CNC Machining Bearing Blocks

[magnets]

Originally Posted by Mk.32

What kind of mill do you have?

The cost of materials, tooling and then machine time (that does "cost" something) versus $20 a piece from vex is sometimes hard to justify.

It's quite a bit of material removal as well.

2 robots with 6 wheels is 12 of the $20 bearing blocks, or $240.

Aluminum for all the bearing blocks is under $40, and I don't expect to purchase any new tooling for this. That's a savings of $200, or the price of one drive gearbox.

We're still not sure how much money we'll have by the start of build season, but every dollar counts!

[Chadfrom308]

What I would do is turn most of it, then use a mill to cut the flange and drill the holes. You could use round stock and probably cut less off. Also, you could order a long bar of aluminum rather than small pieces and keep turning them.

For turning you would need a good contouring tool (at work, we use all ISCAR tooling) and then a good boring/ID tool. Then slap it on a mill and cut the rest with a drill and an end mill. Then as a bonus, chamfer it because chamfered things look nice!

And if you have the big bucks and have a really nice lathe with a live tooling, then you can do it all on the lathe

Originally Posted by Chadfrom308

And if you have the big bucks and have a really nice lathe with a live tooling, then you can do it all on the lathe

If you have the big bucks it would make more sense to just buy them and save yourself from having the spend time machining bearing blocks.

[Mk.32]

Originally Posted by Andrew Lawrence

If you have the big bucks it would make more sense to just buy them and save yourself from having the spend time machining bearing blocks.

Yes... but if you already have the $80,000+ Lathe/Tooling and I would "hope" the experience to take full advantage of it, it would be a trivial part to make. Heck you go into production and sell them at the same price as vex almost....

Personally I've made the same design before, I would just machine them from billet on a tormach in two operations... basically start with a hunk of alum in the vice and then machine all of one side. Make some soft jaws (alum jaws bolted into the vice so I can machine a profile them to grab circle parts) and then flip them over and face off the otherside/other operations.

Good practice probably, especially if you haven't done parts this complicated before.

[Gray Adams]

Originally Posted by Mk.32

Yes... but if you already have the $80,000+ Lathe/Tooling and I would "hope" the experience to take full advantage of it, it would be a trivial part to make. Heck you go into production and sell them at the same price as vex almost....

Personally I've made the same design before, I would just machine them from billet on a tormach in two operations... basically start with a hunk of alum in the vice and then machine all of one side. Make some soft jaws (alum jaws bolted into the vice so I can machine a profile them to grab circle parts) and then flip them over and face off the otherside/other operations.

Good practice probably, especially if you haven't done parts this complicated before.

Definitely what I would do. Given that its FRC, you could easily get away with circular interpolation and not even reaming the bearing pocket, though a 1.125" reamer is a pretty good investment. It's easily worth the cost of doing it yourself if you're doing 4 or more, especially when you consider student time to be as cheap as it is.

[waialua359]

Originally Posted by Cory

Do you have a reason for not buying these?

That's the best bang for the buck you're gonna get. If you want to learn to use your mill to the fullest extent of its capabilities, that makes sense. But this might be a problem to throw money at.

My first reaction was what Cory has mentioned.
Is this a learning exercise opportunity or you just need these?
I can see making it, if alternatives werent available, but these days, you have lots of other options which are much cheaper and less time consuming.
When we used bearing blocks in the past, our program made them as a learning experience for students and aesthetic reasons.
These days, we dont even use bearing blocks at all for our drivetrain.

[Joe G.]

Why are so many people in this thread so hostile towards the idea of a team making these themselves? It's clear that they have the resources and valid motivation to do so, a plan to schedule production so that it does not come at the expense of other things, and are seeking the knowledge to back these resources up.

In the past few years, after my team acquired decent manufacturing resources, it became quite rare for us to purchase something that we could make ourselves in a reasonable time frame. It's not just about providing students with a learning opportunity either. Cost was by far our most limiting factor; 1687 robots were always built on tiny budgets, but we acquired enough manufacturing ability and dedicated students/mentors to work around this and build quality robots. We also liked having to learn how to produce as much of our robot as possible inhouse. This meant we were less at the mercy of vendor availability and lead times if something broke or needed to be changed last minute (which saved us more than a few times), and gave us the ability and confidence to design in slight variants to parts we could have gotten COTs, when it was beneficial to our design. In many places, this actually saved us lots of time, since we did not have to spend time designing to fit the dimensions of COTs parts.

I know that these bearing blocks are a pretty standardized item, but I can definitely see why a team would like to save money in their fabrication, and maintain freedom to modify them as it suits their design.

Buy a man six bearing blocks, he'll have a sweet west coast drive next season. Teach a man to fixture properly on his CNC machine...


In regards to the original question, this may be of use.

[techhelpbb]

I think it is admirable to have interest in making something yourself.
Hence I side-stepped the issue of 'working smarter' until now because it expands the topic beyond where I think it was asked to go.

One should consider that merely having CNC experience and CNC access does not mean that the job will be as efficient as possible.

One might not have the correct tools for that CNC machine and therefore the job might take longer or buying the tools might drive the cost of a job way up.

CNC machine tools are full of consumables in the form of coolant, tools and sometimes fixturing and all those should be considered in the costs.

One might not have the experience to operate the machine themselves and are therefore burning a mentor or donor's time to make these items. If you are asking for someone to give you time, materials, possibly tools and labor it can quickly become very much not in their interest to make something they could more cost effectively buy.

I certainly will not discourage the adventure of learning but there are important lessons in efficiency to be had here.

I would be interested to know how long it takes teams to perform this manufacturing process that have done it themselves. In the past I helped Team 11 make much less complicated bearing mounts and it took a few hours to turn 6 of them out. In that case these parts needed to be custom as it was an after thought and the design was not made with these commercial alternatives in mind.

Brainstorming here - I would love to see a gallery that shows some of these CNC jobs and some simple details like the time and expendables needed to make them. I think it might help those learning set expectations more effectively.

[Isaac501]

Originally Posted by magnets

Normally, we would do that, but money for buying robot parts has become an issue for us, and we'd still like to have a practice robot. We could save some money making these and the little cams ourselves.

My plan was to make some of the parts we usually buy on our CNCs on the first few days of build, as the CNC is not usually in use on these first days, and we have students who would do this.

I was originally planning to just do a row of 6 or 10 on one piece of aluminum, but I'm thinking it may be easier to take one, machine the top (the tube end that fits into the other side) all the way down to the base, then put it in a fixture plate with a 1.375" wide square with radiused edges to locate it to machine the outline, the mounting holes, and the bearing flange c-bore.

If $20 for a properly machined and toleranced part is a concern, a pair of drilled and tapped holes the appropriate distance from the bearing is "free".

Use button head screws as retention for a flanged bearing.

[Cory]

If you want to make these for the season, make a few now to understand the best sequencing and workholding.

They're relatively simple parts but they do require good tolerancing and concentricity to be held between the bearing bore and the OD/ID of the sleeve on the backside (depending on which side you're making). You can easily create scenarios where the shaft is hard to get through both bearings, or doesn't spin nicely if you don't do it right.

I would caution anyone from blindly copying the fixture that 254 made that was linked by two posters in this thread. That was the first fixture I designed and is suboptimal in a number of ways. It took 4 years to get a combination of good fixturing/setup practices and CAM tweaks to get programs we can confidently run for good parts the first time, every time, with minimal operator intervention and without ejecting parts from the fixture/breaking tools/not holding tolerances.

Far better to start out with soft jaws, as you will need to know how to cut them for far more parts than just bearing housings.

[Jared]

Our team is considering a similar process.

Cory pointed out that runout can be an issue with these parts, so you may want to consider creating the sleeve OD and bearing bore on a lathe, then cutting this part in half and using each as one side. This way, you have matched sets of bearing blocks. Also be sure to use one of these round fixtures to locate the part when setting up on the CNC.

These likely wouldn't be difficult parts to make by hand if you were willing to give up the neat shape on the outside and the rounded corners on the 1.375" square.

The procedure I've come up with involves one 0.25" thick 3" x 3" fixture plate with a 1.375" wide square with .250" radiused corners for the square feature on the bearing blocks. The plate would also have two threaded 10-32 holes for the two mounting bolts in the bearing block.

My most efficient setup is as follows:

3" x 12" x 1.25" aluminum stock for 6 blocks (stock is extra tall so that the vice can grip the bottom 0.250")

1. Machine bearing bores, sleeve OD, rounded square feature on base, .196" holes for six bearing blocks.
For this operation, hold the stock in a vice (or two) so that the lowest point of the stock is about 0.25" below the face of the vice jaws.
Because the stock is a little bit oversize, zeroing only needs to be accurate to .050" or so.
Also, the bearing bore and the sleeve should have little runout as they are machined in the same operation.

2. Cut out individual blocks on bandsaw
Again, this doesn't need to be super accurate. It's good as long as you don't cut into another bearing block.

3. Put in lathe and face off 0.250" (the stock was .25" too tall)

4. Flip upside, put in fixture. The 1.375" square should fit nicely into the fixture plate, and you can put two 10-32's through the bearing block into the fixture plate to hold it down.
Now, locate part with a dial test indicator using the bearing bore. You should be able to get within .001" with a decent dial indicator. You shouldn't have to re-zero between bearing blocks if your fixture plate has good tolerances. It may be worth making the 10-32 clearance holes in the bearing blocks the next size down (0.1935") so that the fit more accurate with this fixture.

Now, machine the bearing flange c-bore and the outside profile. Use a countersink bit on a drill press for the two 0.196" holes.

You could honestly skip this last step and put in the bearing flange counterbore on a lathe and just the cut the outside profile on a bandsaw, and you could skip the fixture plate.

If my poorly written description doesn't make sense, let me know, and I'll put up a picture.

[AdamHeard]

I'd argue that more time should be put towards fundraising, and COTS should be used more.

The in house CNC is more useful during that time for prototyping, or for machining things that can't be bought.

Imo teams should only be making bearing blocks now if they are continuing a legacy design of their own (and therefore can use previous year's for practice bots).

[Richard Wallace]

Originally Posted by AdamHeard

I'd argue that more time should be put towards fundraising, and COTS should be used more.

I agree with this principle generally. I will also agree specifically for these components, when my team has become convinced they fit and perform well. We have come to expect that of VexPro/WCP parts.

That said, the OP did ask how to make bearing blocks, not where to buy them, and this thread has been an interesting read on that topic.

[magnets]

Originally Posted by AdamHeard

I'd argue that more time should be put towards fundraising, and COTS should be used more.

Fundraising isn't the issue. We've got people ready to give us money, but we may be limited to spending an incredible $1,000 during build season (or at least during the first few weeks) due to some school policy issues.

The in house CNC is more useful during that time for prototyping, or for machining things that can't be bought.

Imo teams should only be making bearing blocks now if they are continuing a legacy design of their own (and therefore can use previous year's for practice bots).

I'm not sure I agree. I don't know about you, but our team has never been able to design something in CAD, generate a toolpath, set up a fixture, and make a part on a mill on the first day of build season. We have access to two CNC mills, and a guy who programs CNC's for a living, so we can make these parts easily in 6 hours on day one of build, including setup time. There's no reason not to do this on the first day of build season to save some cash. The last few years, we've ended the season with less than $100 in our account, and we would've spent more if we could.

A few seasons ago, I invited three students and another mentor to my house on the first evening of build season, where we machined 10 4" aluminum wheels, a frame, and drive gearboxes. This took us from 5:00 pm to 3:00 am.

Could we have bought those parts? Of course. But the team paid absolutely nothing for this drive base. The aluminum, and sprockets were donated, and we made the rest (including gears!), and got bearings and CIM pinions in the kit. That was our most successful build season ever.

[Travis Covington]

Originally Posted by magnets

We have access to two CNC mills, and a guy who programs CNC's for a living, so we can make these parts easily in 6 hours on day one of build, including setup time. There's no reason not to do this on the first day of build season to save some cash. .

If so, I would ask the guy who will likely program your machines how he would want to hold these pieces during machining. Seems like a good first step? The guy doing this for a living likely has more experience than all but a few people who have responded.