Hi Everyone,

I thought I might start a lil build log of my 5 axis CNC router. I've been lurking around on these forums for a while, gaining knowledge and inspiration from all the excellent threads around.

This is my first time building something CNC (and my first time building something of this scale too!). Learning how to use my new RF45 clone along the way as well.

About the machine itself, its a moving gantry, 5 axis CNC machine with XYZ, and AB axis (incorporated in a bi-rotating head). Planned travel is 500x600x300 respectively. Ballscrews and linear guides on all linear axis. So far I've been constructing the machine solely out of aluminium, due to the ease of machining, and availability. The price is a bit of a bummer though.

Aim is to be able to route complex shapes out of wood, plastic, and hopefully aluminium. Being able to mill brass and thin stainless steel in 3 axis might be possible as well. As such, I've been designing everything to be as beefy as possible.

Attached are two pics of my half completed Z axis. Travel is 300mm, 5mm pitch ballscrew, 495oz.in stepping motor. My intended design is to create an open 'box' shape, to stiffen up the structure. Not too aesthetically creative, but its cold functionality than I'm after.

I'm aware that this is a rather ambitious design, especially for a first timer. I've seen a lot of threads were people have started off on the 5 axis path, and then promptly vanished. But I do intend to finish building this machine- too much money has been spent on parts and tooling to turn back now.

**I've got a quick question with regards to Mach3. Its installed, and I've restarted the computer before connecting a breakout board from CNC4PC to it. However, when I try jogging the system, there doesn't seem to be any output from ANY of the parallel port pins...I've tried reinstalling the program, as well as trying on a second computer. Even tried Mach2! (I'm sure there is nothing wrong with my wiring)
Both computers were PentiumIII, running WinXP.
Is there something small that I'm missing? Most people on the forums seem to just install it and it functions without any hitches? The parallel port is enabled, and the address in Mach3 is correctly referenced.

Besides that, the build is going smoothly. My rotating head should almost be finished by tomorrow (just making some stepping motor mounting plates and couplings at the moment). I'll post up some pics tomorrow night if I get the opportunity.

-Wilson-
(Anyone here in Sydney btw?)

have you tried checking the pins with a meter to see if you are truly not getting output?

Make sure you're not using a printer cable, they won't work. You need a cable with all the wires going straight thru. Easiest way to check if Mach is working is to check the direction pins at the port while jogging back and forth. Voltage should switch from 0 to 5V, depending on which direction your jogging.

I am using a male to male parallel port cable (not a printer cable)

So far, I've tried 'scoping the step pin, with nothing meaningful.
In the diagnostic tab in Mach3, are the pins in the representative parallel port supposed to light up when you jog the system? Because nothing lights up on mine... >_< (Same on both computers)

(I've almost finished the bi-rotating head. Will post up pics soon!)


Thanks,
Wilson

Gerry is correct. Actually if you are purchasing them they are called parallel cables (used for serial transmission) in the old DB24 serial connectors. Some had a crossover pin configuration for modems (called null modem cables)...this is NOT what you want. Printer cables were centronics which had a funky connector that wouldnt connect to anything but a printer or centonics device. If you already have the cables you can check pin numbers. I have attached a site to reference the pins for the Null modem cable. http://www.nullmodem.com/NullModem.htm.... the correct one the pins will not cross over.

Just a quick check....do you have jog enabled on the run or diag. screen... led must be on

Thanks for the all the replies everyone ^_^
The jog LED is on - the numbers on the DRO are moving as well. I think it might be the cable's fault. After searching through the whole house, I finally found a cable that wasn't parallel-printer...but it had a serial cable branching out from one of the parallel ends.... Since it was all I had at the time, its I tested the system with. If indeed the pinouts have been remapped to suit a modem, then that is the most likely source of all this mess >_<
I shall test it out with a proper cable tomorrow.

Besides that, I've essentially finished the birotating head. It only needs a few finishing touches here and there, and a quick fix to an mechanical interference problem.
It is made from 2 rotary actuator units that I purchased on ebay. Bolted on are 495ozin Nema23 stepping motors. Fabricated some custom couplings to attach the motor shaft to the gears used. (Note to self: Purchase some V-blocks!)
The actuators originally had a servo motor, coupled to a planetary gearhead, and then to the final gear. Since the stepping motor has much more torque, i'm only using the final gearing, which consists of simply a small gear driving a much larger one!

The assembled head will be bolted onto the Z axis with a 500mm long arm of 80-80 extrusion.

A bit worried about the weight of the gantry at the moment. Z axis is 10kg, head is 6kg. The arm will be around 3kg. For the X axis, I have some large NSK, 25mm rails set aside. Bringing the estimated gantry weight close to 40kg. Would a 495ozin Nema23 motor, driving a 20mm, 5mm pitch ballscrew be able to handle the inertia? Or should I consider upgrading to a Nema32?

And some pictures:
(Pic 1: Layout of components used; from the top, disassembled rotary actuator, mounting brace, mounting blocks, motor mounts, stepping motors, custom couplings)
(Pic2: Mock assembly of the 2 actuators - they can't be bolted together yet due to me forgetting to chamfer an edge.)

Enjoy~

-Wilson- (2:48am!)

Just curious - did you make or buy the adapter(?) for the small gear. When I made my 4th axis using similar drives that was the tough part! I drilled and mounted the small gear. I like your piece better - looks much neater and maybe more dependable. If you purchased - got a part number or source?
:cheers: Jim

I wanted to avoid having to modify the gear at all costs, because I would have no idea where to get another if I messed it up.

The adaptor is similar in shape to the shaft of the planetary gearhead (9.5mm shaft, with 2 opposing flats). Also turned a 6.4mm diameter on the end to help get a tight fit. 12mm of 16mm dia shaft is left on the end for motor attachment.
Next, the opposing end was bored with a 6.35mm, 22mm deep hole for the stepping motor shaft. The shaft is held in place with a M3 grub screw. The overall length was around 32mm. But thankfully, the offsets from the motor mounting plate still allowed the gear to line up roughly in the same position as the old one.

The gear itself is currently attached onto the shaft using a press fit. Somehow I produced a slight taper on the adaptor (probably from lack of skill whilst sanding >_<). Exploiting this unintended bonus, I pressed the gear in gently using a vice, being careful not to crack it in case it was brittle.
Had I possessed more forsight, I would've made the 6.4mm shaft bit longer, and included a groove for a C spring clamp. Improvised a solution by tapping a M3 hole in the end, in case I need to use a countersunk screw to hold the gear in.

I don't have a CAD program on my computer at the moment, so I hope my lengthy explanation will suffice :)

-Wilson-

Nice word picture -- well explained. Seems like that is the toughest part of this "mod." If you'd like to price out a couple of those I'd be interested. Might be a mini market for them considering all the interest in these drives last year or so.
:cheers: Jim

I'm still experiencing problems with Mach3.
Using the C1 breakout board from CNC4PC.com
+5V is connected to both sides, and the enable pin is hooked up to 5V.
Input side is connected with a male to male parallel cable (not modem or printer cable)
I don't seem to be getting any meaningful output on the pins when I try to jog the system. None of the LED's light up when I jog weight, but the pin2 LED stays on ...

**QUESTION: In Mach 3, under the diagnostics pane, if the system is operating correctly, the squares on the representative pins of the parallel port light up when the system is jogged?? Because on mine, they all remain OFF...

I would be great if anyone could shed some light on this.

Thanks,
Wilson

could you send us a few pics of your hookup. picture sometimes is worth a thousand words. Also what amperage is your 5v power supply?

Bob

The setup is essentially just 2x 5V lines, and a parallel cable. I've disconnected the drives, as I'm sure they are not the problem.

The 2 x 5V lines are *both* coming from the computer's USB ports via 2 USB cables. I know this negates the electrical isolation of the system due to the common ground, but I'm not expecting anything nasty coming through and destroying my PC (the drivers are optically isolated as well anyway). With regards to the amperage, one USB port should be able to draw around 700mA, and seeing as I'm only utilising 2 outputs, with no loads on them, it there should be sufficient current available.

The problem is coming from the parallel port - I oscilloscoped the outputs today. Disabling all axis instead for X, and setting pins 6 and 7 for step and direction, and I checked out the waveforms as I jogged the system.

The voltage is either at zero, or perpetually at 5V. Changing the direction did *not* change the logic state on the direction pin, and I could discern no noticable pulse waveform on the step pin with the oscilloscope.

I suspect that this is a software problem of some sort, because I've tried this on two computers, with no avail. I've run the drivertest that comes with Mach3, and both computers have passed.
Tried it on a vista computer as well (after applying a registry fix) with no luck either....

Does anyone have any clues? Its really begining to frustate me, as I anticipated that the electronics would be the easiest part of the build.

Thanks,
-Wilson-

The voltage is either at zero, or perpetually at 5V. Changing the direction did *not* change the logic state on the direction pin,

If changing direction did NOT change the voltage, how was it at both 0V and 5V?

Try changing the port settings in the BIOS. Try all the different settings.
Also, check the port address in the device manager. The setup video on the Machsupport site shows how to get the correct address.

Hi Gerry,
Thanks for your help.
I should've been a bit more clear. The direction pin was stuck at 5V, whilst the drive was at 0V (might've been the other way around), but the important thing was that they remained unchanged regardless of whether I was jogging, or jogging in a different direction.

The parallel port is on 0x378, which I've already checked. The LPT port is currently on ECP mode. I shall try it out with SPP tomrrow and see if that gives any difference.

If all else fails, I'm waiting for a PCI-parallel card from ebay to sort out my woes.

-Wilson-

Set the port in mach3 to 0378, not 0x378.

Have you tried contacting Aurturo at cnc4pc? he is usually very helpful in troubleshooting his products? www.cnc4pc.com

Haven't been able to do much with this project due to uni, but I've managed to acquire a nice piece of aluminium plate. It measures 865x165x40. Weighs 16kg!
The nice thing about it is that it was cut from a very large plate, and isn't extruded material. The surface is much flatter, and seems ground.

I've attached a picture of a mock X axis assembly, which has 2 NSK LH25 linear rails, and a THK 5mm pitch ballscrew. The whole lot will be enclosed with bellows to protect against dust.

This assembly will be bolted onto the gantry, and will be about 500mm from the table surface. Previously I was worried about the gantry crossplate flexing under load... not any more. If I put my whole weight on the plate, calculations predict that it will flex 35 microns :D

I've also done some calculations for the gantry uprights. I intend to make two square columns from 80x20 aluminium bar. Milled and bolted together. I had considered using 8080 extrusion, but the moment of inertia with the plate column is around 3 times higher than the extrusion, and at roughly the same cost / meter!
After doing some calculations (ignoring any inertial effects - assuming slow cutting), a 500N load on one of the uprights produces 0.02mm of deflection.

What are the prerequistes of being able to machine steel? Would an aluminium machine, sufficiently rigid be ok? Or would there still be some nasty chatter / resonance effects given the light weight of aluminium?
It seems like I might be able to machine steel plate given a reasonable spindle? (Not that I intend to).

Has anyone seen an adjustable gantry in a DIY design? ie, being able to lower / raise the gantry bridge depending on the cutting application. I've searched the forums without too much avail. It seems on these hobby machines, having an adjustable gantry would really increase the versatility of them. Being able to lower the gantry to increase rigidity whilst machining flat materials, whilst still being capable of doing 3D contouring by raising the gantry bridge to achieve the required height clearance.
I have a few ideas regarding this -
i) Using a round column and using adjustable clamps (think of a drill press)
Perhaps with 2 round columns / side to increase rigidity
ii) Using a square column and regularly spaced bolt holes
iii) Make 2 dovetail columns :D
Adjusting the height can be as simple as adding a vertical allthread/acme screw to raise/lower the gantry. Only disadvantage would be having to dial in the gantry everytime, unless you could add some accurate dowel pins.
I'm considering option 2 at the moment, adding a 16mm linear shaft + 16mm leadscrew on each side to support and raise the column.
Ideally, I would like to try the dovetail method, but I'm afraid there will be some racking problems over the lengths involed (~800mm) Dovetail cutters are also very expensive :(

What do you guys think?

-Wilson-

Has anyone seen an adjustable gantry in a DIY design? ie, being able to lower / raise the gantry bridge depending on the cutting application. I've searched the forums without too much avail. It seems on these hobby machines, having an adjustable gantry would really increase the versatility of them. Being able to lower the gantry to increase rigidity whilst machining flat materials, whilst still being capable of doing 3D contouring by raising the gantry bridge to achieve the required height clearance.

What do you guys think?
-Wilson-

I am presently building an adjustable bridge into my DIY mill design. You are right
it is a very nice feature of a cnc design, unlike the useless tilting vertical column found on X2 mini-mills, which I own. I have two back to back INA 20mm
rails going vertical at left/right handsides on my design, with the stage
areas linked together to get the rigidity that I feel is needed to support a heavy stainless steel x-axis; my y-axis is moveable at the base of design.

I will be watching your build log closely.
Thanks for sharing.

Jesse
(Java77man)

This machine will have swarf protection on all axis. The A and B axis are fully enclosed, and the X,Z axis will receive bellows to protect against swarf. (Haven't decided how to protect the Y axis yet)

I've posted up a picture of a set of homemade bellows. It is folded from a section of canvas like material. The material is used for roll up blinds (The type where you tug the string, and if you're not careful, the blind rolls up all the way to the top). The material is fairly springy, but it seems to hold its shape very well. I intend to paint it and coat it with some varnish.
Seems like a good cheap DIY way for swarf protection when coolant isn't being used. I hope material stands up to the constant creasing.

If anyone is interested, I'll post up the geometry needed to make a generic bellow of a certain height / width, but it should be fairly easy to figure out.

Does anyone have any thoughts regarding the adjustable bridge idea?


-Wilson-

Wilson:

Okay, you asked for it; thoughts about implementation. I like your 3 possible
methods, actually I considered all 3. (i)The Round column was a favorite in my thinking as I thought I might use two cheap $50 drill presses to obtain the
components necessary to make a slipping round column of sufficient strength
to both support and also to lock the gantry bridge at a specific height. It even had two headunits to use as a good right angle for use at the top of the bridge. I was thinking 500mm bridge to span however.
(ii) square column with spaced bolt holes, also simplified the height adjustment by having unique step height settings. If enough cheap steel
was available and a counter weight to make locking the set levels easy, it
could also work.
(ii) two dovetail columns, boy that would be nice, to just rack/pinion them in
place or dial adjust them with a heavy weight screw, something like the screw in a mill vise.

I had difficulty making a choice among the 3 methods: it wasn't until I got a
eBay purchase that forced a decision: my choice was dual square column using linear rails/bearings on each left/right hand side. A counter weight will be absolutely necessary. A heavy screw will be the next addition to each
side for easy height adjustment, and finally perhaps a rod/gears to drive both
screws at the same time.

My thinking was rigidity, rigidity, rigidity as much as possible to make any of the 3 choices work. When steel doubled in price, I started looking for complete assemblies to buy on eBay rather than raw materials to build with
because it was actually cheaper to buy than build each axis.

It is tough to make a choice, I am just glad that I now am bumped into action
on what to do, and can now start to work on the implementation.

Best of everything in your choice,
Jesse
(java77man)

Hi,

I've posted a preliminary CAD design drawing here to spark some discussion.

All 3 linear axis are positioned at machine zero in the drawing. The drawing still has several errors - the columns and table both need to be lengthened, and the X axis drive components are still missing, but the main look of the machine is there.

I've purchased a slab of 570x640x40mm aluminium for a good price to use for the table. Whilst I'm at it, I will be adding provisions for flood coolant.

How do people usually design the base on large CNC routers?
When I look at my CAD drawing, the base seems very flimsy compared to the rest of the structure. It is composed of 100x25 aluminium plate, with 50x50 3mm thick square tubing in the middle. The base will be bolted to the bench.
I've seen that the designs from tormach and Haas tend to have a welded steel subframe...however, this is something I would like to avoid, due to the inaccuracy of welding, and the need to remachine / epoxy the mounting surfaces for accuracy

Unlike many other CNC routers, my design has a very high centre of gravity, which is worrying.
In response to this, large industrial 5 axis routers tend to have a fixed gantry and moving table...which isn't practical for me given the space constraints of my garage.
I've considered building a router similiar to Rainnea, but that would necessitate 2 leadscrews and parallel drive...something which I don't have. It would also have much less versatility when it comes to switcing between milling aluminium and 5 axis plastic routing because the linear bearings will be permanently suspended quite high above the work surface.

I've settled on having regularly spaced tapped holes up the column onto which the Y axis can be bolted at various positions. The Y+Z axis weight will be mostly offset by two 250N gas struts working in tandem.

It would be awesome if some builders here could shed some light how to construct a solid router base or just some general advice :)

I shall hopefully start work on the Y axis this weekend. I've made a lot of effort to design it to be as modular as possible. (Much like a "bolt-on" linear actuators). Should be interesting trying to machine a 865mm plate on a machine with only 600mm travel.

-Wilson-

Just a quick update, the Y axis is almost finished. The 40mm aluminium plate has been machined with 5mm deep channels to accurately locate the linear rails (as recommended by NSK). Additionally, there are some large 12mm holes for mounting, and were also used to clamp the piece on the mill table. M6 holes were tapped in the ends to fix on the end plates + act as mounting points for cable carrier, gas struts, etc.
Just need to finish the end plates, the carriage plate and to fold the bellows. The axis now stands at 22kg and is 800mm long.

Disappointing, I found out there was a bow in the piece i was using. It was bent around 1.5mm in the middle...I clamped that area as lightly as possible as not to cause excessive deformation, but I feel that the axis won't be as accurate as I had hoped. Probably will add shims under the rails in the future as needed. Had to machine the plate in 2 passes, because my mill only has 570mm travel.

I think if I ever build a second machine, contrary to most builders here, I will build a *smaller* machine than the first!
[Tip to new builders - start with a tabletop version and work your way up!]

Comments / suggestions would be greatly appreciated :)

-Wilson-

Don't know if you saw my post containing a pdf of a US Patent of a 5 Axis Router.

The thread is here:

http://www.cnczone.com/forums/showthread.php?t=54401