as I have a design I want to try out.

It's a mix of design really:
- build area 400x400x400 mm
- square cube design frame of 40/40 mm aluminium extrusion
- static heated printbed(insulation, heater, aluminium heat-spreader and glass on top)
- dual Bowden extruders(if Bowden setup can handle the distance, +800 mm)
- moving gantry in xyz of 20/20 mm aluminium extrusion with stiff aluminium cutout square plate frame on top, to ensure stiffness)
- with corexy horizontal movement(two motors with spectra line)
- with corez vertical movement(one motor with spectra line) like Seward sketched in post of August 6, but with 5:1 ratio and balanced out.
- Smoothieboard 5X

I guess I will get massive "peppers" for my moving gantry, but I really hate moving my printed object around at high speeds.

So I have a couple of questions:
- at what degree can one combine different coordinate systems, ie plain Cartesian-XZ, Core-XY, Core-Z(if I can call it that) and so on without having to write a new operate system?
- I guess if it will be a problem software vice I can revert to plain Cartesian-XY horizontal movement with two stepper motors and Spectra line and CoreZ for vertical movement(this should be, when put together, plain Cartesian-XYZ), right?
- how would a schematic drawing of CoreZ balanced out with 5:1 ratio and one motor on the base look like?
- I guess I would need a rather large spool to take up all the needed line for a movement of +400 mm?

And so the million dollar questions, that I've been scratching my head for a long time about.
Since the gantry is moving up and down, I guess there are no hope of getting rid of the motors from the gantry, and get them stationary on the frame, or is it possible with spectra line?
Could the slack be taken up, spooled in/out on a separate spool?

To make the gantry stable I've been thinking of using four slides, one in etch corner.

Hope this wasn't to much for a first post.

Thank you so much
Terje

So anyone care to comment on my take on trying to balance out the CoreZ?

See the schematic drawing attached.

The middle square is the lower part of the printer, and the left and right is hence the left and right side of the printer folded out.
The pink lines are my gantry, that I want to lift.


Terje

Hi Nicholas,
I'm a complete noob in the 3d printing world... Last week I started looking around in order to get informations on how printer works, designs and all thre stuff needed to make and use a 3d printer.
I ended up in your design... I like it very much. I think that IF I'm going to make a 3d printer I will follow your design.

Being a noob I've lot of questions...
* In order to reduce the width of the X (having the same printable area) do you think it is possibile to bring bearings close to each other? (like in my attached file)... I don't se any drawbacks in doing this, do you?
* You used 625VV bearings... any disadvantages in using little ones (like 624VV)? They should be cheaper (and also will nuts, bolts, washers will be cheaper)...
* As far as I have understood you developed a new firmware in order to drive the printer... any info?
* Electronics: does your corexz rely on the standard arduino mega / ramps to do the magic?
* I like very much the use of guitar tuning peg to tension the strings... will any standard guitar peg works or I've to look for special characteristics? Which one did you use?

Thanks,
Paolo

Edited 1 time(s). Last edit at 12/09/2014 10:17AM by tulindo.

Paolo,

With your configuration the string length changes as the motor turns. Compare a drawing with the pulleys very close vertically with one where they are very far, and you'll see the angles between the strings and vertical are quite different in the two extremes. The strings are along the hypotenuse of a triangle rather than being along the vertical side.

That could be compensated for by various methods, like a spring, or by routing the string around other features designed to have compensating behaviors. However, in any case the relationship between the motor angle and the pulley positions will not be linear, and you'll have to compensate for that mathematically in your driving software or the things you print will be distorted somewhat.

It's possible to make the system more compact with smaller errors, such as moving some of the pulleys to a second plane but on the same axes as the first ones--then the angle changes as things move can be smaller. However, there will still be some distortion unless all the vertical pieces of the string are always precisely vertical, parallel to the vertical guide rails etc. It's generally worth keeping the geometry as true as possible--there are enough other error sources as it is! (E.g. cable spool not precisely concentric with motor shaft, angles of strings between spool and first pulley changing as the spool winds, etc.)

Dave

Thanks Dave, looks like I forgot that trigonometric function are not linear...
Your answer is very clear.

Quote
tulindo
* In order to reduce the width of the X (having the same printable area) do you think it is possibile to bring bearings close to each other? (like in my attached file)... I don't se any drawbacks in doing this, do you?

There are lots of ways to reduce width but as DaveGadgeteer mentioned you want to keep the lines vertical. Imagine doing a 90 degree fold along the z axis. I kept it all flat in one plane for simplicity. Width isn't really a problem. It is about as wide as most 200mm printers would be.

Quote
tulindo
* You used 625VV bearings... any disadvantages in using little ones (like 624VV)? They should be cheaper (and also will nuts, bolts, washers will be cheaper)...

I just like the 625VV better. The 623VV or 624VV bearings would be great. The 625VV bearings are actually thinner in their axial direction. 625VV can also use normal washers for spacers unlike the others. At the end of the day, they are all good.

Quote
tulindo
* As far as I have understood you developed a new firmware in order to drive the printer... any info?

PM your email and I will send you my hackish MarlinXZ.

Quote
tulindo
* Electronics: does your corexz rely on the standard arduino mega / ramps to do the magic?

I use a RAMPS-like controller.

Quote
tulindo
* I like very much the use of guitar tuning peg to tension the strings... will any standard guitar peg works or I've to look for special characteristics? Which one did you use?

Most any should work. You will have to figure out how to mount it so pick one that you like the form factor. For me, they are too long so I have to do lots of work to get the tension area in the right spot.

---

ConceptFORGE
Wally, GUS Simpson, LISA Simpson, THOR Simpson, Sextupteron, CoreXZ

Hello all,

I'm loving the Core XZ concept and wondering about peoples experiences with printed spools?

Cheers,

@DennyP: I have used printed spools with no truing on GUS Simpson. It is not hard for the error to exceed 250 microns. For a standard serial bot this would be much of a problem because the period of this error is spread over 60mm. With correct tolerances on parts, that would never be a problem. The problem you will run into with the CoreXZ is that the error will keep the prints from being level. You can solve this by characterizing the spool but you will have to redo this every time you restring the machine and you will have to modify firmware to allow you to put correction data in.

However, a printed spool that has been trued should be great. A block, sandpaper, and a drill would be enough to true it.

---

ConceptFORGE
Wally, GUS Simpson, LISA Simpson, THOR Simpson, Sextupteron, CoreXZ

Thanks to everyone for their contributions to this fascinating and informative topic. I'm particularly interested to hear what seems to be very successful uses of spectra/fishing line.

My basic question is this. Is the consensus that (with good design) spectra/fishing line for X, Y and Z gives similar (or perhaps even better?) print precision and speed as GT2 belts?

I'm not building a CoreXZ so sorry for being a bit OT. My plan is: for X&Y either CoreXY or DW-G (still deciding pros/cons) driven by line; Z lifted in 4 corners by line & pulleys too (can this be designed accurate enough compared to belt-driven lead screws?); 400-500mm cubish build area; enclosed to run up to about 80C; based on V-slot with wheeled carriages.

I would be grateful for any suggestions please.

Quote
SydneyDesigner
Thanks to everyone for their contributions to this fascinating and informative topic. I'm particularly interested to hear what seems to be very successful uses of spectra/fishing line.

My basic question is this. Is the consensus that (with good design) spectra/fishing line for X, Y and Z gives similar (or perhaps even better?) print precision and speed as GT2 belts?

I'm not building a CoreXZ so sorry for being a bit OT. My plan is: for X&Y either CoreXY or DW-G (still deciding pros/cons) driven by line; Z lifted in 4 corners by line & pulleys too (can this be designed accurate enough compared to belt-driven lead screws?); 400-500mm cubish build area; enclosed to run up to about 80C; based on V-slot with wheeled carriages.

I would be grateful for any suggestions please.

Quote
wikipedia
It is not advisable to use UHMWPE fibers at temperatures exceeding 80 to 100 °C (176 to 212 °F) for long periods of time.

Belts have a few benefits. They can be temp resistant and you don't get triangulation errors from them walking on the spool. Strings also have a few benefits. More routing flexibility. Cheaper. They can be tensioned with guitar tuners. Etc.

As far as which one is better, it depends. Give me the specs for a target machine and I can probably hit it with either material. I tend to think that fishing line belongs on small to medium sized printers. I would classify yours as a large printer.

---

ConceptFORGE
Wally, GUS Simpson, LISA Simpson, THOR Simpson, Sextupteron, CoreXZ

Thank you for your help.

Quote
nicholas.seward
I tend to think that fishing line belongs on small to medium sized printers. I would classify yours as a large printer.

Why do you think line might not tend to belong on larger printers like my 400-500 mm cubish printer? Is it mainly the triangulation and longer spool lengths? Or is there something else?

It seems to me that one of the main benefits of line over belts for larger printers is the considerable reduction in moving weight. For my printer for example, the CoreXY GT2 belt weight would be about 500g. The belts are not only linearly accelerating and decelerating all the time but and are also constantly changing rotational inertia (they are effectively large spinning lengths of mass). This all requires more forces and torques on the printer. Do you think it is beneficial to reduce these forces and torques by using line instead?

... it's the higher elasticity with longer lines.

When building my first XY-laserplotters I've used high-elastic steel wires with diameters from 0.5mm to 1.5mm and got some elasticity caused resonancies even with the 1.5mm thick steel wires, when the driving length was 700mm to 1m ...

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

@SydneyDesigner: VDX gets at the core of the issues. That said you have to look at the design requirements. As I said, I think I could hit the requirements with either a belt or a string. The problem is that it becomes harder and harder to solve the problem with string as the size of the working envelope increases.

Before I go through the though exercise of designing a printer of your size with string, I want to remind you that string will fail at 80C so it isn't appropriate for you situation.

As the string becomes longer, the system become more elastic to the point that you can see horrible artifacts when you hit the natural frequency and when you make sharp turns. You also have horrible triangulation error due to the longer spool. (The way I do it, the spool has to store the length of travel plus on the spool.) We can solve the elastic problem with thicker string but then that exacerbates the triangulation problem. We could increase the diameter of spool but then the resolution goes down if we direct drive them. You could drive the spool with a belt reduction. (Ironically, one of the few ways to do it cheaply with little backlash.)

So now the system has the rotating mass of the spool, the thicker string, and the reduction belt and pulleys. I suspect that the weight savings isn't as much as you would hope for. Additionally, the complexity has increased and this setup will look very foreign to most technicians that you might have fix your machine in the future.

Don't get me wrong, I love string drive but by my math the sweet spot for its usage is in the 200mm range.

Edited 1 time(s). Last edit at 02/12/2015 12:50PM by nicholas.seward.

---

ConceptFORGE
Wally, GUS Simpson, LISA Simpson, THOR Simpson, Sextupteron, CoreXZ

Thank you for your replies @VDX and @nicholas.seward - I think you have helped me to avoid going off in the wrong direction. But just to clarify, a couple more questions please.

Is it impossible to get line with a stiffer elasticity modulus than the typical quoted 0.8 - 1.5 GPa for UHMWPE (Spectra, Dyneema)? And are there no lines at all that can be operated at 80C?

Even if we could get line with a 20GPa elasticity modulus (which I calculate would give it a similar stretch as a typical GT2 belt), am I right in thinking that line would still cause worse elasticity resonance issues compared to the more heavily damped construction/profile of a GT2 belt? (VDX's steel cable experience seems to suggest this.)

Are the aramid/kevlar cored GT2 belts the stiffest and most heat resistant? Where can I get them?

I can't give you a specific recommendation with out much more time and more detailed specification.

The natural frequency and amplitude of that frequency is a function of... modulus of elasticity, diameter of cord, length of cord, moving mass in the system, max jerk setting, and max acceleration setting.

I use 100lb test fishing line but I have seen 300+lb lines. The modulus is a function of the material but if you increase the diameter you increase the natural frequency and decrease the amplitude.

UHMWPE also goes through a transition around 80-100C so you will have to use another material. I am not sure what the availability is of other high tensile synthetic cords and what their properties are. See what you can find and look up their specs. Trying them in a printer will be the ultimate test so please share your findings.

Steel wire can with out a doubt be made to work for you. The question is whether belts would be a simpler or cheaper solution.

Bottomline: Define what you want your printer to do and then find a solution that works for you. There are an infinite number of ways to do it so don't get too wrapped up in finding the right solution initially. Try something. If it doesn't work, identify what is wrong and address it. (Math, physics, and engineering could save you money but trial and error will get you to a solution just as well.)

---

ConceptFORGE
Wally, GUS Simpson, LISA Simpson, THOR Simpson, Sextupteron, CoreXZ

... I've first used steel-reinfoced fishing line, but the hull weared off, so changed to pure steel wires - 'high-elastic' steel-wires with 19 or 47 strands and diameters of 0.5 to 1.5mm are much more stiff than fishing line with the same diameters.

Instead of wire you can use steel stripes too, but then you're right on the way to toothed belt

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

Thanks for your help. I'm trying to find some good examples of printers built with steel wires. But haven't found any yet.

I am experimenting with corexz, but as I research Spectra line it appears that it has a problem called "creep" where it elongates when held under tension over time. For this reason they don't like to use it for standing rigging on performance sailboats, but use Kevlar or Vectran instead. I don't want to have to re-tension and re-level my printer after I have gone some time without using it. Has anyone experienced this problem?
I was thinking of using Kevlar, but it supposedly has more problems with abrasion and fatigue. Vectran seams to be be perfect on paper as it has lower creep and better abrasion resistance. Has anyone tried either of these fibers and are there any problems I should know about before I buy some?
I am planing on experimenting with 120lb test Kevlar with PTFE coating available from twinline for $6.75. The other option is 80 or 200 lb. Vectran available in 250 ft spools for $5 or $30 respectively.

Edited 1 time(s). Last edit at 02/18/2015 02:57PM by robomaster.

I can't speak about the others but I have use spectra. I had to tension/move-through-range-of-motion/tension/move/tension/move/... at first. It quickly hit a point where the tension didn't decrease after an iteration. I have let the 100lb test line sit under 5lbs of tension for months at a time (under heavy use) with no noticeable lose of tension. If it is creeping now, it is on the time scale of years. (I have also be operating it for around a year with no sign of the string fraying.) I suspect that keeping the tension to 5% of its rated load is what keeps the creep to a minimum. (Unfortunately, I do not have any hard measurements for you. My theory is that the creep under constant load will either asymptotically approach a value or slow down over time. I think I will put a kid on this research project.)

That said, I have two tuners for the XZ stage. One is for level and the other is for tension. A quick turn on the tension peg will release or reapply the tension without messing with the level of the gantry.

I would be curious how the other string do. I wouldn't worry about abrasion resistance. A good design will not have the string sliding over any surfaces. I would worry about fiber fatigue. Wrapping around the pulleys will quickly start breaking fibers of more brittle materials.

Quote
robomaster
I am experimenting with corexz, but as I research Spectra line it appears that it has a problem called "creep" where it elongates when held under tension over time. For this reason they don't like to use it for standing rigging on performance sailboats, but use Kevlar or Vectran instead. I don't want to have to re-tension and re-level my printer after I have gone some time without using it. Has anyone experienced this problem?
I was thinking of using Kevlar, but it supposedly has more problems with abrasion and fatigue. Vectran seams to be be perfect on paper as it has lower creep and better abrasion resistance. Has anyone tried either of these fibers and are there any problems I should know about before I buy some?
I am planing on experimenting with 120lb test Kevlar with PTFE coating available from twinline for $6.75. The other option is 80 or 200 lb. Vectran available in 250 fl spools for $5 or $30 respectively.

---

ConceptFORGE
Wally, GUS Simpson, LISA Simpson, THOR Simpson, Sextupteron, CoreXZ

Thanks for the answer. I teach a high school physics class, so I think I will test these fibers as an example science project next year. I just need to find some good suppliers that have genuine materials for cheep. Does anyone know what brands/suppliers of Spectra are most commonly used for 3D printers?

Quote
robomaster
Thanks for the answer. I teach a high school physics class, so I think I will test these fibers as an example science project next year. I just need to find some good suppliers that have genuine materials for cheep. Does anyone know what brands/suppliers of Spectra are most commonly used for 3D printers?

Any place that has fishing supplies will have UHMWPE braided fishing line. I personally got 1000m of an off brand from china. Who knows what the variance is from brand to brand? I found that the name brand stuff lost its color quicker. The off brand stuff smelled horrible. I trust that 100lb test fish line from any brand can withstand the measly 5ish lbs required.

---

ConceptFORGE
Wally, GUS Simpson, LISA Simpson, THOR Simpson, Sextupteron, CoreXZ