[NIC 77]

NIC - One disadvantage of this design is I'll likely lose maybe 50-100mm in x-direction. I guess that's just one disadvantage of this solid design?

That's true. But on the other side you have the motor sticking out, so it doesn't make as much of a difference to the overall size on that side.

If you look at what I drew up it's similar to ThomConcept but with the addition of the chip guard. I am more of a fan of the high rail design as opposed to tall gantry risers. ThomComcept's design looks like it has tall gantry risers, but it doesn't, it's a high rail design, he just spaced his gantry bearings very far apart.

I really think that it's a good idea to look at what has been proven to work, in particular because you said that you want to cut mostly aluminum. Can you find some examples of designs that cut aluminum well and have really tall gantry risers like yours where the ballscrew is under the machine? It's not a rhetorical question...I don't know, have you found any?

I personally don't like anything driven from under the sides of the machine because it prevents the machine from being bolted down along the sides, which increases the rigidity on the long axis.

Also there is nothing preventing you from doing a dual drive under the machine if that is what you want to do. It's not what I would do, but it's not my choice. Ultimately, all anyone here can do is offer an opinion. Sometimes those opinions are good, and sometimes they are not.

Since it is so important to you to keep the footprint as small as possible....I suggest starting to draw in things like your drag chains, etc to see how this effects the size of your machine.

Designing a machine from scratch takes so much work and rework in the CAD files to get it right. That's why I've been saying to hold off on buying any parts until everything is finalized the way you want it. It's not a time to rush things.

You should check out the Avid CNC benchtop PRO design. It looks like they have made it to take up as small of a footprint as possible for the amount of travel. Ballscrews driven on the sides, as opposed to from underneath, and that would be my choice. I will have more to say on this later.

https://www.youtube.com/watch?v=ZUNpi0lTg5c

I have a bunch more things to say, but that will come over a few more posts. Hopefully my comments are helpful, and not too annoying. I'm certainly not trying to come off as a know it all, because I certainly don't know it all.

CAD CAD and more CAD
Order material
Shop for parts
Wait for parts
Hmm... need more tools,,,shop for tools
Cut material
Oops didn't think of that.... back to CAD
Damn customs lost my parcel.... shop for parts
Don't have that tap....shop for tools
Can't find that fastener local....wait for fasteners in mail
What ? I need to understand electricity ? I'm doomed !

Or at least that describes my projects

That describes mine as well. So true. I added the bold to CAD.

This is an interesting build as well, but has a moving table design.

https://www.youtube.com/watch?v=WvOhgM5uBF4

[peteeng]

Hi Redding - Good time spent is finding a commercial machine that does what you want. A considerable amount of time has gone into it to get it to that development point. So this gives you a real point of reference. If it does what you want then copy it saves 100's of hrs of work or buy it...

Designing from scratch there are two paths 1) Design from the envisaged footprint inward and you usual end up with less working envelope to what you wanted 2) Design from the Z axis out and you get the envelope you want and the minimum size footprint to achieve this. Footprint and envelope usually conflict. Spend lots of time in CAD at a "rough" level then jump to a detail level when all the stars line up. Maybe 400 hrs plus into the work...

High rail designs are considerably stiffer then column designs....Peter

[v2tiller]

I agree with everyone that if you have a definite work area size you need then design is critical.
When i built my cnc i had in mind a 500 x 500mm cut area. I was on a budget as most people and was building using both new and used components.
I ended up with 400 x 320 cut area by the time i had finished.
My machine dimensions are 1100 x 750 and its a moving table type.
Please keep this in mind as its something i learned at the end of my build but wish i knew earlier, especially because you want to cut aluminum.
Most diy builds you won't see any way covers of any sort.
You want protection for your rails and screws.
So design them in at the beginning.
I added mine last and lost about 100mm travel.
You can get them custom made in china for a good price just be aware that you need to add roughly 50-100mm each side of your axis rails so they will fit or you will lose travel.

[v2tiller]

Here's a pic of my cnc.
I got the bellows made in china and they are pretty good so far.
If you send them the dimensions you want they will let you know how much space the bellows take up when compressed. This is where you will lose travel if your rails aren't long enough.

Attachment 446372

[NIC 77]

V2tiller,

Nice looking machine.

You bring up some good points.

One thing that I was thinking about today and have been meaning to post about is that I would do things totally different from the OP, based on mistakes that I have made in the past.

Here's what I mean......

I am currently collecting parts and finalizing the design of a new machine build. It has started out as an industrial 3d printer, but I am keenly aware that it is so overbuilt that it would make a very handy general purpose machine.

I have so far collected 3 surplus linear stages and am looking for a fourth one that fits my needs for a fully enclosed dual column moving table design. Like you and redding I am fighting with myself over the final size of the machine.

The linear stages all have covers to keep out contaminants, and the quality of the components is simply great. And I realized, why in the heck am I proposing designs, that I wouldn't do myself on a machine of this size?

redding, you should consider a 180 on this design...all you need are 4 good linear stages and you could make something really special with way less effort. Even things like play in the fixed end bearing supports or backlash in the ballscrews.....if you can find the right stuff, those won't be a problem. Yes, you still bolt some aluminum on there to stiffen things up.

Unless you were going to go all out, like Linux_Fan, or ThomConcept, and make something filled with Epoxy, IMO, it's a good option.

If you look at the Avid CNC Benchtop Pro, it's basically 4 linear stages bolted together.

3 of these

https://www.ebay.ca/itm/ROBOSTAR-Use.../112918527136?

and 1 of these

https://www.ebay.ca/itm/LPK-Used-LP1.../112583395364?

and you've got a machine.

Just watch out for that seller, as he only ships UPS, and in Canada at least, you get hit with some hefty shipping fees at the door.

Other options

https://www.ebay.ca/itm/2-400W-AC-Se.../191819112105?

https://www.ebay.ca/itm/Used-Linear-.../200996916404?

https://www.ebay.ca/itm/Cartesian-Ro.../190989236225?

https://www.ebay.ca/itm/ROBOSTAR-Use.../112850931167?

You get the idea.

[NIC 77]

This would be very very useful - I'd love to learn how todo do this also. Yes I have got these NEMA34 Motors "86HS11860A4" driven by DMA860 controllers (same as MA860).

Ball Screw Length (between Floating/Fixed ends - inner distance):
Y: 628mm (700mm total length)
X: 528mm (600mm total length)
Z: 185mm (250mm total length)
Rotor Inertia: 1800g/cm^2 which I converted to 0.00000018 kg/m^2
RPM v Torque: Attached - Similar motor (in RPM vs Nm @ 48V), but also attached this motor PPS with 60VAC. I think just refer to the similar motor graph will be fine.

OK....that's 0.00018 Kg*m^2 no worries.

I think you got the X and Y reversed? I'm only interested in the total length of the ballscrews for this calculation....so I will use 700mm x 2motors for the long axis and 600mm x 1motor for the gantry.

The motors you posted, there are definitely two different motors there, One weighs 3.8 Kg, and the CNCRP 960 weighs 3Kg. Both seem to have about the same inductance of around 2 mH (CNCRP 960 of 2mH and 86HS11860A4 of 2.1 mH). One has a body length of 100mm and the other of 118mm. More holding torque on the ones you bought, but more torque at low RPM can't help you.

Looking at the torque chart for the 86HS11860A4 @ 60V, I am assuming that the pulse per second is the same as steps per second with no microstepping. So at 4000 PPS, that would be 4000 / (200 Steps per rev) = 20 Revs per second or 20x60 = 1200 RPM. That's about 1.75 Nm of torque on that chart, while the Avid CNC 960 Oz In (I call it that, because it's the Nema 34 they use and that's where the chart came from) is at 1Nm at 1200 RPM and 48V.

Yeah, it's probably fair to use the second graph at 48V as a comparison, and I assume the 2000 pulse in that chart means that it was tested at 10x microstepping. Good news, because I have several of these motors and I've already inputted the torque chart.

It's not really going to matter though. I can tell you right now that for a dual driven gantry using these motors, a gantry weight of several hundred pounds is no big deal with a lead of 10mm. We'll see if your Y axis can keep up at 5mm lead, and what the effects of weight and lead have on performance when I do up some graphs.

Once I've done up those graphs, I'll take a back seat for a while and stop crowding your thread.

[v2tiller]

Thank Nic.
You bring up a good idea with the linear stages / slides for redding
My z axis is a linear stage i got on ebay.
It has 190mm travel, full machined from a 50mm billet of ally, 15mm nsk rails, nsk ground ballscrew, remote grease ports came with the bellows and is way more solid and accurate than i could ever build myself.
Cost me about 200us with stepprt motor.
Its definitely worth looking at these as it can save a lot of time and money.
Here's a pic of inside my z axis.
Attachment 446398

[redding]

Thanks very much for the feedback so far everyone. I have taken your points on board and thinking about leaning towards this design as attached. I have taken note of the Z-axis design (rails on spindle side), went for a fixed/stationary bed approach, added ribs to the side gantry supports with rail and ball screw right next to each other. I will now add some covering and guards to protect everything. The X-axis two motors sit 5mm below the entire CNC, I couldn't think of a way to fix this without going to 80160 - just going to ensure the table I make for it allows for this (or just motors hang off the edge).

What do you all think?



Key things below have been mentioned and here are my thoughts on it (both are great ideas):

Linear Slides
- I haven't been able to find any in Australia which appear to be as cheap as new rails that I can get at the moment. I'll keep looking, and if find a good one, perhaps it can replace the Z-Axis

Gantry Sides resting ontop of Bearings
- I designed and tried this, I lost way too much side travel to justify the advantages of this change.

Z-Axis
- I am able to buy a cheap right angle bracket (right angle, 6mm steel with welded tabs). I will use this at first, just to get running. I will then make some kind of block out of aluminum to support the stepper here.

[cyclestart]

There are all kinds of ready-made mounts available. Included is a photo of one type. If they fit or can be made to fit it's a permanent solution.
(sorry, too lazy to shrink the photo)
Attachment 446422
If your CAD software does motion it's a good idea to check for interference on every design change, even small changes.
If you're sourcing from China take advantage of the services offered. For example the seller I use for ball screws will end machine them to any drawing sent to them.

[NIC 77]

I did up those graphs for X and Y.

I used a motor armature of 0.00018 Kg*m^2

I used a ballscrew inertia of 0.000000123 (Kg*m^2) / mm of length. This number comes from a catalog I have from THK for their rolled 20mm diameter ballscrews, so it should be close to whatever you get.

I didn't bother trying to guess at a coupler inertia, I just added an extra 50mm to the screw lengths.

750mm x 0.000000123 (Kg*m^2) / mm = 0.000092 Kg*m^2. So you can see that on a 20mm diameter screw, the motor armature has just a bit more rotational inertia than the screw. You'd also be fine with going to 25mm diameter screws.

I used a ballscrew efficiency of 0.9, and things like preload torque and bearing friction are not included.

It's simply a theoretical comparison of performance based on the torque chart for the CNC router parts 960 oz-in motor graph at 48V considering only the aspects of rotational inertia for ballscrews.

In practice, you may have to tone things down a bit from the graphs to not loose steps.

For comparison, in the video from ThomConcept (post #2), he is running at 0.5G and 590 IPM (15m/min) rapids. You can see how lightening fast the rapid moves are in the video. With the servos he's using he can go up to 1G, but he toned it down.

In the video from Berend Lucas van der Wide (post #14) he's using

2600mm/min feed for adaptive clearing
1500mm/min for finishing-pass
Rapids and acceleration is tamed down here (12m/min 0,25G)
Tested earlier today with 0,9G and 20m/min, but that's just scary

Those are both servo driven machines, but it gives you some numbers to put to the videos.

I'm guessing you would be fine with a maximum of 200 IPM and 0.2G to do fast light cuts with a 6mm single flute bit on your machine to get a decent material removal rate, but you can see how the 10mm lead is better, and for good accelerations, 20mm. You shouldn't need very much cutting force at all. I even threw in the 32mm lead because a couple of the linear stages I posted links to are 32mm lead.

https://www.cnczone.com/forums/linea...118-cnc-2.html

Post #21 tells you what equations were used. For anyone curious, I did modify this slightly to include the efficiency of the ballscrew when considering it's rotary inertia.
So for this example Jballscrew = 0.000092 x (1/0.9) is the equation I used. All of the other equations are there if you want to give it a try. If you think I've made any mistakes, which is possible, please come back with a worked example.

Attachment 446432
Attachment 446434
Attachment 446436
Attachment 446438
Attachment 446440
Attachment 446442
Attachment 446444
Attachment 446446

[NIC 77]

The X-axis two motors sit 5mm below the entire CNC, I couldn't think of a way to fix this without going to 80160 - just going to ensure the table I make for it allows for this (or just motors hang off the edge).

What do you all think?

If you go for the 80160 (or more) on the sides, you could make your gantry risers less tall and still have the same travel. What is the cost of the plate compared to the extrusion?

Have you thought about putting a spoilboard on your bed?

The plate for your Y axis...you might want to make it a bit taller...you will need to connect it to your drag chain somehow.

Some ribs on the actual Z plate will make it stiffer

I like the changes you've made when compared with the start design.

[peteeng]

Hi Nik - good work - are they imperial or metric G's? Peter

[NIC 77]

Hi Nik - good work - are they imperial or metric G's? Peter

You almost got me there . G's don't have units other than "G", lol. Now I am second guessing myself. Honestly, I haven't modified or looked too closely at my spreadsheets since I revamped them a few years ago.

It's NIC, there's no "k". It comes from a Saturday Night Live skit starring the Rock where he plays the character "Nicotrel". It was the callsign given to me by my peers a long time ago.

The big thing I take away from the charts is that more weight, if it can make things stiffer, or more damp, is good, and if you want better speeds and acceleration, increase the lead of the screw....(for a stepper at least).

One of many things that was not considered is dynamic rigidity and how this is affected by increased weight at a CofG that is offset from the bearing placement under higher accelerations.

There are just so many variables for an entire machine design, it's easy to understand why commercial vmc's do big cast iron castings with powerful servos and large linear bearings.

redding,

One last thing I thought about, you were talking about possibly using taller extrusions on the sides....so if you look at the video I posted about the Avid benchtop machine (post #40) starting at 2:30 to 3:20, it mentions being able to set the table at three different heights.

If you did go with taller sides you would be able to modify your machine later if you wanted to move the table lower to add a vise or if you realized that you only wanted to cut out plate and move it higher. The less your Z axis moves, the more rigid it is and the harder you can push it. Just a thought.

[peteeng]

Hi Redding - To throw a spanner in the works about configuration. Since your in the early design phase now's a good time for an alternative. Your building a std 3 axis moving gantry config. The major design hurdles with this config are: 1) The saddle is a complex very tight space and usually pushes the Z axis way out in front of the gantry so there is a big eccentric moment from these conditions. Also the wide saddle due to the bearing spacing limits your transverse gantry motion. 2) The Z axis being cantilevered twice (out and) now downwards creates even more wobble factor. I have spent years grappling with this to make it smaller or remove the saddle with no success. This all contributes to poor machine stiffness...Until recently;

I have come across lifting gantry designs. There are a couple in the forum if you look but I came across them in very large mills.

https://www.youtube.com/watch?time_c...ature=emb_logo

https://www.waldrichsiegen.com/news/...comes-to-life/

1) The saddle is removed so the stickout of the Z axis is minimal and you can have tightly spaced bearings 2) the Z axis is always the same set depth to the gantry so can be short and stiff 3) The columns can be elephant legs with no conflict with other areas.

You still use the same bits just in different places. You do need a "bridge" across the top of the columns but there is one lifting gantry in the forum that does not use a bridge. All my machines going forward will be this style. I think it covers all the right bases. I'm building a conventional router at the moment and when that's finished I'm building a router and a mill based on the lifting gantry config. By the way they call this axis the W axis. Have a think gives you high Z if you want with no drawbacks I think (other than its different, but it may become the norm) .. Keep Making Peter

[peteeng]

Hi Redding - I'm not proposing you have a Z axis and a W axis by the way. The lifting column would be the Z axis. Cheers Peter

[cyclestart]

peteeng
Mechmate places the router inside the gantry. Those are large machines so the wide X bearing spacing isn't the same compromise as it would be on a smaller machine.
Attachment 446486

[NIC 77]

I'm currently working on my own dual column machine...

I see some potential problems for a moving gantry (in the X direction) configuration from a DIY perspective and I'm of the opinion that they are best suited for a moving table design.

For the moving gantry design, optimally you might need an extra linear rail on each side for the X axis, the bearings would need to be spaced further apart along the X, the structure would need to be very stiff, and counterbalancing the gantry could be more problematic.

For the moving table design the structure supporting each Z axis can be made as an interconnected box, eliminating the cantilever aspect from the two tall Z axis altogether, giving more stiffness, and it's far easier to do simple counterbalancing using counterweights for the gantry. You could also use elephant legs.

I see these machines as being very useful for applications where you may incorporate a rotary axis or even a 5 axis machine. Or where you have a substantially varied stock size. For example, I want to drill some holes in some 7" x 7" steel tubing on my small mill. I had to buy some special stubby bits so I don't run out of clearance.

The OP wants to cut mostly flat plate and wants as absolutely small of a footprint as possible, and it appears that this build is also on a tight budget. Those 3 things go against the dual column concept.

Have you found any dual column DIY projects that have been proven to cut aluminum really well that he could use as a template for his design?

I'm not saying that he shouldn't consider it as a concept or even explore the idea with some CAD work. It does add far more complexity to the design.

I think for what he wants to do that he would be better off doing a smaller copy of Linux_Fan's work if he was willing to spend a bit more time, money, and accept a larger size or smaller cutting area. But that's just my opinion.

I absolutely agree that the dual column design has a ton of advantages, that I wish I had noticed sooner, and I could see it becoming very popular as a replacement in the DIY community for those who might normally build a single column mill from scratch, or as mentioned, something with rotary or 5 axis design concepts.

[peteeng]

Hi Cyclestart - Shopbot started out with a twin horizontal gantry as well but soon went to a conventional layout. Twin gantries whether vertical or horizontal still have a saddle and have the structural issue of when loaded which gantry becomes dominant? So you end up with one side doing the work then in a change of direction the other side does the work. Twins are used because they are smaller but because of the issue mentioned this means its undersized and indeterminate in structural nature. I haven't formally analyzed a horizontal twin gantry design but it does not solve enough issues to investigate it in my mind. The lifting gantry however solves many issues so is worthwhile to pursue.

Hi NiC - Most routers are "dual column" or "bridged" I do not see the lifting column design to be any more complex then a std moving gantry design. Dual columns are used to improve the working envelope over single column designs. They are quite common as mills. Datron, Hass, Mori Seiki and many others have them. Moving table with fixed gantry solves many problems but still has limits. Yes it is up to Redding to shift thru all of this stuff. In the wash up its his machine...Keep at it Redding....Peter

[redding]

I did up those graphs for X and Y.

Thanks so much NIC for these graphs. It tells me indeed running 10mm lead is critical. I will certainly go dual 10mm pitches for X Axis. The Y-Axis, since I will need to upgrade to 10mm later for sure!

If you go for the 80160 (or more) on the sides, you could make your gantry risers less tall and still have the same travel. What is the cost of the plate compared to the extrusion?

Have you thought about putting a spoilboard on your bed?

Indeed it will fit 20mm spoil board MDF. Using a 50mm length end bit, it can sink approx 10mm into the 20mm spoil board. This is just in case I want to do an aluminum plate, using a shorter endmill (incase 50mm is too long).

The plate for your Y axis...you might want to make it a bit taller...you will need to connect it to your drag chain somehow.
Some ribs on the actual Z plate will make it stiffer

Good ideas - I will add these, and add some L bracket

If you go for the 80160 (or more) on the sides, you could make your gantry risers less tall and still have the same travel. What is the cost of the plate compared to the extrusion?

I actually started this design before the previous, but quickly found the z-axis plates would hit the 80160 if doing aluminun plate on a 20mm spoilboard. Or do you think there's a better way around that?

I was also worried that the top of the 80160 may not be stiff enough as all of the weight is at the unsupported top section of the 80160, but I guess some ribs could improve that.

Hi Redding - To throw a spanner in the works about configuration.

Thanks for this idea - I do 100% agree it's a very stable and nice design, however I think like NIC said I do prefer a smaller footprint right now. If I am successful with design, perhaps I will certainly look into it (and have a machine to make parts for it!). Thanks for the idea anyway!

[NIC 77]

peTer,

I am so very annoyed with you right now, and I'm asking myself if it's a fair thing to feel or if I'm being overly sensitive?

You're going to make a DIY "dual lifting column rising gantry" router where the gantry moves in both Z and X and isn't "any more complex then a std moving gantry design"? And it's going to do a really good job at cutting aluminum? That's what I took away from your last post.

And redding will be your test subject to see if this works? Is that what we're going for?

It seemed pretty obvious to me in my last post that when I described a "dual column" machine that I was talking about a "dual lifting column rising gantry design". But you ignored that and talked some BS anyway.

For someone who wants to cut mostly flat plate, what you propose is just a bad idea, I don't mind saying so. Please don't talk down to me. You are not the only one on these forums with an Engineering degree. It means little. You don't have to always be right. I know that I'm not, and I'm not afraid to admit when that's the case.