Looking through what I had bookmarked on Ebay for this work, I came across the item that started it all off. An A8 printer for £86. I did want to go for a mill rather than a 3d printer, mainly because eventually I felt that I could use a mill more. But considering I'm looking at that sort of money as a 'test' for the mill, I am now wondering whether my 'test' should be that printer. At least I'll be able to figure out the CAD side of things, get familiar with the layered approach, etc.
Edit: OK, this is on pause for now, I bought the 3d printer... I might come back to this in the future, possibly with some rants and complaints about how I should have gone straight for the CNC mill and not bothered with additive manufacturing...
Wednesday 15 May 2019
Tuesday 14 May 2019
Shrinking the bearings
The first task I have given myself is to figure out how to shrink the carriages, and therefore there is a need to shrink the internals.
Bearings
The standard bearing is a skateboard bearing, 8mm inside bore, 22mm outside diameter, 7mm width. The code is 608zz.
6 is the bearing type (single row deep groove)
0 is the material strength (extra light),
8 is the inside bore, and
zz means both sides are enclosed.
By the looks of it, these skateboard bearings are as small as they can get without compromising on strength. The best I can see is a 7mm internal bore, but because M7 is then a peculiar size for the bolts it would make a huge mess for the sake of a 1mm saving. They're also a narrower bearing which probably wouldn't be good. So the bearings cannot be shrunk.
Bolts
The standard bolts here are M8. This seems to basically line up with the bearing internal diameter. Unfortunately as described above there is no option of reducing the size of the internal diameter, so the bolts also need to stay at M8. I think this is a good thing though, the next size down would be M6 and I wasn't comfortable with an M6 bolt. So no shrinking here either.
Motor
This is an interesting one. Basically the motor can be any size as it sits outside of the carriage. Therefore I might as well stay with the NEMA 17 or 23 motors.
Belt
Erm, no. I don't want this shrunk. It's only 10mm already, it would have to be reduced quite considerably to actually make a saving.
Frame
At last, something that can be shrunk!! The standard frame is 40mm x 40mm. The biggest steel I have is 25mm by 25mm. Therefore it makes sense to shrink it down to 25mm.... or does it?
I believe one of the reasons for the frame size is to reduce rotational errors. This comes back to an earlier post where I said the bearings ride near the edges of the frame. If they were in the middle, there wouldn't be anything stopping the whole carriage rotating on the frame. The same applies here, if I bring the bearings in to ride on 25mm square section, I'm losing 15mm of twist resistance. OK, I'm simplifying things but it makes sense in my head.
But... there can't be rotational forces on the Y movement, because the gantry and the two motors prevent it. So in this case I could have the bearings centrally mounted and nothing would happen. Therefore I can use 25mm square section and it should be fine. I'll have to revisit the X movement maybe, but I can do that after.
Conclusion
Well, the conclusion is that it's not going to come down by much!! Each direction inside the carriage needs to accommodate a 22mm bearing, 25mm steel, then another 22mm bearing. 1mm either side for clearance, and 3mm thick steel. This gives a total of 77mm. I could use 80mm steel if that was it, but it's not. I need to add 13mm for the belt drive (10mm belt, 1mm shoulders, and 1mm clearance). So the one direction would need to be 90mm. And of course, it would have to be 90mm square as it's almost impossible to get 80 x 90.
I think this ties in with what I expected, I'm reducing the frame by 15mm so I wasn't going to gain much more on the carriages. I'll go for 90mm square section.
Then all it leaves is the length of the carriage, currently 130mm. I'll have a bit more of a think about that one...
Bearings
The standard bearing is a skateboard bearing, 8mm inside bore, 22mm outside diameter, 7mm width. The code is 608zz.
6 is the bearing type (single row deep groove)
0 is the material strength (extra light),
8 is the inside bore, and
zz means both sides are enclosed.
By the looks of it, these skateboard bearings are as small as they can get without compromising on strength. The best I can see is a 7mm internal bore, but because M7 is then a peculiar size for the bolts it would make a huge mess for the sake of a 1mm saving. They're also a narrower bearing which probably wouldn't be good. So the bearings cannot be shrunk.
Bolts
The standard bolts here are M8. This seems to basically line up with the bearing internal diameter. Unfortunately as described above there is no option of reducing the size of the internal diameter, so the bolts also need to stay at M8. I think this is a good thing though, the next size down would be M6 and I wasn't comfortable with an M6 bolt. So no shrinking here either.
Motor
This is an interesting one. Basically the motor can be any size as it sits outside of the carriage. Therefore I might as well stay with the NEMA 17 or 23 motors.
Belt
Erm, no. I don't want this shrunk. It's only 10mm already, it would have to be reduced quite considerably to actually make a saving.
Frame
At last, something that can be shrunk!! The standard frame is 40mm x 40mm. The biggest steel I have is 25mm by 25mm. Therefore it makes sense to shrink it down to 25mm.... or does it?
I believe one of the reasons for the frame size is to reduce rotational errors. This comes back to an earlier post where I said the bearings ride near the edges of the frame. If they were in the middle, there wouldn't be anything stopping the whole carriage rotating on the frame. The same applies here, if I bring the bearings in to ride on 25mm square section, I'm losing 15mm of twist resistance. OK, I'm simplifying things but it makes sense in my head.
But... there can't be rotational forces on the Y movement, because the gantry and the two motors prevent it. So in this case I could have the bearings centrally mounted and nothing would happen. Therefore I can use 25mm square section and it should be fine. I'll have to revisit the X movement maybe, but I can do that after.
Conclusion
Well, the conclusion is that it's not going to come down by much!! Each direction inside the carriage needs to accommodate a 22mm bearing, 25mm steel, then another 22mm bearing. 1mm either side for clearance, and 3mm thick steel. This gives a total of 77mm. I could use 80mm steel if that was it, but it's not. I need to add 13mm for the belt drive (10mm belt, 1mm shoulders, and 1mm clearance). So the one direction would need to be 90mm. And of course, it would have to be 90mm square as it's almost impossible to get 80 x 90.
I think this ties in with what I expected, I'm reducing the frame by 15mm so I wasn't going to gain much more on the carriages. I'll go for 90mm square section.
Then all it leaves is the length of the carriage, currently 130mm. I'll have a bit more of a think about that one...
Monday 13 May 2019
Starting to form a picture...
So I have the start of an inkling of a thought of what is going to happen. It's mainly driven by the 'goodenoughCNC', a strangely simple designed CNC device. It uses bog standard bearings and steel frames to do the job.
The main units are a set of three carriages. Two for left and right Y movement for a gantry, and a third for X movement. Well, I should say there are four carriages as one is required for Z movement, but that's the next subject.
The carriages are currently 100 x 100 blocks, that will run on metal bearings keeping it fixed on a 40x40 frame. Metal bearings, simple frame, square blocks, all adds up to nice and cheap!!
To keep things still inside the carriage, there are three bearings per plane (six in total). Two on one surface, one on another. Typically in a triangle pattern. All bearings act on the edge of the 40 x 40 metal, I'm guessing if they were driven on the middle surface the whole carriage would be able to rotate which isn't a good thing. The bearings are fixed in the carriage on threaded rod, a total of six lengths going through the carriage.
The NEMA motors sit on the side of the carriages, and are belt driven. I was initially going to avoid belt driven as I thought they were inaccurate. Technically they are... they're 0.2mm tolerance vs 0.1mm tolerance of a linear bearing!! The belt is kept on the drive by four further shouldered bearings, and fixed on the end by a screw driven tensioning device.
What's more, a quick price up of the relevant parts suggest each bearing is about 50p, the toothed drive for the motor is £2 and the belt itself is another £2. Threaded rod is about £5 for a length, with nuts, washers and stuff maybe making up another £3. Given the steel is effectively free, that makes each carriage;
Carriage bearings 6 x 50p = £3
Belt bearings 4 x 50p = £2
Toothed Drive = £2
Belt = £2
Rod = £5
Hardware = £3
That's £17 per carriage. Maybe the same again for a decent torque stepper (one per carriage), that's about £100 to get from nothing all the way to a working transport that just needs a Z movement. That's a nice budget for me.
The only thing is the size, 100 x 100 is quite large considering the entire device is only 500 x 500. One of those either side would leave a cutting bed of maybe 250 x 350. Is that enough? I'm not sure. Reducing everything by 50% would be better, that would leave me with 350 x 450. It would also mean the frame could go to 25mm or even 19mm, which I have a lot of. I'd have to figure out how much to shrink the bearing and threaded rod, and unfortunately the belt and pulley couldn't shrink.
Hmm, prototype time I think!!
The main units are a set of three carriages. Two for left and right Y movement for a gantry, and a third for X movement. Well, I should say there are four carriages as one is required for Z movement, but that's the next subject.
The carriages are currently 100 x 100 blocks, that will run on metal bearings keeping it fixed on a 40x40 frame. Metal bearings, simple frame, square blocks, all adds up to nice and cheap!!
To keep things still inside the carriage, there are three bearings per plane (six in total). Two on one surface, one on another. Typically in a triangle pattern. All bearings act on the edge of the 40 x 40 metal, I'm guessing if they were driven on the middle surface the whole carriage would be able to rotate which isn't a good thing. The bearings are fixed in the carriage on threaded rod, a total of six lengths going through the carriage.
The NEMA motors sit on the side of the carriages, and are belt driven. I was initially going to avoid belt driven as I thought they were inaccurate. Technically they are... they're 0.2mm tolerance vs 0.1mm tolerance of a linear bearing!! The belt is kept on the drive by four further shouldered bearings, and fixed on the end by a screw driven tensioning device.
What's more, a quick price up of the relevant parts suggest each bearing is about 50p, the toothed drive for the motor is £2 and the belt itself is another £2. Threaded rod is about £5 for a length, with nuts, washers and stuff maybe making up another £3. Given the steel is effectively free, that makes each carriage;
Carriage bearings 6 x 50p = £3
Belt bearings 4 x 50p = £2
Toothed Drive = £2
Belt = £2
Rod = £5
Hardware = £3
That's £17 per carriage. Maybe the same again for a decent torque stepper (one per carriage), that's about £100 to get from nothing all the way to a working transport that just needs a Z movement. That's a nice budget for me.
The only thing is the size, 100 x 100 is quite large considering the entire device is only 500 x 500. One of those either side would leave a cutting bed of maybe 250 x 350. Is that enough? I'm not sure. Reducing everything by 50% would be better, that would leave me with 350 x 450. It would also mean the frame could go to 25mm or even 19mm, which I have a lot of. I'd have to figure out how much to shrink the bearing and threaded rod, and unfortunately the belt and pulley couldn't shrink.
Hmm, prototype time I think!!
Monday 6 May 2019
Pictures I like
A bit of reference material, ideas that look good but need a bit more analysis before I try and use them.
This is from https://openbuilds.com/builds/hd-cnc.6277/ a heavy duty CNC mill that is apparently 'build complete' but there's not much info between this picture and the final one.
Another one from the same build, this time the Z axis;
This looks to be quite a common design, so most likely what I'll build first.
Last thing, the gantry plate;
This is a key part of the build, variations of which are used for each axis in various forms. Basically it's a mounting plate for the support bearings (in this case the four sets of four holes towards each corner), the drive bearing (central four bolts), and mountings for the z axis (two sets of smaller four holes at the top and bottom middle). This bit has to be cut and drilled extremely accurately, with each set of holes being in perfect line. For example, if the support bearings are crooked, they won't travel smoothly on the rail. And since I'm aiming for the rail to have less than 0.1mm tolerance, I need the same tolerance here. I might practice creating one of these...
This is from https://openbuilds.com/builds/hd-cnc.6277/ a heavy duty CNC mill that is apparently 'build complete' but there's not much info between this picture and the final one.
Another one from the same build, this time the Z axis;
This looks to be quite a common design, so most likely what I'll build first.
Last thing, the gantry plate;
This is a key part of the build, variations of which are used for each axis in various forms. Basically it's a mounting plate for the support bearings (in this case the four sets of four holes towards each corner), the drive bearing (central four bolts), and mountings for the z axis (two sets of smaller four holes at the top and bottom middle). This bit has to be cut and drilled extremely accurately, with each set of holes being in perfect line. For example, if the support bearings are crooked, they won't travel smoothly on the rail. And since I'm aiming for the rail to have less than 0.1mm tolerance, I need the same tolerance here. I might practice creating one of these...
Random links...
OK, time to add a page of pages...
https://openbuilds.com/builds/openbuilds-minimill.5087/
Openbuild minimill, this does seem to be using a lot of things I have seen in my travels and considered 'a good idea'. And it has videos, lots of videos. I kind of prefer to read text (hence these blogs) but a video might be worth watching. I'll put them with all the Udemy videos I'm planning to watch...
https://all3dp.com/1/best-diy-cnc-router-kit/
A page of the 'best' CNC routers. By 'best' I assume they mean 'most expensive'.
https://www.youtube.com/watch?v=K9UA9ZRFwWU&t=362s
I can't go much further without referencing This Old Tony. I'll no doubt be pinching a lot of his fixed gantry CNC mill as that seems to be very effective. Massively over-engineered for what I need it for, but still effective.
https://openbuilds.com/builds/openbuilds-minimill.5087/
Openbuild minimill, this does seem to be using a lot of things I have seen in my travels and considered 'a good idea'. And it has videos, lots of videos. I kind of prefer to read text (hence these blogs) but a video might be worth watching. I'll put them with all the Udemy videos I'm planning to watch...
https://all3dp.com/1/best-diy-cnc-router-kit/
A page of the 'best' CNC routers. By 'best' I assume they mean 'most expensive'.
https://www.youtube.com/watch?v=K9UA9ZRFwWU&t=362s
I can't go much further without referencing This Old Tony. I'll no doubt be pinching a lot of his fixed gantry CNC mill as that seems to be very effective. Massively over-engineered for what I need it for, but still effective.
Sliders
The main consideration for the gantry and Z motion is the sliding motion. It needs to be smooth, accurate and strong. I'll save the drive mechanisms for a separate post, I want to consider the supports first.
The normal approach is linear bearings or bushings on polished round bar or tube. This is a cheap option, and good for 3d printers. Problem is as the strength is turned up, so does the price. Linear bearings in particular can get very expensive very quickly. China specials are £5 each for a 20mm tube, 30mm are £14 each, 40mm are £30 each. I'd need four minimum for each axis, that's 12 bearings. £360 just for bearings!!
The one option that did occur to me was a 'rollercoaster wheel' approach. Three nylon wheels at the top, middle and bottom of the mounting bar. It would be a larger set up than the linear bearing, and more complex to set up, but it would be very rigid. It would also give me an option of making the bars more rigid as I could reinforce them midway across.
Having said all that, apparently nylon bearings on a steel or aluminium platform does cause things to wear and lose tolerance. This is a criticism of the openbuilds cncmill, where they use nylon wheels on the v slot of the extrusion.
And then on to This Old Tony. TOT has built a fixed gantry CNC mill from scratch, and while the build is massive overkill (it is built to cut steel and the construction and cost reflects this), some of the techniques can be miniaturised. His approach is linear rails;
This is what he's used for his CNC mill. 15 mm doesn't sound very big but apparently it's enough. Unfortunately while they are Ebay specials (TRH15B for my future reference) it's £71 for two rails and four bearings. That's basically enough for a single axis, so multiply that by three and I'm back up to bearing prices. I'd like to go down this route though, if I can find a smaller version then that would work. But I need to make sure that it can handle loads in all three axes. For example, the rails above have quite a deep incut on the side which suggests they are quite strong in every direction. Whereas smaller ones seem to have less of an incut which might be an issue. Which brings me back to my rollercoaster plan as that would have good strength in every direction.
Oh, and while I remember... I was planning on having this machine get close to 400 mm of axis range, but I've just realised. I can't do a 400 mm depth, that would be useless. I reckon I only need 100 mm of travel. And even then, most cuts will be 10 mm deep at most, with metal cuts being 1 mm. But I'll keep 100 mm so it can cope with deeper materials. And it's only this shorter Z axis that needs to cope with forces in three directions, the other two axes only need to worry about two directions of force.
The normal approach is linear bearings or bushings on polished round bar or tube. This is a cheap option, and good for 3d printers. Problem is as the strength is turned up, so does the price. Linear bearings in particular can get very expensive very quickly. China specials are £5 each for a 20mm tube, 30mm are £14 each, 40mm are £30 each. I'd need four minimum for each axis, that's 12 bearings. £360 just for bearings!!
The one option that did occur to me was a 'rollercoaster wheel' approach. Three nylon wheels at the top, middle and bottom of the mounting bar. It would be a larger set up than the linear bearing, and more complex to set up, but it would be very rigid. It would also give me an option of making the bars more rigid as I could reinforce them midway across.
Having said all that, apparently nylon bearings on a steel or aluminium platform does cause things to wear and lose tolerance. This is a criticism of the openbuilds cncmill, where they use nylon wheels on the v slot of the extrusion.
And then on to This Old Tony. TOT has built a fixed gantry CNC mill from scratch, and while the build is massive overkill (it is built to cut steel and the construction and cost reflects this), some of the techniques can be miniaturised. His approach is linear rails;
This is what he's used for his CNC mill. 15 mm doesn't sound very big but apparently it's enough. Unfortunately while they are Ebay specials (TRH15B for my future reference) it's £71 for two rails and four bearings. That's basically enough for a single axis, so multiply that by three and I'm back up to bearing prices. I'd like to go down this route though, if I can find a smaller version then that would work. But I need to make sure that it can handle loads in all three axes. For example, the rails above have quite a deep incut on the side which suggests they are quite strong in every direction. Whereas smaller ones seem to have less of an incut which might be an issue. Which brings me back to my rollercoaster plan as that would have good strength in every direction.
Oh, and while I remember... I was planning on having this machine get close to 400 mm of axis range, but I've just realised. I can't do a 400 mm depth, that would be useless. I reckon I only need 100 mm of travel. And even then, most cuts will be 10 mm deep at most, with metal cuts being 1 mm. But I'll keep 100 mm so it can cope with deeper materials. And it's only this shorter Z axis that needs to cope with forces in three directions, the other two axes only need to worry about two directions of force.
Thursday 2 May 2019
Frame materials
As far as I can tell from what I've been reading, the key part to an accurate high resolution cut is a rigid frame. if the frame doesn't move (apart from when it's supposed to, IE the gantry) then there is a fair chance the cutting tool will be able to perform at it's best.
Lets have a look at the options...
Wood
Cheap, plentiful, easy to work with... but the idea of having a frame material that is 'softer' than the material being milled seems like a bad idea. So I'm dismissing this one straightaway.... well, apart from using a sheet for a sacrificial bed, that seems like a good idea.
Aluminium extrusion
More expensive, still plentiful, almost like it was designed for the job. But is this strong enough? And is it cheap enough? Chinese suppliers are around £5 per foot for 20mm, £9 per foot for 30mm.
Aluminium bar
And the price goes up... aluminium bar at 40mm x 10mm is £16 per foot, and based on my limited knowledge of material strength may not be any stronger than the extrusion. So another no here.
Steel tube
Now this is an interesting one. Price? Almost zero, I have a fair bit of leftover lengths from the car build. Even if I have to buy more, a 19mm square section was £20 for 4 metres. Strength is more than aluminium, and from the car build I am very familiar with triangulation. Construction would be easy as well, the parts could be welded together instead of bolted (although I'd have to be careful about warping). The extra parts would just require drilled and tapped holes, even the adjustable parts could be done with slots in the steel. It would also be a fair bit heavier which I am under the impression is a good thing. The only question mark here would be the gantry, a steel gantry would need a decent amount of torque in the motors. And support in the support bars. Although with only a 400mm span I would expect the rigidity to be fine.
If I go down the steel tube route, I might also pick up on a tip from This Old Tony. His gantry CNC is filled with sand to give it some extra 'stuff'. Obviously this would only apply to the gantry and base, the moving parts will be left as is.
Subscribe to:
Posts (Atom)