Carnie Cutter – My first cnc build... Whoooh! Help needed!!!

[jestah]

It is painfully clear that the point of no return flew by a long time ago but I have not got around to documenting my first adventure into building my own cnc so here is a bit of a catch up so far:

I started with the idea of a small desk top router, convinced my self that I should build a Joe 2006 that soon progressed into starting a 4x4 hybrid and then next thing you know it I have dropped my life savings on enough parts to build this:



The 2.4x1.2x 0.3-0.35m cutting envelope should be able to cut mdf like butter and hopefully be stiff enough for slow cutting of aluminium and composite parts. I hope to achieve around a 0.05mm accuracy but will be more than happy with any thing up to 0.1mm.

I'm building from the ground up so it started with a steel frame made from

X main beams - 200mm Square tube with a 9mm wall
Support beams 200x100 and 150x100 6mm beams that will support the MDF spoil board.
Legs 150 square tube 5mm wall
Bracing 75x50x5mm

The table will come in at around 900kg but I may still consider filling the frame with sand... The feet will be bolted to the floor. I plan to spend a decent amount of time clocking in the rails as I want silky smooth movement and good repeatability. The legs and bracing will be welded together this week and then bolted to the supports and x rails. I have decided not to weld the x rails to the supports as my boss who is helping me with the steel fabrication ( I am some what clueless to how to work it as im a surfboard shaper by training) is worried about the stress from the welding causing the machine to move over time. So far the we have just bolted the x rails to the supports but should we consider using a thickened epoxy bed between the two? The upright legs are each capped off with 12mm plate and have a 24mm stud so we can shim them level before lowing the supports and x rails on top. Once again will a bolt and shim stock do or should we use an epoxy bed?

On top of the support beams sit a pair of 20mm rails on al supports. They will have 50mm long cars spaced apart 250mm (center to center of the cars) that will support the gantry. I have hoped that this distance will be long enough to give a stable base for the gantry. Each side is driven by a 10mm pitch 25mm ball screw and 960oz stepper. By my calcs I am over powered but after they arrived on my door they look tiny for something that will push close to 150kg of gantry back and forth. Each end of the ball screw will be held in place by a pair of angular contact bearings so I can slightly tension the screws and hopefully be able to drive them safely up to 400rpm. The build was never looking for top end rapid speeds but more low speed torque so I used nema 34 steppers from homeshopcnc.com over servos for ease of construction and tuning. I hope that this should be able to accelerate at a decent speed and reach rapids of 100ipm.

The build so far:



The gantry has two 130mm 9mm wall box support legs to spread the bearing cars 250mm apart and 200x200x6mm gantry beam to mount the Y rails to and also provide a big section for added stiffness. This drawing has the rails mounted 90degrees apart in opposite corners for z clearance but have since realised that a rail mounted above and below will give the same rail distance (once again around 250mm apart) and be much faster to clock in and the small loss of clearance will not be an issue as the tool will hang down about the same length too. I am a bit worried about my weight distribution to each of the cars as it seems that the rear car will be getting a lot more load due to the offset of the larger box. Should I be worried about racking?

The y carrage is also being redesigned to be a full surround tube for added stiffness and will be made from carbon fibre. The ball screw will be mounted and tensioned the same as the x axis.

I have looked at a lot of different designs for the z and like the idea of mounting the rails to a box section that the spindle will be mounted inside (using a box due to the fact that stiffness is a product of cross sectional area and therefore a 150x150x5mm box will deflect less than a 15mm plate) The cars will be mounted 250 mm apart. The total z movement will be around 350mm with a touch more clearance that will depend on spoil board thickness. The motor mounting is becoming a point that we can not decide on. I would think that mounting it with a toothed belt and spinning the nut would allow the motor to stay as close to the bearing as possible and prevent racking from having the motor mounted 360mm above as to clear the full travel.

I have a Chinese 2.2kw water cooled spindle that will be turned on and off my mach but for now will have speed set manually on the VFD. Is there any need for a safety to detect if the coolant flow is some how stopped?

The motors all feed from 4 G203 plugged into the back of a Pmdx-132 Bob. This is powered by a 48v 1000va toroidal transformer and 40000 milli farads of capacitances for smooth supply. This should give an output of just under the 70vdc that I was recommended to run the motors at.

I have to be honest and admit that I am not an electronics expert by any means and am struggling to see how to connect all the required systems together to create a safe machines. I will use a set of n/o switches for homing (each to there own imput) and a set of n/c limit switches wired in series to one input. I also understand that the limits should also be linked to the E-stop in a way that it kills the ac line power. The estop should also trigger relays for the spindle power, water pump and vac line and be such that an other switch needs to be depressed as well as the e-stop button be turned and released and the estop switch closed again before the machine can be used again. I can see that there is an estop input on the pmdx and have assumed that this is only there to warn mach of an E-stop condition an is its only function as the E-stop circuit itself triggers all of the need power kills and not be affected in case of computer/software problems. Can any one point me in the direction of a good “to code” estop circuit diagram with the relays so I can study and understand the system as a whole better? I have searched quite a bit and not found any helpful info that uses the 132 board. I have had a bit of luck at the scrap yards the past few days and managed to get enough drag chain and a really decent box to mount all of my electronics in! Should I mount the geckos to an al heat sink and cut a hole in the side of the stainless box to out the other side of the heat sinks out of the box as its made from stainless and will not conduct heat that well.




Just a quick recap of my biggest burning questions:

What should I paint the steel frame with?
Is 250mm bearing car centres be enough separation for good solidness?
Should I be worried about racking due to the gantry main beam being off centre?
How should I mount the z motor?
Coolant flow check circuit?
Basic help with my controller.

Any other points that I have forgotten or misunderstood. Please the more feed back, good and bad the better as I am very new to this and really would like help the huge vault of skill and knowledge that is out there!

I am sure I have left out many details but this will have to do for one night.... I will try to add posts as the project progresses! Thanks to all the builders who have inspired me and helped me get this far!

Jestah
Uglie Carnie Longboards

[CarveOne]

Your Z axis box/router assembly will be heavy if it is steel. If you mount the box tube to the Y axis slider assembly and mount the router to another slider assembly that is attached to rails on the fixed box tube you can significantly lower the weight that the stepper or servo has to lift. This will also reduce the tendency for the Z axis to back load the Z lead screw and drop down after turning off the power. A brake is used on some machines because of this. Adding dead weight is not always a good feature.

CarveOne

[jestah]

Hi carver,

In steel box with rails mounted to the spindle box mount it ill come in around 30kg. With the 5mm pitch and 960oz motor I am using my calcs show that it is once again well over powered.

I have read your concept a few times but want to check that im seeing it right. Are you suggesting that i mount the rails to my carbon L plate and have the cars on the box section to reduce weight?

[harryn]

Hi - That is quite an impressive project you have taken on.

I need to read the details a few more times to capture all of them, but a few quick comments:

a) Epoxy + bolts vs weld
- Welding tends to warp things
- The bolts + epoxy is a good way to go

b) Leveling
- Consider using some SELF LEVELING EPOXY to help true things up

Examples
- The top of the X main beams
- The top of the cross beams that will support the MDF spoil board

Doing this might require you to make a temporary "bridge" so the epoxy can flow evenly and level all of the beams to the same point. IMHO, this is the absolute easiest way to level something like this out, and the results are amazing.

c) Filling the tubes
- I don't think mass will be your problem, but ringing / noise might be from the hollow tubes
- Look at IRFAN's build for some ideas there
- I am considering filling mine with the ground up rubber from shredded tires (pretty cheap here)

d) Plan for dust and noise management - now
- There is not a router build out there where noise and dust were thought about either "before" or "after".
- Plan for dealing with it NOW while you still can.
- An enclosure is nearly a must IMHO
- Also plan to deal with dust that falls below the table. There is a thread here somewhere where the builder put a slanted pan under the table to capture dust.

e) 100 ipm
- Honestly, for the money you are spending, 100 ipm for MDF cutting is a little slow, unless I guess you are cutting 25mm thick stuff in one pass.
- With a 1.5 kw bosch router, hand held, I can cut 20mm thick plywood, full thickness at over 120 ipm. (tested it this past weekend)
- Of course for Al, 100 ipm is fast - normally with cutting fluid.

Since you seem to be good at calculations - what is your expected torque load on the rails from the gantry mass "static" and from the "routing". Honestly, I don't see a problem with your method of mounting the Y gantry to the cars.

You do need to drive the Y gantry from both sides to avoid racking - no question. I am mixed on the need to have a belt drive connection between the two x axis ball screws, but it is certainly something to allow for "in case".

Water cooled spindles - if they run with a coolant flow failure, they will fail. This is a good place for a flow / no flow interlock.

It's might seem strange, but if you had also used the 200mm tubes as your MDF spoil board support, you could almost skip the leg build portion of the project. I wonder if you would be as well off just laying some 200mm tubes cross wise under your bed, flat on the floor vs the time to build that table. I guess it depends on how high you really need the table for loading it. Maybe 3 tall I beams could replace the table base - unless you already the the money spent.

Keep it coming - great project.

HarryN

[ger21]

As Harry said, 100ipm rapids will be painfully slow for a machine that large. Imo, you should be aiming for at least 300ipm cutting speeds, and even faster rapids.


There's some good E Stop info here.
E-STOP Action

[jestah]

Thanks for the input guys!

I have double checked my calcs and it seems that I had entered something a bit funny so on rechecking according to the Nook I should be able to spin a 20mm root ballscrew supported at each end with a Double Bearings at a distance of 1585mm at around 530 rpm. This should give a speed of 208ipm in the x axis. While i can see that maybe under some tension i could bump this up to 570-600rpm but at that point i think it is all going to come down to the motors toque at that speed.They are wired in parallel but due to the lack of torque curve i can not assume more than 100ipm for the moment.... I am happy enough with that but am leaving room in the budget for a bottle of bubbles to celebrate any thing faster :>

Epoxy leveling... We use a VERY thin epoxy for our infusions at our workshop that we considered using but then wondered how do you get them level in relation to each other? do you need to make a cross dam so they all feed into each other? We were considering just bolting it together and shimming as close to level as we could then using a spoil board we surface with the cutter to give a flat bed, studding with t nuts (plugged with wax) so we can then infusing the table with epoxy to harden up. over kill?

I'm just sitting down with a fresh cup of coffee and going to go over that estop thread again. the first time I read it it was all a bit much info... second or third time lucky i say!

"Since you seem to be good at calculations - what is your expected torque load on the rails from the gantry mass "static" and from the "routing". Honestly, I don't see a problem with your method of mounting the Y gantry to the cars."

I'm ok with equations once i get an understanding of the concept. Please note that im only 23 years old, no formal training past collage (i think you crazy north Americans call it high school as i was only 17 when i left) and am basing my build almost purely on the HUGE wealth of information that is the zone and interweb. The bearing loading confuses the heck out of me and I need to look into that asap as the guy who is doing all the metal work for me (i recently shattered my ankle into a few different parts and can not lift much more than my lap top...) is about to start. I'm a bit worried about racking and would like to have the information backed up before we get cutting.

My guess as to how to work this out is weigh up all the parts or make some rough gustimates, plug this into solidworks and find a centre of mass and then use this point to work out how much toque each bearing is receiving when sitting still and when cutting in each axis... I would really appreciate help with this. Can any one suggest any good online toque calcs i can use to speed up this process?

As to force i used the calc spreadsheet based on CNCMechanics v1.1 that gives me a required toque needed of 625oz (based on my screws and a cutting force of 1000N and gantry weight of 150kg) I am hoping that my 2 motors running at 300rpm should be able to achieve this with a bit of room to spare. I am taking a WILD stab at 1000N cutting force as i have trouble finding the equations to calculate cutting force in relation to cutter size, feed speed and material hardness.

thanks SO much for the feedback, really really really helpful!
Jestah

[jestah]

so the score of the drag chain was good but they were coved in crud so time to get my gloves on and do some scrubbing....

[jalessi]

Jestah,

Why not use rack and pinion instead of a ballscrew?

http://www.stdsteel.com/gr_stock.htm

Jeff...

[jestah]

i decided to go ballscrews as they are simpler to impliment at slow speeds and have less backlash.

I am in no need for huge speeds from this cnc. I would rather go slow and take twice as long but be 4 times as accurate because often we build plugs that need to mate VERY closely with parts made in other workshops. If our part is to spec we dont have to rebuild, BIG problem when you sink a few grand into a carbon part to have to rebuild because you were a mm out.....

All the parts are purchased now so no turning back any way :P

[CarveOne]

Hi carver,

In steel box with rails mounted to the spindle box mount it ill come in around 30kg. With the 5mm pitch and 960oz motor I am using my calcs show that it is once again well over powered.

I have read your concept a few times but want to check that im seeing it right. Are you suggesting that i mount the rails to my carbon L plate and have the cars on the box section to reduce weight?

Think of rotating the tall gray box tube and the rails on it by 180 degrees. Then attach the tall gray box tube to the inverted gray 'L" shaped carriage (carbon L plate) so that the tall gray box tube only moves horizontally with the Y axis carriage (carbon L plate). Now mount the router (or spindle) to the Z axis carriage so that only the router and Z axis carriage moves vertically. The weight of the tall gray box tube is now mounted on the carbon L plate and is not one of the weights that the Z axis motor has to raise and lower. I see a number of designs here that do it this way, and I have recently used it on my second machine build.

If you have a motor that can easily move it the way you have it designed, that's ok. I was just suggesting an alternative method that will transfer the weight off of the motor's weight loading.

CarveOne

[jestah]

Think of rotating the tall gray box tube and the rails on it by 180 degrees. Then attach the tall gray box tube to the inverted gray 'L" shaped carriage (carbon L plate) so that the tall gray box tube only moves horizontally with the Y axis carriage (carbon L plate). Now mount the router (or spindle) to the Z axis carriage so that only the router and Z axis carriage moves vertically. The weight of the tall gray box tube is now mounted on the carbon L plate and is not one of the weights that the Z axis motor has to raise and lower. I see a number of designs here that do it this way, and I have recently used it on my second machine build.

If you have a motor that can easily move it the way you have it designed, that's ok. I was just suggesting an alternative method that will transfer the weight off of the motor's weight loading.

CarveOne

I think there may be a confusion about my set up, ill post an other angle so you can see whats going on in my z. At the moment the tall gray box IS the z plate. Inside this is the spindle and to the back side of this is the rails that engage the cars that are bolted to the carbon L plate.

From your description i see that the grey box will be moving side to side, as well as the L plate on the gantry and i will need an other set of slides to get z movement.




this is the same style of set up i will be using but i have chosen a 150mm box section to mount my spindle inside of rather than a 15-20mm plate to mount it to. They are both about the same weight but a box with that much cross section will be many times stiffer and with a deep 350mm of z i need as much stiffness as i can get!

[harryn]

The box structure will make a nice Z axis with much less chance of the spindle bouncing around than a flat plate axis. I sort of did something similar in my proto build (not as far along as yours)

There are several posts on the forum - just do a search "self leveling epoxy". Another resource is www.precisionepoxy.com.

Self leveling epoxies usually are rather thin and have wetting agents to make them really flow out flat. Hand leveling is a skill / patience thing. Self leveling epoxy approaches curvature of the earth kinds of numbers.