A new hope: Epoxy granite and steel.

[hardenum]

Hey,

Preface.

I've done a fair bit of research over the last few months and I decided to open a build thread. There's a bunch of things I haven't read about yet, hopefully I'll have a clearer picture of everything in the coming months.

Materials.
I'm gonna limit myself to 2 materials. Epoxy granite and toolsteel sheet up to 30mm thick.

Structural.
There are 2 options I'm considering:
A. Epoxy granite bed(Y-axis) and steel welded gantry(X-axis). Z-axis steel.
B. Epoxy granite bed(Y-axis) and epoxy granite gantry(X-axis). Z-axis steel. I'd prefer this option due to the free
form advantage of casting epoxy granite.

Motors.
Lichuan's A4ST 80's.

Linear Rails.
Hoping to snipe some mint Rexroth roller rails from ebay. I've seen a bunch of them come and go over the last few months so there seems to be a somewhat steady supply out there.
Size 35 for Y, 25 for X and Z

Belt driven ballscrews.
Hiwin 2510 on x_axis and y_axis(2x), Isel 2510 on z. For backlash 2x screw nuts on each ballscrew. 25mm Gates GT3 5M belt. 18 teeth on the motor and 36 teeth on the ballscrew,

Spindle.(max. 2x380V)
BT30 with a custom 'hydraulic' drawbar pushed be an axial pump driven by a 40/60 servo, mounted to the side.
driven by a servo size 130/140. RPM up to 5500.

alternative Teknomotor C60/67-D-DB-P-ER32. It's a 6000-9000RPM 3.3kW spindle for EUR 1000,00 https://www.damencnc.com/userdata/artikelen/c60-67-d-db-p-er32-3-0kw-6000-9000rpm-124513-en-G.jpg


EDIT 2.

Some pics as promised.

Spindle

BT30 cartridge.

Thanks!

[jarjar]

Hi,
That sounds like a heavy build!
But you are missing some crucial informations: what would be the working area and what do you plan to make with it??



- There are also holes on the side of the carriages, is that so that I can move the lubrication port from the front to the sides?

Yes but you have to pierce a hole with a needle and tap the thread.




I'm thinking of a BT30 and a size 130 servo(180 might be too heavy for a moving gantry), something at the range 12-24Nm and able to get to 9000RPM. I got a google problem with this though, I'm not getting any matches for "spindle servo" or even "spindle cartridge", the results almost always show up regular router spindles.

Check out CTB servo in China. They make nice spindle servos, but they are quite bulky.




Alignment:
- aligning ballscrew with the rails, I know how to do rail-to-rail alignment, but how do I do it for the ballscrew ?, the old fashioned way of moving the carriage and then tightening the ballscrew mount? Is there a more foolproof way of doing it? How do I make sure there are no forces acting perpendicularily on the ballscrew?

https://www.youtube.com/watch?v=mLCHT2ywk5A

[JAZZCNC]

Motors.
Lichuan's A4ST 60's or 80's. Probably 60's due to lower inductance.

These are AC Servo's not DC steppers so inductance doesn't play such a large part in the performance as your using mains AC voltages so pushing volts/current through them isn't a problem, it's the rotor inertia you are more concerned with when it comes to Servo's.

[hardenum]

These are AC Servo's not DC steppers so inductance doesn't play such a large part in the performance as your using mains AC voltages so pushing volts/current through them isn't a problem, it's the rotor inertia you are more concerned with when it comes to Servo's.
Duly noted.


That sounds like a heavy build!
But you are missing some crucial informations: what would be the working area and what do you plan to make with it??
Target build volume 320(Y)640(X)180(Z). Primarily aluminium(AW-6060 / AW-6082 / AW-2007/ AW-1050A/ AW-5754/ AW-5083/ AW-7075), occasionally mild steel. Mostly sheet metal, sometimes a mold for silicone injection(low pressure).


Check out CTB servo in China. They make nice spindle servos, but they are quite bulky.

Thanks for the tip, I found one that I like. https://ctbservo.com/product/servo-spindle-motor-2-2kw-14-0n-m-1500-8000rpm-a3%ef%bc%9a175x175x360-mm/

I'm really not an EE guy, despite many attempts. What's the deal with Rated Speed and Max Speed.

Does this mean I can run any servo over it's Rated Speed or are "spindle motors" a different motor type that allows it?
ie. synchronous for positioning, asynchronous for spindle? Or can I take a eg. an oversized servo and use it to drive the spindle cartridge?

[JAZZCNC]

Does this mean I can run any servo over it's Rated Speed or are "spindle motors" a different motor type that allows it?
ie. synchronous for positioning, asynchronous for spindle?

Both types of motor can be run above their rated speed however the difference is how long for.? An Axis servo can typically be run 100-200% above rated speed and torque but only for a few seconds. A spindle Servo motor is nearly always running above its rated speed but with a loss in torque.

How does it work.?
Now, I'm no Motor expert so don't quote me, but I think it comes down to the Motor type and how it's constructed.

An axis servo motor type is (synchronous) and uses permanent magnets to drive the Rotor, this is why it can give constant torque through its speed range.

The Spindle Servo type is (Asynchronous) which has no permanent magnets and relies on the rotating field to produce a magnetic field. This means it can run at higher than rated speeds constantly but at the cost of lower torque than at rated speeds.

There are other differences in how they work regards the drive. An Axis servo is typically run using POSITION mode whereas a Servo Spindle is run in SPEED Mode. (There is also TORQUE mode but POSITION and SPEED are the most commonly used.)

This is why you can rotate the shaft of the Servo Spindle motor when powered up but you cannot spin the shaft of an Axis servo motor. The Axis servo drive is constantly monitoring the rotor position and applying corrections to keep it POSITION, if you turn it by hand too far it will fault the drive.
Now, if you ran the Axis servo in Torque mode then this is similar to POSITION but now the drive monitors the TORQUE and tries to maintain it and if the torque falls below a set range it will fault.

SPEED mode you can probably guess by now.! The rotor position isn't monitored at all and is free to spin without any faults in the drive but the Speed is monitored and the drive will try to maintain this speed and again if it falls below a preset threshold the drive will go into a fault.

Now Servo spindles can position the rotor shaft for things like tool changing, they do this by switching into POSITION mode, and using the encoder they can rotate the shaft to an angular position.

So in Recap, Axis Servo can run above rated speed/torque for limited time periods without loss of torque. Spindle Servo, mostly, constantly run above rated but with some loss in torque as speed increases. This due to motor construction, (synchronous) and (Asynchronous)

Hope this helps.

[m_c]

There generally isn't that much difference between an axis servo motor, and a spindle servo motor. The main difference is in how they're utilised.

A servo motor rated speed/torque is what it should be able to happily produce 100% off the time without any overheating problems.
In terms of an axis, you're never likely to sustain using 100% of that continually, so you can intermittently drive the servo harder, which is where the instantaneous torque figures come in.

The rated speed is where the laws of physics really kicks in.
Regardless of being AC or DC, motor torque is directly proportional to current. More current = more torque.
In the case of DC, motor speed is proportional to voltage. More voltage = more speed.
In the case of AC, motor speed is proportional to frequency. Higher frequency = more speed.

As a motors speed increases, the back emf (electromotive force) increases. Think of emf as if the motor was acting like a dynamo/alternator being spun, in that it produces power, aka emf. The faster the motor spins, the more emf it produces. Now this emf fights 'back' against the voltage being applied to the motor, so the faster you spin the motor, the more voltage you need to maintain a given current through the motor. This applies to both DC and AC motors.
At a motors rated speed, this is the point the rated supply voltage can still overcome the back emf to provide the rated current.

Above this point, with an AC servo, as speed is not reliant on voltage, you can continue to increase the frequency, but you then start to lose torque, as you don't have the voltage to force enough current through the motor.
The motor essentially goes from being a constant torque source, with power proportionally increasing as motor speed increases to the rated speed, to a constant power source, with torque proportionally dropping as speed increases.

This applies to pretty much all electric motors.

In terms of a standard servo, and a spindle servo. The main difference is likely to be the spindle servo is derated, with better cooling to reduce the risk of overheating. There is nothing stopping you from using a standard servo, and running it above it's rated speed, to drive a spindle. Most servo manufacturers will list a rated speed, but they'll also produce speed/torque graphs that show motors performance above the rated speed, up to the speed they deem possible to run the servo.


In terms of drives.
Torque mode is the most basic, and gives the most responsive control over the servo. The big downside is torque mode is inherently unstable. Anytime the load on the servo changes, the servo speed also changes, so the controller has to be very responsive and tuned very well to maintain position.

Speed mode gives almost as much control over the servo, but due to the additional filtering involved, it's not quite as responsive. Speed mode was the mode any old school DC servo with a tacho fitted used, as at the time, controllers just weren't responsive enough to use torque mode. The tacho and servo drive (well technically amplifier) essentially acted as a damper to reduce the responsiveness needed from the controller to maintain position.

Position mode sits on top of speed and torque modes. It's basically a closed loop controller within the servo drive.

How any servo and drive responds, is entirely down to settings.
I'll just correct Jazz on torque mode. If you set a servo to torque mode, and command it to produce say 50% torque. With no load it'll spin up to it's rated speed. Stall it, it'll sit and produce 50% torque against that stall. Load it so it spins the opposite way, and it'll still happily sit producing that 50% torque against how it's being spun.
The servo drive doesn't care how the motor is turning/being turned, as long as it can produce the request torque without triggering a fault, it'll continue to do so.

Servo tuning does vary between axis and spindle though.
In an axis you want positional accuracy, so they get tuned to hold that position as best as possible. The downside is you end up working the motor harder, as it continually changes output trying to hold position, and you'll often find what works well at low speed/stationary, doesn't work so well at high speed, so you need to compromise to minimise following error under all conditions.
With a spindle, you're not as concerned about position, so you can tune them to be far more sluggish to respond to changes. A spindle being a few turns from position at 3000rpm isn't a major issue, as long as the speed remains constant. The last thing you want is a spindle that surges when it encounters a sudden load change, so you'll generally tune them to be more sluggish to respond.
As somebody asked me when tuning the servo spindle on my little mill, why do you need a spindle that only has a 20 count position error?
Even when rigid tapping, all you need is a stable spindle that you can reliably stop, it doesn't have to be on position to achieve that. It's up to the Z axis is ensure the thread is where it should be.

Although I still love playing with the servo spindle on my lathe, but it highlights how sluggish it's tuned. Put it in C-axis mode, and you can rock the chuck, before the drive just sluggishly pulls it back to where it should be. It could probably be tuned to respond faster, but it wouldn't make any difference to the parts it's makes.
You have to consider the loads the spindle actually sees. My lathe has a 5.5KW 3000rpm spindle servo, so it only produces 17Nm of continuous torque at the chuck. Given it's got a 8" chuck fitted, that means it only takes 170N pull at the edge of the chuck to overcome the spindle. Doesn't sound a lot, but when you're drilling/milling within the capacity of the lathe, it's more than enough to keep things where they need to be.

[hardenum]

Again many thanks to JAZZCNC and m_c!

A dedicated direct drive spindle is too long for my build, the shortest I found is 385mm (https://sc02.alicdn.com/kf/HTB1l9RHX16sK1RjSsrbq6xbDXXaC/234643450/HTB1l9RHX16sK1RjSsrbq6xbDXXaC.jpg_.webp) with a motor that puts me at almost 700mm Z axis and that's just too long, so unless I design my own spindle I don't see how I can fit in max 500mm length. I guess that's off the table then.

[hardenum]

gone.

[jarjar]

Can't see the picture :(

[hardenum]

Is the "Attachment 29166" not clickable?

[jarjar]

It links to "invalid attachment"

[hardenum]

It links to "invalid attachment"

I tried opening it in edge and it also didn't work, I do not know how attachments work here so I changed the host. it should be visible now.

[jarjar]

Why the big cutout for the motor? Looks like a big waste of material and it certainly won't help with the stiffness of the fixed ballscrew bearing block.

For the gantry I would make a much bigger section, ideally square. EG is not very stiff at about 30 GPa.

[hardenum]

Why the big cutout for the motor? Looks like a big waste of material and it certainly won't help with the stiffness of the fixed ballscrew bearing block.

For the gantry I would make a much bigger section, ideally square. EG is not very stiff at about 30 GPa.

Actually that's because I wanted to make a universal casting for the belt assembly, and that meant all axes would have the same belt config, belt length etc. like this: https://i.imgur.com/5XeRLwY.png

The x_axis actually sits on a structural steel plate that provides the necessary stiffness and connects the x_axis with the supports.

[Nikolaguca]

Did I miss it, what is the gantry section size?

Sent from my SM-G970F using Tapatalk

[hardenum]

Did I miss it, what is the gantry section size?

Sent from my SM-G970F using Tapatalk

it's 280x140mm

[Nikolaguca]

I'm building a steel gantry router and I wanted to go with 200x100 steel tube but was given the advice to go with 200x200 because you're not just fighting gravity but cutting forces as well (and torsional forces when the z axis is lowered down)...

I'm using 250x250x10mm steel tubing for the gantry at 1300mm lenght to get the stifness I need.

Sent from my SM-G970F using Tapatalk

[hardenum]

Power supply

I've seen a toroidal PSU is a popular choice. I've also read that it can be quite dangerous to work with, as in getting electrocuted. I have octopus fingers, they get into all sorts of places without my permission, which usually results in cuts and wounds, I'd rather avoid any potential dangers, I'm quite clumsy in general. What are my options here?

AC/DC

I'm also confused about the fact that these (eg. https://www.cnc4pc.com/unregulated-linear-1440w-72vdc-20a-toroidal-psu.html) output DC, if I got AC servos, how does that work? Do I convert it back to AC?

Thanks!

[Doddy]

There's nothing any more dangerous about a toroidal PSU versus any other PSU of similar voltage and wattage. However, I think you might be slightly confused though - I'd expect that you'd be buying the servo with a matching servo driver. Looking at lt Lichuan 60's on Aliexpress - that's certainly a common bundle. In which case - check the datasheet, but the servo drivers would likely be 220VAC mains driven, and in which case you wouldn't need the high power, high voltage toroidal PSU. Or maybe I'm too tired and I've missed a point in your post?

[hardenum]

There's nothing any more dangerous about a toroidal PSU versus any other PSU of similar voltage and wattage. However, I think you might be slightly confused though - I'd expect that you'd be buying the servo with a matching servo driver. Looking at lt Lichuan 60's on Aliexpress - that's certainly a common bundle. In which case - check the datasheet, but the servo drivers would likely be 220VAC mains driven, and in which case you wouldn't need the high power, high voltage toroidal PSU. Or maybe I'm too tired and I've missed a point in your post?

I will definitely be buying a matching driver(amplifier) from lichuan. I really have no idea about powering electronics, I probably don't even know the questions I should ask. What's confusing me is that the servo's are AC and the recommended PSU outputs in DC.

[Doddy]

Link to information that recommends a DC power supply for the servo driver?

[hardenum]

Link to information that recommends a DC power supply for the servo driver?

No such thing, it was my assumption, so in other words I don't need any PSU at all, just plug the drivers directly to mains?

[Doddy]

Read the specification for you selected drivers.

You will always need part for control logic, switching etc., but if you have mains driven servo drivers then this would be a much smaller power supply, at lower voltage and power.

[hardenum]

You will always need part for control logic, switching etc., but if you have mains driven servo drivers then this would be a much smaller power supply, at lower voltage and power.

ah, that makes sense, okay, thanks.

[jarjar]

Some servo drives need only a AC supply, others need AC + 24V DC for the control logic.
In all cases you will still need a DC supply for the motion controller and other things. Basically a small switching PSU like this: https://uk.rs-online.com/web/p/din-rail-power-supplies/0428477/

[hardenum]

gone

[pippin88]

Why direct drive?

Won't it be much harder regarding drawbar etc?

[hardenum]

gone

[jarjar]

The Z looks complex. I assume you have a complete machine shop to build it. Ideally it should be as wide as the gantry carriage.

I would go fixed gantry for the relatively small working area you planned.

Maybe also consider simplifying the shape of the columns, but in the end you'll be the one to build the molds :)

[hardenum]

The Z looks complex. I assume you have a complete machine shop to build it. Ideally it should be as wide as the gantry carriage.

I would go fixed gantry for the relatively small working area you planned.

Maybe also consider simplifying the shape of the columns, but in the end you'll be the one to build the molds :)

I do not, all I have is a hand drill and a dremel, I'm outsourcing it. The sheet metal cutting and welding. It's not more complex then a straight or a C plate. It's just a square steel tubing in the end. I've actually received a quote for the whole piece from the local shop EUR600 total for the thing.

Besides the Z, nothing is final yet. I'd go with a fixed gantry but I haven't got a clue yet as to how I'd deal with the bed. What I mean by that is I want to be able to mount the 640x320 work piece vertically. With a moving bed, I don't have any ideas how to do that though.

[jarjar]

I do not, all I have is a hand drill and a dremel, I'm outsourcing it. The sheet metal cutting and welding. It's not more complex then a straight or a C plate. It's just a square steel tubing in the end. I've actually received a quote for the whole piece from the local shop EUR600 total for the thing.

What about the pretty much mandatory stress-relieving after welding, the precision milling/grinding for the rails, boring for the spindle body, milling for the spindle flange?

[hardenum]

What about the pretty much mandatory stress-relieving after welding, the precision milling/grinding for the rails

stress-relieving is in the quote. No under-rail machining, epoxy leveling them. If I were to have them machine all the required surfaces on it, might as well find a different hobby. Machining prices in germany are off the charts. One time I wanted to bring my 300x200mm aluminum heatsink for a shop to mill 3x 10mm wide 200mm long cannals on the underside for heat pipes, the quote was EUR2000.


boring for the spindle body, milling for the spindle flange?

You mean the cartridge? I'm not doing it from scratch. I'll be (or rather the shop will be) making a different drawbar for one of these https://ae01.alicdn.com/kf/H064296257e25438595acb0dc015c524f3.jpg, and slightly modifying the spinning part on their lathe. You are right though, that is not a part of the quote above.

[jarjar]

Steel-filled epoxy molding against a surface plate, yes, that can work. But not epoxy leveling.

I meant the bore to fit the spindle and milling the mounting surface flat and perpendicular to the rails where the spindle flange will be bolted.

[hardenum]

I meant the bore to fit the spindle and milling the mounting surface flat and perpendicular to the rails where the spindle flange will be bolted.

This sounds overcomplicating? The mounting surface that will be touching the cartridge will be cut with a saw so pretty straight to me if you ask. I mean I'm not making swiss watches, I don't need that kind of precision.



Maybe also consider simplifying the shape of the columns, but in the end you'll be the one to build the molds.

A smart man once said "people should put more work into molds".

[pippin88]

It's a gantry style with a lifting z axis, no way to do belt drive the traditional way.

https://i.imgur.com/NfWXHS3.png

There are dedicated direct drive cartridges that have a hydraulic thing that moves the drawbar like these:
https://sc02.alicdn.com/kf/HTB1l9RHX16sK1RjSsrbq6xbDXXaC/234643450/HTB1l9RHX16sK1RjSsrbq6xbDXXaC.jpg_.webp

The thought of modifying a $500 belt drive cartridge into a direct drive with a hydraulic drawbar fills me with excitement.I don't understand why you cant have a belt driven spindle. Just put the motor beside or in front of the spindle instead of behind.

I'm not saying direct drive is a bad idea. I just suspect it will be difficult dealing with drawbar issues etc

[hardenum]

Steel-filled epoxy molding against a surface plate, yes, that can work.


This, thanks for this, another piece of puzzle for my project. Really cool idea, having a hardness gradient on the cast piece.

[JAZZCNC]

Jar Jar and Pippin are making some very valid points and If you are not careful you are going to waste a lot of euros.

You are over complicating the design around the spindle and underestimating the importance of precision required in surfaces, etc. The fact you haven't even considered or looked into Fixed gantry design or even understand how it works tells me that you haven't done enough research into what is required for the machine you require.

I suggest you STOP and go take a look at fixed gantry designs because it's the BEST design for the size of the machine you are planning to build.

[hardenum]

Jar Jar and Pippin are making some very valid points and If you are not careful you are going to waste a lot of euros.

You are over complicating the design around the spindle and underestimating the importance of precision required in surfaces, etc. The fact you haven't even considered or looked into Fixed gantry design or even understand how it works tells me that you haven't done enough research into what is required for the machine you require.

I suggest you STOP and go take a look at fixed gantry designs because it's the BEST design for the size of the machine you are planning to build.

I cannot build a fixed gantry, there is no way for me to mount the workpiece vertically if I went with a moving table. There would be no point in building any cnc if I can't mill all 6 sides of the work piece.

I have looked over hundreds of designs over the past half a year and only a moving gantry with a fixed bed with a cutout ala Datron allows me to mount the workpiece in all orientations.

There is another design, a horizontal cnc mill, that would allow me to mount the workpiece so that I can mill all 6 surfaces. But to be honest it looks quite daunting to go down that route.

The spindle choice is pretty easy though. One day your bank account is full, the next day it's empty. If I went with a motorized spindle, I'd be screwed if either of the components broke. Cartridge + motor is far safer if something were to brake and needs replacing.

I'm happy to be proved wrong though.

I was thinking of having the front mold piece machined instead of plates that are cast into the casting. Cast it after having everything aligned to the machined front mold piece. There is no way for me to transport the casting to a shop for machining after it's been cast.

By the way, I'm months, if not years, away from actually starting doing any of this work, so I really appreciate any and all insights!

[JAZZCNC]

I cannot build a fixed gantry, there is no way for me to mount the workpiece vertically if I went with a moving table. There would be no point in building any cnc if I can't mill all 6 sides of the work piece.

I have looked over hundreds of designs over the past half a year and only a moving gantry with a fixed bed with a cutout ala Datron allows me to mount the workpiece in all orientations.

Only a 5 Axis machine can allow you to cut on all 6 sides in one operation. If your meaning to re-fixture and cut the other 2 sides then that's easily fixed, it's called an Angle plate bolted to the bed. Because that's all the Datron is providing with it's cut out, it's just an inverted angle plate built onto the end of the bed.

Other than that there is no difference between Fixed or Moving gantry other than Strength and size, Both have their own strengths and weakness but in your working envelope then the Fixed gantry wins hands down if you are wanting a strong machine.

If you want to work on 5 sides in one operation then Bolt a 4th axis to the table, again this would work on both designs but the Fixed gantry will do it better because it's stronger.




The spindle choice is pretty easy though. One day your bank account is full, the next day it's empty. If I went with a motorized spindle, I'd be screwed if either of the components broke. Cartridge + motor is far safer if something were to brake and needs replacing.

I'm happy to be proved wrong though.


I wasn't meaning the choice of separate Spindle + motor, that works good, but more your implementation of it is too complicated. Also, the Z-axis design is narrow and weak which will make all the trouble you'll be going to a pointless exercise.
There are plenty of examples of ATC spindle setups, just about every Milling machine with ATC uses the same setup. You just need the same setup and just rotate it 90deg so it's at the side if you don't have the room behind the spindle.

Also, I'm not sure your choice of Servo motor is correct as it looks just like an axis Servo rather than a Servo spindle motor.?

These are just some of the things I'm meaning when I say you are going to waste lots of Euros because none of them is cheap and easily done wrong if you haven't thought about and planned every last detail.

Theory and Cad are great, however, Reality is a Mother - F@~ing -Bitch just waiting to slap your face and kick you in the Nuts who never plays by theory's rules and changes them to suit Her whenever she wants.

[jarjar]

If your meaning to re-fixture and cut the other 2 sides then that's easily fixed, it's called an Angle plate bolted to the bed. Because that's all the Datron is providing with it's cut out, it's just an inverted angle plate built onto the end of the bed.

I can see how this would be problematic if you want to mill on the side of long pieces that won't fit under the gantry. With a fixed bed the cutout allows the work piece to extend below the bed. If OP wants to do a lot of those pieces then the moving gantry seems indeed the way to go.

About the sawn surface to bolt the spindle on, I think this is a big no-no. I'm not talking about precision watch making, just basic good practice for mating surfaces. If the contact area is bad, you will lose most of the stiffness and also distort the spindle housing, possibly damaging the bearings very quickly.

[hardenum]

the Z-axis design is narrow and weak which will make all the trouble you'll be going to a pointless exercise.

I thought it's pretty strong, stronger then what I'd get with a U plate and a spindle bolted to it. I based it on dmu 340. In which direction is it weak?


I can see how this would be problematic if you want to mill on the side of long pieces that won't fit under the gantry. With a fixed bed the cutout allows the work piece to extend below the bed. If OP wants to do a lot of those pieces then the moving gantry seems indeed the way to go.

That is what I meant.

[JAZZCNC]

I can see how this would be problematic if you want to mill on the side of long pieces that won't fit under the gantry. With a fixed bed the cutout allows the work piece to extend below the bed. If OP wants to do a lot of those pieces then the moving gantry seems indeed the way to go.

Pro's and cons to both but IMO the cons for having the slightly longer length end machining the price is a too high price to pay in strength terms. There are other better ways. . . For instance.!

For high clearance requirements, I have a design that will work and provide high clearance under the gantry with minimal loss in strength. I've just built a small version with a cutting area of 500 x 500 x 350 that could easily have the clearance and strength increased if required without any impact to it's cutting ability.


https://www.youtube.com/watch?v=M-hbeujeuRg

[hardenum]

I did a mockup of this design and I had concerns about the depth of cut. It does solve the Z axis problem you mentioned. There's actually this start up that's doing the exact design called Vulcan, that's meant to be sold for around 8k. Here's their structure:

https://scontent-frx5-1.xx.fbcdn.net/v/t1.0-9/107018603_737475153672782_2949729719032547967_o.jp g?_nc_cat=105&ccb=2&_nc_sid=9267fe&_nc_ohc=I6QX3inr0hUAX-7LMEF&_nc_ht=scontent-frx5-1.xx&oh=58ebd5547c427b75efb554c2f33838d2&oe=5FF33F20
https://scontent-frx5-1.xx.fbcdn.net/v/t1.0-9/105958491_736662403754057_6884295613843831119_o.jp g?_nc_cat=110&ccb=2&_nc_sid=9267fe&_nc_ohc=m_orI3xtXGgAX-0Zbw9&_nc_ht=scontent-frx5-1.xx&oh=a50327f0477547deb305b45fb41f4558&oe=5FF208F9

[JAZZCNC]

I did a mockup of this design and I had concerns about the depth of cut. It does solve the Z axis problem you mentioned.

Why did you have concerns about DOC.? It doesn't get any better than with this design for a machine with a larger cutting area as the Z extension is never any longer than the tool length, and that is at any Z height.
Some of the largest Gantry mills in the world use this exact design. If you build it strong enough the only limit on DOC will be spindle power.

[hardenum]

Why did you have concerns about DOC.? It doesn't get any better than with this design for a machine with a larger cutting area as the Z extension is never any longer than the tool length, and that is at any Z height.
Some of the largest Gantry mills in the world use this exact design. If you build it strong enough the only limit on DOC will be spindle power.

The main DOC concern was that when I put different pieces with different height onto the vacuum table I'd hit the higher ones with the vertical beam. Like machining the bottom and top piece (http://www.jt-precision.com/uploads/201816770/cnc-milling-aluminum-telecom-parts37293225867.jpg) one after another without actually having to go and change the piece. Sure I could put the higher one to the front, it would work, but knowing me, I'd switch the 2 unconsciously and mess some part of the machine.

Anyway, do you recommend a ballscrew on each column or one thicker in the middle?

[JAZZCNC]

The main DOC concern was that when I put different pieces with different height onto the vacuum table I'd hit the higher ones with the vertical beam. Like machining the bottom and top piece (http://www.jt-precision.com/uploads/201816770/cnc-milling-aluminum-telecom-parts37293225867.jpg) one after another without actually having to go and change the piece. Sure I could put the higher one to the front, it would work, but knowing me, I'd switch the 2 unconsciously and mess some part of the machine.

Anyway, do you recommend a ballscrew on each column or one thicker in the middle?

I suppose it's possible gantry brace could hit if your having very different height materials but I don't think it will be a massive issue and how often are you likely to do that.? It certainly won't be an issue cutting the part you show.
Also there is nothing stopping you having both a lifting gantry for very high parts and fitting a conventional Z axis to this. This way you have the best of both worlds. Again this is commonly done on Large gantry mills, and when I say large I mean something that could mill a full size Car from a billet of steel.!!

Regards the ballscrews then it only works with one on each column.:cower: . . . How could you fit a central ballscrew and still have the material move thru the opening.:stupid:

[pippin88]

The main DOC concern was that when I put different pieces with different height onto the vacuum table I'd hit the higher ones with the vertical beam. Like machining the bottom and top piece (http://www.jt-precision.com/uploads/201816770/cnc-milling-aluminum-telecom-parts37293225867.jpg) one after another without actually having to go and change the piece. Sure I could put the higher one to the front, it would work, but knowing me, I'd switch the 2 unconsciously and mess some part of the machine.

Anyway, do you recommend a ballscrew on each column or one thicker in the middle?That piece would not be machined assembled. It would be done in two pieces and then the hinges connected.

Your design won't machine that work in one piece anyway. Your cutting tool will be say 5mm, but 50mm higher in Z you will have your full spindle cartridge/Z axis to hit the piece standing up.

How much experience do you have machining / making parts?

Some of the issues you are trying to solve with machine design are actually fixturing issues.

No machine design will save you from forgetting to do something correctly.

[hardenum]

I suppose it's possible gantry brace could hit if your having very different height materials but I don't think it will be a massive issue and how often are you likely to do that.? It certainly won't be an issue cutting the part you show.
Also there is nothing stopping you having both a lifting gantry for very high parts and fitting a conventional Z axis to this. This way you have the best of both worlds. Again this is commonly done on Large gantry mills, and when I say large I mean something that could mill a full size Car from a billet of steel.!!

Regards the ballscrews then it only works with one on each column.:cower: . . . How could you fit a central ballscrew and still have the material move thru the opening.:stupid:

I guess some clever positioning? https://scontent-frx5-1.xx.fbcdn.net/v/t1.0-9/105958491_736662403754057_6884295613843831119_o.jp g?_nc_cat=110&ccb=2&_nc_sid=9267fe&_nc_ohc=m_orI3xtXGgAX-0Zbw9&_nc_ht=scontent-frx5-1.xx&oh=a50327f0477547deb305b45fb41f4558&oe=5FF208F9

Anyway you have me convinced(https://www.k-mm.com/wp-content/uploads/2014/12/large_machining_121_guideways_picture4.jpg).

Gonna take a long deserved brake before I dive into the new design. Thanks

[pippin88]

Regards the ballscrews then it only works with one on each column.:cower: . . . How could you fit a central ballscrew and still have the material move thru the opening.:stupid:

I like your new machine, though I think hiding the ball screws like that reduces stiffness too much by cutting into the tubes.

You can have a single central screw on a lifting gantry machine. The screw moves up and down with the gantry. Nut it attached to the fixed top cross beam. Either rotating but with servo attached to fixed cross beam, or rotating screw with Z servo moving up and down with Z.
Pros: Single screw (screw mapping, servo tuning, cost). Cons: A bit trickier to implement. Will have screw sticking right up high when Z fully up.

[JAZZCNC]

I guess some clever positioning?

That's not clever positioning, that's restricting movement and height.!!

[JAZZCNC]

I like your new machine, though I think hiding the ball screws like that reduces stiffness too much by cutting into the tubes.

Well that depends on application and what stiffness is required. In this application it doesn't need high stiffness but does require the height. However, I can tell you it's still very stiff.

I will be building a much larger version that will work the same with lifting gantry but with much more bracing using 10mm wall tubes rather than 5mm used on this.



You can have a single central screw on a lifting gantry machine. The screw moves up and down with the gantry. Nut it attached to the fixed top cross beam. Either rotating but with servo attached to fixed cross beam, or rotating screw with Z servo moving up and down with Z.
Pros: Single screw (screw mapping, servo tuning, cost). Cons: A bit trickier to implement. Will have screw sticking right up high when Z fully up.

You could but it will look crap and require you have a high ceiling.
Also it won't work very well because the gantry could and IMO will definately rack when plunge cutting at the outer edges. Disaster waiting to happen IMO.

The design I've shown only uses a single motor connected to the screws with belt/pulleys so it's only one extra screw so no big expense in the grand scheme.

[hardenum]

The design I've shown only uses a single motor connected to the screws with belt/pulleys so it's only one extra screw so no big expense in the grand scheme.

Was just about to ask whether I have to switch from belts to direct drive. Do I need brakes on the vertical servo's? It's gonna be lifting 150kg after all?

[JAZZCNC]

Was just about to ask whether I have to switch from belts to direct drive. Do I need brakes on the vertical servo's? It's gonna be lifting 150kg after all?

You have to design to your own requirements but lifting 150Kg then yes I would go with dual motors with brakes, doesn't need to be servo motors thou, Closed loop steppers will do the job perfectly well enough.

[hardenum]

Retracted.

[hardenum]

Quick servo question

Can I mix and match servo sizes? e.g size 60 for X and size 80 for y etc. ?

[jarjar]

Quick servo question

Can I mix and match servo sizes? e.g size 60 for X and size 80 for y etc. ?

Of course.

[Clive S]

Quick servo question

Can I mix and match servo sizes? e.g size 60 for X and size 80 for y etc. ?
yes

[hardenum]

It's not looking good, the change to new design doubled EG by volume. I'm gonna have to look into doing some FEA to know where I can shave off some material.

Question

Are there any handbooks that talk about forces acting on a machine during machining that I can use as a guide?

[hardenum]

What do you all think about this setup? Moving the bed to the side(for easier side milling)? I'm not sure if this is a machining center or a grinding or whatever though. By the way that's not a workpiece but the fixture block. Can I mount a workpiece to a vacuum table sideways?

http://www.pillevaran.com/images/rezin02.jpg

https://i.ibb.co/1Mvc6bx/asdff3ws.png

So who's slapping their forehead?

[pippin88]

Well that depends on application and what stiffness is required. In this application it doesn't need high stiffness but does require the height. However, I can tell you it's still very stiff.

I will be building a much larger version that will work the same with lifting gantry but with much more bracing using 10mm wall tubes rather than 5mm used on this.


I can't see how the stiffness loss with a slot is worth it.

The below shows two identical columns, except one has a slot cut in it. There is 2.36x as much deflection with the slot.
29240

Bellows don't compromise rigidity.



You could but it will look crap and require you have a high ceiling.
Also it won't work very well because the gantry could and IMO will definately rack when plunge cutting at the outer edges. Disaster waiting to happen IMO.

The design I've shown only uses a single motor connected to the screws with belt/pulleys so it's only one extra screw so no big expense in the grand scheme.

Who cares what it looks like?

High ceiling - maybe. Not that high. Depends on what level the base of the machine sits...

Racking is a consideration. Depends on width between rails and bearing spacing along the rails.

I agree, ball screw expense is not the biggest consideration (though starts to get much more when you use high quality bearing blocks etc). I'm thinking about screw mapping and linear encoders - much harder with two screws.
Belt stretch is a concern for a long enough belt to link those two ball screws.

[JAZZCNC]

I can't see how the stiffness loss with a slot is worth it.

The below shows two identical columns, except one has a slot cut in it. There is 2.36x as much deflection with the slot.
29240

Bellows don't compromise rigidity.

Like I said the first time, it depends on the application if it's worth it or not. This machine has a specific usage (which I can't mention because of customer confidentiality) that doesn't require high strength so deflection won't be any concern even if it deflected 10x.
This is not my first rodeo, I wouldn't build something that wasn't more than up to the task it's been designed to do by a large factor.

If I was building a machine for cutting more aggressively and I will be very soon, then you'll see it's designed very different with much more bracing and thicker wall tubes, but it will still have the slots because they very much protect the screws. Also, you are making a big assumption that there is nothing inside those tubes.!!

I'd also be interested in the forces you applied to those tubes.? Are they realistic or would we need 30Kw servos with 2mm pitch 30mm ball-screws to repeat...:joker:



Who cares what it looks like?

More than you think do care, also usually IMO when something looks wrong, it usually works wrong or awkwardly.




Racking is a consideration. Depends on width between rails and bearing spacing along the rails.

Racking can't NOT happen IMO when it's plunging into hard materials for operations like drilling, it's just by how much.?

The extra expense of making it more massive, etc to counteract this far outweighs buying another ball-screw as it doesn't stop at just the width, spacing, etc. As it gets heavier, wider, etc, it requires larger and more costly everything.
Doubling up means you share the loads so can use a smaller size and less expensive components which cancels out any cost offsets but gives a much better-balanced machine.!. . . Which looks like it was designed by someone who knows what there doing and not a tight-arse who's trying to save a few ££.


I agree, ball screw expense is not the biggest consideration (though starts to get much more when you use high quality bearing blocks etc). I'm thinking about screw mapping and linear encoders - much harder with two screws.
Belt stretch is a concern for a long enough belt to link those two ball screws.

Well first I wouldn't use belts on something any wider than that little machine I did and I wouldn't use belts if I was chasing accuracy to the degree of mapping ball-screws as it defeats the point and like you say makes it very difficult.

Regards Screw mapping then very few do that at the DIY level and if using linear encoders rather than rotary encoders then you shouldn't need to map the screw. Your only concern then is keeping the two sides in sync and I'd assume if using linear encoders the control system will be fully closed-loop so will handle that side.!

[hardenum]

If I was building a machine for cutting more aggressively and I will be very soon,

Looking forward to that!

[pippin88]

Like I said the first time, it depends on the application if it's worth it or not. This machine has a specific usage (which I can't mention because of customer confidentiality) that doesn't require high strength so deflection won't be any concern even if it deflected 10x.
This is not my first rodeo, I wouldn't build something that wasn't more than up to the task it's been designed to do by a large factor.

If I was building a machine for cutting more aggressively and I will be very soon, then you'll see it's designed very different with much more bracing and thicker wall tubes, but it will still have the slots because they very much protect the screws. Also, you are making a big assumption that there is nothing inside those tubes.!!

I'd also be interested in the forces you applied to those tubes.? Are they realistic or would we need 30Kw servos with 2mm pitch 30mm ball-screws to repeat...:joker:


The absolute value of the deflection (and the absolute load) is not the point. The point is the comparison between slot and no slot under the exact same conditions.

29255

You can double the thickness of the tube and you will still be behind. Better just to get rid of the slot.
(If you put the slot in the opposite side of the tube, away from the rails / load, deflection is roughly in the middle between 'no slot' and 'slot')

If you want to cover the screw, put it within another tube attached to the side, but don't compromise the stiffness of your main structure.

What is inside the tube that negates the loss of the major load path?




Racking can't NOT happen IMO when it's plunging into hard materials for operations like drilling, it's just by how much.?


Any axis where the cutting force is not directly centered and you are only using one screw is subject to some degree of racking. Every cartesian design has one axis where the cutting force is not centered. Your machine above has the Y axis (table axis) with one single screw. Are you concerned about that racking?

The risk of racking is about the proportions / aspect ratio between the spacing between the rails, and the bearing spacing along the rail.



The extra expense of making it more massive, etc to counteract this far outweighs buying another ball-screw as it doesn't stop at just the width, spacing, etc. As it gets heavier, wider, etc, it requires larger and more costly everything.
Doubling up means you share the loads so can use a smaller size and less expensive components which cancels out any cost offsets but gives a much better-balanced machine.!. . . Which looks like it was designed by someone who knows what there doing and not a tight-arse who's trying to save a few ££.


Adding a second servo etc becomes a non-trivial expense. Tuning two servos with two different screw errors (unless you are paying $$$ for very high precision screws) is non-trivial.

For my router I have used dual drive and would never go for a single drive on a wide gantry. For a wood router driven by steppers and cheap ball screws I totally agree with dual drive and strongly encourage everyone to drive both sides of the gantry.



Well first I wouldn't use belts on something any wider than that little machine I did and I wouldn't use belts if I was chasing accuracy to the degree of mapping ball-screws as it defeats the point and like you say makes it very difficult.

Regards Screw mapping then very few do that at the DIY level and if using linear encoders rather than rotary encoders then you shouldn't need to map the screw. Your only concern then is keeping the two sides in sync and I'd assume if using linear encoders the control system will be fully closed-loop so will handle that side.!

Knowing the error and compensating before it happens (screw mapping) is generally better than chasing it after it happens.

(Obviously the machine I have in mind with AC servos, linear encoders, 45mm roller rails is more than the average hobby build).




Machine design is always a series of compromises. I am not saying your machine design is wrong, or that mine is right. I am merely pointing out possibilities and weak points. For me, the slot in a tube to hide a ball screw is not worth it at all. There are other ways to protect ball screws. Less experienced people than you may not understand the design choices / compromises you have accepted and copy that feature without understanding the substantial impact it has.

[JAZZCNC]

The absolute value of the deflection (and the absolute load) is not the point. The point is the comparison between slot and no slot under the exact same conditions.

29255

You can double the thickness of the tube and you will still be behind. Better just to get rid of the slot.
(If you put the slot in the opposite side of the tube, away from the rails / load, deflection is roughly in the middle between 'no slot' and 'slot')

If you want to cover the screw, put it within another tube attached to the side, but don't compromise the stiffness of your main structure.

What is inside the tube that negates the loss of the major load path?

I understand what your saying and I'm not arguing that you are wrong, my point is that it doesn't matter if the loads are never going to be high enough to deflect the tubes.
The application of the machine shown is cutting plastics with minimal cutting forces, so even at the top of travel, there will be no significant cutting forces that will come close to deflecting those tubes.
End of the day it's about the application and if I was chasing strength then I would take measures to cancel out the effects of cutting slots in tubes, but in this application, which is the only application this machine will do all it's working life then it allows me more scope to full fill other design requirements. Like, compact a design, etc.



Any axis where the cutting force is not directly centered and you are only using one screw is subject to some degree of racking. Every cartesian design has one axis where the cutting force is not centered. Your machine above has the Y axis (table axis) with one single screw. Are you concerned about that racking?

The risk of racking is about the proportions / aspect ratio between the spacing between the rails, and the bearing spacing along the rail.

Again your correct and no argument from me. But there is a big difference in the direction of the cutting forces and like you say proportions/aspect ratios working together. So in this direction and at this width then the racking will be much higher unless the bearing spacing was proportional and this isn't practical for a Z-axis as it would need to be huge so increasing height or lowering travel, take your pick. Using two ball-screws takes all this away and reduces any risk of racking.




Adding a second servo etc becomes a non-trivial expense. Tuning two servos with two different screw errors (unless you are paying $$$ for very high precision screws) is non-trivial.

For my router I have used dual drive and would never go for a single drive on a wide gantry. For a wood router driven by steppers and cheap ball screws I totally agree with dual drive and strongly encourage everyone to drive both sides of the gantry.



Knowing the error and compensating before it happens (screw mapping) is generally better than chasing it after it happens.

Erm.!! tunning and mapping are IMO two completely different things, but if you are using Linear scales then screw mapping isn't required so no chasing is being done. This is why all high-end machines use linear scales ( and Very expensive ballscrews)



Machine design is always a series of compromises. I am not saying your machine design is wrong, or that mine is right. I am merely pointing out possibilities and weak points. For me, the slot in a tube to hide a ball screw is not worth it at all. There are other ways to protect ball screws.

Again we mostly agree, but we will have to agree to disagree about the relevance of the slot because while you may see weakness from your Armchair in the VIRTUAL example you provided I can tell you the REAL WORLD example I provided doesn't deflect to any significant degree and is a magnitude stiffer than is required to do what it as been designed to do.
My test of every router/mill I build is to cut the Aztec calendar in 10mm Aluminum because of the very fine detail in it. Then straight after I cut it again over the top to make sure there are no errors. I do this because if there were any errors or significant deflection then the second pass would destroy the fine details cut in the first. (See the poor pic)

29261


Less experienced people than you may not understand the design choices / compromises you have accepted and copy that feature without understanding the substantial impact it has.

Well first I haven't compromised anything, Also I can't be responsible for others' assumptions and if like you they assume I've just cut slots in the sides without knowing or caring about the effects of doing that then that's on them not me.!
Anyone who is building should design based on their own application and research or ask questions based on those needs. If they just copy from a single picture or video without asking or getting in touch with the designer then more fool them.

[hardenum]

Solved.

[hardenum]

Are these regular fixture clamps you'd use to hold the workpiece or some special types?

https://i.imgur.com/gdaktuO.png

[hardenum]

After much consideration I've decided to no longer build a cnc, any cnc for that matter.

The reason is the large discrepancy between what I want(EUR20k) and what I'm willing to pay(EUR10k).

You can go ahead and close this thread. I better find myself a cheaper hobby.
Thanks