Extensions soldered to motors. No molex connectors this time, no motor or extrusion swapping is planned in the near future.
And then I thought to give them a test spin.
And then I remembered what Jeremy said.
And then I thought that twinkle twinkle little stepper would be perfect to put the cnc children to sleep (haha… cnc children).
And then I thought to give it a try.
And then cnc daddy went to sleep at 2am.
just made my day!
awesome machine. cant wait to see it done.
One of the most important pieces of the build is not going to be part of it.
It’s a sample ballscrew machined to the bearing block specs, which btw came today. I am not building the machine around specific/factory ballscrew lengths. I am sizing the ballscrews to fit the machine. Its purpose is to test fit, prove the theoretical/estimated dimensions and prevent a $150 mistake. If I order custom size ballscrews and they do not fit, unlike shoes, I cannot return them and ask for a different size. I call this the insurance piece.
My wiggle room is what I gain by swapping the fixed side spacers (shown below), i.e. their thickness difference: 0.8mm.
I could have picked a more forgiving type of bearing blocks to mount on the extrusions (BK12/BF12), but I opted for the ones that mount on the plates (FK12/FF12) in order not to sacrifice X & Y travel.
The spacers themselves seem 5mm so there is a lot of error margin on the length if they are too long. Alternatively you can make shims between end plates and bearing block if the shaft is too short. This should be enough to prevent that $150 mistake.
Seems a pretty short ballscrew that insurance piece, maybe for a speed clamp?
If I recall correctly, they are 6.3mm & 7.1mm. So, theoretically yes, I have more room. I can think of two approaches in case I miscalculate (despite my sample) and fall out of range.
Hopefully, I wont need to.
The setup I had in mind (blocks outside of plates) when I ordered the insurance piece required in paper X+63mm length; i.e. 63mm extra length of ballscrew for X inner wall-to-wall distance. So I got 163mm to test for 100mm inner.
I have since switched to another setup (blocks inside of plates; easier for motor mounting, less motor protruding, different discussion), which in paper requires X+26mm. The insurance piece shows X+25.5mm.
Made some stickers for the gamepad.
The relevant discussion and some set up hints are over at another thread, where I also posted a demo video. This will soon belong to the B-Carve, so I thought to share a picture here as well.
The ballnuts came.
I have to confess that this was a big gamble. Well, big in terms of the delay (it almost took 2 months), small in terms of cost. I was willing to test if I could get three ballnuts with metal deflectors for a total of $32 (shipped) vs the plastic deflector kind I could find locally for $40 each!
The gamble was matching ballnuts & ballscrews of the same spec but from different factories. With such a tight fit there was no guarantee that they would work together.
To take it one step further and to satisfy our curiosity of how things work (and make sure that nobody messed up in the assembly line) let’s take it apart and repack it. I had an aha! moment earlier today after googling images like this
and was confident about trying it. Here are all the components: body, deflectors, balls, seals and a plastic pipe to keep everything together when a screw is not inserted.
I’ve seen videos of people feeding them from the top with the screw in place and twisting gradually. It’s scary not to be able to see what exactly happens and to make sure that no balls fall in the between the circuits zone (which will cause a jam). The method I followed was one ball at a time, one circuit at a time. Grease for greasing and for beating gravity too.
First circuit done. Btw, it’s a 1605-3 ballsrew, i.e. 16mm OD, 5mm pitch, 3 circuits.
All circuits in place.
Sleeve fitted. This facilitates inserting and removing the screw without knocking the balls out of place.
Ballscrew in, seals in place. Turns smoothly, no play, great fit! Second proof provided by the the insurance piece.
Plus ballnut mount.
Test Y axis drive assembly with temporary Y plates that were cut a couple of days ago.
Btw, here’s making sure that there is enough clearance for the flexible coupler’s spider between the motor and ballscrew shafts. Third proof provided by the the insurance piece.
I believe we are now ready to make a drawing and order the ballscrews in appropriate lengths.
Got to say, I really appreciate you making these posts. Engineer p…n!
Order placed for 3 final lengths with this end machining.
The only tolerances mentioned refer to the various shaft diameters. Hopefully this will save us from having to sand the shafts where the bearings sit (which we did with the sample piece and took us forever). Speaking in first person plural kind of sounds like a TV chef, doesn’t it.
And then we chop the onions…
For a good machinist this should not be a problem. Normally you would specify a h7 fit (ISO) which is a tolerance related to the shaft diameter, for a 10mm shaft +0/-0.015mm and for a 12mm shaft +0/ -0.018mm.
When oversize, put the shaft in a drill and a 600 grit wet and dry sandpaper on the spot where it is too large, a drop of oil and slide the sandpaper slowly back and forth while turning. The black slurry will assist in polishing the surface.
@StuartPearson
Thanks, I’m really enjoying this.
@ErikJanssen
That’s funny, I first noticed the “h” notation on the low profile clamp’s specs mentioned earlier in this same thread. I even googled about it but nothing relevant came up in the top results. Or maybe I did not recognize it, as I was wondering what “h6 steel” is. Duh. Thanks for teaching me something new.
I like the drill & sand idea too. With a shaft about 750mm long though, I would need to also add a second, far end, support to basically emulate a lathe. Hopefully the shafts will be h7 grade and I won’t have to 
Capital letters are used for holes, lowercase for shafts the combination tells the fit. H7p6 is a fixed combination, for a sliding fit you would select H8f7.
The idea is that a sliding fit for a big shaft has the same tolerance reference as a sliding fit on a small shaft although the actual tolerances are smaller.
No google, that came from an era where we had to learn things, not lookup 
Let’s cross an item off the list.
I’ve been wanting something better than this for attaching accessories to the T slot.
The closest thing commercially available for a 30 system profile is a hammer bolt, but comes with 3 disadvantages:
So, let’s make our own. Since there’s an excess of these
we can turn a few of them to this
debur/reform the first threads from the other side
and create a T bolt on steroids
which is a perfect fit for the T bed profile.
Then we can mount accessories like this
which btw is one from my crappy clamp set that I keep for sentimental reasons (they were the very 1st project I cut with the X-Carve).
EDIT:
I forgot to mention that the T bolts are meant for variable height attachments (e.g. clamps). For fixed height attachments (e.g. wasteboard strips or the low profile vise) the T nuts work best.
That counts for most of us
The ballscrews came. We have good news and bad news.
The good news is that the sizes are perfect, the fit between the plates is great. The bad news is they messed up the threaded part.
But we will still make them work.
We have two options. It might take a while, depending on what cure we pick and where the medicine will come from (EU or Far East). Then it’s all downhill.
So, after the guy at the hardware store saw me for the third time in 3 days getting samples of nuts of all M12 pitches (M12x1.00, M12x1.25, M12x1.50, M12x1.75) and still failing, he asked if I wanted to borrow his thread gauge. I said sure!
The pitch is coarser than 0.8 .
The pitch is finer than 1.0 (which was the given spec) .
Yep, 10 threads can fit in 9mm. Holy #$^#, this is 10% off!
Now, here’s the most interesting part. Due to the backlash in metric threading, collision between a 0.9 and 1.0 pitch (i.e. a 0.9 screw and a 1.0 nut) occurs after the 4th thread. If you look closely, you can notice that the first thread touches the right side of the gauge’s tooth and the last thread touches the left side of the respective tooth. The next one will collide.
Which gives us solution #1.
Even though the thinnest standard nut I can find (DIN439) is 6mm thick => 6 threads => collision, some people do make even thinner ones. These are 3mm thick => 3 threads => no collision.
And here’s solution #2.
Forget the nuts. Use a shaft collar.
Solution #1 pro: better fit, can use pairs as jam nuts (or even fill the whole threading all the way up to the coupler as a big nut/spacer => bombproof)
Solution #1 con: slow delivery
Solution #2 pro: fast delivery
Solution #2 con: comes second in fitting tight (axially)
Or I can get both, they are pretty inexpensive, and see what I’m happy with.
And, btw, since I’m gonna have to wait some more, I took the opportunity to also order some 0.1244" dia, grade 25 balls to replace the 0.125" ones that came with the ballnuts. They do fit, but thought to make them a little smoother in order to (a) not overwork the motors trying to beat some extra friction and (b) extend the ballnuts life. More on that when they come.
Also no load bearing. Even without collision the nut will only touch at one specific spot and not transfer the load of the leadscrew. In this case a collar is a better solution.
Have you verified the insurance piece?
Best solution is to contact the supplier of the rods, explain them the problem and ask the whether they can make you some specific nuts with the same tolerances or send you a set of new rods with the proper thread.
I think it is fair to give the supplier a chance to rectify what they messed up.
Since this is suggested by two royal advisors, who am I to say otherwise
Also, since in ascending order of optimal engaging and load balancing the shaft collars come in these types
grub screw type < split type < two piece type
let’s try the latter, and specifically this one.
I believe I might have been too straight forward when I told them that the ballscrews look like they have come out of a slaughter house and not a machinist shop.
I have already asked for a refund only for the machining fee, which in effect will cover the additional cost of the new shaft couplers (which is caused due to the fact that the machining was 10% out of spec to begin with). They guy did not want to talk to me today, he told the secretary to tell me they are thinking about it.
In their defence, in the past they have been great and very fast in their after sales involvement when they needed to correct some minor issues (a broken bearing here, missing drag chain links there…) . I also want to continue being their client, it’s nice to have a local option and someone you know their name and phone to talk to (even when they are too pissed to talk to you cause you called them butchers).
True but I doubt Phil is royal.
If you can find a spacer you can use the coupling to be your collar.
Heeeey… spoiler alert for solution #3.
You gotta be kidding me! I woke up this morning with this idea. I have even found a gas copper pipe of 15mm OD, 1.2mm wall (=> 12.6mm ID - closest to 12mm I could find). I have pipe cutters at home from my bike tinkering times.
I was planning to use this as a temporary step till I get the super duper $60 couplers
Seriously now, I don’t think I would keep this for longer. 1.2mm wall and double duty for the coupler (which is essentially a split type shaft collar) might be too much.
When searching around for parts I found this company https://www.hardware-cnc.nl/nl/winkel/kogelomloopspindels/kogelomloop-spindels-16mm-spoed-5mm
They carry a large range of spindles (spindels) in different sizes and can customise the length of the spindle but only do that on the short uncomplicated side of the spindle. https://www.hardware-cnc.nl/nl/winkel/kogelomloopspindels/kogelomloop-spindels-16mm-spoed-5mm/custom-aanpassing-spindel-detail
