Precision 3D Printer ... from scratch (Page 1 of 2)

Looks cool, but have you thought about a ballscrew on each end of the x axis? otherwise inertial weight may be a problem at higher speeds.

Looks good.

My only suggestion would be don't disregard using a RAMPS setup instead of the Printrboard. It offers everything the Printrboard does, but has a lot more pins available for additional functionality, and a much wider user base.

Also, the AT90USB1286 (mega equiv on the Printrboard), whilst outwardly attractive for its inbuilt USB functionality, isn't as nice from a modular or expansion perspective as the ATMega + USB Chip approach fo the Arduino MEGA...

Cost wise, they are virtually identical...

I haven't received it yet.  Funny enough, it took 4 days from the day I bought it to get from China to San Francisco, and it's been stuck in the USPS system now for a week.  If I don't get it by tomorrow, I'll have to declare it lost and get USPS to start looking for it.

The other 3 miniature screws are coming from Korea.  They're Japanese made Issoku screws and they're used, pulls from CNC machines.  According to the vendor, they are in very good to excellent condition.  We'll see when they get here.

I looked at the RAMPS  boards, and I'm not a big fan of how it's laid out.  I guess if you're an Arduino fan, you'd probably appreciate the design with the plugin header connectors for the shields and such, I for one kind of like the simpler, more compact design of the Printrboard.

As for price, the cheapest i can find a similarly configured RAMPS, with fake Arduino board is about $120.  With real Arduino about $150.  I can get the Printrboard for $85

Compatible Firmware

    Sprinter Sprinter: Supported, use MOTHERBOARD == 9.
    Marlin Marlin: No official support yet, but works with Lincomatic's fork.
    Repetier List_of_Firmware#Repetier-Firmware: Supported, use MOTHERBOARD == 9.
    grbl List_of_Firmware#Grbl: No official support yet, but works with Lincomatic's fork.

Just know that if you switch to the printrboard then you will not be able to use the same firmware as the rest of us. Not that it should be too big of an issue.

Being able to make a cool machine is one thing. Being able to take it to market, promote it tirelessly, design support and inventory systems, hire good techs, deal with angry frustrated customers, cash flow problems, spurts of overwhelming problems... there's more to starting a new business (if that's what you're thinking) than you'd imagine if you've not done it. But, if you're talking about contributing your ideas to the RepRap community, then by all means...

Hey Lawsy,

I got the ball screw last week, and sent it to a machine shop to have the screw cut to length, and the ends machined.  I got it back, but unfortunately, the machine shop forgot to cut it to length first, so I sent it back, I should have it early next week.

I'm having both ends machined, the lower for the 8mm to 5mm Flex coupler to the motor shaft, while the upper end will be riding inside a 608 bearing which will be mounted at the top of the frame.  This should get rid of wobble once and for all.

I did play with it a little before I shipped it out, and I like it.  The movement is smooth as silk and surprisingly free.  I can take the screw, hold it vertically, and give the nut just a little nudge and it will roll down the screw length on its own for 3-4 inches before stopping.  That's pretty impressive for something that it tight enough to not have any play/backlash, so I expect the vertical travel to be nearly perfect, which should translate into prints free of banding.

If I still have banding after this, then it's not hardware, it's something else.

Looks sweet!  but I echo ronsii's concern about the y-axis drive mechanism.  When it moves, I suspect the whole x-axis will want to swing around the bearing at the right end rather than slide, causing backlash and possibly binding problems.

All this is conceptual right now, at this moment I got the 2 Issoku miniature ball screws and they're sweet.  These things are virtually frictionless.  I also have the long 8mm linear bearings (LM8LUU) as well as the 6mm end support ball bearings

Next step it to find a machine shop that will make the aluminum blocks for me.  I also need to decide if I'm going to use NEMA17 or NEMA23 motors, I'll probably stick to 17's.

Once everything is together, I'll be able to see if the "Y" rack will bind on the opposite of the ball screw.

All the blocks will be custom machined out of aluminum to my specifications.  The linear bearings are standard LM8LUU (45mm) or LM8UU (24.5mm).  They will be either pressed in or secured with set screws. 

Fortunately, these are simple to machine, so the costs of having these made should not be very high.

Here is an example of the blocks as they fit together and the bearings in side them.

I'm not sure I understand what you mean by irreversible.  The ballscrew turns bidirectlionally, so each ball screw is driven by one motor only. 

At this moment, I haven't decided if the motor will be in-line and drive the screw directly, or if I will be mounting it on top of the end block and using a gear drive.

Precision screw should work better then belts for backlash reasons, however i suspect that it will also be much slower; since it usually has very high mechanical force advantage. An idea i played with in my mind and on the back of a napkin is to add 3 linear position sensors and have the head location controlled in real time via the position sensors instead of just linear calibration of drive belts.

Problem with optical positioning is that is usually works in a short range of distances. I was thinking more along the lines of http://www.celesco.com/linearpot/