[thegolfer]

Hi All,

This is my first post on CNCZone! I hope it's not too silly.

I have a question regarding machining tolerances, and in particular, realistic flatness and parallel tolerances.

I am in the process of designing parts for a CNC router that I hope to have professionally machined. One of the parts (image attached) requires a plate to be machined flat on one side (the back side) and then have channels machined into the other side that will hold linear guides and a ballscrew support unit. The depth and coplanarity of these three channels needs to be pretty accurate, but aside from that the thickness of the rest of the part is negligible.

If I asked a machinist to machine these channels to be be parallel and coplanar with each other and with the back of the part to within +/-0.05mm (0.001"), would this be relatively easy to accomplish on a CNC mill? Or, would grinding need to be employed?

Any input would be greatly appreciated!

Toby

[Bwana Don]

I say grinding. How much surface area are we talking?

[christinandavid]

What's it to be? +/-0.05mm or +/-0.001"

The main factors I suppose would be the size of the part and nature of the material.

+/- a thou is a bit tight unless the part is lapped.

+/- 0.05mm is not too unreasonable to mill, if the part is under 500mm in length.

I'm talking flatness, not parallelism. Parallelism should be a doddle.

DP

[thegolfer]

What's it to be? +/-0.05mm or +/-0.001"


DP

I should have rounded up +/-0.002" It makes a 100% difference to my question!

The part is 156mm wide x 360mm high, that's a surface area of 561.6cm^2 (87.05 inches^2). The channels are 20mm wide by 360mm high (x 2) and 60mm wide by 360mm high. The material is 6061 T6 aluminium (or similar alloy with comparable modulus of elasticity, weight, etc).

I'm guessing that a grinding operation, if needed, is going to increase the cost of this part dramatically?

[keebler303]

I would recommend using MIC-6 or similar precision cast tooling plate. It will be flat from the start and has little chance of warping once it's machined. I think that a decent machining job would be fine and grinding is not necessary for the accuracy you require.

Matt

[rowbare]

A good machinist should be able to handle that. You might want to start with something like MIC 6 tooling plate to get a uniform thickness. Downside of that approach is $$$. However it may end up cheaper than paying for a surfacing operation.

I don't know if you have seen this before: https://tech.thk.com/en/products/pdf/en_b01_089.pdf

bob

[txcncman]

I would recommend using MIC-6 or similar precision cast tooling plate. It will be flat from the start and has little chance of warping once it's machined. I think that a decent machining job would be fine and grinding is not necessary for the accuracy you require.

Matt

A good machinist should be able to handle that. You might want to start with something like MIC 6 tooling plate to get a uniform thickness. Downside of that approach is $$$. However it may end up cheaper than paying for a surfacing operation.

I don't know if you have seen this before: https://tech.thk.com/en/products/pdf/en_b01_089.pdf

bob

+1
+1

[fizzissist]

I've got a problem with this. The approach.

This is really a design problem.... What's the tolerance that you NEED, not what you want. From there you find the manufacturing method to meed the requirement.
If that method is prohibitively expensive, or not available, then you change your design.

What I'm seeing here is a "I think this is what I need" sort of mindset (nothing personal, just the observation) while what is important to any project that will ultimately dictate the tolerances you will be able to achieve (as in this case).

The tolerances here aren't terribly difficult to achieve when the process is done properly. What I see though is a design approach that is perhaps more complex than it needs to be, while a simpler design may well provide a more accurate end product.

I don't know what the end design looks like, but this looks like a mounting plate, and rather than surface the front and back, buy the pre-surfaced tooling plate, clamp it properly, then machine ONLY the mounting surfaces, not entire channels or large surfaces. You'll get less flexing, less internal stress, and you can worry about only that which needs to be worried about.

Forgive me if I'm wrong, this is just the way it strikes me.

[txcncman]

As fizz points out, consider the design. How much are you going to flex/warp this when you mount it during install? How much stress will it be under? At 12.5 mm thick under 100 pounds of stress, you can probably toss your +/-0.002" out the window.

[keebler303]

...
I don't know what the end design looks like, but this looks like a mounting plate, and rather than surface the front and back, buy the pre-surfaced tooling plate, clamp it properly, then machine ONLY the mounting surfaces, not entire channels or large surfaces. You'll get less flexing, less internal stress, and you can worry about only that which needs to be worried about.
...

It is best practice to have a machined shoulder to mount your rail to. The rails are not guaranteed to be any kind of straight when you get them. With proper design, you can push the rail into the shoulder and tighten it down so the rail is forced straight by the shoulder.

I would guess at least half of all hobby builders say they want X, they need Y but will settle for Z if the price is right. I'd be surprised if that is not the case here. Just having professionally machined components will be far ahead of a lot of hobby builds.

If you truly need 0.002" accuracy, significant extra expense will be incurred to make sure you hit that. If it's just what you'd like to have, I'd say just build it with a reasonable amount of care and you should be happy.

Matt

[thegolfer]

Thanks for all your replies everyone! Basically what I was looking for was, can a tolerance of +/-0.05mm be achieved in a relatively straightforward manner? That question has been answered, but you guys have also raised a lot more questions for me to consider

Fizz, I understand where you are coming from. A design needs to start at what is required, and then work towards achieving that goal. If that goal is too complex/costly, change the design to fit what is achievable.

Matt, I think you've hit the nail on the head with my approach thus far. I would like to design my parts so that the components I fit to them are fitted within recommended tolerances, however I also know that there are 1000s of hobby machines out there running similar components (i.e. linear guides, ballscrews) way out of tolerance and they are working ok (in a hobby setting).

I am a first year mechatronics engineering student, so I would like to design this machine as "professionally" as I can. I think I'll go back to the drawing board (given that I've got a lot of ideas from you all) and see what I can come up with

[thegolfer]

A good machinist should be able to handle that. You might want to start with something like MIC 6 tooling plate to get a uniform thickness. Downside of that approach is $$$. However it may end up cheaper than paying for a surfacing operation.

I don't know if you have seen this before: https://tech.thk.com/en/products/pdf/en_b01_089.pdf

bob

Bob, I have not seen that document before, but I have seem similar from another company (PMI). I'll dig it up.

It has given me a lot of food for thought!

[keebler303]

Indeed, in a hobby environment there are many shortcuts which can be taken without obvious or immediate consequence. Don't be too proud to copy what someone else has done. Most machines are quite similar in construction so you can get a pretty good idea of what works by studying existing designs.

Matt