My first ever built CNC router.

[Timatronics]

Hi to everyone here,

I've been working on a design for a DIY CNC Router for a little while and recently I said damn it, I need to build it. I've kinda gone headlong into it and started to order bits (probably a mistake) but I wanted to get some professional feedback to hopefully put my mind at rest a little.

I've attached 2 pictures of the frame work and also the completed (well almost) design. Its designed to take 400x400 materials with a rough cutting height of about 120mm. I plan to use NEMA 23's @ 2.2NM and 1.5kw spindle (er11) collet and it will hopefully be used to cut aluminium, delrin, brass (Maybe) etc. 1605 ballscrews and 20mm linear rails for the guildance.

28331
28332

My goal was to try and make one for about £1000 but as I've added all the extra bits its about £1300 at the moment without any control costs. I think its more than likely £1500 by the time its all there.

These are what I have ordered so far as well as the ali extrusion and metals, nuts and bolts etc.
https://www.aliexpress.com/item/32955478058.html?spm=a2g0s.9042311.0.0.17f54c4ddlT n1H
https://www.aliexpress.com/item/4000188787306.html?spm=a2g0s.9042311.0.0.17f54c4dd lTn1H
https://www.aliexpress.com/item/33057647776.html?spm=a2g0s.9042311.0.0.17f54c4ddlT n1H

I have some concerns after I ordered these bits, more so as I started researching more whilst waiting for them to arrive. The linear rails and ballscrews have arrived and seem great, some of the rail blocks need looking at but I see that's a common thing. My spindle also arrived with the VFD which sadly doesn't have a ground point so I need to figure that out at some point. Still waiting for the steppers but after reading up about the drivers and power supply I am a little concerned they won't be up to the job.

I am currently learning as much as I can about the wiring and I think I am getting there with that but I do have a little confusion about the number of power supplies I need in total especially as I plan to use homing swtiches. Each day is definitely a learning curve.

Originally I was planning on using the Breakout board that comes with the stepper kit but adding up the extra costs of getting a serial computer and mach3 or even using LinuxCNC I am kinda now leaning towards a DDCSV2.1 for offline control. With that in mind I use Fusion for CAM so I'll need to sort a post processor out.

Anyway, enough of my ramblings. I'd be very much interest in any feedback or suggestions you may have.

Thanks

Tim B

[driftspin]

Hi to everyone here,

I've been working on a design for a DIY CNC Router for a little while and recently I said damn it, I need to build it. I've kinda gone headlong into it and started to order bits (probably a mistake) but I wanted to get some professional feedback to hopefully put my mind at rest a little.

I've attached 2 pictures of the frame work and also the completed (well almost) design. Its designed to take 400x400 materials with a rough cutting height of about 120mm. I plan to use NEMA 23's @ 2.2NM and 1.5kw spindle (er11) collet and it will hopefully be used to cut aluminium, delrin, brass (Maybe) etc. 1605 ballscrews and 20mm linear rails for the guildance.

28331
28332

My goal was to try and make one for about £1000 but as I've added all the extra bits its about £1300 at the moment without any control costs. I think its more than likely £1500 by the time its all there.

These are what I have ordered so far as well as the ali extrusion and metals, nuts and bolts etc.
https://www.aliexpress.com/item/32955478058.html?spm=a2g0s.9042311.0.0.17f54c4ddlT n1H
https://www.aliexpress.com/item/4000188787306.html?spm=a2g0s.9042311.0.0.17f54c4dd lTn1H
https://www.aliexpress.com/item/33057647776.html?spm=a2g0s.9042311.0.0.17f54c4ddlT n1H

I have some concerns after I ordered these bits, more so as I started researching more whilst waiting for them to arrive. The linear rails and ballscrews have arrived and seem great, some of the rail blocks need looking at but I see that's a common thing. My spindle also arrived with the VFD which sadly doesn't have a ground point so I need to figure that out at some point. Still waiting for the steppers but after reading up about the drivers and power supply I am a little concerned they won't be up to the job.

I am currently learning as much as I can about the wiring and I think I am getting there with that but I do have a little confusion about the number of power supplies I need in total especially as I plan to use homing swtiches. Each day is definitely a learning curve.

Originally I was planning on using the Breakout board that comes with the stepper kit but adding up the extra costs of getting a serial computer and mach3 or even using LinuxCNC I am kinda now leaning towards a DDCSV2.1 for offline control. With that in mind I use Fusion for CAM so I'll need to sort a post processor out.

Anyway, enough of my ramblings. I'd be very much interest in any feedback or suggestions you may have.

Thanks

Tim BHello Tim B,


Welcome to this forum.

This one of the best places to get information on diy cnc builds.

Most people post their ideas and design concideration in a ( pre ) build log.

So people can help you narrow down any options you have for size machine, materials to use, purpose of the machine, trusted suppliers, budget and so on.

Any 1st reply to a build log starts with don't buy electronics kits.
Finish you design before buying anything.
Post design pictures
Ask for critique


So you don't waist any money.

I have one cost nutral tip.

Swap the bearing blocks and rail on the z axis.
So the rails are on the moving part of the z axis.


You will find here what you are looking for.

Grtz Bert.




Verstuurd vanaf mijn SM-A505FN met Tapatalk

[Timatronics]

Hello Tim B,


Welcome to this forum.

This one of the best places to get information on diy cnc builds.

Most people post their ideas and design concideration in a ( pre ) build log.

So people can help you narrow down any options you have for size machine, materials to use, purpose of the machine, trusted suppliers, budget and so on.

Any 1st reply to a build log starts with don't buy electronics kits.
Finish you design before buying anything.
Post design pictures
Ask for critique


So you don't waist any money.

I have one cost nutral tip.

Swap the bearing blocks and rail on the z axis.
So the rails are on the moving part of the z axis.


You will find here what you are looking for.

Grtz Bert.




Verstuurd vanaf mijn SM-A505FN met Tapatalk

Thanks for the reply.

Yea, I was expecting a few "don't buy anything yet" responses and completely understand. I tend to rush into things which could possibly be a massive downfall but I also like learning from mistakes although potentially expensive ones.

Thanks for the advice on the z axis. I will consider it for sure. Whats your reasoning behind it out of interest?

Thanks

Tim B

[driftspin]

Thanks for the reply.

Yea, I was expecting a few "don't buy anything yet" responses and completely understand. I tend to rush into things which could possibly be a massive downfall but I also like learning from mistakes although potentially expensive ones.

Thanks for the advice on the z axis. I will consider it for sure. Whats your reasoning behind it out of interest?

Thanks

Tim BHello Tim B.

Well the best thing is your committed now.
You will own a CNC router.

I think it can be a running system in the end.
You can upgrade parts later.



For the z axis.

You want the blocks in a square on a solid plate close to each other.

That is the stiffer setup.

The rails and spindlemotor will all add to rigidity of the z axis plate.

What ballscrews did you buy?
Are they straight and all?

Grtz Bert.

Verstuurd vanaf mijn SM-A505FN met Tapatalk

[RobC]

Hello Tim, I did exactly the same as you but posted my build progress as I went so I got help before I made any purchases for expensive wrong items. I'd strongly recommend if you can to cancel the air cooled spindle and get a water cooled 2.2kw spindle and the reciprocal VFD for such. You will only ever hear the motors and cutting instead of a constant noise which for me kills the fun of watching your machine make something :emmersed:

Breakout board, so many to chose from. You can go all out cheap and buy a generic 5 axis one for £6, they work perfectly in my experience only you likely will need to buy a Uc100 to convert from serial to USB if you do not have a serial capable motherboard. They're around £100 so take your pick wisely, I'd recommend going for a newer ethernet one with gecko drivers.

The big thing which I can't say enough which I didn't do but now try in all my designs is one very clear statement. If you can't make it perfect, at least make it adjustable! Makes like so much easier when later you think, why can I not get a perfectly smooth face and you then need to start cutting shims and making cuts so you can tram an axis in if you are not able to (What muggings here had to do because I'm impatient).

Here is my first machine all be it on its second iteration with a new gantry milled from 20mm ali instead of 10mm steel which it previously had. The first image is my complete disaster if im honest, I relied too much on the steel being true of which it wasn't so here is the latest version in the second picture In an old enclosure.

28339
28338

I'm still not sure if its worth making a new one when I can because I'm holding good tolerance .01mm most of the time and it works well.

Kind Regards
Rob

[Doddy]

you likely will need to buy a Uc100 to convert from serial to USB if you do not have a serial capable motherboard.

"you likely will need to buy a Uc100 to convert from USB to Parallel if you do not have a parallel capable motherboard."


Fixed that for you

[JAZZCNC]

I've been working on a design for a DIY CNC Router for a little while and recently I said damn it, I need to build it. I've kinda gone headlong into it and started to order bits (probably a mistake) but I wanted to get some professional feedback to hopefully put my mind at rest a little.

Ok well saying don't buy anything yet would be saying it after the horse as bolted so instead I'll say SEND IT BACK.!!

The rails and ball-screws are ok-ish thou they are the budget range but still better than round rail so your ok there.
However the drives and motors you need to dump as they will cripple the machine, esp if using parallel port.

The spindle is open to debate, some like Air cooled for it's simplicity. Me personally I prefer water cooled not so much because of the noise, which they are much quiter, but for the fact I can run jobs that are 20-30hrs and not worry. Try running a cheap air cooled spindle for more than 8hrs continuously and it will melt down.
Now your thinking aghh but I won't ever cut that length of time.!!.. … But you'll probably find you will at some point, esp if you ever want to cut 2.5D/3D ie: relief carvings etc which even a small-ish job can easily take 8hrs if it's detailed.

Regards the controller, if your going with a parallel port then go with Linux CNC. Mach3 is rubbish with a PP and flaky.
My advise is spend a little more and buy a decent ethernet motion controller like UC300 or AXBB because parallel port is slow and frustrating. The motion controller is the heart of the machine so it needs to be good.
Regards the Stand alone DDCV controllers then all I'll say is I know lots of people who have taken that route only to still end up fitting a PC based controller because they find them restrictive or don't like the fact they can't do the things a PC based one can.?



I have one cost nutral tip.

Swap the bearing blocks and rail on the z axis.
So the rails are on the moving part of the z axis.

No that's the case with this design and it will restrict the machine rather than make it better unless the whole gantry is re-designed to account for it.!

I've said this several times now but there is very little difference between the two methods in real world use depending on gantry design. The type and design of machine/gantry is a better selector for which method is best.
In this case it would restrict the travel of the Z axis which would require the gantry is made taller to get the travel back. This negates any advantage that rails on the front plate would have and actually turns it into a disadvantage because tall gantry sides introduce more vibrations at the spindle which negatively affects finish quality and tool wear.

Also because the gantry is short n stubby with a short front plate, which is a good thing, the cutting forces are transferred straight back into rear plate at all times.
Yes there is a constant length lever affect but it's short and not an issue when cutting, however the advantage in cutting height it gives is well worth the tiny difference.
The short stubby gantry makes this possible.

[RobC]

"you likely will need to buy a Uc100 to convert from USB to Parallel if you do not have a parallel capable motherboard."


Fixed that for you

I have no idea why serial was in my head, haha thanks!:culpability:

[driftspin]

No that's the case with this design and it will restrict the machine rather than make it better unless the whole gantry is re-designed to account for it.!

I've said this several times now but there is very little difference between the two methods in real world use depending on gantry design. The type and design of machine/gantry is a better selector for which method is best.

Hi JAZZCNC.

I will study this one,

I Always thought mounting the horizontal and vertical bearingblocks on the 4 corners of a single piece of rigid plate would always beat any other setup and would result in the most stif setup.

No influense from flex there.

The flex in the moving zplate could be made less by putting on 2 big motor mounts using round type spindle motormounts.

In this case the spindle motor is square which makes that one a bit easier.

I would have thought having a bigger lumb of metal (rail rizers, square rails, z plate, motor mounts, bolted to the spindle would lessen the vibrations.


So I have something to think about.

It is good to have al these experienced guys here.


Grtz Bert.




Verstuurd vanaf mijn SM-A505FN met Tapatalk

[JAZZCNC]

Hi JAZZCNC.

I will study this one,

I Always thought mounting the horizontal and vertical bearingblocks on the 4 corners of a single piece of rigid plate would always beat any other setup and would result in the most stif setup.

No influense from flex there.

The flex in the moving zplate could be made less by putting on 2 big motor mounts using round type spindle motormounts.

In this case the spindle motor is square which makes that one a bit easier.

I would have thought having a bigger lumb of metal (rail rizers, square rails, z plate, motor mounts, bolted to the spindle would lessen the vibrations.


So I have something to think about.

It is good to have al these experienced guys here.


Grtz Bert.




Verstuurd vanaf mijn SM-A505FN met Tapatalk

Like I say it depends on the design but in most cases the difference is negligible and other areas of the machine often overlooked will have a bigger impact.
In this design, the difference between losing Z-axis travel and raising the gantry to regain it will far outweigh any benefits that stiffening the front plate might make.
Also, it's a mistake to think that the rails stiffen the plate because they don't to any great degree. Plus the weight of the longer front plate and rails have a negative effect in that they require more power to give the same acceleration which is important in a Z-axis.! . . . these are the little things which are often overlooked but can have a big impact on performance.!

[Timatronics]

The rails and ball-screws are ok-ish thou they are the budget range but still better than round rail so your ok there.
However the drives and motors you need to dump as they will cripple the machine, esp if using parallel port.



Thank you for the reply JazzCnc. Fortune paid on me with the steppers and drivers as after a mix up I got my money back and for some reason they also delivered them anyway. What would be your recommendation of Steppers and drivers (budget wise)? Currently I'm looking at the Nema 23, 2.2NM (8 wire) ones from CNC4YOU with the CW5045 stepper drivers with a 48v power supply.

I'm pretty happy with the spindle and got it wired up to the VFD and it works great so far. I chose air over water cooled as I prefer the convenience of them and this machine will probably end up travelling with me when I work at different workshops.


advise is spend a little more and buy a decent ethernet motion controller like UC300 or AXBB because parallel port is slow and frustrating. The motion controller is the heart of the machine so it needs to be good.

I looked up your suggestions of controllers and will take them into consideration, more so the UC300 as the AXBB is going over my budget. I see alot of Chinese ethernet controllers - would they be any good for budget reasons? Currently I have an old laptop given to me by a friend with a parallel port so I'm planning on installing Linux CNC just to get me going for now.

Thanks
Tim

[Doddy]

I looked up your suggestions of controllers and will take them into consideration, more so the UC300 as the AXBB is going over my budget. I see alot of Chinese ethernet controllers - would they be any good for budget reasons? Currently I have an old laptop given to me by a friend with a parallel port so I'm planning on installing Linux CNC just to get me going for now.

Thanks
Tim

I have one flavour of Chinese Ethernet motion controller - a Novosun NVEM 5 axis controller. Used it for a couple of years on one machine. Replaced with a UC300ETH. It's collected dust for the next couple of years as I bought a Mesa Ethernet controller for the Lathe (Lathes present a problem with control software which pushed me down a LinuxCNC route, which pushed me at the Mesa). All-in - around £500 extra because I DIDN'T want to use the NVEM. The NVEM work, on a good day, but I've had to modify stepper drivers to work reliably. There was all sorts of problems ensuring Mach3 and NVEM synced correctly - you might find for example that if you exit Mach3 (crash, etc) that you have to repower the machine to get the NVEM ready to accept the connection on restart.

My experience - I can afford to throw money at a problem to make the NVEM go-away (although its sat on a shelf, somewhere). For the price of a UC300ETH, and a parallel BoB, to me it's a no-brainer which to buy. Plus, I've kind of fallen in love with UCCNC rather than Mach3, so that's a more clear choice for me (the NVEM cannot be used with UCCNC... last I checked).

[Timatronics]

I have one flavour of Chinese Ethernet motion controller - a Novosun NVEM 5 axis controller. Used it for a couple of years on one machine. Replaced with a UC300ETH. It's collected dust for the next couple of years as I bought a Mesa Ethernet controller for the Lathe (Lathes present a problem with control software which pushed me down a LinuxCNC route, which pushed me at the Mesa). All-in - around £500 extra because I DIDN'T want to use the NVEM. The NVEM work, on a good day, but I've had to modify stepper drivers to work reliably. There was all sorts of problems ensuring Mach3 and NVEM synced correctly - you might find for example that if you exit Mach3 (crash, etc) that you have to repower the machine to get the NVEM ready to accept the connection on restart.

My experience - I can afford to throw money at a problem to make the NVEM go-away (although its sat on a shelf, somewhere). For the price of a UC300ETH, and a parallel BoB, to me it's a no-brainer which to buy. Plus, I've kind of fallen in love with UCCNC rather than Mach3, so that's a more clear choice for me (the NVEM cannot be used with UCCNC... last I checked).

Thanks for the information Doddy,

So regarding the UC300ETH, its not a complete package then, I need a separate BoB to work with it? Would it also work on LinuxCNC? I guess its not a real issue as UCCNC is pretty cheap.

Thanks
Tim

[JAZZCNC]

Thank you for the reply JazzCnc. Fortune paid on me with the steppers and drivers as after a mix up I got my money back and for some reason they also delivered them anyway. What would be your recommendation of Steppers and drivers (budget wise)? Currently I'm looking at the Nema 23, 2.2NM (8 wire) ones from CNC4YOU with the CW5045 stepper drivers with a 48v power supply.

This is easy to answer.!! . . . I wouldn't fit a standard stepper system these days because for the same or similar money you can better get a closed loop system like this one below. Closed loop system gives you more piece of mind that your not losing steps because the encoder tracks position and the drive corrects any following errors.

https://www.ebay.co.uk/itm/Hybrid-Servo-Stepper-Motor-Driver-57HSE2N-D25-Closed-loop-2N-m-57HSE-Motor-Kit/174299397313?hash=item28950dbcc1:g:o7cAAOSw4vVdgef 7

The Cnc4you drives are ok but nothing special and you could probably buy the same things cheaper but if your happy with a standard setup and want a Uk supplier then they are ok. However, the drives are rated for 50V max so I wouldn't recommend you use 48v PSU, Ideally you need a 10% safety factor on the PSU due to the motor feeding power back into the drives. I think you'll find Cnc4you have a warning on there 48V PSU saying not to use with CW5045.! . . This is why.!)




I looked up your suggestions of controllers and will take them into consideration, more so the UC300 as the AXBB is going over my budget. I see alot of Chinese ethernet controllers - would they be any good for budget reasons? Currently I have an old laptop given to me by a friend with a parallel port so I'm planning on installing Linux CNC just to get me going for now.

I would give the Cheap Chinese controllers a miss, I've never used one but seen others with lots of problems from them.

The UC300 needs a Breakout board so it's not any cheaper than the AXBB and to get it's full potential you want a good breakout board that can use all it's ports etc.

I've seen you mention LinuxCNC.! You cannot use LinuxCNC with UC300 or the AXBB it's UCCNC or Mach3/4 only.
If your going to use Linux CNC and want to use Ethernet then you should be looking at Mesa cards. However for a simple setup then Linux CNC on the parallel port will be fine if your prepared to accept it's a little more complex than say Mach3 or UCCNC to setup.

[Timatronics]

Hi,

I just wanted to say a big thank you for everyone's help here especially Jazzcnc for pointing out some purchase choices. I knuckled down and finally finished it a few weeks back, a few days before returning to work. I've not had that much time with it since but so far (apart from some settings issues) I've been super pleased with how its going. I had to think a lot about the controller and eventually settled on the DDCSV2.1 for now. Anyway here is the finished build and pretty much the first cut I made.
28839
28840
28840

All the best

Tim

[RobC]

Good job, looks rather quite nice and compact! Will you be putting it in any enclosure or fences around to prevent the material from being thrown all over the place? I have mine in an enclosure with a 24v 100mm fan pumping the airborne dust outside. Not to mention it also is a damn sight better on the ears as the noise is reduced by a good near half of what it would be without an enclosure.

[Timatronics]

Good job, looks rather quite nice and compact! Will you be putting it in any enclosure or fences around to prevent the material from being thrown all over the place? I have mine in an enclosure with a 24v 100mm fan pumping the airborne dust outside. Not to mention it also is a damn sight better on the ears as the noise is reduced by a good near half of what it would be without an enclosure.

Thanks Rob,

Quite possibly yes as I was testing some cuts today and I'm pretty my whole street could hear it, not great on a Sunday. Its mainly gonna be used for aluminium and delrin so that tends to just gather around the machine so I'll need to look at extraction close to the spindle at some point. At the moment its my trusty old hoover :)

Tim

[RobC]

Yeah mine too is very noisy when doing deep cuts, I thankfully have a water cooled spindle so its just the cutter noise. I don't have any setup for collection, I clean up on each new tool change and then the end with a 100mm dust collection system hose. Again all down to the noise of the dust collector :thumbdown: