They are listed on the 4th step of the build manual.

 

I have plenty of lengths of 2040 but no 2060. Can I use 2040 instead for the base? If so, time to build the Sphinx. Also, my C-Beam is 600 long, does that give any issues having a longer Y axis?

 

Man...I was so happy to see this!!! That's what I've been needing for such a long time. BUT...I started to read reviews on this, thought I'll do it before I go out and buy it, and the majority is negative since the center ring pops out on them after a very short use. They are WAY to expensive to be used only a short time and then discarded.

Too Bad!

 

Justin:

I just use the screw mandrel that comes with the Dremel:



Along with the reinforced 1.25" cutoff wheels which are much stronger than the standard wheels that come with the Dremel:



Shop around, if you buy generic brand these can be as little as $10 for 100 pcs.

And yes, they are fragile, but at 10c each I dont care if I break them.

MG

 

Hey, noob here to Openbuilds projects. I have started building my Sphinx "XL" (1 meter C-beams) and also saw the Y-gantry mini V-wheel M5 bolts needed to be cut down to 27mm. I do not have a belt sander or Dremel, just a hacksaw. I was worried about accuracy of getting the M5's at 27mm, and not taking a long time to cut multiple 30mm M5's.

I 3D-printed a small block with 4 holes @ 1" spacing, then soldered in brass M5 threaded inserts to use as a cutting jig. I threaded the 30mm M5 bolts through, used a 3mm spacer as a depth gauge on the cutting side, and on opposite side tightened flanged nuts to lock bolts in place. The PLA did get chewed up, but it worked OK -- all were about 27mm length. I might do this a little different next time, but I guess that's how you learn.




I would like to say that Kyo did an awesome job on Sphinx documentation (BOM, PDF, video, etc), Chris Laidlaw's Sphinx plates are of great quality workmanship, and this and other Openbuilds forums have been a great source of guidance and definitely inspiration to build. (But, my wallet has suffered...).

 

Do you have a BOM for 1500x1500? Or 1000x1500? I love this design. I think I want to go bigger though, I plan on cutting out cabinet pieces and furniture.
Thanks a ton!

Tony

 

I know it probably says somewhere but how big a project can this CNC cut?

 

Thanks for the info. This will be my first CNC build and I am doing my homework on the subject. I am thinking I would like to be able to do projects up to approximately 24" x 36". So that being said how much longer and wider does the frame need to be.

 

I've been running my 1x1 Sphinx for a while now. One annoyance that I've found is that the lock collars have a tendency to become loose and are hard to tighten.

I have been upgrading the 5/16" (about 7.94mm) clamping collars. These seem to be much more robust. (eBay link: (4) 5/16" 2PC FULL DOUBLE SPLIT NEW SHAFT COLLAR BLACK OXIDE CLAMP SC31D SC031D | eBay) Perhaps this is something for the parts store to stock?

In the future I plan on upgrading the screws and nuts used for the wheels on Z to class 10.9. Another upgrade that's planned is to use springs to take out the slack on the lead screws.

 

Truly one of the weakest links in the system if you ask me. Openbuilds really needs to go with 8mm Clamping Lock Collars and get rid of the set screw version.

I will say in my opinion, but I know other members that share the opinion as well that the set screw versions are poor for the task at hand. Unless you make a flat in the leadscrew, and that would have to be placed perfectly, the set screw will always shift the collar to be at an angle, this introduces backlash in the system.

I have recently replaced all of my collars with a clamping style and things have improved significantly in terms of accuracy.

 

@Kyo Thanks for awesome documentation, sharing and general opensoucedness (is that a word? not according to Chrome spellcheck )



Hats off to all of you who have actually built a CNC machine - I've been promising myself that I wold build one for about 7 or 8 years, and even started collecting parts from old printers and scanners. Now I find myself in a position where I need to make some double sided PCB's, and I think that I can justify a CNC mill if it meets two requirements :

1) The precision / accuracy is sufficient to etch proof-of-concept PCB's.
2) Versatile enough to also cut and mill aluminium and wood.

Now this Sphinx CNC can certainly fulfil part 2, so that just leaves me wondering about part 1.

So after having read each post in this thread, and having tried a Google search to see if this question has already been answered, I hope I don't embarrass myself by asking a question that has already been answered:

Has anyone used this mill for PCB etching? If so, what were your results like? Comments?

Or any other comments re PCB's and Sphinx CNC.

I am also wondering if @beardyblair 's XL will do the job, but I will await answers here before asking on his build blog.

 

I think a CNC for doing circuit boards is a bit of overkill. It is certainly possible, but the level of accuracy needed to do .008" traces on .008" spacing is pretty high. Also, the machining surface must be perfectly flat to within a couple of thou, which is difficult to achieve with a large CNC router. Finally, even though you could do double sided boards, registration would be fairly difficult. Careful design is needed, and jumpers from side to side are inevitable. No chance of solder mask or silkscreen, either.

Don't get me wrong, it certainly is possible, but to pay off a $2000+ CNC machine by doing circuit boards would take a looooong time. Especially at the current cost of prototype PCB's. I have used EasyEDA, and double sided plate-through boards with solder mask and silk screen on both sides, along with tin plating, comes out to less than $0.25 per square inch, and took a week to deliver. There are tons of other prototype board services out there that are so cheap they might as well be free.

You can't even come close to this making your own boards. (Fun as it may be)

MG

 

@Metalguru, thanks for your feed back - much appreciated.

Yep. Understood. Unfortunately, I was a little unclear in my earlier post. I'm looking for an *excuse* to dedicate some time and money to a build, not justification ;-). I know it will probably never return a financial profit, but think of all the fun making stuff!! So many uses flying through my head ....

Hehe, not to mention tools required for that spacing.I'm only considering prototypes, and my components aren't that small (yet). What positional accuracy can be achieved with this CNC? 1mm? 0.5mm? 0.1mm? 0.05mm [I work in mm, 0.394", 0.197", 0.039", 0.02"]?
I understand that it will come down to how well everything is constructed and maintained, but can I move the spindle 0.5mm accurately?

Yep. Thought the well supported bed in this (relatively) small build would make this achievable. I would initially mill a holding frame into the waste board. This would serve 2 purposes - 1 to ensure that the bed is flat to the mill, and 2, allow me to drop the PCB into it and have it restrained accurately. With this mill, if I found that the bed had to much flex, I could of course, mill my own set of plates for a smaller mill. I know, backwards from where everyone else wants to go *shrug*.

BTW, you still haven't put me off my build yet

 

Go for it! My intention is not to "put you off", but just to vent my opinion.

With a belt drive machine like the ox, depending on what microstepping you use on the motor, you can get a position resolution of about 26-52 steps/mm, which is about 0.08-.04 mm per step. That's not bad, only a couple of thou. Resolution is not the same as accuracy, however. With a screw drive, you can get 4 times that or more.

You numbers for decimal equivalents were off by an order of magnitude, 1mm=.0394", not .394".

Milling a flat surface with the machine is a good way to get a perfectly flat surface pupendicular to the bit. Compensates for all the little misalignments of the machine.

I'm not saying you can't make PCB's with a mill, but I certainly could find better uses for one. Precision work is best done with a smaller machine, less flex, etc. Signs don't require high precision, but need a large table. It's very hard to make a machine that does everything well.

Also keep in mind that in order to mill metals like aluminum, the machine benefits from a small, stiff structure. Also, aluminum is quite difficult to machine, it requires lubrication and a lot of practice. And a spoil board that doesn't turn into oatmeal when it gets wet.

 

With 8mm leadscrew and 8 micro stepping you would get a accuracy of about 0.005. That’s in theory. But <0.1 mm is very much achievable with a Sphinx.

 

*Blush*

At least I waited for my second post to embarrass myself.

Colin.

 

Thanks.