New CNC router design - Gantry

[yngndrw]

Hi all,

I'm planning a new build and I think I have come up with a general idea of how the gantry will be, but was hoping to get some input on the rigidity of it all.

So first of all, some requirements:

• Ability to cut aluminium very well
• Ability to cut steel, slowly
• Ideally the ability to face mill, slowly is fine
• Working area of 600-700mm in X and Y, 200-300mm in Z
• ATC BT30 spindle
• Enough space for a Haimer 3D probe (About 100mm tall)
• Fixed bed with vertical clamping area

The general idea is a gantry supported by high-level linear guides, with a ballscrew on each side - Something like this:



Starting with the spindle, I was considering one of these: https://www.ebay.co.uk/itm/113626455544 These are 100mm diameter, 400mm long and seem to weigh about 15-20kg from what I have found. I'd use something like this (https://www.ebay.co.uk/itm/113361541312) for the Z axis, mounting the spindle nose about 100mm higher than the bottom of the gantry at its highest position to give space for the Heimer probe.


Originally I was considering a 1m long 80x160mm heavy aluminium extrusion (https://www.zappautomation.co.uk/mc1...x160-7520.html) for the gantry, strengthened by adding a 20mm aluminium plate onto the front, top and back - The extrusion alone is 11kg and the plates would add about 20kg. The plate would also add a nice flat surface for the linear rails and is easily made using my drill press. A chunk of steel box-section is out of the question as I don't have the capabilities to mill it flat, and don't know anyone with a milling machine large enough to do a 1m span.

The next idea was to use a THK KR65 linear guide (https://tech.thk.com/en/products/pdfs/en_a02_122.pdf) - The 1m version (700-800 stroke) with a single block and a cover which is 36kg alone. I'd add a large (30-50mm thick) aluminium plate to the back which would be about 200-300mm tall (20-40kg), with an extra linear rail along the top for additional support.


Thick plates would mount the gantry to the Y axis rails, with the blocks spanning about 200mm in length for stability. These would be mounted on aluminium extrusion, strengthened with thick Aluminium plates as required. The rest of the frame can then be made with a mix of steel box section and aluminium profile, as it's not as critical as the gantry.


I was hoping to use some existing 180W servo motors for all axis (http://www.jmc-motor.com/product/980.html), with two motors on the Y axis. A partial (10kg?) counter-balance on the Z axis probably wouldn't be a bad idea due to the weight of the spindle. If power is an issue, I'd go for larger AC servo motors.


So I suppose the main questions are:

• Will my plan for the gantry be rigid ? Is a large pre-made linear guide like the THK KR65 ever going to be enough, even with strengthening plates ?
• Will my gantry be too heavy for its own good ?
• Is (fast) aluminium milling realistic for this: a) Strength / span of gantry ? b) Type of spindle ?
• Is (slow) steel milling realistic for this: a) Strength / span of gantry ? b) Type of spindle ?

Thanks,
-Andrew.

Similar Threads:

• Build Thread- Hybrid Design - Fixed Gantry Router for High Speed Aluminum Milling - Metal Nibbler 5
• Gantry design
• Gantry: planning to design moving open gantry
• New design, gantry or something else
• Gantry design???

[Andreiir]

With that spindle u can't cut steel , u need low rpm , and high torque for steel .

[yngndrw]

What are the general requirements for steel in terms of RPM and torque - I had always assumed that you could just take very light cuts (Within reason, the chip size still needs to be reasonable) but I don't know what the numbers are, even ballpark. The spindle graph suggests that it has a flat torque rating of 1.9Nm right down to 0RPM, but I'd have to confirm that with the seller.

[Andreiir]

If this spindle works on 1.9 kw on low rpm , then u can cut steel with .

[ger21]

What are the general requirements for steel in terms of RPM and torque - I had always assumed that you could just take very light cuts (Within reason, the chip size still needs to be reasonable) but I don't know what the numbers are, even ballpark. The spindle graph suggests that it has a flat torque rating of 1.9Nm right down to 0RPM, but I'd have to confirm that with the seller.

Probably well below 5000 rpm.
That spindle says 12,000-24,000, but probably doesn't have much torque even at 12,000 rpm. I wouldn't believe that chart.

Light cuts in steel with a router spindle would be a few thousandths per pass. Not practical at all.

[yngndrw]

Thank you both, it is helpful to be able to put some numbers down and calculate things. So even if we were to blindly trust that spindle graph (I'll verify it with the seller, there are enough people selling those that one is bound to be in contact with the manufacturer), the lowest speed to get 1.9kw from the spindle is around 10,000RPM which is a bit too high. Having said that, I'd be quite happy with a few thou cuts (4 thou is 0.1mm) if I could get a face mill with a decent width. I'd mainly use the steel capacity for facing long parts, E.g. For mounting linear rails onto a welded frame.


While I was searching around, I have just found this video which was interesting:


It shows a 1.5kw HF spindle being used to mill steel - Initially: 4.75mm stepover, 0.3mm (12 thou) depth of cut, 10,000RPM - Then later on: 4.25mm stepover and 1.3mm (50 thou !) depth of cut. Am I missing something obvious here, because the video doesn't match up with what I've read both here and elsewhere about milling steel ?

[ger21]

You'll quickly wear out tools and your spindle if you use that spindle to cut steel. It's just not made for cutting steel. There's lots of things on YouTube that you shouldn't do.

Having said that, I'd be quite happy with a few thou cuts (4 thou is 0.1mm) if I could get a face mill with a decent width.

You're not going to be able to run a face mill, unless it's very small.

[peteeng]

Hello Andrew- to cut steel as you've been told the machine needs to be rigid and your heading in that direction. The video is a machine that is a moving table machine which by nature is rigid and mills are 99.9% always moving tables. Gantry machines have the issue of "walking" which means one side always is ahead of the other and then the elastic response of the system catches up the other side. On rigid machines this is very small but there, on soft machines it can be a really big problem. This issue is amplified by the bearing clearance. Linear bearing blocks come in clearance classes. Normal cars are either zero clearance or light preload (check what you have). Mills and routers should be medium or heavy preload to remove the clearance, create bearings with no backlash or play and minimise wear. If a car has zero clearance when new and is used heavily it will quickly become a clearance bearing and that's not good. Similarly with ballscrews they come in various preloads for similar reasons. If you want a machine to do a job find a machine that does this and copy it. If you "hope" a light machine can do a heavy job then you are going down the wrong road. It may do something more than its built for once or twice but then it will start to do its normal job badly. The other thing about gantry stiffness is to make the bearings as far apart as possible. However this chews up the working zone. They need to be at least one car apart. I generally start at one hole spacing apart so if I get stuck I can move the bearing into the middle holes. Two hole spacings is very good if you have the room. There is no clear calculation or guideline for this spacing that I can find. It's up to you. There are geometric guidelines for spacing bushes due to stick slip issues being driven by geometry and they usually start at an aspect ratio of 3x or even 4x the bush length for precision or high loaded guides. I have had an application where I used a 5x aspect ratio for the bush and it still slip-sticked then stuck so I had to convert the design to a linear rail. Keep Making... Peter

Here's a good article on bearing aspect ratios https://pages.pbclinear.com/rs/909-B...Phenomenon.pdf as a gantry machine is a cantilever it's important to understand stick-slip behavior to create a smooth operating machine. And in regards to the last video that's a mill not a router. It would be better to use a slow speed spindle vs the 10000rpm. The thing to check is the chip load which you need to feed speed and the rpm to calculate. If the video is running a really small chip load then the tool is rubbing and will wear fast. Recommended chip loads are published by the tool manufactures. Interestingly this is relative to machine stiffness. Aluminium tools are recommended to cut at say 0.1-0.25mm chipload. So if your machine is conventional cutting with a sharp tool and the tool deflects more then the chip load it can't cut. So in this case we need to design a machine that under the cutting loads does not deflect more than say 0.1mm other wise it just slides past the material. Chip loads for timber are around 0.5mm so if the machine deflects 0.25mm it will cut but probably make a rough edge. Food for thought? Peter

Hi Andrew - I looked at the video and noted some of the feeds/speeds. Its cutting free cutting steel at heavy chip loads. Up to 0.48mm which is a heavy cut. It's a super rigid machine much more rigid then yours can be. cheers Peter

[yngndrw]

Gerry:
Ah I see, I didn't factor in the life of the bearings etc under the additional load. Aluminium only it is then, which hopefully relaxes the rigidity requirements enough for my design to work.


Peter:
Sadly moving table would take up too much space and space is my main limitation, so I'm stuck with moving gantry. I could probably find heavier pre-loaded bearings for the Y (table) axis, but if I'm to use a THK KR guide for the gantry I'm going to struggle. I suppose the other solution there is to reduce the load by over-sizing them, which is why I was planning to use the KR65 over anything smaller.

Thanks for the link on bearing aspect ratios, that was a good read and something I hadn't considered. I suppose that in this use, the gantry has two separate axis which are cantilevered and therefore has cutting forces applied upon - The vertical distance between the end of the tool and the center of the bearings, and the horizontal distance along the Y axis between the center of the spindle and the center of the bearings. The largest for me by far is the height which would be around 300mm at its worst.

Now if I'm to use the THK KR series, I'm have a bit of a dilemma. If I use the KR65 for the increased rigidity, the carriages are 145mm in overall length. Factor in the additional gantry length lost due to the end bearings and supporting the gantry on each end and realistically I'd want to limit my carriage width to about 200mm - Dual blocks are therefore impractical. A single carriage, if we assume a bearing length of 130mm would bring the ratio to around 2.3:1 which seems quite reasonable according to that article. The alternative would be to use a smaller KR series, such as the THK KR46 with an overall carriage size of 110mm. With a shorter carriage I could two blocks, however according to that article, the bearing length should be taken as the center to center distance which is 110mm, so shorter than with the KR65. The other thing that I could do is to spread out the carriages on the extra upper linear guide that I was planning on using. This doesn't have the same length limitations due to its placement and therefore the carriages could be spread out further - Possibly giving an effective bearing length of around 200mm at the top, but at a greater vertical distance to the end of the tool. (Let's say 500mm, so a ratio of 2.5:1) This clearly doesn't help for the force generated at the tip of the tool, but would help for the momentum of the heavy spindle and z-axis when accelerating and decelerating.

For the Y axis, I could use heavier-preloaded bearings with an overall distance of around 300mm, for an effective length of around 200mm. I believe the effective moment arm length would be half the gantry span, so 500mm - Again giving a ratio of 2.5:1.

Am I on the right track there with my bearing length calculations, or have I misunderstood the article / mechanics of the problem ?

As for the rigidity of the machine as a whole, or specifically the gantry - I have no idea where to start with the deflection calculations. I couldn't find any information on deflection in the THK KR document and I don't know how this changes when you start bolting thick aluminium plates to the back. Having said that, my gut says that 0.1mm seems massive for deflection of such a large chunk of steel and aluminium - But I'm not sure how to calculate it to be sure.

[peteeng]

Hi yngndrw (young andrew?) - Stiffness of recirculating ball linear bearings is not related to its size. Size relates to strength. Either static, dynamic or life are the three numbers of concern. The stiffness of the bearing is in its support structure and the efficiency you connect to it. The 4 rows of bearings in a car are just like a metal key on a shaft it's effectively rigid (unless it has clearance or wear) . So pay attention to the support structure and its connections.

I have been running a router on light preload bearings for just over 1.5 years and they have been given a tough time. They have travelled no where near their design life of 50kms yet they have become clearance bearings which has allowed the z axis to wobble a little. I highly recommend using heavy preloaded cars if you can get them. They don't feel any different to the hand but having preload means that if they wear I suspect they will still stay preloaded vs std bearings with light preload. Looking at two bearing manufacturers design docs they both recommend heavy preload for routers and mills. you won't see the issue at day 1 but sometime down the track it will become visible if you start with light/std preload.

As for the geometry I can't check your numbers but I expect cars with 2x spacing will be fine and 3x will be better. Spacing and various geometries are all compromises you have to juggle for your own circumstances.

In terms of stiffness if you know the tool load or estimate it by looking at other similar machine specs you can estimate deflection using simple beam theory or by getting hold of a section and putting it on two bricks place the tool load on top and measure the deflection using a dial gauge. I use FEA to do all of this from solid models (see attached, sorry trying to convert MP4 files to WMV files). The gantry is torsion critical so a round tube would be stiffest but a round tube is difficult to connect rails to! So keep considering and make the best guestimate you can. G-wizard estimates tool loading as well as it estimates tool deflection at the set cut parameters. Peter

[yngndrw]

Yes that's right, my name is Andrew Young - I have that username so that it's never taken, although it is quite hard to read out over the phone.

It looks like used THK KR linear guides are ruled out for two reasons, ability to find pre-loaded guides and the linear rail lengths being linked to the ballscrew length which in turn limits the overall travel and the carriage spacing I can use.

So from that, new design for the gantry: (I'm somewhat limited by the tools and equipment I have here, ultimately my main machinery is my trusty drill press, so I'm best making this out of flat plate and getting as few parts machined as possible)
- Linear Guides: New HIWIN parts - HIWIN so that I can afford new parts to the preload spec that I need: https://motioncontrolsystems.hiwin.c...2r1000zbhii-zz (~18kg)
- Ballscrew: Probably a used part from Ebay, there's loads on there so I'm sure I can find one - Belt drive so that I can mount the motor on top and reduce the over-all length.
- Gantry: 20mm thick 300mm tall Aluminium plate on the front (16kg) bolted to 20mm thick 200mm wide Aluminium plate for the base (11kg), strengthened by 6x 20mm aluminium gussets bolted in every 200mm down the length (17kg). Milling required: Two faces of the gussets, one long face of the front plate. (I'd prob screw some thinner (4mm?) plates to the top and back (~5kg?), for the cable chain / motor mounts and to neaten it up - The alignment of parts isn't critical.)
Total weight for this axis without the ballscrew / motor: ~67kg

By using linear rails which span the full length, I can afford to have a total block length of 300mm without losing much travel. The ballscrew can then be shorter, allowing space for the end bearings and drive pulley.

I'll have to have a think about it before I do some proper calculations, but I think it's doable. The other option is to weld up some steel box-section and get it milled, but I have no idea what kind of cost that would be - Having said that, the milling is probably a lot cheaper than all of that aluminium.