Sorry I missed this message earlier. Looking at my notes on the thingiverse design page, I had 370mm of distance on a 500mm length. If you change that to 1000mm lengths, you should be able to get an additional 500mm which would make that a total of 870mm travel along that axis. I do not have that machine currently put together, so am not able to answer your 2nd question.

 

I don't think the accuracy problem is related to the bearings but will be interested to see if you get better accuracy with them. From my testing, the accuracy problem seems related to the wheel travel and having better grip on the wheels would help the most. The other two similar rolling plotter designs I have seen on here have a different type of wheel that I could not find a source for & they did not give a source for them. For all the wheel gripping surfaces I have tested, the O-rings gave me the best accuracy. My latest idea was to put 6 O-rings on each wheel and use the double wheel plate on each end to support the wheels on both sides. I have had other projects come up & haven't gotten around to testing that idea yet. The double wheel plate will lose some travel distance along that axis because the V-slot will extend to 2nd wheel plate.

 

I put this design away for a few months & am considering going further down this rabbit hole using one motor for both sets of wheels with a connecting M8 smooth rod between them. I had been thinking about doing something like this for a while & after seeing this belted Z-axis design I could see how to do it. GitHub - kevinakasam/BeltDrivenEnder3: All parts including Voron files for the Belt Driven Ender 3 V3 1

My biggest problem was being able to adjust the belt for the motor to M8 rod & also be able to adjust the belt between the M8 rod & wheels. Looks like this should work. I won’t have to reprint the wheels at least. I am also going back to the O-ring wheels as those worked the best for me & since the wheels are supported on both sides now, I added as many as would fit (6 on each wheel).

I had to order another closed loop timing belt. Shortest I have is 188 & it needed to be in 148-152mm range. 148 is what I calculated, but when I drew all the teeth in the belt, comes out closer to 152. I ordered 146, 148, 150 & 152mm lengths since they were fairly inexpensive. One of those should fit. I had to change the wheel spacing from 110mm to 115mm for the wheel belt to fit. I changed the 2060 brackets to 2 parts on each side with a T-nut connecting to the 2040s underneath. I have a +/-3mm adjustment range for the belts. Maybe I will have the new parts printed by the time the new belts come in.

 

That seems like it would complicate too much for me & not sure if I would know how to implement that. Thanks for the suggestion though.

 

I am finally getting back to what will probably be the final design of this machine. I have it mostly assembled now. I thought getting that long m8 rod would be a problem getting lined up properly, but using a simple 3d printed shim 26mm wide while tightening each bearing position made it quite easy. I also changed the F625zz bearings to a F695 2RS ABEC-7 which were about $12 for 20 of them (only 8 required).

After looking at this a couple of days, I realized I needed a little more distance between the battery PS & the top mounted carriage motor. This photo shows using 2 - 2040x1000mm V-Slots & a 2060x250mm V-Slot in the middle. To give me more distance between the PS & motor, I need to add 2 short 2060s, either cut a 250mm in half or use 2 of them. I then realized that this part does not need much support. It is just there to support the weight of the battery PS & the carriage motor. The only stress it has is the belt tension on the motor, so I think a 3d printed custom 2060 will work here. I say custom because it will incorporate the connection mounts to the 2040s on each side and look like this assembled.

The 1st print feels quite solid printed with voxel PLA+ (I still prefer eSun PLA+ to this). This part I printed with 5 perimeters, 50% rectangular infill & .12mm-.28mm adaptive layer height with PrusaSlicer & it took 4 hours & 40 minutes. My short test print with .2mm layer height didn't need any filing, but with the .28mm layer height I had to use a jewelers file a little to smooth it out to fit between the 2040s. I added a little bit of plastic to the back of the base and a 5mm brim to give it better adhesion on the build plate.

Since this was so solid, I am going to back off my settings to see if those will be sufficient. Second print is using 3 perimeters, 30% rectangular infill & .16mm-.24mm adaptive layer height. That should cut about an hour off the print time. The .24mm layer height max might be enough to not need to file it.

 

Almost there. While assembling the new 3d printed custom 2060 mounts I realized the bearings needed to be close to where the motor belt connects to the M8 smooth rod via that 20T gear pulley, otherwise it can flex that M8 rod too much. I redesigned the 2060s to use T-nuts & made a separate part for the idler bearing mounts. That way I can put them as close as possible to the motor gear. Might could just use 1 bearing there, but 2 will give it support on both sides of that 20T gear pulley & that is what I have done on the outside wheel gears. This should work better. Here is a view from underneath the plotter.

 

It feels pretty symmetrical now & am liking how this latest version is finally coming together. I did find one more change to the Wheel motor mount. Before I had it mounted with 2 M5 T-Nuts on the inside top & side. In order to get the other 608-2RS bearing as close to the drive gear, I moved all 4 mounting holes to the top. Might only need 3 mounting holes but left it at 4 for extra support. After I replace this part, I should be ready to put the carriage belt on & test run it. Here is a top & bottom view where that motor is. I can get the bearings even closer than what it is drawn but can visually see it better this way.

 

Here is the 1st movement test of this latest configuration.

 

Here is an interesting laser engraver that rolls along a flat surface that someone pointed me to on another forum. Looks like their accuracy is better than mine.

 

After watching that video, it seemed like adding some extra weight to this machine might help it, so I started looking at where I could add it easily. The obvious place is the cavities of the V-slot 2040s & the simplest choice for that seems to be BBs. Before buying these, it seemed like a good idea to have a close approximation of how many I might need. I asked ChatGPT how many BBs would fit into the cavity of a V-slot 2040 & it wasn't able to do that. Here is the excerpt I used from ChatGPT with using a cylindrical cavity.

how many daisy bb’s will fit in a 13.834 mm diameter 1 meter long cylinder and how much would it weigh

To calculate the volume of the cylinder, we need to use the formula:

V = πr²h

Where V is the volume, r is the radius (half the diameter), and h is the height (in this case, the length of the cylinder).

First, we need to find the radius (r) of the cylinder:

r = d/2 = 13.834/2 = 6.917 mm

Now we can calculate the volume of the cylinder:

V = πr²h = π(6.917 mm)²(1000 mm) = 150,295.33 mm³

Next, we need to determine the volume of one Daisy BB. Unfortunately, I don’t know the exact dimensions of a Daisy BB, but let’s assume it’s a sphere with a diameter of 4.5 mm. The formula for the volume of a sphere is:

V = (4/3)πr³

where r is the radius (half the diameter). Thus, the volume of one Daisy BB would be:

V = (4/3)π(2.25 mm)³ = 38.485 mm³

Now we can divide the volume of the cylinder by the volume of one Daisy BB to find out how many BBs would fit inside:

150,295.33 mm³ ÷ 38.485 mm³ ≈ 3,904.7 BBs

Note that this is an estimate, and the actual number of BBs that could fit inside the cylinder may vary depending on how tightly they are packed.

Finally, to calculate the weight, we need to know the weight of one BB. Again, I don’t have this information, but let’s assume it weighs 0.36 grams (which is the weight of a different brand of BB).

Thus, the total weight of the BBs would be:

3,904.7 BBs × 0.36 grams/BB = 1,405.1 grams

So the BBs would weigh approximately 1,405.1 grams (or 1.4 kg).


It came up with 3904 BBs. Someone on the V1e forum pointed me to this page Sphere packing - Wikipedia & said that number should be more like .63 factor of that. Turns out to be even less than that, probably because that the odd shape of that cavity. I calculate around 1500 BBs. I weighed a 100 of them to get a good average weight. The weight of 1 BB came up to 0.14oz or .397 grams. I did two tests of adding the BBs. The 2nd time I shook the V-Slot after adding each few hundred & was able to add another 30 BBs than the original test. This gives me about 1lb 5 oz of extra weight per V-Slot 2040.

Here are some photos of my experiment. of course I had to 3d print a funnel & a couple of simple endcaps.

 

I have not tried it with the extra weight yet. I plan to do an Assembly video of it and have it all apart at the moment.

Yea, but can you get a funnel that closely matches the V-Slot cavity? Vase mode print. 0.104oz. Cost a nickel in plastic & had more fun factor in designing rather than going to dollar store.

 

I started going out of my comfort zone & made an Assembly video for this design. Here is the link to Part 1 that is all the Subassemblies. The final assembly will be in Part 2.

 

I added Mostly Corrected Wheel Plate Assembly video since I had a few mistakes in Part 1. Still had some mistakes in this video but added overlay text & graphics to clarify this assembly. I have the final Part 3 done also but will wait until tomorrow to upload that, since that will take a while to upload.

 

Here is Part 3 which is probably the final part of this assembly build.

 

Interesting. I had not seen weights made from Tungsten. Thanks for the reference.