[Maccydu]

Well, I've finally (almost) completed my first Router build. Still waiting for my inverter, and some electrical parts (E-Stops, Limit Switches, Panel Enclosure) but managed to run all the axis for the first time today.

I was given the THK linear rails and carriages, so the design was built around those. Overall travel is roughly 400mm x 500mm x 100mm

Can't wait to start cutting something with it!

Really wish I'd taken a load more pics during the build - I will on my next one - especially now I can machine my own parts!

Have also uploaded a video to Youtube showing the 3-axis running. Very slow at the moment (2000mm/min) - need to work on speeds and feeds. (Video cuts short by about 10 seconds, all it does is return to centre of bed)

https://www.youtube.com/watch?v=kX5o...ature=youtu.be

I managed to get the laptop off eBay for 99p! - Needed a harddrive (I had one lying around), but it runs pretty fast, came with a Windows XP sticker, and the all-important parallel port!

Would love any feedback - my next design will be a moving gantry design I think.

Thanks

Matt

[jfong]

Congrats on your first CNC. THK HSR30's, Nothing but the best!! I have some SHS30's on my router, they are certainly heavy duty rails.

I would beef up the uprights if you can for added rigidity. I used 80/20 series15 1545 size on mine plus a 5/8 aluminum gusset plate for extra strength.

Nice work!!

[RicknBeachcrest]

That is a nice looking machine. Like mentioned above, I would add some diagonal bracing to the uprights.

[A_Camera]

Very nice build, congratulations. I have also fairly recently finished mine and mine is also a fixed gantry version, but my vertical beam has much more stability than yours, and you can't have too much of that, so I would advise to increase that considerably, especially if you plan to mill aluminum as well.

Another thing... you say yours is "Very slow at the moment (2000mm/min) - need to work on speeds and feeds." Mine is topping 450mm/min and I find it far too fast when something goes wrong and the machine needs to be stopped. OK, I do wish I could go faster when everything works fine, but 2000mm/min is I think fast enough for small table machines, like yours and mine. Remember that milling at that speed is not possible anyway, so the benefit of fast rapids isn't as much as one would expect, except if you often need change tools and move far away from the work area, the benefits are limited. Just for curiosity, what voltage/current are you driving your steppers with?

A last advice... whatever happens, don't run without properly calibrated limit switches! ...especially at that high speed.

[jfong]

I currently have my CNC set to 250ipm Rapids (6300mm/min) and I'm switching over to servo to get closer to 500ipm. I do a lot of carving in wood so faster the machine runs the better. I don't know what the OP is using his machine for but for me 2000mm/min is slow. My CNC is only 2'x3' size.

[A_Camera]

I currently have my CNC set to 250ipm Rapids (6300mm/min) and I'm switching over to servo to get closer to 500ipm. I do a lot of carving in wood so faster the machine runs the better. I don't know what the OP is using his machine for but for me 2000mm/min is slow. My CNC is only 2'x3' size.

Fine, but have you ever REALLY measured the speed? Also, when you are carving, it can hardly matter, since I don't think you can accelerate/decelerate and cut that fast, and rapid moving speed mostly matters when you move across the table. With small table that high speed is in my opinion not handle-able, but perhaps I am wrong. Another thing... are you using belt drive and perhaps a gear box as well? It is quite different to what I prefer, since I need VERY high accuracy, and that costs speed, or comes at a very high price. Also, if you mill/carve fast, you also need a very fast spindle (perhaps over 20k RPM), which I assume you have. I am milling with 0.3mm end mills as smallest, and NEVER larger than 6mm (only used for surfacing), so I don't need that kind of speed, but of course, all things being equal, if you have a very high speed system that is better, but not if you lose accuracy, at least not for me. I guess for carving wood, that kind of accuracy is not important, and I guess your tools are considerably larger than mine so it may work for you, but never the less, 6300mm/min is difficult to stop when you see something went wrong... even at my very low speed I was not there in time twice, once the tool run trough the work piece and the mill drilled a hole in my precision wise, destroying the work piece but luckily, the wise is still OK, another time I managed to stop, but too late to save the work piece. Other times I stopped the movement in time, but not always in time to save the end mill... so, perhaps for a newbie, like I am, as well as the thread starter, the highest possible speed is not the goal until we are 100% sure of everything, and I don't think you can be 100% sure with a DIY machine. Once again, I'd love to have a machine which is extremely fast, as well as extremely accurate, but I am pretty happy with the speed and the accuracy I get, even if that sounds pathetically slow to some of the forum experts.

[jfong]

I custom built my router with high end linear components. THK linear rails and NSK ballscrews throughout. 2.2kw 24000rpm spindle. The Z and Y have been converted to servo. The dual X axis have not. I have some high performance brushless servos that need to be installed. I'm using Applied motion brushless drivers. Machine ran fairly fast for a stepper but will go even faster once I finish the servo conversion. The X axis steppers can't keep up with the Y and Z servos so the machine is detuned to 250ipm Rapids.

Of course when carving, it's not going to reach max speed because you still have to accell/decell but it still runs really fast.

I've seen some small table travel 3d printers move blazingly fast with high acceleration. I have way more mass so that would be impossible but those printers sure do fly.

As fast as my CNC can go, it's slow compared to a machine like CNCrouterparts that use rack and pinion. I've seen videos of them going 1000ipm. I'm pretty sure there are plenty of DIY CNC's made by CNCzone users that run 250ipm or faster.

Oh I know about mistakes. Been running the machine for 2 years now. Some very stupid but you learn from every one of them. I'm fairly confident on the machine now that I walk out for a few minutes and come back. I do have a video camera that I can view the CNC running on my phone. Sometime you just need another cup of coffee.

[A_Camera]

OK, I did some calculations, and 6300mm/min (250ipm) is not impossible with steppers, microstepping set to 8 if you have a lead screw with at least 5mm pitch, which I assume you have. That will set a pulse engine requirement to pulse at 33.6 kHz.

In my case, I made my own lead screws and nuts and the pitch is 1,75mm, so that speed is not possible for me because it would demand a 96 kHz pulsing and Mach3 is limited to 45 kHz. I am also not sure that my steppers could manage 3600 rpm, which would be needed. Currently the stalling speed is around 500rpm. I know, that sounds pathetic. Never the less, I am happy with the results because it is working fine and the speed is high enough for what I am doing. Of course, I do have upgrade plans, and those include real ball screws with low friction and larger pitch, higher stepper voltage (currently I only have 36V) and also higher spindle speed (currently I am limited to about 10k rpm, so that is pretty low as well as changing the computer to something with higher performance. I have some other upgrade plans as well but these are the main performance improver. Never the less, I built the machine on budget and earning money on it, even though this is just a hobby for me.

[jfong]

First time I heard about someone making there own leadscrews. That's a challenge in itself to do right. It's very hard to keep the pitch constant throughout the entire length. How was it done???

I bought all my ballscrews on eBay so not that expensive. Keep away from the cheap ones. Look for ground ballscrews from THK, NSK, kuroda, steinmeyer etc. C3 or better.

This is a old video, first time testing the axis at 300ipm. Stepper motor.

Gantry first movement test - YouTube

Testing the repeatability using the best test indicator I own

DIY gantry accuracy test #2 - YouTube

These NSK ballscrews are very good.

[A_Camera]

First time I heard about someone making there own leadscrews. That's a challenge in itself to do right. It's very hard to keep the pitch constant throughout the entire length. How was it done???

Sorry, I expressed myself a bit wrong, or confusing. I used ordinary stainless steel rod, cut off a piece from each rod and made a matching tap out of each, which I used to make a delrin nut with. It works very well, really zero backlash and it slides nicely, even if it is not like a real ball nut. It will of course eventually wear out, but it is easy to make new nuts if I need to.

Here is a short video of my machine:

https://youtu.be/Va0h5YpRqKE

I bought all my ballscrews on eBay so not that expensive. Keep away from the cheap ones. Look for ground ballscrews from THK, NSK, kuroda, steinmeyer etc. C3 or better.

No, ballscarews from eBay are not that expensive, but everything included costs at least $500 for my machine, and when I started my CNC building I was not sure that I'll be able to finish it so I wanted to save as much as possible. Of course, now that I know I can, it is easy to say that the decision was a mistake but on the other hand, I have not lost anything in the process, quite the opposite, I learned a lot. Upgrading with real ballscrews is peanuts and now I know it is going to work and that it's definitely going to improve speed, even if the resolution will be a bit less, 0.0031 mm/step instead of 0.0011 mm.

[jfong]

Your CNC is much better than my very first one I made. That must be atleast 15 years ago.

Waiting for the OP to post some videos of his CNC cutting aluminum. We're kind of off topic and this is his thread. Very few DIY CNC's have used such high end linear rails in 30mm size. They may look small in his pictures but if you ever held one, they are huge and extra rigid. Makes me want to find some on eBay but I know it would be very $$$$

[Maccydu]

So, finally made a bit more progress on my router, which includes building a stand, putting the panel together (temporarily), adding another panel for the inverter to go in on it's own (far too much electrical noise from those things!) and used the machine itself to make the new aluminium plate uprights.

Next steps are to tidy everything up, properly finish the panel, and then start adding limit switches etc. Wil then probably sell it and start again!!

Will keep you all updated as it goes on. At some stage I plan to replace the supported round rails on the z-axis for some 20mm square rail. I have the parts, just not the time!

Have a video of milling out the pocket, will upload to youtube and post a link.

Have fun!

Matt

[Maccydu]

Video below. I really need to get myself a stand for my phone!

[gfacer]

How is the Hertz spindle? I was thinking about buying one but went Chinese instead.....how warm does it get? Thinking if I ever need to replace a old spindle with a new one, the Hertz is probably a better choice than the chinese which I have using to replace porter cable routers with.

[Maccydu]

The Spindle seems pretty good, although I've definitely not pushed it hard yet. There was quite a bit of noise at around 15000RPM, but I've not fully bolted everything down everywhere so I'm not sure if there was any issue with the spindle or not.

After running it for about 30 mins doing the pocket (I know, I know, that's VERY slow) the spindle didn't feel warm at all, so I'm taking that as a good sign. What was more important to me was the accuracy of the pocket, and the surface finish, and both were spot on.

The spindle cost me about £430 all in with shipping and import duties, which is expensive compared to the Chinese ones, but the price has since dropped, most likely because of the strength of the £ vs the Euro, and also some of the unrest in the region (It came from Turkey).

At the current cost of about £330 all in I'd definitely buy one again. Only thing that bothers me a little bit with it, and I'm sure many people have the same issue, is that at low speeds - in particular conventional drilling speeds for aluminium - the system has very little torque and the spindle will stall a lot, so I'm having to peck drill with 3-flute cutters.

Other than that, very happy with it.

[asniex]

What is the total cost of the build mate ? I am very much interested in building something like this. I have been researching a lot about the prices of the linear glides and comparing the prices from the local vendors and online. Also I even considered getting a 3040CNC(very pricey with shipping and duty ) to machine out parts for another CNC :P
I have never built one before. Is it possible for a complete beginner(I have built a 3d printer) to build something like you have ? I wont be milling much aluminium on a daily basis I need it to make better parts for a desktop CNC I have in mind one that could mill PCBs and get really minute details for miniature builds like palm sized quadcopter etc.
I might be milling enclosures and some parts for various future projects of mine. what would be the accuracy of this machine ? Will I be able to mill PCBs ? 10mil tracks atleast.
@A_Camera Really nice CNC you got there. I have started reading your blog post. Would have been great if you could post your design.
Thank you.
Asniex

[Maccydu]

The cost so far is probably just a touch of £2k, although that doesn't include the THK rails and carriages on the X & Y axis that were given to me. They're obsolete now, but the equivalent from THK would cost around £1k. I'd never buy them, but I would use profile rail and bearings, which I reckon would cost me about £500 for some semi-decent branded of the same size.

With regards to the 3040CNC, I know people have lots of opinions about them, and how good they are or aren't, but I wouldn't touch one at all. I think they have their uses for small bits of work like PCB and some light engraving, but for accurately cutting aluminium I wouldn't be so sure, although YMMV.

If you did ever want some parts machining, please let me know and I'd be happy to have a go. I think I could get an accuracy of around +/- 0.1mm as the machine stands now, but that really is just a finger-in-the-air guess (and based on a few bits I've done already)

Matt

[asniex]

OMG !!! 2k pounds that's like 3000$ ! So this goes on my things I can only dream and not afford list :P
I have a budget of less than a $1000 and where I live its really hard for me to source the parts where I live (India) and quite a few machinist I met here only make generic parts no custom parts and services which offer laser cutting of parts wont do it for a one off part, so china is the only source for my parts. Just when I taut I could stay away from china and buy some branded ones likes HIWIN or even THK they cost a fortune.
I am looking at various designs made by many people it inspires and makes me wanna build one myself more and more. Right now I am just collecting old tools like saws and some of the used Machinist grade drills etc (I only own a hack saw and a Bosch drill currently :P).
I came across 2 or 3 designs that one can build just with the help of hand tools but they were made of wood. And I dint wanna buy X carve even though I was really tempted to get one. It dint seem very rigid and structurally sound to me.
It was really kind of you to offer me to machine parts mate. Thanks a lot for sharing the project it really inspired me(there were other projects too but yours were simple and nicely designed)