PDA

View Full Version : BUILD LOG: Aluminum profile and alu plate 1500mm x 600mm x 150mm working area, first cnc.



pierantoine
17-05-2015, 05:50 PM
Hello,
i've read many posts but know i would like to build and run th machine.
i already have the low part of the frame, and now i'm building the gantry and the Z axis, i had few years ago bought a tb650 kit, some steppers that i will replace with a plc4x-g2 and am882 drivers, and some round supported rails, that i could replace with hiwin 20 mm profile rails for the Z axis.

i don't need extra quick cutting speeds, and want to cut wood, plastics, and aluminum (slowly) if i can.
i don't have a cnc on the workshop, so the design is done to limit the work that i'll have to do myself after ordering aluminum profile and plates that are cutted for me by the seller. http://www.mycncuk.com/attachment.php?attachmentid=15363&stc=1
http://www.mycncuk.com/attachment.php?attachmentid=15364&stc=1 the new gantry design with hiwin rails and http://www.mycncuk.com/attachment.php?attachmentid=15365&stc=1 the prevous design with round supported rails

do you see any mistakes that i can arrange ?

Thanks for your help.

Boyan Silyavski
17-05-2015, 07:26 PM
So you are again going to make something that again will need upgrade?

if you haven't found already that the linear rail choice takes the machine to different category of ability to cope with materials and jobs, too bad.

I am not sure about that gantry and Z design at all. That gantry makes sense with square rails and more importantly made from steel and Z from thick aluminum plates. There are very good other design here at the forum and over internet using round supported rails.

Generally my advice will be - design the best you can , save money and buy the rails and screws, then the frame and slowly build it if money is an issue. Plus life helps you and money comes.

And before you think, hey this guy is a jerk- tomorrow i am going to work for a week something i said i would not do anymore, so i could make some $$ to finish my current build. So its not only you to aim higher.

pierantoine
17-05-2015, 08:38 PM
Hi !
Thanks for your answer, when i said previous i mean the one i've designed before, but it hasn't been built.
Every parts are not made of steel, but the idea is that some parts in steel can be bolt on aluminum profile.

About thicknesses, the plate supporting the spindle is 20mm thick and the plates on the z axis are 10mm.

pierantoine
12-12-2015, 05:53 PM
Hello,After the Boyan's answer, I've redisign it; this one is inspired from those on this forum This project is to really puts hands into the cnc world with a simple machine that will work, I've no lathe, milling machines or welding equipement at home and the main idea was to do a DIY project using the maximum of aluminium profile or aluminium small plates
I'would like to cut mainly MDF, wood, PMMA, and aluminium if i can, but I'don't need a Fast machine.
16785

16783 16784


Work area: X-axis: 600mm, Y-axis: 1500mm Z-axis: 150mm

Frame : i've started to make a frame with 30x60 aluminium profile from motedis.co.uk (http://motedis.co.uk/) wich is perfect for sbr20. this aluminum frame could be cocrete epoxy filled and straightened with (alu or steel) plates.A friend could help me to make a steel frame, in this case i will go with the epoxy leveling technique. (i've seen the useful boyan post in his build log).
Gantry made of aluminium 80x120...

all the plates are 20mm thick maybe i can use less somewhere...

i could remove the orange plate and replace the yellow profiles with a 200mm plate.


Mechanical parts:



Ballscrews: RM2005: X-axis: 1650mm x 2, Y-axis: RM2005 800mm x 1, RM1605: Z-axis: 340mm x 150mm



Ballscrew supports: BK15 (fixed end), BF15 (loose end) BK12/BF12 for Z axis.
Linear rails: 20mm Hiwin profile rails hgw20ca Z0 or Z1 Preload.


i've seen that a lot of designs use pulleys to avoid resonances, i can use it later but now, i want a simple machine to start with.

Electrical parts:


Motion controler Purelogic PLC 4x-g2

http://purelogic.ru/shop/elektronnye_komplektuyuwie/platy_kommutacii_bob_konvertory_shim/ustrojstva_kommutacii_drajverov/ http://www.mycncuk.com/threads/4931-Purelogic-CNC-breakout-board-PLC4x-G2


Stepper motors:

for y-axis 2x Nema 23 270 Oz/in ref 23hs7430 from chinese "longs motor"
for x-axis (to buy, wich torque i need) mybe the same ref or 23hs7430 equivalent


Stepper drivers; Leadshine AM882 x 4
PSU Linear, toroidal transformer (rectifier, capacitors)
Spindle motor: 2.2KW with VFD and water cooling.



what do you think of this design ? do i need to change things ?
thanks for your help

happy christmas builds !

Boyan Silyavski
12-12-2015, 07:23 PM
Hi,
looks ok to me with exception being the Z. There i believe is better to use 2 pulleys and belt so you move the motor out of the wayof the Z plate. How much travel you achieve the way you have drawn it?

Also, if you have time, read in my signature build 1, how i made the Z moving plate. Basicaly you use 2 pieces of steel that raise the rails so the ballnut housing can pass freely and at the same time this makes the Z plate much stronger without additional hassle . That instead of making pocket for the ball screw, the way you have designed it. Not that its bad, simply what i say makes stronger Z plate


Also design it so that the spindle itself strengthens it , that why i think double bracket is better. SB here from the forum can make you 2 brackets for the price of 1.

Good Luck!

Clive S
12-12-2015, 07:29 PM
If you are mainly cutting wood the screws would be better with X and Y with 10mm pitch and 5mm pitch for Z.
The PLC-4X-g2 I don't think is a motion controller it is a BOB so has to run from the PP. or buy the matching motion controller as well to run it from the Ethernet port. Or look at the smoothstepper card.

Check the inductance on the motors before buying. Good one can be got from https://www.cnc4you.co.uk/Stepper-Motor-Plus/Nema23-3.1Nm.
In fact don't buy any electronics until you have the final design done and built.

I think you will also need some diagonal bracing on the frame.

Good luck with the build.

pierantoine
12-12-2015, 11:24 PM
Thank you for your answers..

Boyan,

the 2 spindle bracket is a very good idea, but I've also already seen your spindle block mount on your 2nd design (quite unusual...) similar to what jonathan has also done and for me it's perfect ! i'm tempted to follow this route in the future...

I have 150mm of travel
up 16793 down16794



i've seen your video: if i'm right between the black front plate and the rail there is another black steel part (red arrow), did you screwed the rail on this part (brace) or in the front plate ?
16791

i can do something like this the rails are screwed in the front plate, the green parts acts like a brace, maybe green parts and front place could be epoxy glued together ?

16792

Boyan Silyavski
12-12-2015, 11:32 PM
yes, its bolted, see bellow

http://www.mycncuk.com/attachment.php?attachmentid=12684&d=1404078308

pierantoine
12-12-2015, 11:49 PM
thanks again, what a fast answer !
ok here is the 2 brackets and the bolted rails from the front...

16796

pierantoine
09-01-2016, 03:23 PM
Hi,

I have
Do you think there is an improvement in mounting carriages on the side (position A : on the green plate on the pic) or behind the yellow plate (position B), I would say the position A is better as it is wider it doesn't reduce the working area and the external distance is 340mm instead of 200mm .but i can be wrong...

question 1 : For you what is the best position for carriage A or B ?

17155 17156


Question 2 for this axis 800mm wide do you vote for a 1610 or 2010 ?

Thanks for your advices

P-A

njhussey
09-01-2016, 09:10 PM
Definitely A...my axis (I call it Y axis) is 850mm wide and I use a 1610 ballscrew.

pierantoine
07-03-2016, 11:51 PM
i will take the steel frame route.
there is a useful post where i've seen nice frames from Dean and Boyan... i've made a mix :-)


17874
I've seen a post where Dean suggest 2 fixed ends for the ball screw
http://www.mycncuk.com/threads/8490-Steel-Gantry-Machine-1500x750x200/page3
"To give you an idea I recently built a Slaved axis 4x4 router with 20mm Dia 10mm pitch screws 1700mm length Fixed/Fixed ends with 2 x 3.1Nm steppers"..

the 2 beams supporting the rails can be slightly not parralels so an adjustment for fixed ends must be adjustable.
if you have made something like this i'm really interested.. thanks

pierantoine
16-05-2016, 02:23 PM
Thanks to every one after many hours of reading here is what I will make.
Here is the final version before starting the build and ordering the parts...
18464
the green frame could be made on a bench before the red / yellow...

adjustable mounting of the ends...
18465 18466


the green plate is 15mm ground plate maybe. i'have found 20mm but is not ground...
18467



between collet and frame (without bed)
218mm Z up an 68 mm Z Down.
1846818469


18470

For the plates 40, 50, 51 supporting the carriage would it be better to have 20mm not ground or 15mm ground ?


Do you have any comments before I start the build ?
Many thanks..

Lee Roberts
17-05-2016, 09:55 PM
For the plates 40, 50, 51 supporting the carriage would it be better to have 20mm not ground or 15mm ground ?


Do you have any comments before I start the build ?
Many thanks..
15 Ground, if you can't machine the mating surfaces.

Orange plate not needed, punch down from plate 51 to the nut (see pic example).

http://uploads.tapatalk-cdn.com/20160517/011b71eaa91f165bb6b3debe1b66616c.jpg

Or, if you really want to create a box, punch down from plate 51 to plate XX but make orange plate an X shape, locate nut in the center of the X (less material than current orange plate).

http://uploads.tapatalk-cdn.com/20160517/0f9e0844ab859a19f365021005bf5759.jpg

.Me

pierantoine
08-01-2017, 11:54 AM
last version with 80x120 aluminium profile.
20306

I have shorten the long axis to improve acceleration / whipping

i've read the forum and it seems that for 1500mm rm1610 could be a better choice than 2010 for acceleration reasons. do you agree ?

JAZZCNC
08-01-2017, 01:05 PM
i've read the forum and it seems that for 1500mm rm1610 could be a better choice than 2010 for acceleration reasons. do you agree ?

Yes for acceleration reasons but at that length your still high chance of whip with 16mm screw. And why would you need Higher acceleration than 20mm could provide.?
At this length 20mm is better suited and to improve things more go with 20mm pitch and use 2:1 ratio on the pulleys. Also have Fixed bearings at both ends.
This gives the advantage of lower screw speed and higher torque. The Fixed bearings make everything that little bit stiffer.

pierantoine
08-01-2017, 01:48 PM
Hello Dean ! thanks for your answer, yes both fixed and 2:1 with htd5 15mm was planned with ali pulleys 2 grubscrews and i've seen that you also advises that using Angular Contact Bearing from Fred worth the price so i'll do like that..

JAZZCNC
08-01-2017, 09:22 PM
Hello Dean ! thanks for your answer, yes both fixed and 2:1 with htd5 15mm was planned with ali pulleys 2 grubscrews and i've seen that you also advises that using Angular Contact Bearing from Fred worth the price so i'll do like that..

Would only use 2:1 ratio with 20mm pitch. If done with 10mm pitch the speed will be too low for router.

Boyan Silyavski
09-01-2017, 03:17 PM
Those green bars at front and back are not necessary. Dont worry about them. Its better to be open frame.

pierantoine
10-01-2017, 12:13 AM
Would only use 2:1 ratio with 20mm pitch. If done with 10mm pitch the speed will be too low for router.

for info i have nema 24 4nm steppers from cnc4u,
This is my first router it will be used for multicut materials (aluminium, wood, polycarbonate, etc...) and i'm looking for the best resolution before speed, but on the other end i understand that for some materials.. a fast feedrate could be need, so maybe an rm2010 could be an intermediate solution...

1 - we always see Z axis with rm1605, i assume with 1:1 or another ratio ?
2 - For my long axis to get a 1400mm travel i could use a 1550mm rm2010 with an 1:1 ratio but will it work correctly ? (whipping...)
if it wont work wich length would you use maximum ?
3 - If using RM2020 for cutting aluminium do you think i can have a good resolution with a pulley change to something like 4:1

Many thanks for your help..

Boyan Silyavski
10-01-2017, 12:40 AM
200 steps per revolution will move the axis 10mm if using screw xx20 and if 1:2 as Dean suggested. so for 1mm -20 steps. But as you will use 1/8 microstepping, then 160 steps per mm. So 1/160=0.00625 resolution.

pierantoine
10-01-2017, 10:23 PM
Hello !
okay that's clear thanks..

JAZZCNC
10-01-2017, 11:04 PM
200 steps per revolution will move the axis 10mm if using screw xx20 and if 1:2 as Dean suggested. so for 1mm -20 steps. But as you will use 1/8 microstepping, then 160 steps per mm. So 1/160=0.00625 resolution.

Boyan is correct that what I suggested gives same 10 movement per rev like 10mm pitch because of the ratio but don't be fooled by the micro step numbers because they are unrealistic in real world use.
Micro stepping is best thought of as way to increase smoothness not resolution. Yes the resolution is increased to some degree but not anything like the numbers suggested.

If you take it that Full step 200 is the minimum resolution then it's safer bet. What you actually get will be better but how much does depend on several factors like motor quality etc. If you work on calcualting using half the micro step used this will probably be more realistic figure.

Now if your going to use 2:1 ratio on the RM1610 ( I didn't realise you was doing this!) then whip won't be problem because the same apply's. Your screw will be spinning at half the speed than if 1:1 which is when whip occurs on 10mm pitch. Down side is that Max speed is halfed and with 10mm pitch this means for materials that require higher feeds you could struggle.

There are other factors that come into play which will dictate which way is best for you. The relation ship between Motors, drives, psu and Controller will greatly affect the kind of performance you'll get from machine.
Just saying I'm using 4nm motors (even if low inductance) isn't enough to determine overall performance.

For instance same 10mm pitch with 4nm High inductance motors with only 45v using Analog drives controlled by the parallel port with 25khz will perform much much slower than same size motors with Low inductance using 65v with Digital drives and Ethernet controller at 125khz. Upto half the speed.!!
Even if motors are low inductance the remaining factors still play big part in overall performance.

So this relationship between components affects how you go about it.

However I can safely tell you that on machine like yours if using 4nm motors with 65v with 80v digital drives and good motion controller, ideally Ethernet based that you will, provided build quality is to decent standard with good alignment of rails ballscrews etc with 2:1 ratio on 10mm pitch achieve 7500mm/min without too much trouble.


If using 2:1 on 10mm pitch gives what you require in speed terms then I'd go with 16mm.

pierantoine
12-01-2017, 08:40 PM
Boyan is correct that what I suggested gives same 10 movement per rev like 10mm pitch because of the ratio but don't be fooled by the micro step numbers because they are unrealistic in real world use.
Micro stepping is best thought of as way to increase smoothness not resolution. Yes the resolution is increased to some degree but not anything like the numbers suggested.

If you take it that Full step 200 is the minimum resolution then it's safer bet. What you actually get will be better but how much does depend on several factors like motor quality etc. If you work on calcualting using half the micro step used this will probably be more realistic figure.

Now if your going to use 2:1 ratio on the RM1610 ( I didn't realise you was doing this!) then whip won't be problem because the same apply's. Your screw will be spinning at half the speed than if 1:1 which is when whip occurs on 10mm pitch. Down side is that Max speed is halfed and with 10mm pitch this means for materials that require higher feeds you could struggle.

There are other factors that come into play which will dictate which way is best for you. The relation ship between Motors, drives, psu and Controller will greatly affect the kind of performance you'll get from machine.
Just saying I'm using 4nm motors (even if low inductance) isn't enough to determine overall performance.

For instance same 10mm pitch with 4nm High inductance motors with only 45v using Analog drives controlled by the parallel port with 25khz will perform much much slower than same size motors with Low inductance using 65v with Digital drives and Ethernet controller at 125khz. Upto half the speed.!!
Even if motors are low inductance the remaining factors still play big part in overall performance.

So this relationship between components affects how you go about it.

However I can safely tell you that on machine like yours if using 4nm motors with 65v with 80v digital drives and good motion controller, ideally Ethernet based that you will, provided build quality is to decent standard with good alignment of rails ballscrews etc with 2:1 ratio on 10mm pitch achieve 7500mm/min without too much trouble.


If using 2:1 on 10mm pitch gives what you require in speed terms then I'd go with 16mm.

Really thanks for your complete answer !

pierantoine
12-01-2017, 08:43 PM
to resume for everyone from slower to faster if i'm right...
rm1610 with 2:1 pulley ratio 7.5m/min feedrate
rm1620 with 2:1 pulley ratio 15m/min feedrate
rm2020 with 1:1 pulley ratio maybe 30m/min

for information
stepper : 4nm cnc4u
controller : MESA 5i25 + 7i76
driver : AM882 Leadshine@70V

and for resolution : 200 steps with 2:1 ratio and 10mm pitch -> 400 steps for 10mm -> 40steps/mm ->25 microns resolution (more with the micro step effect...)

Boyan Silyavski
13-01-2017, 11:51 PM
I believe ~10m.min is very good for a DIY machine, not servo driven. i would not trust steppers for faster speeds. Also there is that thing that the machine to benefit from the fastest speeds it must be big. See the formula at the bottom of that page http://www.prusaprinters.org/calculator/

My machine is servo driven and as snappy as a bullet , acceleration 3000mm/s2 and at >15000mm/min looks bloody dangerous and i am very very careful not to make a mistake. Stepper driven machine i am not sure that could achieve that. Think more 1000mm/s2, if i am wrong sb correct me.

What i am saying is you need distance to achieve that theoretical speed. Your aim should be to achieve higher acceleration, than speed.

PS Plus all must be perfectly rigid, especially Z to benefit from that speeds

JAZZCNC
14-01-2017, 12:57 AM
I believe ~10m.min is very good for a DIY machine, not servo driven. i would not trust steppers for faster speeds. Also there is that thing that the machine to benefit from the fastest speeds it must be big. See the formula at the bottom of that page

Fact using stepper doesn't mean can't have higher feeds. The feeds depends on pitch. Could have 50,000mtr/min quite easily and still have acceptable acceleration with the right pitch.
What stepper does restrict is having High Feeds and High Resolution. The price you pay to achieve high speed is Low resolution.

Also To answer your question if Stepper can achieve 3000mm/s/s and 15,000mm is yes.!! . . . . . . Could it do this and hold the same resolution at your size/weight of machine then No.!! . . . Hence why you wouldn't fit steppers.

However on smaller machine with lighter gantry then yes it's possible with careful selection of pitch and motors etc.

Good machine will have nice balance of speed and Acceleration. IME for Medium machine size which this is, 10,000mm/min and 1500mm/s/s is all thats required.

Boyan Silyavski
14-01-2017, 01:15 AM
Good machine will have nice balance of speed and Acceleration. IME for Medium machine size which this is, 10,000mm/min and 1500mm/s/s is all thats required.


I think also that one should be very happy he found this forum and all information here. I will give a reward if sb finds me commercial machine that is not servo driven that could actually achieve that, speaking about cheap Chinese knockoffs and other ebay style machines.


Actually for intricate carvings on aluminum i lower the acceleration to 1000-1500 for better results. Its difficult to have all in one. Its not a big deal, just change one number in the controller. For wood i could push it even higher than that 3000 but not a real need.

JAZZCNC
14-01-2017, 02:55 AM
I will give a reward if sb finds me commercial machine that is not servo driven that could actually achieve that, speaking about cheap Chinese knockoffs and other ebay style machines.

Not quite sure what your saying here Boyan.? . . . . Are you trying to say this isn't achievable with steppers.? Because if so then your very wrong.!
Just about every machine I build will or should say is designed to achieve these speeds or very close to them. Some Smaller machines will achieve well in excess of 10mtr/min and 1500mm/s/s.
Don't actually run them or tune them at those speeds for reliabilty reasons but they will do it if required.

I find above 2000/s/s gives negative affect to machine as it's far too jerky on short moves leading to poorer surface finishes.

Davek0974
14-01-2017, 09:28 AM
Your aim should be to achieve higher acceleration, than speed.



This is the vital fact, I have found the negative side of this equation - low accel/high speed and what it does with Mach3 is give terrible contour following because of the constant velocity engine Mach uses. Its not just a little bad, it's the difference between turning out scrap and turning out parts.

I have retuned with lower speed as higher acceleration was too jerky on my old Bridgeport conversion and its all working nicely now.

JAZZCNC
14-01-2017, 11:19 AM
This is the vital fact, I have found the negative side of this equation - low accel/high speed and what it does with Mach3 is give terrible contour following because of the constant velocity engine Mach uses. Its not just a little bad, it's the difference between turning out scrap and turning out parts.

I have retuned with lower speed as higher acceleration was too jerky on my old Bridgeport conversion and its all working nicely now.

Exactly Dave. Like all things in life Balance is the Key. Good machine is tuned with nice balance of speed and Acceleration.
Getting wrong either way will have negative affect on machines performance in different areas of work type.

For instance your case of high feed but with low accelration caused corner rounding. This was on relatively small parts with tight radius on small machine. On larger machine like router with larger parts say doors or shaped work tops etc with no or large radius sections then high feeds with Lower Accel will be ideal.

Different strokes for different folks, Or machines in this case.!

This is why Mach3 and some other Control software will let have you have different profiles which can be used to have machine tuned or setup differently to optimise the type of work your doing.

Boyan Silyavski
14-01-2017, 11:24 PM
Not quite sure what your saying here Boyan.? . . . . Are you trying to say this isn't achievable with steppers.? Because if so then your very wrong.!


I am saying you are our hero and i am happy that i found this forum and we learned to build machines that are superb for the money and better than crappy commercial ones where usually they have some major flaw in design.

Clive S
14-01-2017, 11:57 PM
I am saying you are our hero and i am happy that i found this forum and we learned to build machines that are superb for the money and better than crappy commercial ones where usually they have some major flaw in design.creep :cower:

JAZZCNC
15-01-2017, 12:41 AM
I am saying you are our hero and i am happy that i found this forum and we learned to build machines that are superb for the money and better than crappy commercial ones where usually they have some major flaw in design.

Ah ah now your just blowing smoke up my jacksy.! :hysterical: . . . . . I understand.:thumsup:

pierantoine
01-04-2017, 02:29 PM
Here is my Wiring diagram done with tinyCad and some Dean pictures :orange:

the Bob is a Mesa 5i25+7i76 (LinuxCnc)

1 A Pilz S4 is not in use but it could be if needed
2 a second Contactor could be use if necessary for the 70v steppers supply
3 The limits directly controls the safety S5 pilz relay.
4 The S5 cuts all signals immediatly and after a delay cuts the main on the VFD
5 Home could be connected to one input but i have 32 inputs so more than needed.

i've already tested the inductive sensor action on pilz s5, it works.. that may not be the best solution ?

Comments are Welcome !
thanks for your help

Clive S
01-04-2017, 06:07 PM
Its nice to see other Mesa users on here. Do you need the 10V power supply ie can't this be got from the VFD if it is only for spindle speed?

I am not sure I would connect the 24V -ve to the star Gnd. I would however connect the 5v and 24v -ve together.

Make sure you have the switches on the 7i76 in the correct positions, W2 needs to be in the right hand position if you are using external 5v supply Also on the 5i25 W1 and W2 should be in the down position (default) when using external 5v to the 7i76) otherwise the magic smoke will appear. If you know this already then fine.

PNP sensors work best on the 7i76 although I do use NPN.

You might find this useful https://forum.linuxcnc.org/media/kunena/attachments/3278/7i76_Anschluss_2016-12-03.pdf

I could not open your drawing in acad so find it a bit hard to read.

pierantoine
01-04-2017, 11:33 PM
Its nice to see other Mesa users on here.
:beer: Hi Clive thanks for your help !

Do you need the 10V power supply ie can't this be got from the VFD if it is only for spindle speed?
I didn't know this solution, I will take a look.


I am not sure I would connect the 24V -ve to the star Gnd. I would however connect the 5v and 24v -ve together.
Ok thanks that's a question i had in mind..


Make sure you have the switches on the 7i76 in the correct positions, W2 needs to be in the right hand position if you are using external 5v supply Also on the 5i25 W1 and W2 should be in the down position (default) when using external 5v to the 7i76) otherwise the magic smoke will appear. If you know this already then fine.
yes that is correct but thanks again, it could be useful for another reader of the post.


You might find this useful https://forum.linuxcnc.org/media/kunena/attachments/3278/7i76_Anschluss_2016-12-03.pdf
Oh Yes ! Really interesting !


I could not open your drawing in acad so find it a bit hard to read.
I've added a pdf it will be more clear