PDA

View Full Version : BUILD LOG: starting from scratch- router 1st project



idefix
01-11-2013, 11:11 PM
I really appreciate the sharing of experience of the community through this forum.
I've been following the various projects exposed here for some time already and have decided finally to step in by myself. My english is not perfect (sorry for that) but I will make all my best to explain my project.
For sure I will have many questions... but I'm here to learn :) from all your experience. So please don't hesitate to comment and give me advice to avoid letting me going in the wrong direction.

First, I already own a lathe and a little milling machine (round column type) plus a big column drill (1.90m heigh with 30mm drilling capability) salvaged from srap... and totally refurbished. Let's say they are my basic tools.
10562

I'm planning now to design and build a CNC router/milling machine able to machine light material but also aluminium. I already start from an existing welded frame (50x50x3 tubes) which needs some reinforcement; but however this gives me the overall size of the future machine: 1100mm x 650mm.

My first investigations are giving the following axis travels: X 750mm Y 350mm Z 150-200mm
The distance under the spindle attachment and the table will be adjustable to accomodage eventually the use of a milling wise or the use of a horizontal axis divider

Here are some pictures of the project :

105591056010561

Screw ball types are 1605 for Y and 2005 for X (maybe 1605 sufficient ?)
Next question is the use of simple or double separated screwball nuts or dual screwball nut (combined type). what should you recommend as I want to be able to machine large holes with H7 accuracy...

The main beam of the steel gantry will be filled with concreat and same for X axis support beams.
The steel tubes sections are 50x50x3, 50x50x4 and 100x50x3
Currently the axes guides are round types (diam 20 and diam 16 respectively for X and Y axis)

The Z axis will be fixed on the Y box (made of 20mm machined aluminium)

Thank in advance you for all your future comments.

Philippe

mekanik
02-11-2013, 12:49 PM
Hi Philippe
Welcome to the forum, Nice workshop.
I am a noobe to the CNC field so can't offer much in the way of advice but jazz or one of the guys will be pitching in shortly with something more constructive,one thing is for sure you will probably have to up the budget to include profile rails and i would imagine you will require zero backlash ballnuts for the kind of accuracy you are looking for.
Regards
Mike

Boyan Silyavski
02-11-2013, 12:53 PM
Hi,
i think the gantry ball screw that moves the Z left right is very high, it should be middle or lower, cause there will create a racking force.

Same with the gantry ball screw nut housing. it should be in the middle of the bearings, not at the end, even better will be to be offset exactly where the gravity/force center will be, which is a bit further in front direction from that center.

idefix
02-11-2013, 02:42 PM
Thanks for those first comments.
I agree with your statments and I am already considering changing the round chafts for Hiwin type rail. Would the type HGR15 be sufficient or should I move straight to larger type ? (don't plan to mill hard steel)

The choice of the position of the y axis above the guides was dictated by the attempt to reduce dust pickup by the screw.ball
For the X axis, I had in mind the possibility to mount 2 screwball nuts to be able to reduce backlash. Currently only one is shown on the drawing.
However, I'll try to move the ballscrew upward to reduce the arm length.

Thanks

JAZZCNC
02-11-2013, 04:40 PM
Hello Philippe,

I see several things really.

Yes change to profiled linear rails if you can afford them has they make much better and accurate machine. 15mm will be fine but 20mm often don't really cost much more and the larger bearing size helps.

Ball-screws then I'd use 1610 not 1605 if your mainly cutting soft materials has they allow higher feed rates which you'll need. They also provide enough resolution for most Aluminium work unless working to fine tolerances. If you connect to screws with timing belts/pulleys, and I suggest you do, then it's very easy to change pulleys and increase resolution and torque but at the expense of speed,
I would change the Position of the Long axis (X axis) ballnuts to mid point of gantry and have supporting braces at both sides of drop bracket not just one like you have drawn.
While the Y axis nut being high up is not the ideal it will be fine there and better there than further back but lower down.

Regards the rear plate of the moving carriage then you don't really need it has it's adds very little strength and more weight to the moving carriage reducing acceleration, also adds expense.
It makes building the carriage more difficult has the connecting surfaces and edges must be perfectly 90deg to each other else it turns into a parallelogram, any error will increase binding on the bearings and it doesn't take much to bind them solid. It also interferes with other aspects of the machine which you haven't drawn and maybe not taken into account.? Like cable management. The rear of the gantry is perfect place for Cable chain and with the plate nothing can be fasten at the rear.!! . . Simple thing but can be annoying.
Not having it takes nothing away from machine and makes building easier and less work. Just machine the top bearing plate so the ballnut mounts direct to this and is central, the bearing plate will extend back above the ballnut then you can bolt down direct into the ballnut mount from above so doing away with the drop plate and potential weak spot.

Regards the Gantry then I'd consider having some Triangle gussets around the back strengthening the connection between sides and gantry cross piece, these braces are simple things which really help with resonance dampening.

One last thing.? The Z axis rear plate doesn't look very tall so you may find you need to extend this from how you have it drawn now.? How do you plan to build Z axis.?
With the short back plate I'm getting a feeling your thinking to have the motor travel with the front plate but I'd recommend you don't has it's harder on the motor and reduces acceleration. It makes using Belts to connect motors to screws harder and I suggest you do use timing belts/pulleys has it's better for few reasons. Mainly smoother working motors due to better resonance handling and flexibility regards fine tuning if needed.

Oh and think about capping ends and using Dry sand rather than Concrete has it can be removed for when moving and dampens vibrations very well.

idefix
03-11-2013, 11:40 PM
Hello, Jazz
Thank you for these valuable comments and suggestions. I have rebuilt the gantry with Hiwin style rails and blocs. size 15 for the Z axis and 20 for the Y axis and future X axis.
You are definitely right, the size 20 is almost same cost than size 15 for a much better strength...
I have tried my best to keep the Z moving plate as close as possible to the gantry beam (not so easy). Z axis travel is 170mm
Following your advice I have plugged the beams with covers (but do not decide yet where to place the plugs to fill/remove the dry sand)
To achieve correct geometry of the sitting planes for the rails, I can have access to a large milling machine at the company workshop (this will minimize the efforts of the alignment step)
The ballscrew type has been changed to 1610 for Y and Z axis and will be driven by timing belt ratio 1:1 (with space left to replace for 2:1 ratio whenever necessary)

Some pictures of the changes:
1057710578105791058010581



Again a few more questions :

Do you think it would be better to design a cover on the Z sides (to limit dust and debris penetration) ?
For lubrication, should I use grease (with the nipple) or should I already design oil pipes from now ?
For the Z axis stepper motor; should it be facing downward or facing upward ? (would depend mainly on the position of the ballscrew...)
Regarding the long X axis driven by two separate ballscrews and separate timing belts + motors, I plan to link the steppers on 2 different axis via mach3 sofware. Does its sound ok ?
Last question related to ballscrews: does the chinese ones so called "antibacklash" realy fit the requirements ? or should I go for the coupling of 2 separate nuts as exposed in Jonathan thread (a sufficiently strong machine) ?

Again thank you for your very valuable support.

JAZZCNC
04-11-2013, 12:45 AM
Hi

Ok first my mistake and should have made clear, Only use 1610 on X & Y axis use 1605 on Z axis.

Yes I would have cover over Z axis has it protects everything.
I would only go to the trouble of oil system if access is difficult and really wouldn't go to all the trouble of Oil pipes when access is easy like yours would be.

Motor position doesn't matter really other than ease of fitting and working on pulleys etc, I tend to always position them with motor down and thread the motor holes has it makes for easy adjustment and fitting. Either works fine thou and I've done both ways.

Single or double Nut will depend on your requirements but for most router type work then single is plenty good enough and even for moderate accuracy milling work they are ok. If your wanting very fine tolerences, say for PCB or Engraving work then yes 2 nuts are best but then I'd say you've built the wrong machine anyway.!!


Regarding the long X axis driven by two separate ballscrews and separate timing belts + motors, I plan to link the steppers on 2 different axis via mach3 sofware. Does its sound ok ?

This I'm not 100% sure what your meaning.? . . . . If your asking is it ok to use 2 motors for one axis and slave in Mach3 then yes that's fine and is done all the time. You just have to be aware of it's limitations and tune the motors correctly other wise you can stall one motor while the other keeps going which can be very messy and destructive.!!

idefix
04-11-2013, 09:01 AM
Ok fine and clear ! Thank you so much Jazz for these directions.
Will stay with the grease nipples and will correct the Z screwball type for 1605
For the X axis with 2 motors yes I plan to slave them in Mach3.
Let's finish the design work and will come back in a few days.

EddyCurrent
04-11-2013, 05:59 PM
Regarding the long X axis driven by two separate ballscrews and separate timing belts + motors, I plan to link the steppers on 2 different axis via mach3 sofware. Does its sound ok ?

Philippe, It may not be obvious but if you use two motors you will need 2 drivers, a slightly bigger power supply, more wiring. In my case it added about 100 GBP to the cost. I'm not complaining it's just that I did not fully think about it at first but I still like the 2 motor idea best. The alternative of course would be to use one motor and longer drive belts to each ball screw.

I see you have been following the frame construction threads closely :joyous:

Boyan Silyavski
05-11-2013, 06:21 PM
Hi

. . . . If your asking is it ok to use 2 motors for one axis and slave in Mach3 then yes that's fine and is done all the time. You just have to be aware of it's limitations and tune the motors correctly other wise you can stall one motor while the other keeps going which can be very messy and destructive.!!

Dean,
i have always wanted to ask that question, hope you know the answer- if the motors that drive the gantry don't drive it directly but with pulleys 1:1 ,if the gantry is racked by chance, wouldn't the belt give away and just start skipping instead of destructing something?

JAZZCNC
05-11-2013, 07:03 PM
Dean,
i have always wanted to ask that question, hope you know the answer- if the motors that drive the gantry don't drive it directly but with pulleys 1:1 ,if the gantry is racked by chance, wouldn't the belt give away and just start skipping instead of destructing something?

Well yes and no dependant on motor type used.?
With steppers In reality what happens if one motor stalls is the other keeps going until the other motor stalls, by which time a number of things can happen depending on speed it was travelling at time. Often motors usually stall at higher feeds has that's where torque is lowest and machine incorrectly tuned for high velocity and high Acceleration just to compound the matter.!

Now often when first motor stalls due to low torque the other motor isn't far behind on available torque so it doesn't take much to bring both to a halt but that's not what causes the damage.? . . . . It's the 50-60Kg gantry travelling at 150mm/sec (9mtr/min Av rapid speed) just at one side which does the damage.!! . . . Doesn't matter how fast you are at reacting to seeing the motor stall you can't stop the machine fast enough and in the few seconds it takes to hit the "Oh shite" button before one side has travelled 300+mm further than the other side.!! . . . I'll leave the rest to your imagination has to what can be damaged or screwed up.:hopelessness:

Now if you were using Servo motors then yes possibility it would strip the belt teeth but chances are it wouldn't stall either and the drive would detect any following error and E-stop anyway.

So the moral of the story is when slaving either be conservative on the tuning, use drives that detect stalling or use a different approach.?
I've done all 3 and my preference is single motor with belts connecting screws but I'm also ok with slaving you just have to be aware of the limits and this is why I like drives that help and take some of the risk away.!

idefix
04-01-2014, 09:05 AM
Hi everyone,
I'm back after some time...
First of all let me wish you and you beloved relatives all the best possible for this new starting year. May 2014 be placed under new creative spirit, and wonderfull designs (as usual !).

I have taken some time to digest all the valuable informations posted on this thread and finally started the construction after some revisions on my design.
The parts of the main gantry beam (supporting Y and Z axis) have been welded together and the tube sections filled with vibration damping material. It has been coated with epoxy paint for durability and ease of cleaning and it is now ready for machining to generate the reference planes and associated straight edges. The overal weight is now 17Kg.
Some pictures:
111731117411175

1117611177111781117911180

As you can see, I've finally chosen to drive the X axis with its 2 ballscrew sets with one single stepper plus HTD 5M-25mm timing belt and 24 tooth pulleys. The belt tension device is not yet defined.
The beams will also be filled with antivibration material compound.
The rails supporting X axis will be leveled using epoxy technique.

There will be a cover for the whole Z axis assembly and an other one for the front belt drive system.

Need your advice for the axis end sensors: Do I need 2 sensors for each axis or only one sensor activated by 2 separate fingers (one at each end) ? What are you using on your own machines ?

Looking for your comments.

Philippe

Boyan Silyavski
04-01-2014, 09:46 AM
Nice development.

To be honest i have similar size machine and though i have bought the sensors/the smallest mechanical ones, that people around forums say work fine/, i have never mounted them. May be others will disagree but for this size machine, driven by steppers i don't see a reason to put sensors at all. I use the soft limits and with some care am ok. The worst that could happen is for the motor to stall at one end of the axis. Yeah, it happens some times and nothing breaks.

JAZZCNC
04-01-2014, 09:46 AM
Need your advice for the axis end sensors: Do I need 2 sensors for each axis or only one sensor activated by 2 separate fingers (one at each end) ? What are you using on your own machines ?

Looking for your comments.

Philippe

For the limit switches I use 1 x Proximity switch per axis that travels with the axis and is triggered at each end with adustable target. The switches are PNP NO(normally open) and wired in parallel sharing just one input.

Home switches are Fixed and Use proximity.

TonyD
04-01-2014, 07:16 PM
The CAD above is beautiful.

Neale
08-01-2014, 10:39 PM
Philippe,

I'm designing a machine which is similar to yours, although slightly larger capacity, and trying to use as many ideas discussed in this forum as possible.

I am working on the Z carriage design and, as you said earlier, trying to keep the moving and fixed plates as close as possible. I cannot see from your CAD drawings how you have managed to fit the ballscrew nut in between the plates. The nut is wider than the profile rail + bearing height - have you put packing under the bearings to increase the space? I can see that you have machined steps into the plates and a recess where the ballscrew nut is fixed.

idefix
09-01-2014, 08:13 AM
Hi Neale,
I'm happy to share related information back with you.
I have broken the design so that you can see the area as detailled as possible. Let me know if this is what you are looking for.
As a comment note that for the Z axis my rails are type HGR15 with Hiwin HGW15CA blocks and the screw is a 1605.

Distance from Gantry plate to Z plate (plane supporting the blocks) is 24mm

11222112231122411225
112261122711228

Neale
09-01-2014, 10:01 AM
Many thanks for the details. I had started with 12mm plates but I think that you are using 20mm which allows you to machine recesses for extra clearance, and there are also packing pieces under the rails. Back to the drawing board!

Are you using something like Solidworks for the drawings? The drawings of slides, ballnuts and so on are very detailed and look as if they might have come from the manufacturer!

idefix
09-01-2014, 01:18 PM
Yes Neal, using SW for the drawings with some parts already loaded from the web such as rails, blocks and M4 and M6 screws plus the timing belt pulleys. All the rest is own design.

I do confirm that base material thickness for the plates is 20mm (initialy for ridgidity purpose) but helps making the cavities...

Jonathan
09-01-2014, 01:24 PM
For a small gain in strength and machining time, you should be able to make the recesses in the Z-axis not cover the whole length, as the ballnut doesn't traverse the whole rail.

Neale
09-01-2014, 04:05 PM
That's the kind of thinking that leads to something that looks great in the drawing but is impossible to assemble! In this case, I don't think that that would be a problem but I've painted myself into a corner before now...

Jonathan
09-01-2014, 04:10 PM
So think harder! Admittedly I did recently make something that turned out to be impossible to assemble...

Have you checked that the Z-axis belt clears the posts upon which the motor is mounted? If you decide to change the ratio it might not. Also, I wouldn't reduce the thickness of that motor mount beneath the motor, partly because it needs to be strong but also because it adds an extra machining operation which will increase the cost of the part unnecessarily.

JAZZCNC
09-01-2014, 08:49 PM
Hi Neale,
I'm happy to share related information back with you.
I have broken the design so that you can see the area as detailled as possible. Let me know if this is what you are looking for.
As a comment note that for the Z axis my rails are type HGR15 with Hiwin HGW15CA blocks and the screw is a 1605.

Distance from Gantry plate to Z plate (plane supporting the blocks) is 24mm



I wouldn't use those spacer plates and just buy thicker Material because your introducing several more operations into the job but more importantly increasing the chance for error.!! . . . . Unless Your getting the work done for free won't be saving any money and will probably be cheaper because of less cycle time involved, also it won't be much lighter.

idefix
09-01-2014, 10:24 PM
I wouldn't use those spacer plates and just buy thicker Material because your introducing several more operations into the job but more importantly increasing the chance for error.!! . . . . Unless Your getting the work done for free won't be saving any money and will probably be cheaper because of less cycle time involved, also it won't be much lighter.

I was not enthousiast with those spacers but this is in my opinion the only way to adjust the position of the rails afterward... Because the lower Hiwin blocks are fixed first with positon guaranteed by the machining of straight edges, I still need to be able to adjust the rails... And here is the role of the spacers.

Related to the reduction of the thickness of the motor holder, this is necessary due to the relative short length of the axes. Mechanicl strength will still be plenty because belt tension effort combined to pulley position is only generating low flextural torque.
However, have you some suggestions (little drawing appreciated) about how to design this part in a different way ?

For the Z axis the position of the posts have been checked for sufficient clearance for the belt path. Of course this introduce limitations with the size of the pulley but I don't expect to change from ratio 1:1

Thank you very much for all your valuable comments making this thread alive.

JAZZCNC
09-01-2014, 11:53 PM
I was not enthousiast with those spacers but this is in my opinion the only way to adjust the position of the rails afterward... Because the lower Hiwin blocks are fixed first with positon guaranteed by the machining of straight edges, I still need to be able to adjust the rails... And here is the role of the spacers.

Well that's worse than expected then because really you want referenace edges in the Front plate not just the back plate.!!

One way I use to maximise clearence with thin material and low rails etc is to not cut the referance edges into the material and sit the rails and bearings direct on the plate face.!!. . . . To give the reference edges you want for the rail and bearing alignment you then machine and press aluminum dowels into the plates half way on the reference edge you then machine these Dowels half way and flush with material to create the reference edge. This way you save machining and keep the strength of the plate plus have an accurate referance edge to work from. It's also much quicker than milling full length slots.

If you use Ground machine plate then this technique ensures you get the best accuracy from the plate because even the best milling can't match a ground surface to mount the rails on.!



Related to the reduction of the thickness of the motor holder, this is necessary due to the relative short length of the axes. Mechanicl strength will still be plenty because belt tension effort combined to pulley position is only generating low flextural torque.
However, have you some suggestions (little drawing appreciated) about how to design this part in a different way ?

Regards the Motor mounting then just offset the motor and flip.!
See the pic of Cad model for one of the little machines I build and if you want to see the real thing I can take a picture for you.! The bracket is wide and square because I put a cover over the Z axis and it fastens to this plate, to reduce weight pockets are machined in the under side along with a pocket for the Motor.