Hello,

I've been losing more and more time lately reading up about the DIY CNC movement, and I'm working up the nerve to start putting together a mill of my own. I'm still in the research phase right now, but I'd be interested if anybody could answer a few easy questions. If there is a nice DIY-CNC FAQ kicking around somewhere, I'd appreciate a link.

1) How do you manage to cut out parts using these kinds of router mills? I understand that you just plunge the bit in and carve out the shape (a little deeper each time) but doesn't that gouge up the surface supporting the piece? Are the machines so precise and well calibrated that they skim along the top without damaging it?

2) I hear threaded rod described in terms of two numbers (1/4-20). What do those numbers represent? Is it diameter and thread-count?

3) I (think I) understand the pipeline that takes the design in one end and burps out g-code from the other. How challenging is it to calibrate your CNC mill to accurately produce your designs? I mean, if for some reason my x-axis translates slower than my y-axis, my software needs to know so I can draw a square, or else everything might come out squashed. Is this a concern? Or is this why we use the same hardware for both axes? :)

I appreciate your time in reading this. I'm thinking about buying the plans and attempting to make the 7th Sojourn. I'm interested in doing relatively small bas relief projects in wood. Its cutting area, precision, and price seem to fit and I've found favourable reviews of it around the internet. I hope to come back soon with some more technical questions regarding the actual building!

Cheers,
Rob

Hello and welcome, my router has been a lot of fun, I'm sure it'll be the same for you.

1) You use what's called a spoil board. I screwed down some 3/4" MDF between the "T" track the ran a leveling program to make it nice and flat. I'll probably get 3 leveling opportunities before the spoil boards need to be replaced.

2) Precisely!

3) I used TurboCNC now Mach3 and both those programs allow each axis to be set separately so you dobn't have to use the same hardware on each.

Curtis

Yes - just to clarify, the calibration is done using maths i.e. if you know your threaded rod pitch (threads per inch or mm) you can convert this to lateral movement of the leadnut per motor step. The software takes care of the rest. Interestingly when I trialled mine I found the movement to be a smidge out, due to the inaccuracies in threaded rod manufacture (about 0.25% in my case). But it's easy to compensate for this by adjusting the steps per unit slightly. Using the Xylotex controller and metric M8 rod, the maths works out at 1260 motor steps per mm. This is the resolution of the machine (1/1260mm), however in reality the accuracy is considerably less than that due to backlash, wooden consturction and so on. Not bad though still :) The hardware for the axes doesn't need to be identical, but I guess it simplifies things slightly if it is.

In terms of cutting a perfect rectangle, the other thing to watch out for is that the x and y axes are at exact right angles. Mine weren't, but you can check this easily by using the machine to make a mark in the four extreme corners of the movement range. If the diagonal measurements don't match, then the angle isn't 90 degrees and things need adjusting slightly to compensate.

Hope that helps :)

Thanks guys, I'm feeling smarter already!

if you know your threaded rod pitch (threads per inch or mm) you can convert this to lateral movement of the leadnut per motor step.

Hah, I guess "thread count" is the term used for linens, isn't it?

Maybe you can answer another question. As a confused newbie, I get a little bit mixed up reading some of the DIY CNC pages out there. On this forum, I've seen it said a few times that building a machine will cost somewhere between $500 and $1000. But Cranky's plans boast less than $200 for parts.

Is this because it's a starter model? Or is it a difference of opinion regarding what is necessary (maybe using rescued printer steppers instead of the higher torque ones)? I mean, I have no intention of cutting aluminum. I'd be surprised if I routed anything harder than sugar maple.

Cheers,
Rob

it may cost $200 in materials, but you have to include the electronic parts on top of that, which can be easily $300-500 for a complete kit (control board, power supply & motors). Then more $$ for minor electronic parts (wire cables, limit switches, etc), spare computer to run the program, the program themselves, etc...

On an "average" first DIY build I'd expect $1k easily not including PC, software, and the router/spindle itself

I'd agree with that kind of figure - it's really hard to put a price on things since you're buying little and often i.e. a few nuts and bolts here and there etc. However I'd put the total cost of mine at around £500-£600 for everything including router and 2nd hand PC (double that for USD), and I got most of the wood for nowt. Depends how far you want to go with the build - lots of my parts are 'cosmetic' and not essential to it working as such. I think anyone selling plans is likely to intentionally give a figure on the lower side of what the reality might be though!