First Build - Yet Another Alu Profile Router - 900 x 800

[CaptainBarnacles]

Hi All,

Here's the first ramblings and outpourings of my mind as I embark on my first CNC router build.

Obviously I have already fallen into the trap of having purchased much of the gear before finalising the design. Here's what I have so far:

4 qty Hiwin 20mm Linear Rails and Carriages @1000mm (X & Y Axes)
2 qty Hiwin 20mm Linear Rails and Carriages @400mm (Z Axis)
2 qty 2010 Ball Screws @ 1000mm (Y Axis)
1 qty 2005 Ball Screw @ 1000mm (X Axis)
1 qty 2010 Ball Screw @ 400mm (Z Axis) - May look to replace this with a smaller, lighter screw??
2 qty 40 x 160 Alu Profile @ 1000mm
2 qty 80 x 80 Alu Profile @ 1000mm
4 qty 40 x 80 Alu Profile @ 1000mm
4 qty NEMA 23 motors
4 qty TB6560 Drivers
24V 250W PSU
4 qty Flex Couplers
6 qty Proximity Sensors
Misc Fasteners and fixings etc.

Engage sarcasm mode: I'm having so much fun learning Fusion 360 :Disengage sarcasm mode. Here's my part finished design:

25016
25017
25018

As you can see from my partial design I have opted to use tall profiles for the Y axis sides of the frame in order to reduce the height of the gantry support plates as much as possible. My thinking is that it will keep the CoG of the router as close as possible to the same height as the Y axis linear guides. On the X axis I have opted for 2 qty 80 x 80 profiles in the hope that it provides way more rigidity than I need. X Axis guides are mounted top and bottom of the gantry to minimise the router's extension from the gantry.

My current design considerations:

Direct or belt driven X and Y axes? I am currently leaning towards belt driven.
Reduce the Z Axis ball screw size to reduce overall Z axis assembly weight and provide a finer pitch, perhaps a 1204 or 1604 screw?
Use universal connectors or use L brackets to join the 40x80 of the bed to the 40x160?
Are my stepper going to have enough grunt if I use a 1:1 belt drive? They are 270oz rated, perhaps a 2:1 ratio?
Whether to make my own aluminium plates or have them made for me? I think I am plenty capable of making them but I have no mill or CNC machine so it'll all be hand work (I have woodworking tools such as mitre saw, jigsaw, router table, drill press etc.) so I question whther I can make them to the tolerances required for a really strudy machine.

I have given very little thought to the Z axis at this point but I am treating it as a 'module' that I'll bolt on towards the end of the project.

I think that's it for now. As I have more thoughts and ideas I'll vent them here.

Cheers,
Paul.

[Clive S]

4 qty TB6560 Drivers
24V 250W PSU

These will stunt the machine, if you can go with AM882 (not easy to find now) or https://www.ebay.co.uk/itm/2DM860H-2phase-NEMA23-NEMA34-Stepper-Motor-Driver-32bit-DSP-DC80V-1-5-6-0A-/401403682399 with a 68V power supply

[routercnc]

Agree the 24V psu and TB6560 drivers give very low performance and possibly lost steps under load.

You can hand make the large end L plates using a template and router with a carbide fluted end mill. Accuracy is not critical as the outline is mainly cosmetic and the holes can be drilled oversized. If you are fussy they can be the first things you remake with the machine.

Belt drive is often neater as the stepper can be tucked away. At 2010 you would usually 1:1 drive for wood and 2:1 ( step down) for aluminium.

Z axis should be sketched soon as the ballscrew location, ballnut, and stepper location at the end of the gantry all need to fit together. You also need to add the spindle plus a cutting bit to check the gantry is at the right height to give the range of movement you need on the Z axis plus clearance to machine over a vice.

1605 is often fine for Z but if you have bought the other size then use that.

Draw as much as you can before cutting anything because the details can catch you out !

[CaptainBarnacles]

These will stunt the machine, if you can go with AM882 (not easy to find now) or https://www.ebay.co.uk/itm/2DM860H-2phase-NEMA23-NEMA34-Stepper-Motor-Driver-32bit-DSP-DC80V-1-5-6-0A-/401403682399 with a 68V power supply

Thanks for that, I'll take a look at those drivers and an uprated PSU. I bought the motors and drivers as a kit and just assumed that they would all work well together. At the time I bought them I did have a smaller machine in mind but you know how it goes! As per your signature; as I get deeper into this I'm starting to realise just how little I know :rolleyes:

[Clive S]

I bought the motors and drivers as a kit and just assumed that they would all work well
Yes kits never work well, they just seem to put things together that don't match.

Have a look through Joe's thread https://www.youtube.com/watch?v=pZo2LVQA9UI&list=PL1FIADAKba_uTgFU5qqS3i705fuSogBXT and see his build log on here.

He also has a very good vid on a power supply build.

[CaptainBarnacles]

Agree the 24V psu and TB6560 drivers give very low performance and possibly lost steps under load.

You can hand make the large end L plates using a template and router with a carbide fluted end mill. Accuracy is not critical as the outline is mainly cosmetic and the holes can be drilled oversized. If you are fussy they can be the first things you remake with the machine.

Belt drive is often neater as the stepper can be tucked away. At 2010 you would usually 1:1 drive for wood and 2:1 ( step down) for aluminium.

Z axis should be sketched soon as the ballscrew location, ballnut, and stepper location at the end of the gantry all need to fit together. You also need to add the spindle plus a cutting bit to check the gantry is at the right height to give the range of movement you need on the Z axis plus clearance to machine over a vice.

1605 is often fine for Z but if you have bought the other size then use that.

Draw as much as you can before cutting anything because the details can catch you out !

Some good points there, many thanks.

I think I am leaning towards cutting my own plates and, like you said, if I'm not entirely happy with them I'll machine replacements when I'm up and running. I normally tear through wood based projects because I am so comfortable and familiar with the material and I know how what tolerances I have to play with. Clearly working with metals is going to be a whole different ball game and I'm going to be working much more slowly, methodically and precisely (hopefully!). I'll also need to avoid errors as the raw material cost is so much higher than most woods - being a Yorkshireman I'm going to be sqeezing every last penny :joyous:

I like the idea that the motors are tucked into the footprint of the frame as there's less chance of them getting knocked. I'm operating in a fairly confined space (my garage) so anything that saves space is a bonus.

Is it possible to have two sets of pulleys and belts and swap them over for specific materials? I imagine that I'll be cutting wood 95% of the time so could I run a set of 1:1 and then swap it out for a 2:1 when cutting acrylic and aluminium? Would I just need to change the parameters in my software to account for it?

Looks like I'll be on the Fusion 360 for a few days! I have only had a few hours experience of it so far and after many years of using Sketchup I am finding the learning curve pretty steep.

Thanks again for all the advice.

[CaptainBarnacles]

Yes kits never work well, they just seem to put things together that don't match.

Have a look through Joe's thread https://www.youtube.com/watch?v=pZo2LVQA9UI&list=PL1FIADAKba_uTgFU5qqS3i705fuSogBXT and see his build log on here.

He also has a very good vid on a power supply build.

I sat down last night and started watching Joe's YouTube videos, wow! Some brilliant info there, I learned a lot and got some great ideas about how to proceed with my own project. Thanks for that link, I thought I had watched every meaningful CNC video on YouTube, how I missed Joe's series I don't know.

[CaptainBarnacles]

These will stunt the machine, if you can go with AM882 (not easy to find now) or https://www.ebay.co.uk/itm/2DM860H-2phase-NEMA23-NEMA34-Stepper-Motor-Driver-32bit-DSP-DC80V-1-5-6-0A-/401403682399 with a 68V power supply

Just been looking into AM882 drivers. They offer three variations - AM882, AM882H, DMA882S. According to the details the AM882H has "fun" but the AM882 has "no fun" and it seems both are being replaced by the DMA882S. Whichever version I choose I am looking at about £200 for four drives. I can get four 2DM860 drives for about £150.

Is it worth spending the extra £50 or so for the AM882s over the 2DM860s? If I go for the AM882s, is it better to get the tried and tested AM882 or AM882H, OR do I get the latest DMA882S?

AM882 Drives: (https://www.aliexpress.com/item/Driver-AM882-input-24-80v-DC-output-current-1-0-8-2A-match-with-motor-NEMA/32597359605.html)
2DM860 Drives: (https://www.aliexpress.com/item/2DM860-digital-stepper-driver-86-stepper-motor-drive-voltage-AC30-80V/32739391138.html)

[Clive S]

Just been looking into AM882 drivers. They offer three variations - AM882, AM882H, DMA882S. According to the details the AM882H has "fun" but the AM882 has "no fun" and it seems both are being replaced by the DMA882S. Whichever version I choose I am looking at about £200 for four drives. I can get four 2DM860 drives for about £150.

Is it worth spending the extra £50 or so for the AM882s over the 2DM860s? If I go for the AM882s, is it better to get the tried and tested AM882 or AM882H, OR do I get the latest DMA882S?

AM882 Drives: (https://www.aliexpress.com/item/Driver-AM882-input-24-80v-DC-output-current-1-0-8-2A-match-with-motor-NEMA/32597359605.html)
2DM860 Drives: (https://www.aliexpress.com/item/2DM860-digital-stepper-driver-86-stepper-motor-drive-voltage-AC30-80V/32739391138.html)

The drive I linked to is the H version and shows : Supply voltage 30VAC (DC40V) ~ 80VAC (DC110V) this means you could just put a toroidal transformer connected to it because the drives can accept AC as well as DC

Your link shows 48-75VDC, power supply, which is a bit low for a 68V power supply.

I have used the 2DM860H several time and not had any problems

[CaptainBarnacles]

The drive I linked to is the H version and shows : Supply voltage 30VAC (DC40V) ~ 80VAC (DC110V) this means you could just put a toroidal transformer connected to it because the drives can accept AC as well as DC

Your link shows 48-75VDC, power supply, which is a bit low for a 68V power supply.

I have used the 2DM860H several time and not had any problems

Aaaahhh, I see. Thanks Clive, I'll get the 2DM860H drives that you linked to.

Cheers,
Paul.

[CaptainBarnacles]

Just before I push the button on a transformer could I just have a quick sanity check please?

Here's my thinking:

I have 4 qty 3A max steppers and I've ordered 4 qty 2DM860H drivers. I reckon that it's highly unlikely to have all four motors chugging away pulling 3A each at any given time so I've assumed 10A as the max draw. The drivers will take AC so I won't have voltage drops and capacitance calcs to take into consideration. I am looking at a toroidal transformer with 2x35V AC secondary windings in series ie 70V AC (the drivers are rated up to 80V AC). Therefor 10A * 70V = 700VA (750VA is the closest standard size).

Am I missing anything?

[Neale]

Can't argue with your arithmetic, but the usual choice is 650VA. In fact, I ran mine with a 500VA toroidal for some time, until it failed. I don't believe that the failure was due to overloading - just one of those things - as it never became particularly warm. I went for one size up because it cost very little more and fitted the space available. There are a few factors here which mean that you tend to get more volts than you would expect. One of these factors is that the transformers will often deliver 5% or so over the nominal value on the basis that it will drop to nominal at full load. 2X35V in series will give you more than you expect, and input mains voltage is also often above nominal so the output volts go up a bit there. Keep this in mind when you are looking at the max voltage input of your drivers.

[Clive S]

Plus one.

[CaptainBarnacles]

Can't argue with your arithmetic, but the usual choice is 650VA. In fact, I ran mine with a 500VA toroidal for some time, until it failed. I don't believe that the failure was due to overloading - just one of those things - as it never became particularly warm. I went for one size up because it cost very little more and fitted the space available. There are a few factors here which mean that you tend to get more volts than you would expect. One of these factors is that the transformers will often deliver 5% or so over the nominal value on the basis that it will drop to nominal at full load. 2X35V in series will give you more than you expect, and input mains voltage is also often above nominal so the output volts go up a bit there. Keep this in mind when you are looking at the max voltage input of your drivers.

Thanks for that Neale. Good point about the output voltage. I have the option of 2x35V, 2x30V or 2x33V (from different suppliers), I think I'll plump for a slightly lower voltage just to be on the safe side.

Is there anything to be gained from using 4 qty 160VA transformers over a single 625VA? My thinking is that I could assign a transformer to each driver. I can get the 4 160's for about £15 more than a single 625VA so the price difference is neither here nor there. I suppose it would be cheaper to replace a 160VA in the even of a transformer failure but I wondered if there was any performance benefit to the system?

[Neale]

Can't think of any particular advantage of 4 individual transformers, except that it enforces good wiring practice - separate connections from PSU/transformer to drivers, with no daisy-chaining from one to another. Yes, could replace one if it fails, but you need more space to mount them. Swings and roundabouts!

[CaptainBarnacles]

Can't think of any particular advantage of 4 individual transformers, except that it enforces good wiring practice - separate connections from PSU/transformer to drivers, with no daisy-chaining from one to another. Yes, could replace one if it fails, but you need more space to mount them. Swings and roundabouts!

Yeah, I like the idea that there are four completely separate channels - I imagine it will help with fault finding too. Space isn't too much of an issue, I'll probably design an enclosure to suit the kit that I have rather than try and shoehorn everything into a given space.

I've been making some (albeit slow!) progress on my design over the weekend. I am getting frustrated with Fusion so I decided to go back to the beginning and watch some tutorial videos to try and understand the basic concepts, I am still thinking in Sketchup! I guess you can teach an old dog new tricks, it just takes a bit longer :wink:

Inspired by Joe Harris' YouTube series I thought it might be fun to try and document my build in the same way so I have started recording my build in the hope that something I have to say may help others. I am not sure I can add anything to Joe's impressive videos but I reckon the more information that's out there, the better. And if nothing else it should be entertaining to watch me bumble my way through the whole project!

[AndyUK]

And if nothing else it should be entertaining to watch me bumble my way through the whole project!

Can't wait! :D

[CaptainBarnacles]

I'm back!

It's been more than a year since my last confession. In that time I've moved house, changed jobs, started new projects and generally transitioned into a new life. The best part is that I now have a huge (but very cold) workshop in which to play and more time to play in it! :beer:

I have been slowly accumulating parts and ideas for my CNC router and I'm ready to pick up my project again from where I left off. I am currently pondering my motion controller choices (AXBB-E, Acorn, old PC with LinuxCNC or other - I think I've ruled out Mach3/4), sourcing some cheap aluminium plate (seemingly no such thing exists!) and refining my design (still struggling with Fusion).

When I get something that you won't laugh at I'll post my latest design for a bit of critique and guidance.

Cheers!

[Kitwn]

When I get something that you won't laugh at I'll post my latest design for a bit of critique and guidance.
Cheers!
That's exactly why there are no detailed photographs of my machine on the forum.:smug:

Re Fusion 360: I know it's very powerful, it's free and I would benefit significantly from mastering it, but every time I try to get competent in it's use I get frustrated instead and return to my old friend CamBam.

Kit

[Neale]

Personal view - F360 is well worth struggling with, although I'm not sure where you start. It is a steep learning curve at first and, to be honest, I suspect that personal tuition from a friend who knows it is the best way to go. Once you are off the ground, then the videos start to come into play. One problem is that based on reputation the best series of tutorials available online are those from Lars Christenson but they were made a couple of years ago and the F360 user interface has changed a bit since then. Not massively - an experienced user would very quickly see and adapt to the differences - but for a beginner, things like saying "Select the Model workspace" is a bit misleading when it is now called the Design workspace. I've done a bit of tuition for local model engineering club members so I'm aware of some of the problems beginners have and I'm not sure that they are addressed very well in the online tutorials. That's all true, anyway, for the CAD and modelling side of it. CAM and gcode production is a different issue where things are a whole lot more complicated but for design work, keep bashing at F360, get a bit of help if you can from someone you can actually ask questions, and you'll get there.

You will find a few photographs of my machine in a thread here somewhere. People did laugh. Nevertheless, the machine works, and it works better than i ever expected. I'm doing a fair bit of detailed machining in steel these days, which isn't bad for a machine designed and built for working with wood. Keep at it - you' ll get there!

[CaptainBarnacles]

Personal view - F360 is well worth struggling with, although I'm not sure where you start. It is a steep learning curve at first and, to be honest, I suspect that personal tuition from a friend who knows it is the best way to go. Once you are off the ground, then the videos start to come into play. One problem is that based on reputation the best series of tutorials available online are those from Lars Christenson but they were made a couple of years ago and the F360 user interface has changed a bit since then. Not massively - an experienced user would very quickly see and adapt to the differences - but for a beginner, things like saying "Select the Model workspace" is a bit misleading when it is now called the Design workspace. I've done a bit of tuition for local model engineering club members so I'm aware of some of the problems beginners have and I'm not sure that they are addressed very well in the online tutorials. That's all true, anyway, for the CAD and modelling side of it. CAM and gcode production is a different issue where things are a whole lot more complicated but for design work, keep bashing at F360, get a bit of help if you can from someone you can actually ask questions, and you'll get there.

You will find a few photographs of my machine in a thread here somewhere. People did laugh. Nevertheless, the machine works, and it works better than i ever expected. I'm doing a fair bit of detailed machining in steel these days, which isn't bad for a machine designed and built for working with wood. Keep at it - you' ll get there!


That's exactly why there are no detailed photographs of my machine on the forum.:smug:

Re Fusion 360: I know it's very powerful, it's free and I would benefit significantly from mastering it, but every time I try to get competent in it's use I get frustrated instead and return to my old friend CamBam.

Kit

Haven't heard of CamBam, I'll take a look. I like 'free' and open source software but I have no objection to paying for good software if it gets the job done. F360 is hugely frustrating at the moment, especially so as I normally pick up new concepts pretty quickly, but I would like to persevere and reap the rewards as I've seen just how great it can be. While I'm CNCing as a hobby (for the time being) I still value my time and if I just can't get the results I'm looking for in a timely fashion I may well look into alternatives.

[CaptainBarnacles]

I have two different Y axis arrangements which I'd be grateful for some input on please. I originally designed my machine with top mounted linear rails but then thought I would give side mounted rails a try in order to lower the gantry by about an inch. The trade-off is that my bed will have to shrink by about 80mm from side to side as I've already got the gantry profiles (I know, design first, shop later - I've now got this tattooed on my forearms!).

Any opinions about the pros and cons of either would be much appreciated.

[pippin88]

On top of the side beams is fine.

Using 2 separate extrusions for the gantry is a lot weaker than one big one. If you already have them, it is worthwhile using a plate on the back to join the two gantry extrusions.

[AndyUK]

Either of those two setups will work, but unlike pippin I'd favour the second one in which the gantry beam is bolted to the same plate as the bearing blocks. In the first you have an extra joint to complicate things.

Totally agree on the joining of the gantry beams though, they should be one solid unit, preferably in an L configuration.

[CaptainBarnacles]

Either of those two setups will work, but unlike pippin I'd favour the second one in which the gantry beam is bolted to the same plate as the bearing blocks. In the first you have an extra joint to complicate things.

Totally agree on the joining of the gantry beams though, they should be one solid unit, preferably in an L configuration.

I think that I am leaning towards the second option, it does make sense that the gantry beams and the bearing blocks being fixed to the same piece should be more rigid.

Great call on joining the gantry beams together, I hadn't considered that at all. I guess a few pieces of aluminuim plate or profile offcuts would do the job.

[CaptainBarnacles]

I've been paying some attention to my X and Z Axes design this week and I'm starting to question whether I have built my gantry too high. The fully extended Z Axis looks like it could be vulnerable to excessive deflection given the amount of extension from the gantry.

27097
27098

What I realised is that most of the work I'll be doing will be on relatively thin material (12-40mm) so the router will be operating at full extension most of the time. I also realised that this must be a problem for a lot of machines; do most people raise the bed in order to minimise the deflection of the cutter? I would like to work material up to 100mm thick and I currently have about 120mm clear travel on my Z axis.

Rather than compromise my design to accomodate thick material do you think I would be better to design a machine to work thin material well and accept that I just won't be able to work thicker material. Being able to work thicker material was a bonus rather than a necessity.

[JAZZCNC]

Well, first I think you need to rethink your Z-axis design because that cannot possibly work how you have it shown now.?

Regards the Gantry height then it depends on what your cutting. All the machines I build have at least 100mm of Z travel and most have 150mm and will happily cut aluminum at full extension with a reasonable finish quality and accuracy using correct feeds n speeds. If you require a higher finish quality or deeper cuts or higher feeds then yes it's better, even necessary to have a lower machine or if not then a machine that is built much much stronger than your typical router needs.

To be honest I'm not a fan of the open gantry design you are using with the profiles orientated in the vertical position which is there weakest orientation. Those same two profiles could make a much stronger gantry if you made it an L-shape and put the screw at the rear.

[CaptainBarnacles]

Well, first I think you need to rethink your Z-axis design because that cannot possibly work how you have it shown now.?

Regards the Gantry height then it depends on what your cutting. All the machines I build have at least 100mm of Z travel and most have 150mm and will happily cut aluminum at full extension with a reasonable finish quality and accuracy using correct feeds n speeds. If you require a higher finish quality or deeper cuts or higher feeds then yes it's better, even necessary to have a lower machine or if not then a machine that is built much much stronger than your typical router needs.

To be honest I'm not a fan of the open gantry design you are using with the profiles orientated in the vertical position which is there weakest orientation. Those same two profiles could make a much stronger gantry if you made it an L-shape and put the screw at the rear.

OK, I've gone back to the drawing board with my Z Axis and had another stab at it.

27139
27140
27141

With regards to the gantry, the profiles that I have are 80x80 heavy which I intend to brace across the back with perhaps 3mm aluminium sheet or some flat bar. I had thought that would be sufficiently strong for my needs (I'm cutting wood, MDF, ply, plastics and possibly the odd bit of aluminium), do I need to beef it up further?

Is it the case that you tune the machine's feed rates to the strength and stiffness of the machine?

I intend to use timing belts to drive 2010 ballscrews on my X and Y axes. Is it worth having two sets of pulleys and belts to gear it down for cutting harder materials?

[AndyUK]

I had assumed that your previous drawings with the spindle back-plate were making it shorter so we could see past it, but your latest drawings seem to continue the theme. Are you genuinely planning to have the top of the linear rails for the Z axis hanging in mid air like that? Seems like quite a flimsy design decision for the minimal cost of some extra aluminium plate.

I'd also doubt that the ballscrew needs to be that long - but I gather you've already purchased it?

[CaptainBarnacles]

I had assumed that your previous drawings with the spindle back-plate were making it shorter so we could see past it, but your latest drawings seem to continue the theme. Are you genuinely planning to have the top of the linear rails for the Z axis hanging in mid air like that? Seems like quite a flimsy design decision for the minimal cost of some extra aluminium plate.

I'd also doubt that the ballscrew needs to be that long - but I gather you've already purchased it?

Yeah, I see what you mean about those rails being unsupported when the Z is lowered and yes, I have already got the ballscrew.

Three options then:

- Add more plate to lengthen the router mount plate.
- Consider shortening the whole linear travel by cutting the rails and ballscrew.
- Buy another ballscrew. (perhaps a 1605 would be sufficient?) Looks like I can get one for about £33 with the supports, nut bracket and coupler so that won't break the bank.

I had previously wondered about shortening the ballscrew using my ML7. I'm not a very accomplished machinist and my biggest concerns (apart from completely stuffing the job up!) were about the hardness of the ballscrew and whether the action of cutting into the thread might make the cut wander.

[phill05]

I've been paying some attention to my X and Z Axes design this week and I'm starting to question whether I have built my gantry too high. The fully extended Z Axis looks like it could be vulnerable to excessive deflection given the amount of extension from the gantry.


What I realised is that most of the work I'll be doing will be on relatively thin material (12-40mm) so the router will be operating at full extension most of the time. I also realised that this must be a problem for a lot of machines; do most people raise the bed in order to minimise the deflection of the cutter? I would like to work material up to 100mm thick and I currently have about 120mm clear travel on my Z axis.

I have 120mm clearance under my gantry and have the rails fitted to the back plate on the Y axis and bearing blocks fitted to the plate holding the spindle and travels no lower than the gantry, if you lower your spindle housing to around 30mm of bottom of the plate you have the option then to cut 100mm high or lower the spindle in the housing to cut the smaller jobs, this is how I work mine with no problem cutting ally and have cut some steel with the correct cutters.

Phill

[AndyUK]

Yeah, I see what you mean about those rails being unsupported when the Z is lowered and yes, I have already got the ballscrew.

Three options then:

- Add more plate to lengthen the router mount plate.
- Consider shortening the whole linear travel by cutting the rails and ballscrew.
- Buy another ballscrew. (perhaps a 1605 would be sufficient?) Looks like I can get one for about £33 with the supports, nut bracket and coupler so that won't break the bank.

I had previously wondered about shortening the ballscrew using my ML7. I'm not a very accomplished machinist and my biggest concerns (apart from completely stuffing the job up!) were about the hardness of the ballscrew and whether the action of cutting into the thread might make the cut wander.

Nothing wrong with the ballscrew being that length; its just that you're not going to use its entire travel, so if the ballscrew hadn't been purchased you could save some money by getting a smaller one.

I'd argue the easiest option is just to extend the plate.

Second option is to swap the rails and carriages (as I think Phil suggests) but its quite nice to be able to remove the spindle plate in one go with just four screws on the ballnut; it can be a pain to remove the other way round.

[phill05]

27142

Here you go

[CaptainBarnacles]

To be honest I'm not a fan of the open gantry design you are using with the profiles orientated in the vertical position which is there weakest orientation. Those same two profiles could make a much stronger gantry if you made it an L-shape and put the screw at the rear.

Do you reckon it would be stronger to use a single 40x160 rather than the 2 qty 80x80s (with some bracing across the back of them)?

I have 2 qty 80x80 and 2 qty 40x160. I was using the 40x160s as my Y axis rails but it occours to me that there's no reason I can't swap them out for the 80x80s and put a 40x160 on the gantry. Heck I suppose I could put 2 40x160s in the gantry, would that be overkill? The image in my head of two 40x160s bolted to make an L-shape, as you suggest, is scary big - it would be a behemoth - I like that but would the rest of the build cope with throwing all that mass around? I do also have a load of 40x80 so perhaps a 40x80 bolted to a 40x160 for a L-shaped gantry?

I'm going to take a look at it in Fusion and see how it looks.

[CaptainBarnacles]

Nothing wrong with the ballscrew being that length; its just that you're not going to use its entire travel, so if the ballscrew hadn't been purchased you could save some money by getting a smaller one.

I'd argue the easiest option is just to extend the plate.

Second option is to swap the rails and carriages (as I think Phil suggests) but its quite nice to be able to remove the spindle plate in one go with just four screws on the ballnut; it can be a pain to remove the other way round.

I did try it with the rails the other way around and I wasn't keen. I think that you're right, extending the plate is the easiest option.

[CaptainBarnacles]

I've been looking at different gantry arrangements that I can build using the material that I already have.

27203

From left to right they are:

A: 3 off 40x80
B: 2 off 80x80, 1 off 40x40
C: 2 off 80x80 with 40x40 or 40x80 bracing
D: 1 off 40x160, 1 off 40x80 'L' arrangement
E: 1 off 40x160, 1 off 40x80 'L' arrangement with lower linear rail front mounted
F: 1 off 40x160, 1 off 40x80 'L' arrangement with greater gantry clearance
G: 1 off 40x160, 1 off 40x80 'T' arrangement

My original arrangement was 2 off 80x80, one above the other without any bracing. On the feedback I've had so far I think that any of these new arrangements should outperform my original. I like A, B and C as they allow the ballscrew to sit in the gap but I know that strength and stiffness are more important. D offers good spacing between the two linear rails but I think that it may be a step too far as it limits the material thickness too much. E is my current favourite. The only thing I don't like is that the Z axis will protrude out further from the gantry but I think I can live with it.

Of these arrangements which do you think offers the best performance?

[JAZZCNC]

My original arrangement was 2 off 80x80, one above the other without any bracing. On the feedback I've had so far I think that any of these new arrangements should outperform my original.

Your orginal would been ok if the profiles where bolted together without the gap. The gap wasn't required so bolted together the profiles are stronger.



I like A, B and C as they allow the ballscrew to sit in the gap but I know that strength and stiffness are more important.

The gap would just fill up with crap and shorten the screws life. Like wise being exposed to chips is why I don't like open gantry's along with being weaker. The L shape gantry helps protect the screw as it's behind and is also stiffer.


E: 1 off 40x160, 1 off 40x80 'L' arrangement with lower linear rail front mounted

E is my current favourite. The only thing I don't like is that the Z axis will protrude out further from the gantry but I think I can live with it.

This would be my choice with the profiles you have.

[Voicecoil]

How about a variation on "C", bring the 2 80x80s close together (ditch the crap collecting gap) and plate them front and back to make say a 90 something x 160 section? You could even slip a drop or 2 of epoxy where the extrusions meet if you were feeling keen :whistle:

[AndyUK]

You could even slip a drop or 2 of epoxy where the extrusions meet if you were feeling keen :whistle:

You sir are a brave, brave man.

[Voicecoil]

He who dares wins.

[JAZZCNC]

How about a variation on "C", bring the 2 80x80s close together (ditch the crap collecting gap) and plate them front and back to make say a 90 something x 160 section? You could even slip a drop or 2 of epoxy where the extrusions meet if you were feeling keen :whistle:

Doesn't need to go that far because there's more than enough strength in profile, he'd just be adding weight. The epoxy would just be a waste of money and offer nothing in terms of strength.

If he had a little more length and could afford slightly wider bearing/gantry plates I'd suggested going with 160 flat with 160 on top and with rails in same front and top positions. You have to think about the direction of cutting forces. The wider bottom profile makes the gantry stiffer in the direction of cutting forces.

[CaptainBarnacles]

If he had a little more length and could afford slightly wider bearing/gantry plates I'd suggested going with 160 flat with 160 on top and with rails in same front and top positions. You have to think about the direction of cutting forces. The wider bottom profile makes the gantry stiffer in the direction of cutting forces.

Like this?

27206

That's certainly more beefy than I had planned but I'm in this to build the best machine (supposedly within my budget but that vanished ages ago!) I possibly can. I don't want to start cutting and immediately regret compromising on something critical.

My gantry riser plates are currently 210mm width so this arrangement would fit. My concern is that the CoG of the router would then be well in front of the forward Y-axis carriage. Is this what you mean when you said about wider bearing/gantry plates? ie. I'll need to widen them to bring the CoG back within the two Y-axis carriages? Which would then mean a reduced cutting area?

[pippin88]

C would be best IMO.

The ideal gantry profile is square.

The gantry is subject to vertical force, horizontal forces, and torsional forces.

When the Z axis is all the way down, it acts as a lever and tries to twist the gantry.

[JAZZCNC]

Like this?

My gantry riser plates are currently 210mm width so this arrangement would fit. My concern is that the CoG of the router would then be well in front of the forward Y-axis carriage. Is this what you mean when you said about wider bearing/gantry plates? ie. I'll need to widen them to bring the CoG back within the two Y-axis carriages? Which would then mean a reduced cutting area?

Yes but not so much because of COG but more the loss of travel. COG isn't a big deal as your probably thinking and you are not exactly hanging it out into outer space.
That said I'd also consider looking at 160 flat with 80 on top and rails top n bottom. You'll lose some height between bearings but gain in stiffness.
End of the day it's down to what you need in terms of clearance between gantry and bed, along with travel required. Also what you are wanting the machine to do.?

My suggestions are based on what profile you have, however, if I was building it and wanted a very strong gantry at this size I would use 80x120.

Option E will work fine as well if you cannot afford the length. I wouldn't stress over any of them because each will work fine for a router.

[CaptainBarnacles]

I've been taking stock of this project over the last couple of weeks. It occurs to me that I was getting a bit bogged down and getting a bit obsessed with building the ultimate machine first time around. I took a look at my design and looked for ways to simplify it and design a machine that I, with my limited tools and experience, could actually build in a reasonable timeframe.

A few late nights at the computer (I'm finally starting to get the hang of Fusion 360!) and I have this:

27365
27366
27367

The gantry is 2 qty 80x80 stuck together (though I may still add an additional 40x80 on the flat for more of an L shape profile), 40x160 for the Y axis frame, 40x80 for the bed with MDF over.

It's not a million miles away from some of my earlier designs but it feels more mature and I feel happier with it. I have my 150mm Z axis travel, I lost a little on the bed size but I can live with it for now. Time for less procrastination and more building (when the workshop warms up a little).

[JAZZCNC]

Looks ok to me and will work fine. Thou I've got a few suggestions which will make it a little stronger and easier to setup etc.

The height of the gantry doesn't look very high so you could change it a little and gain some height while increasing the strength and making it easier to setup.
I would make the gantry and bearing plates two separate items. Make the Bearing plate and the drop bracket for ball-screw one part. Then use another plate bolted directly to the profile from the underside and into the shortened gantry sides, this will stiffen the gantry a little. Then bolt this to the bearing plate because how you have it now will make accessing the bearing bolts difficult. But also it allows you to square up the gantry without affecting the ball-screws.

As you have it drawn now then if you need to adjust the squareness all you can do is loosen the bearing plate and rotate the gantry as a whole. This then changes the alignment of the gantry sides so they are no longer parallel to the ball-screws, this, in turn, means your ball-nut mount unless shimmed, etc puts a twist on the ball-screws that causes binding and premature wear, etc.
If you do it as described above then you can just loosen the gantry without affecting the ball-nut alignment. The extra bearing plate also adds a little height to the gantry, It's also very easy to add extra height if ever needed.!

Simple changes which will make a big difference to the machine setup.

Edit: You could also gain back some of the lost travel by making the Gantry plate a little longer than the bearing plate and offsetting the gantry back a little.!

[CaptainBarnacles]

Looks ok to me and will work fine. Thou I've got a few suggestions which will make it a little stronger and easier to setup etc.

The height of the gantry doesn't look very high so you could change it a little and gain some height while increasing the strength and making it easier to setup.
I would make the gantry and bearing plates two separate items. Make the Bearing plate and the drop bracket for ball-screw one part. Then use another plate bolted directly to the profile from the underside and into the shortened gantry sides, this will stiffen the gantry a little. Then bolt this to the bearing plate because how you have it now will make accessing the bearing bolts difficult. But also it allows you to square up the gantry without affecting the ball-screws.

As you have it drawn now then if you need to adjust the squareness all you can do is loosen the bearing plate and rotate the gantry as a whole. This then changes the alignment of the gantry sides so they are no longer parallel to the ball-screws, this, in turn, means your ball-nut mount unless shimmed, etc puts a twist on the ball-screws that causes binding and premature wear, etc.
If you do it as described above then you can just loosen the gantry without affecting the ball-nut alignment. The extra bearing plate also adds a little height to the gantry, It's also very easy to add extra height if ever needed.!

Simple changes which will make a big difference to the machine setup.

Edit: You could also gain back some of the lost travel by making the Gantry plate a little longer than the bearing plate and offsetting the gantry back a little.!

Thanks Jazz, some great suggestions. I am not sure that I understand 100%, do you mean something like this? (sorry, really crude drawings just modded from my originals so ignore hole placement etc.):

27368
27369

The red plate fixes to the gantry profile from below and the gantry plate, the gantry plate fixes to the gantry profile in the usual way (and also to the plates on the back of the gantry that carry the BF/BK blocks for the X axis ballscrew). The red plate can then be fixed to the bearing plate and if I either elongate or slightly oversize the clearance holes I can get some adjustment for squaring the gantry assembly.

Am I on the right lines with that?

[JAZZCNC]

Yes, that's exactly what I mean. Makes things a lot easier regards setting up and ball-screw alignment.
Getting the ball-screws aligned correctly makes a massive difference to performance and the life of the screws. Just a little off and you'll get binding which robs power and causes excess wear on the ball-nut.

[CaptainBarnacles]

Yes, that's exactly what I mean. Makes things a lot easier regards setting up and ball-screw alignment.
Getting the ball-screws aligned correctly makes a massive difference to performance and the life of the screws. Just a little off and you'll get binding which robs power and causes excess wear on the ball-nut.

Cheers for that Jazz :thumsup:

It's stuff like this where years of experience really show. It all looked good (to me) on my plan but the practicalities such as being able to access screws or adjust alignment etc just don't become apparent (again, to me) on a model. I would have been stuck into the build before having one of those "d'oh!" moments.

[AndyUK]

Cheers for that Jazz :thumsup:

It's stuff like this where years of experience really show. It all looked good (to me) on my plan but the practicalities such as being able to access screws or adjust alignment etc just don't become apparent (again, to me) on a model. I would have been stuck into the build before having one of those "d'oh!" moments.

One trick I used was once I had the design all sorted, I put it to one side and started the assembly from scratch, bringing in each component at a time and working out how to attach it, and zooming out to see what access I'd have at each stage.

Still doesn't totally work - but it helps a lot. There are still two bolts that I have no idea how to access under the gantry!

I wouldn't wait too long before adding bolts into your model - the heads can be large enough to change things, and then you might want to think about countersinking them. This ties in with the adjustment Jazz is referring to; there might be places like the Y axis ballnut connection where you want a bit more adjustment room. For example, here I used slots so I didn't have to be spot on on the height of the screw vs the rails. How would you adjust each connection in all three axis? How will you tighten the motor belts? etc.

It was all very overwhelming at the start, but now I'm looking back I regret not going that extra mile with the CAD and thinking of ingenious limit switch placements to hide them away from everything. Its a little irritating trying to retrofit them now. Obviously that's not a concern for right now, but I'd at least consider it before starting the build. The same goes for covering the screws and rails somehow - I'd prefer my build if they were protected from chips, and lubrication pipes to make everything easy to lube when its in place.

[Kitwn]

You may like to mock up some parts of the construction in ply or MDF before committing to expensive aly plate. This works especially well for complex structures like the Z axis and is easy to do with hand tools only. In fact I spent yesterday finally building my Z axis in aluminium plate after running it as a plywood 'mockup' for the last 18 months.

I'm also changing from microswitches to inductive sensors for limit switches and they're going to resemble sore thumbs, whether I like it or not. Andy's right about working out these details in advance if you can.

[JAZZCNC]

Cheers for that Jazz :thumsup:

It's stuff like this where years of experience really show. It all looked good (to me) on my plan but the practicalities such as being able to access screws or adjust alignment etc just don't become apparent (again, to me) on a model. I would have been stuck into the build before having one of those "d'oh!" moments.

It's common for the little details to get overlooked and it's the little details that can make a big difference to how the machine performs and ease of maintenance etc.

I model all the major mechanical components individually in CAD right down to the exact mm and with every hole for bolts etc in their correct locations. Then I build sub-assemblies with the components ie: Z-axis, Gantry, Ballscrews with end bearings and ball-nut mount, this lets me check for clashes with bolts, etc.
Then finally I build a full working assembly with the sub-assemblies all mated together accurately. This way I can check every detail and catch if any interference or if access is restricted etc. You'll be surprised just how easy it is to miss quite obvious details that would be major problem after parts are made.

My personal opinion is that making mock-up parts from MDF etc is a waste of time and provided you are accurate with the cad models then it's not required and defeats the purpose of CAD. Every machine I build is done solely in CAD with each part machined straight from the model and they fit together perfectly every time. (Unless I screw up the machining which never happens.!. .Honest.!!)

However for those who don't have good CAD skills or use lower-level packages that don't allow mating and assemblies etc then I agree it's a good and safe way to work.

The best advice I can give is to build in as much adjustment as possible and pay close attention to bolt hole locations/access to prevent clashes. Think about the little details right at the start like Limit switch locations, Energy chain mounting and paths, Cable routing and access to grease nipples, etc.

Break the machine down into segments and think about how each segment interacts with the next. ie: Z-axis and how it mounts to moving axis on the gantry and how the Energy chain runs along gantry and mounts to moving access does it clash with ball-screws or motors or limit SW etc when moving.?
These little details can be a right ball-ache if left until the machine is built only to find out there is no way to avoid clashing. Often the difference between something fitting or not is only mm's and can completely screw the job up or cause major part remake or big compromises on design.

The extra time in CAD is well worth the time and frustration it can save, not to mention expense if screw up big.!

[Kitwn]

However for those who don't have good CAD skills or use lower-level packages that don't allow mating and assemblies etc then I agree it's a good and safe way to work.


This probably applies to most of your readers.

MDF can be the old man's CAD. Perhaps we need a new acronym? MAD..... Model Aided Design:excitement:

Either method will help you avoid creating a TURD... Totally Unbuildable Router Design:hysterical:

[CaptainBarnacles]

Some really amazing advice and words of wisdom, thank you so much guys.

I had a go at mocking up a couple of components in ply and it was useful to help visualise assemblies before commiting to cutting aluminium. That was a while back and since then my CAD skills have moved on a bit so I am more inclined to model it in CAD now than in MDF or ply (although sometimes it would definately be quicker to just go and knock out the part in ply!).

I will certainly be using CAD with a different midset going forward. Back when my CAD skills were very limited I was using it as a way to visualise the overall machine and give me an idea of what the finished thing would look like. From the advice here I can see now how it's real power is the ability to precisely model the real world and to resolve issues and conflicts before I go anywhere near a power tool.

Looks like I'll be spending this rainy weekend in front of the computer and trying to lock down the design as much I can. I'm really itching to get started with the build but I now appreciate how much time and heartache can be saved by spending the time on the design (although it is starting to annoy Mrs Barnacles - every time she walks past I'm staring intensely at a virtual CNC router. I told her it would only take a few weeks to build, that was over two years ago!).