I want to DIY.. some questions please!

[Tenson]

I have a CNC3040 from eBay and it works great. It has got me on the learning curve and given the confidence to try to make my own machine now :)

So I have a few questions if I may:


1) What are the pros and cons of ball-screw vs. lead-screw?

2) How important is supported rail vs. unsupported? I plan to build a machine that will have a 120cm x 60cm cutting area.

3) What factors dictate the max. speed of a machine? Is it the motors or the drivers, and how can I estimate the speed I'll get?

Thanks!

[Jonathan]

1) Ballscrews have far greater efficiency - about 90% vs 50% and significantly lower, but not zero, backlash however unlike leadscrews due to the use of rolling not sliding friction they can be preloaded to obtain practically zero backlash.
2) Extremely important. An unsupported rail will bend significantly more when a force is applied, ultimately leading to tool deflection and thus poorer accuracy. Supported rails are supported along the entire length, hence the force is largely transmitted to the frame which is significantly stronger than the rail and therefore deflects less. Does depend on the orientation, but either way a 12mm supported rail is probably stronger than a 25mm supported... huge difference.
3) Motors, driver voltage, ballscrew pitch, ballscrew diameter, drive ratio, mass of moving parts, type of rails used, preload of rails ... etc. You can estimate it using this calculator:

http://www.google.com/url?q=http://www.mycncuk.com/forums/showthread.php/1524-What-size-stepper-motor-do-I-need&sa=U&ei=HFBrT8WdDqPU0QW2p4jdBg&ved=0CAQQFjAA&client=internal-uds-cse&usg=AFQjCNEhg854ywip3r8mzkjSyaojkuQQKA

[Tenson]

Thanks for the info Jonathan.

Could somebody please explain to me what is meant by 'fixed side' and 'supported side' with reference to ballscrews?

I don't understand why the same simple bearing can't be used on each end.

[Robin Hewitt]

I don't understand why the same simple bearing can't be used on each end.

The ballscrew is going to push axially (lengthwise). It needs something to push against so one bearing must be fixed axially.

If you fix both ends axially you have to wonder what will happen when it expands and contracts due to temperature change.

Fixing at both ends is good if you can hold it in tension with springs, otherwise you have to let one end free to slide.

[JAZZCNC]

Thanks for the info Jonathan.

Could somebody please explain to me what is meant by 'fixed side' and 'supported side' with reference to ballscrews?

Robin explains it very well but a picture paints a thousand words.!!

Pics below show ballscrew with fixed end on right and a floating end on left, you will see fixed end as threads on ballscrew and the block is wider than the floating end because it has 2 angular contact bearings working togehter to handle the axial loads. The fixed end has just a normal bearing which is allowed to move inside the mounting block if any thermal expansion happens, it is still attached to the ballscrew via a small circlip so it can't fall off.

Simplizzz really.!!

[Tenson]

Thank you, nice picture! This machine is getting more expensive ;) Still, good tools earn their value.

[Tenson]

Hi chaps,

Which of these illustrations is the correct way to mount an X axis with supported rails?

(Yes I know there is no Z axis ;) )

http://i608.photobucket.com/albums/tt169/tenson_uk/X-Axis.jpg

Also a question I feel stupid asking, but what is commonly called the X and Y axis? On my CNC3040 the Y axis is the length of the table and the X axis goes across. I've seen people talk about it the other way though.

[Jonathan]

What you have pictured is generally called the Y/Z axis, although as you say it does vary. I always call X the axis the gantry runs on, so generally the longest axis.

The orientation of the rails in the image on the left is considered better since the deflection due to each bearing is more even. For forces in the direction pulling the bearing off the rail, not surprisingly, the bearings are weaker and will deflect more. If the open sides are facing, as you have drawn, this evens out so the deflection will be the same for all axis which is much better than having significantly different deflection which is what will happen if the rails are both the same way round.

[Tenson]

Thanks for the feedback. You mean like this?

http://i608.photobucket.com/albums/tt169/tenson_uk/Y-Axis.jpg

[Tenson]

Hmm.. I think I'll go for unsupported rails on the Y axis to keep the gantry weight down, plus build simplicity.

[JAZZCNC]

Hmm.. I think I'll go for unsupported rails on the Y axis to keep the gantry weight down, plus build simplicity.

BIG BIG BIG MISTAKE if you do this unsupported rails are a complete waste of money.

Don't be affraid of weight, Mass really helps when it comes to cutting, It absorbs resonance and gives far better finish.
If your trying to cut weight to gain speed then you are sizing components wrong. The feeds you require for the machines intended purpose should be got from choosing components to allow it.
Ballscrew pitch and motor selection along with drives and correct voltage all play a part combined with build quality and component quality. Get anyone wrong and you comprimise the machines abilty in some way.

In your case unsupported rails with a light weight gantry of any resonable width would be a disaster for cutting any material denser than softwood, and even then with lower DOC.

DONT DO IT. . . . FORGET UNSUPPORTED RAILS.!!!!!!

[Tenson]

Ooookay, I will use supported :)

What is the best way to drive the 150cm long X axis? Should I use a central ball-screw or two ball-screws, one on each side? If the latter, what is the best way to power them, i.e. 2 motors connected to 1 controller?

[motoxy]

2 ball screws means that both sides are controlled and this will stop the gantry squewing. I have a motor on each side. Mach 3 can link the two together. A single motor linked by belts to the two screws may have been a better choice as the screws are guaranteed to rotate equally.

you will find quite a few threads about this.

Bruce

OO look at me showing off and giving advice:redface:

[Jonathan]

OO look at me showing off and giving advice:redface:

Good advice :)

It's possible to put two motors on one driver if they're mechanically linked, but I wouldn't advise it. Either put one motor on each screw, or one motor and a long timing belts to link the two screws. The former should go a bit faster, since you have twice the input power with two motors, but in practice I doubt it will go that much faster as you have to guarantee neither motor stalls. Clearly if one does then the gantry skews, which depending on the magnitude could damage things. Once the machine is tuned properly either is fine as the motors will never stall.

I used two motors and two drivers since it would require a very long belt and several idlers to link the rotating ball-nuts.

[JAZZCNC]

Ooookay, I will use supported :)

What is the best way to drive the 150cm long X axis? Should I use a central ball-screw or two ball-screws, one on each side? If the latter, what is the best way to power them, i.e. 2 motors connected to 1 controller?

Eh Eh sorry if I shouted a bit load but can't stress enough what a bad idea unsupported rails are.!!

RE Single or twin screws on X. If Y axis much wider than say 600mm then personally I'd go for twin screws.

Powering them you can either have 2 motors each with it's own drive, 2 motors on 1 drive will work but not correct way to do it and could give problems. . Another don't do it.!!
You could use 1 slightly larger motor and connect the screws together with a timing belt, this is my personal favorite and how i've built my machine. It works perfect so don't be put off by timing belts.

Both work ok but you have to be aware of the downsides to each and choose which you prefer to live with.!
Using 2 motors works ok but comes with the issue if one motor stalls the other keeps going which can do major damage to the machine if happens at high feeds.
It can also be a problem keeping the screws in sync if both motors are not tuned perfectly in sync with each other and one runs ahead or behind the other, this will result in slowly twisting the gantry thou can be corrected by regular homing and squaring, but still wrong and not wanted.
The way to avoid these issues is to not stress the motors so you effectively need to run them below there full potential because you can't afford any missed steps or threat of stalling from lack of torque at high feed rates or rapid moves. The quality of build regards ballscrew aliagnment as to be spot on, which it should be on any machine, but even more so with twin screws/motors as no sticking or binding can be tolerated for fear of stalling motors at high feeds where torque is lower.
Mostly this is not a problem if your aware of it before building because you can choose higher lead ballscrews or build in gearing again via timing belts to the motors to achieve the speeds & feeds you require and keep the motors operating well below the point they start running out of torque.
The sync problem is just a slow setup and tuning job which can either take 2mins or 2 weeks depending how fussy you are.!! Warn you thou it can lead to premature baldness.!!. . I'm down to my last few strands. . :naughty:

The 1 motor pritty much eliminates all this hassle with perfect sync and also means you can run the motors closer to there max. It's also very easy to change pulleys to gear the system up or down.
The only downside is it's not as easy to build because you have build in the belt system and it's tensioning/tracking system and also have the ballscrews machined different from standard machining to add extra bearing support for the pulleys which will put greater side ways force on the ballscrews effectively trying to bend them when the belts are tensioned.

In my experience this is all worth the effort because when done and settled down it never loses position or sync and ounce setup thats it forget about it and never needs doing again with absolutly no fear of racking the gantry unless the belt snaps. (which I've only snapped 1 in over 3 yrs and that was because I left the pulley lose)
If good quality belts are used then belt stretch is minimal and with correct tension and good tensioning system never an issue and very very accurate.
It's no cheaper(well slightly) than 2 motors because you have the belts and extra bearing supports to consider.

Like I say both work ok if your aware of there limits and tendency's it's just I personnaly favor belt system.

Oh and at 150cm your in a funny area regards pitch and screw dia.???? I'm pritty sure you'll want 10mm pitch which often (well from china anyway's) only comes in either 16mm or 25mm dia they don't do 20mm dia only in 5mm pitch.
At 150cm 16mm is on the very edge of being to thin and will give big potential for whip at high speeds. 25mm would be better but comes with the cost that it requires more torque and power to over come the extra mass and inertia of the screw/nut when accellerating and deaccellerating. This means bigger motors which require larger power supply's and drives that can handle the power so it basicly cost's more money.
You can over come this by either taking the route like jonathan did and use rotating ballnut on 16mm screws, but again this will come with expense and i'm sure jonathan will tell you exactly how much and all it's bennifits. . . OR . . bite the bullet and pay for the larger screws and use with nema 34's and correct drives, psu etc. . . . OR . . .use 20mm 5mm pitch and gear 1:2 to give an effective pitch of 10mm at the cost of slightly less torque. ( My machine use's this system thou with nema 34 motors) This gives a slight advantage as it's easy to gear back down and double the resolution of the machine without comprimising the area where the motor works good. It can be done the other way round and gear 10mm pitch to give 5mm resolution but the motors work in an area which is not ideal.!
It's horse's for course's really so take your pick.!!!!!

Hope this helps.!

[Tenson]

Thanks for the detailed reply.

So the problem with a single central ballscrew is that the gantry can twist a little back and forth? While the problem with using dual ballscrews + motors is that it can twist should one motor stall? I don't think my budget with allow the custom machining of ballscrew to fit a timing belt. Where did you get yours from JazzCNC?

The X axis will be 80cm wide. How about running a single 2505 ballscrew on the Y axis with a bigger motor? There is a nema 34 12Nm motor on Zapp Automation. However it needs 6.2A per phase, so what is a suitable driver for it? Is it best to use 2x 6A drivers, one per phase?


What is the problem with 5mm pitch on a 150cm axis? I'd have thought it best to use the same screw pitch on both X and Y axis so they have the same speed, provided the motors have enough torque.

Oh, and finally how can I build an auto tool change? I'd really like to have that at least to choose between 2 bits. Hmm.. I just searched that and it looks like I'd need a special chuck with gearing to transfer power from the motor, plus a system to open and close the chuck. PITA.

[JAZZCNC]

Ok the screws come from Chai at linearmotionbearings which you'll find on ebay. He's a great guy and very trust worthy. It's best to let Chia do the machining as he doesn't charge very much and will machine to your dimensions if asked. I do this with him all the time.

Yes single screw can twist the gantry when cutting at the outer edges, the wider the gantry the worse it will be. At 80cm your on the edge IMO.
If you do proceed with single then make the distance between the bearings as wide as possible to help combat this and have a strong wide cross brace undernieth.

The 25mm screw would work but you don't need 12nm motor 6nm would be ok.
I think you miss understand the phase/drive relationship you only need a drive which can handle the motors rated amp's, so in the case you gave 6.5A. you don't need 1 drive per phase.
The thing you have to be aware of is the differnce between nema 23/4 and 34 motors. Larger 34 motors tend to spin at lower speeds than 23/4 motors. typicly a 34 will runout of puff around 1000-1100rpm where as 23/4's will spin 1500rpm and above. The difference comes from how they produce and carry the torque up the rev range, 34's tend to hold higher torque further up the rev range then drop off really quick as it peaks. 23/4's will carry torque further up the rev range all be it a lower rate and gradually drop off to nothing.
Also 34's need really high voltage to get best performance. Voltage is key to speed and torque in both 23/4's and 34's but 34's require much higher to get best performance.

Unless you have a very well equiped workshop or lots of money then forget autotool changer far too expensive.!! . . . . better alternative is to buy Gerry's Mach screen it's about as close to autotool changing as you'll get without the expense, find it here and it will be the best $20 you have ever spent. http://home.comcast.net/~cncwoodworker/2010.html

[Tenson]

Hi Jazz,

On the datasheet for the motor is said 6.2A (phase) which I take to mean each winding will draw 6.2A. So if I wire in parallel it will draw 12.4A. In series will will need about 3A but twice the voltage. Won't it?

Each side column of the gantry rides directly on the centre point of the linear bearings, and I was going to use solid 12mm alu sheet as a cross-brace underneath. In that case I guess maybe a single central ball-screw with a powerful motor would be okay. However you point that a nema 34 spins slower is worrying as I want all axis to be capable of the same speed. What do you think, is it okay to use a nema 34 only on the Y axis?

Why is running two motors from one controller not a good idea? Provided it can output the required current and voltage it should be okay, won't it?

[Jonathan]

On the datasheet for the motor is said 6.2A (phase) which I take to mean each winding will draw 6.2A. So if I wire in parallel it will draw 12.4A. In series will will need about 3A but twice the voltage. Won't it?

No... in series it's 6.2A and in parallel 12.4A. Until I finish my own drivers you're not going to find drivers that will deliver 12.4A, so have to use 6.2A. As you say, that means a higher voltage is required ... so the only driver worth buying for that completely excessive and uncalled for Nema 34 motor is mains voltage:

http://www.zappautomation.co.uk/2m2280n-high-voltage-stepper-driver-p-3.html?cPath=9_3_132


Each side column of the gantry rides directly on the centre point of the linear bearings, and I was going to use solid 12mm alu sheet as a cross-brace underneath. In that case I guess maybe a single central ball-screw with a powerful motor would be okay.

Not at all - how strong you make the gantry with a single screw is largely irrelevant with regards to racking (although in general you clearly want to make it as strong as you can afford) since you're stopping the gantry rotating about the central ballnut, not stopping it bending. To rotate (rack) the rails do not have to deflect much at all, so the only way to do it would be to have a very large spacing on the X-axis linear bearings, however doing that you're just compensating for bad design in the first place. Use two ballscrews - I guarantee you'll regret one.

This happens with almost every build log - starts off thinking one ballscrew will be fine, then by about page 3 we manage to persuade that it's not.


However you point that a nema 34 spins slower is worrying as I want all axis to be capable of the same speed. What do you think, is it okay to use a nema 34 only on the Y axis?

You won't need a Nema 34 motor.

The only way to get a Nema 34 to have decent torque at higher speeds is to use the mains voltage drivers, so lets add up the costs:

Driver, £176
8.6nm motor, £108
Two RM2510 ballscrews from eBay, roughly £190
4 * 30T HTD pulleys, £22
Long (to link screws) and short timing belts, £27
Ballscrew end bearings & mounts, £50 (bit of a guess but shouldn't be far out, maybe more due to the long belt?).

Total £563

The other option is, as Jazz suggested, to use rotating ballnuts. See here for my experience with this setup:

http://www.mycncuk.com/forums/showthread.php/3340-Rotating-Ballnut-design-ideas/

(On the last page you'll find pictures of what I'm suggesting below, but the rest is worth reading for a more in depth explanation of why they work so well.)

To get the same (feederate and acceleration) as the above, probably better and certainly more than you'll ever need from a rotating ballnut setup you'll need:

2* m752, £80 (eBay)
PSU for 2 motors is shared with Y/Z motors so about £15 extra.
2* 3nm motor, £50 (eBay)
Two RM1610 ballscrews from eBay, £150
4 * 30T HTD pulleys, £22
2 * short timing belts, £10
4 * 7206 angular contact bearings, £40
Ballscrew end bearings & mounts, DIY for £10.

Total £376

So if I wanted to I could charge £187 to machine two rotating ballnut mounts and you'd end up with a better system for the same price. Realistically I'd charge less so it would be cheaper overall.* When I've worked it out before the price difference has been greater since I worked it out with two nema34 motors, not one, but either way it's a significant saving.

Or for about the same price (£376) you could just use the RM1610 and spin the screws as normal. I'm confident that will be adequate and a better option than RM2005 since the critical speed (speed at which the screw starts whipping, so the limiting speed) for a RM2005 screw is well under twice the critical speed of a 1500mm RM1610 screw so it wont actually get a better feedrate. Plus it clearly requires much more torque to accelerate a 20mm ballscrew compared to 16mm, so you'd need a bigger motor anyway.

In your situation I'd go for RM1610 and either rotate the screw or the nut. I went for rotating nuts since my machine has 2000mm ballscrews so rotating a 16mm screw wasn't an option...and equally because I wanted to try something new(ish). I've left my machine on 10m/min with 1m/s^2 acceleration on X and 1:1 ratio. It will do about 15m/min, even more if I change the pulleys, but 10m/min is already plenty so no point stressing the components. That's with the same 3nm motors and drivers, so you can expect better.

To give you an idea of how much more stable a system rotating ballnuts are; I recently didn't notice that one of my ballscrews had worked loose. So one end wasn't fixed at all and the other wasn't tight either. I only noticed the problem when after a few hours cutting aluminium I leant over/on the ballscrew to remove the tool and it moved... i.e. it made no difference! Try that with a rotating screw and you'd probably have ended up with a bent ballscrew.


Why is running two motors from one controller not a good idea? Provided it can output the required current and voltage it should be okay, won't it?

Think what happens if one motor has a slightly different load.. then stalls.


*Other vendors are available.

[Tenson]

No... in series it's 6.2A and in parallel 12.4A.

Hmm, surely the parallel connection will draw twice the current of a single winding, not twice that of the series connections. Similarly the series connection will draw half the current of a single winding, not half that of a parallel set-up.



Not at all - how strong you make the gantry with a single screw is largely irrelevant with regards to racking....
To rotate (rack), the rails do not have to deflect much at all...
This happens with almost every build log - starts off thinking one ballscrew will be fine, then by about page 3 we manage to persuade that it's not.


Lol, okay I get it :tup: I'll go for two ball-screws.

I still wonder though; why isn't it better to use the same screw pitch for all axis, so that providing the motors don't run out of torque, each axis will have the same speed and resolution?

I really don't want to go the rotating ball-nut route, nor linking the ball-screws with a belt. Going with dual ball-screw I'd like to use two motors. I'll bear the tedious tuning procedure when I first get it up and running. At least I know to expect that now thanks to you guys!


Think what happens if one motor has a slightly different load.. then stalls.

My question was really about why it is better to power each motor and ball-screw from a separate drive circuit, rather than just using both motors from one circuit. The motors can surely still individually stall, even with two driving circuits.

[Jonathan]

Hmm, surely the parallel connection will draw twice the current of a single winding, not twice that of the series connections. Similarly the series connection will draw half the current of a single winding, not half that of a parallel set-up.

In parallel the current is twice the current for one winding. In series it is the same as the current for one winding as you have the same current flowing through both.
GCSE Physics...


I still wonder though; why isn't it better to use the same screw pitch for all axis, so that providing the motors don't run out of torque, each axis will have the same speed and resolution?

Each axis has a different mass and length of ballscrew, both of which affect the feedrate. So the same screw pitch on both will result in different speeds since other variables are still different. The resolution will be the same.

Your Y-axis is 600mm (?), so I'd still go for 10mm pitch on Y.


I really don't want to go the rotating ball-nut route

Why?


My question was really about why it is better to power each motor and ball-screw from a separate drive circuit, rather than just using both motors from one circuit. The motors can surely still individually stall, even with two driving circuits.

Same answer, but to put it bluntly if one steppers stalls when you have two on one driver then you can say goodbye to the driver. If one stalls with two drivers it's a bit annoying as the gantry is bent, but not the end of the world.

[JAZZCNC]

Tenson Jonathan's right on all counts here. I gave you the options he's giving you specifics about the rotating nut. If your wise you'll go with the rotating nuts at this length for best speeds verse's cost's.
The rotating nut will give you best of both worlds and save you some money.

The other reason for not using 1 drive with 2 motors is that if the pulse timing doesn't get to each motor at exactly the same time then you will slowly drift either infront or behind on 1 axis so all in all it's a lose lose situation and at best you'll be inaccurate and worst be letting out the magic smoke. . . Your choice.!!

[Tenson]

Hi guys,

Jonathan, if you are genuinely willing to machine two rotating ball-nut/motor mount things for me at a good price then I'll go for it :)

I'm sorry to be persistently stupid but I still don't understand the desire for 10mm pitch. This is my thought process: Two screws with the same pitch, turned at the same RPM, will give produce the same speed and distance of movement. In practice, one axis will have a greater load than the other, so it will need more torque for the screw to maintain the same RPM. However, with adequate torque (two screws, two motors) there should be enough torque to do that. In fact I'd expect a larger screw pitch to need even more torque. If the 10mm screw moves twice the distance with one turn, it will require twice the force to turn it, not to mention the greater mass of the load.

Where is my thinking going wrong?

[JamesMoore]

A ball bearing screw is just that: a screw which runs on ball bearings. The screw and nut have matching helical grooves or races, and the ball bearings recirculate in these races. There is no physical contact between the screw and the nut. As the screw or nut rotates, and the rolling balls reach the trailing end of the nut, they are deflected or guided from this "pitch" contact by means of a return tube and returned to the leading end of the circuit.

[JAZZCNC]

I'm sorry to be persistently stupid but I still don't understand the desire for 10mm pitch.

It's all about the speed vs torque. Remember with steppers as the speed rise's the torque drops away.
So with a 5mm pitch screw and usable max torque for cutting feeds translated to rpm is around 1000rpm with nema23 motor will give about 5mtr/min if your lucky . Taking into account the safety margin for twin direct drive screws the safe area will probably be around 4-4.5mtr/min(Obviously this will depend on mass and friction of each axis). Now at these feeds you'll have low torque so any deep roughing or hard material will require even lower feeds.

With a 10mm pitch the same 1000rpm gives you 10mtr/min with a similiar amount of torque. So obviously at the same 5mtr/min the motors are spinning half the speed so far more torque available so deep roughing cuts are no problem if the spindle and machine can handle them.
The other reason as Jonathan pointed out is 10mm pitch means for the same speed the 10mm pitch screws are spinnig half the speed of 5mm so less chance of whip.
Thou to be honest I don't hold much stall by this and Jonathan and I have debated this before.!
While in principle he's right in physics terms but in practice it doesn't quite work that way and my machine is living proof as it use's all the things Jonathan disagree's with like: twin 20mm x 5mm pitch geared 1:2 with nema 34 and non AC drives on 80V.
It will rapid at 12mtr/min if need be thou I restrict it by tuning to around 7mtr/min to save on screw wear as I don't require high rapid feeds and this allows higher acceleration which is much more usefull.
It will happly cut at 7mtr/min and if you check the vertical post I put up you'll see it cut 10mm deep 1 pass in dense resin inpregnated MDF with no problem, it even cut full depth 13mm for half the length untill the very very worn 6mm cutter snapped and even then the machine didn't stall.
So the critical speed issue is a non starter to me and I'm sure by the very fact of human nature most won't have the will to resist using all the available speed the 16mm x 10mm will give them so whip will be an issue because there's no strength in the screw. With 20mm you have more meat so better able to handle the slightly higher spinning if geared like I do.

All that said if you don't require high resolution then I would still advise you to go 16mm x 10mm pitch as it's far more versatile and cost affective. Rotating nut, direct drive or connecting belts is an individual's decision.!! . . . All I'll say is you still have the sync issue whether rotating or spinning screw.!!. . Thou Jonathans obviously then man to tell you exactly how it differs to direct drive spinning screws.

[Tenson]

I see, so the motors drop in torque is not proportional to the increase in RPM? If you get more than double the torque at half the RPM then I can understand the argument for 10mm pitch vs. 5mm.

Jazz, have you tried not using the 1:2 gearing and just spinning the motors faster? If you found the slower motor speed is better then I should just follow your advice and go for 10mm pitch ;)

I'd like to go for rotating ballnuts if Jonathan will be kind enough to make the parts for me, as it would be nice to keep the drive motors within the machine frame, rather than stuck on the end. Also I could use ballscrew without end-machining since it would just be held in a clamp.

I can get 2510 ballscrew from Zapp, so can you make a rotating ballnut assembly for that Jonathan?

[Jonathan]

I see, so the motors drop in torque is not proportional to the increase in RPM? If you get more than double the torque at half the RPM then I can understand the argument for 10mm pitch vs. 5mm.

A good approximation is that the torque is fairly constant upto a particular rpm (the 'corner speed') after which the torque does drop proportionately. We've come to the point where hand waving explanations are insufficient as there's too many variables. It will depend on the chosen motor and thresholds.

The graph in the 3nm motor datasheet is helpful:

http://www.zappautomation.co.uk/redirect.php?action=url&goto=www.slidesandballscrews.com%2Fpdf%2Fsteppermo tors%2FSY60STH88-3008BF.pdf

(They've just gone up in price by 20%!!)

Also Irving's spreadsheet is very helpful to understand which factors affect the feedrate:

http://www.mycncuk.com/forums/showthread.php/1524-What-size-stepper-motor-do-I-need

Even when critical speed isn't a problem (as is the case with the rotating ballnut) you'll find the 10mm pitch screw wins over 5mm as the torque required to accelerate it is less, due to the kinetic energy of the rotating screw being 4 times less if the screw is only rotating at half the speed.


I'd like to go for rotating ballnuts if Jonathan will be kind enough to make the parts for me, as it would be nice to keep the drive motors within the machine frame, rather than stuck on the end. Also I could use ballscrew without end-machining since it would just be held in a clamp.

I'll work out a price and let you know asap. Though it's theoretically possible to get away without end-machining with rotating nuts I wouldn't advice is as the surface area of the screw in contact at the clamp will be very small. It adds very little to the price with linearmotionbearings2008, so I got my ballscrews cut with the standard end machining, except the same on both ends with the threaded portion. That way you can put a nut on either end to clamp it.


I can get 2510 ballscrew from Zapp, so can you make a rotating ballnut assembly for that Jonathan?

Yep I can do any size, that's what I'm using:

55595560

If you want 2510, although I think you'll be fine with 1610, you'd be much better off getting it from linearmotionbearings2008 on eBay. He sells 1605,1610,2005,2505 and 2510.

This is the version I did which is suitable for 16mm or 20mm screws:

55575558

[Tenson]

Though it's theoretically possible to get away without end-machining with rotating nuts I wouldn't advice is as the surface area of the screw in contact at the clamp will be very small.

I thought about that too, but it seems easy to overcome. Double-nut the end of it or pot the thread then clamp. I hate waiting for long deliveries so I'd probably get it from Zapp!



Yep I can do any size, that's what I'm using:

55595560

If you want 2510, although I think you'll be fine with 1610, you'd be much better off getting it from linearmotionbearings2008 on eBay. He sells 1605,1610,2005,2505 and 2510.

This is the version I did which is suitable for 16mm or 20mm screws:

55575558

The bracket on the right of the picture (and in the final picture) looks good for me as I'd want to mount it under the gantry. If you think 1610 will be enough then lets go for that, though the saving seems small (about £15 each screw).

I've started ordering some parts. My plan is to make the Z axis first, then the X and then finally make the machine frame and Y axis. I ordered a 2.2Kw spindle, 250mm ballscrew and some rail and bearings. I'll then make some 10mm plates to hold it all together.

[Robin Hewitt]

Even when critical speed isn't a problem (as is the case with the rotating ballnut) you'll find the 10mm pitch screw wins over 5mm as the torque required to accelerate it is less, due to the kinetic energy of the rotating screw being 4 times less if the screw is only rotating at half the speed.

Okay you have less rotational inertia from the screw, but twice the pitch means half the driving force on the nut so your max accelerations go down.
Sounds a bit swings and roundabouts to me :naughty:

[JAZZCNC]

I thought about that too, but it seems easy to overcome. Double-nut the end of it or pot the thread then clamp. I hate waiting for long deliveries so I'd probably get it from Zapp!

Your joking my last order came special machined from china in seven days. I'll bet you £10 between Zapp and Jonathan you can't get these screws delivered and machined in less time.!!


Sounds a bit swings and roundabouts to me :naughty:

Agreed.!!

[JAZZCNC]

I'll then make some 10mm plates to hold it all together.

Where you going to use 10mm plate.?? Don't use 10mm plate on the Z axis or the gantry sides else again you'll come to regret it. 15mm minimum 19mm better.!!

[Lee Roberts]

I'll then make some 10mm plates to hold it all together.

Got to agree with Jazzmin on that, 10mm thick is to be considered as sheet material, personally i would say 20mm minimum/maximum myself.

[Tenson]

Okie-dokie!

Are you guys sure that gantry weight is not a problem? I was thinking I could use extruded aluminium profile, but it needs to be quite thick to be stiff, so I'd prefer to go with solid 20mm aluminium. Alternatively I could go with thin sheet and re-enforce it with carbon fibre struts.

[Jonathan]

Okie-dokie!

Are you guys sure that gantry weight is not a problem? I was thinking I could use extruded aluminium profile, but it needs to be quite thick to be stiff, so I'd prefer to go with solid 20mm aluminium. Alternatively I could go with thin sheet and re-enforce it with carbon fibre struts.

Yes we're sure. The mass of almost every CNC mill is significantly greater than a CNC router with good reason. It helps significantly to get a good finish and cut with a decent material removal rate. It is far more important to make the gantry strong than to make it fast and it wont make much difference anyway.

Extrusion is good for the gantry, but quite expensive. Consider steel box section or a combination of aluminium flat bar, extrusion or steel. You'll find plenty of examples of ways to do it with these if you search in the build log section.

[Tenson]

Thanks Jonathan. I saw you used steel box section for the frame of your machine. Trouble is I can't weld! I only have a gas bottle torch for doing a bit of plumbing, and I have little experience with it.

My plan at the moment is to use aluminium profile for the frame and 20mm plate for the gantry. Given the cost of aluminium profile I don't want to use it for the machine bed, although it seems ideal for easy clamping, it would come to about £300! I have some 13.7mm aluminium honeycomb (http://www.trbls.com/buy-online.html) panel for another project so I thought I'd use that as the bed, and then fix some MDF on top as a sacrificial layer I can screw the cutting material on to.

[ecat]

That honeycomb stuff looks interesting, ty for the link.

With regards using plate for the bed, is there any advantage in bolting say 20x80 or 40x120 profile to either the top or bottom of the plate? Full cover or just certain strategic locations?

Also this stuff as a bed plate surface? http://stores.ebay.co.uk/Vakuumtisch-de/T-Nutenplatte-/_i.html?_fsub=2900148014&_sid=218645364&_trksid=p4634.c0.m322 , certainly not cheap but very convenient.

The basic question is, just how much benefit does one gain when increasing thickness by bolting two plates etc together?

[JAZZCNC]

Trouble is I can't weld! I only have a gas bottle torch for doing a bit of plumbing, and I have little experience with it.

Learn.!! . . Buy a cheap £40 quid stick(Arc) welder and give it a go you'll be surprised just how easy it really is.!!
DONT buy a cheap Mig thou they are NO GOOD for boxsection, you need good quality high power for material over 3mm. . . . Stick is cheap, relatively easy and eats 3mm boxsection for breakfast. With a few hours practice on some scrap you'll be surprised what you can do and how much you improve.

That said you don't need to weld!!. A steel framed machine can be easily bolted together using plates. You will need an half decent drill press which you'll need regardless whether you use profile, steel or what ever and it's a must have tool for building a CNC machine, and a good grinder with a selection of grinding and cutting disc's.
When you buy the steel buy a strip of 80 x 6mm flat bar and chop it up for the required plates with grinder.
Doing it with plates means you have loads of adjustabilty which you need in a DIY machine to account for the lack of precision grinding and equipment you don't have.!!
It will also save you a shit load of money compared to profile and be stronger.

There are a couple of things with profile just about every body who hasn't used it before doesn't realise.
1: It's even more expensive than they thought because the extra's like T-nuts and brackets etc which you need to keep it easy simple bolt together cost nearly as much as the profile it's self and if you try to skip them to save Wonga your into a world of pain and akwardness that's just not worth the cost.
2: Often it's not just simply bolt together and often needs counter bores drilling so bolt heads fit between slots etc so again requires a good drill press and little jigs so every thing aligns together.
Personnaly I find it far simpler and quicker using steel and the slight plus profile as in that it makes mounting things like switch etc easier is far too costly IMO and with good design every thing can be done with steel and Alu plate for less than half the price and really not that much longer to build.

I would have thought the honeycomb over time would eventually crush under plunging forces which can be very high. With a bit of experience you learn to avoid plunge machining and try to avoid it as much as possible because it knocks the shit out the machine.

Personally I don't like 1 piece beds like the German stuff because one cock up and the whole thing is knackerd. Far better to use individual pieces which can then be replaced relatively cheap if you Foo bar. On my steel framed bed I used 16mm x 38mm strips of Ali to create a T-slot bed and it works very good and is about the cheapist way (at the time) get a good bed.

The other route I'd consider is to use profile side by side with the bennifit being you have 2 shot's at killing it.!!! So if you damage a piece you can turn it over and use the other side before it's knackerd. These guys also sell single sided 80x16 at £14mtr which aint bad. http://www.metallin.co.uk/shop/index.php?route=product/category&path=3_125

[JAZZCNC]

Got to agree with Jazzmin on that

Why don't you agree with me on owt else I say.!!. :whistling: (Would kick you in the balls if I could find it.!!)

[Tenson]

I have a drill press and also a Makita table-saw with a steel cutting blade. So I could drill and cut box steel.

How accurate is water-cutting? There is is place near my Mum's home that does it, would it be precise enough to do the end plates and bits for the gantry? If it cut the outside dimensions and made a few marks where I need to drill for the mounting of rails etc.. then that could work well for me.

[motoxy]

Water cutting can be very accurate. With 60,000psi and a garnet additive you can accurately cut 2" steel.

[Tenson]

I know water is great for cutting thick materials; a friend of my Dad works (oops, worked) at an industrial kitchen makers, and they cut 100mm work tops. That said, I've seen some slightly rough edges on water-cut stainless plate I once got done (free of charge admittedly). It almost looked like the contour was plunge cut without enough swathe.

I also just want to say thank you to the people who have contributed to my thread! I keep getting told I need to use stronger, faster, better parts but at least I am also avoiding some big mistakes. I think perhaps part of the problem for me is that I got a CNC3040 and figured it is easy enough to scale up, but most of the things they have done in that machine don't work on a bigger scale. I think I also have a loose wallet as I don't have a set budget and I think of this machine as a business investment since I design and make speakers.

Just curious, how many of you like good stereos? Not to be rude, but most of the machinists I have met don't seem to worry about wearing hearing protection even though they work in very noisy places. Please protect your hearing, it's not girly!

[Tenson]

I've just had a quick look and actually steel box section is more pricey than the alu profile, though not including connectors. I was going to buy from this place http://www.valuframe.co.uk/Valuframe Price List 27-02-12.pdf

Even more interesting, it is cheaper to cover the entire bed with 30x30mm profile than 16x160mm.

P.S. Would a mod please move this to the DIY section ;)

[JAZZCNC]

I've just had a quick look and actually steel box section is more pricey than the alu profile,

Rubbish it's no where near the price of profile and I'm not guessing at this as I buy steel all the time. I pay Approx £25 for 50x50x7.5mtr lengths of 3mm box section. Just payed £34 for 80x40x7.5mtr. See your local steel merchant and they may well even cut it for you for a small few.

Water jetting leaves a edge slightly angled so it's no good for pieces that need 90deg joining to others like bearing blocks to gantry sides or the Z axis etc.

[Tenson]

Sorry I don't know my local steel merchants. I looked on Google and the first on-line retailer I found sells 30x30mm mild steel box section at £8.30/meter, and the next sells 30x30mm stainless box at £18/meter.

Alu 30x30mm profile is £6.55/meter.

Looking a bit deeper I foun mild steel for £5/meter. Is mild steel suitable or should I be looking for 'engineering steel'?

Thanks for the advice on water-jet. A shame, as most precision engineers seem uninterested in small one off jobs. I worked shortly with a local one who did a bad job and our relationship went bad when I expected them to make it good. I've emailed a few online guys and they tend not to respond for small jobs.

[JAZZCNC]

Yes mild steel is what you want and you will find it a lot cheaper if look for a steel stockist. 7.5mtr is standard length for steel section and it's often sold and priced per length.

Post what you want making on here I'm sure someone will help you out, unfortunatly I can't offer as I'm commited to other people and only so many hours in a day.

[ecat]

Digressing just a little.

As steel is more suitable and lower cost than ali profile why don't we see more low end commercial machines made from it? The only ones I know of are these sexy beasts http://translate.googleusercontent.com/translate_c?act=url&hl=en&ie=UTF8&prev=_t&rurl=translate.google.com&sl=it&tl=en&twu=1&u=http://www.cnc-valmec.it/serie-hobbystica&usg=ALkJrhg-Mw2rRLf46sP0Zrb3V7vkmqihcw , looks more like pressed thin sheet than welded box section :( So, is it tooling costs? Assembly and finishing costs? Weight? The price of paint? Or have we all become victims of the profile looks sexy fashion cult?

[Jonathan]

Thanks Jonathan. I saw you used steel box section for the frame of your machine. Trouble is I can't weld! I only have a gas bottle torch for doing a bit of plumbing, and I have little experience with it.

That frame was the first welding I had done. Practices on some scrap first that's all...


Digressing just a little.

As steel is more suitable and lower cost than ali profile why don't we see more low end commercial machines made from it? ... Assembly and finishing costs? Weight? The price of paint? Or have we all become victims of the profile looks sexy fashion cult?

I reckon it's because it looks fancy and most 'hobby' machine manufacturers are either too lazy or don't know better. They'll probably argue it saves a lot on labour. I got round the price of paint by not painting it - machine not art.

[JAZZCNC]

:( So, is it tooling costs? Assembly and finishing costs? Weight? The price of paint? Or have we all become victims of the profile looks sexy fashion cult?

No it's because lots are into bling and posing. . . . Paint aint expensive just messy.!!!

I'm always suspicious of machines with lots of fat looking plastic covers.!!. . . . Always makes me wonder what there hiding.????

[Shinobiwan]

Just curious, how many of you like good stereos? Not to be rude, but most of the machinists I have met don't seem to worry about wearing hearing protection even though they work in very noisy places. Please protect your hearing, it's not girly!

I quite like audio ;)

I knew you'd get the bug. The 3020 looks like good training wheels but I'd ignore everything about its construction if your wanting to making a decent machine. They're notoriously shite in general and the principles used are barely suited to its own small frame let alone once scaled up.

I think I already mentioned it to you over on DIYA but I bought a 'commercial' 4x4ft machine from a supplier here in the UK. Its been eventful and to be honest I wish I'd built my own now. Its really not that difficult with that right tools and a helping hand from folks on here. Of course I wanted to be up and running quickly so thought an off the shelf solution would be best. That was December and 4 months on I still haven't cut a single useful part yet. Annoying but around £5.5k lighter in pocket I'm almost ready to go. I hate to think what that money would have bought if I'd DIY'd.

Good luck with the machine. If you listen to the folks like Jazz then you won't go far wrong.

PS. Since your building this can't you go with a larger bed? I'd say 4x4ft is the minimum for cabinet work otherwise you'll be swapping sheets in and out as often as you change undies. 4x4 also has the advantage of taking a standard sheet without cutting. I would have liked an 8x4ft really but that would take up most of the workshop. Interestingly you might want to consider wall mounting or making a vertical frame to get the bed size up and the floor space back.

Don't forget about your choice of spindle too. I know its early days and you haven't even drawn up plans for the construction yet but I'd recommending avoiding regular routers, even the larger one's. I went with one of these types and quickly swapped to a 3hp water cooler spindle with vfd. These are around £250-350(inc del) from china via ebay and can vary in quality depending on the seller. Check out linearmotion_2008, he seems to sell the better models. If you prefer to buy from the UK then CNC4you have the same model for £369

http://www.cnc4you.co.uk/index.php?route=product/product&path=24&product_id=147

[m_c]

Digressing just a little.


Moslty fabrication costs/equipment.
Alu profile just needs cut, bolted together with suitable brackets, and with a bit care you have a pretty accurate machine.
Steel needs cut (more expensive/time consuming to do accurately than cutting alu), joined, be that via welding or drilling and brackets, which both take longer than doing similar with alu profile. If welded, you then have to deal with distortion.
Plus alu profile is also more accurate out the box so to speak, as extrusions are usually accurate within a thou, unlike steel which can vary quite a bit.

I've used alu profile for frames before, and cost wise it is more expensive, but takes less time to assemble, and for what I used it for, resulted in a lighter stiffer frame than using steel. Plus it looks cool ;)

[Tenson]

Hi Ant,

I knew you got a CNC machine but only after reading about it on here as I haven't been reading DIYA much lately.When I read about your machine I thought 'oh, you too!'.

I'd love to go for a bigger machine but I just can't find the space. I moved a couple months ago and have a smaller workspace. But I did get a wife and a bigger listening room in exchange. I'm going for 2ftx4ft as sheets commonly come in that size, and also they fit in my car so I can collect myself. I thought about vertical mounting too, but I'm worried about the practicality of clamping work pieces in place, and I also want to make a dust/sound cover.

I got the spindle today actually :) I was unsure if I should get the 1.5Kw or 2.2Kw motor, but decided to go for 2.2Kw firstly because it was available from CNC4U at the same price as Chinese sellers, and also because I don't want to limit my cutting speed for the sake of £100.

So what are you doing for a water-pump, and does it need a radiator to cool the water? If so, I thought I might run the water through one of the alu extrusions.

By the way, Mr. ZappAutomation, please don't use the shredded paper for packing my floor is all messy now. At least get a paper shredder that makes long lines of it rather than flakes. :) Other than that thank very much for your great and fast service!

Fudge me, that's a big motor!! I hope my Z-axis can cope with 16mm rails and ballscrew.
http://i608.photobucket.com/albums/tt169/tenson_uk/BigMotor.jpg

[Tenson]

The 3020...notoriously shite in general and the principles used are barely suited to its own small frame


They are notorious for that, but it works superbly!!

[Jonathan]

I got the spindle today actually :) I was unsure if I should get the 1.5Kw or 2.2Kw motor, but decided to go for 2.2Kw firstly because it was available from CNC4U at the same price as Chinese sellers, and also because I don't want to limit my cutting speed for the sake of £100.

You made the right choice, but for the wrong reason! The main reason to get the 2.2KW spindle is that it takes ER20 collets which hold up to 13mm (i.e. 1/2") cutters, whereas the 1.5KW spindles are limited to ER11 which is only 7mm. I'm surprised CNC4YOU was cheaper than china - did that include the VFD?



So what are you doing for a water-pump, and does it need a radiator to cool the water? If so, I thought I might run the water through one of the alu extrusions.

Just get a generic brushless submersible pump, something like this one:

http://www.ebay.co.uk/itm/Water-Pump-Hailea-HX-4500-2000L-H-Aquarium-Fish-Tank-Hydroponics-Koi-Pond-NFT-/260980273070?pt=UK_HomeGarden_Garden_PlantsSeedsBu lbs_JN&hash=item3cc3a2efae#ht_2757wt_1037

(seach around - you'll find cheaper ones)

No need for a radiator since if you have a decent volume of water the natural heat conduction to the atmosphere is sufficient. Just find a good size (20L ish) tank and ideally use de-ionised water with antifreeze.


By the way, Mr. ZappAutomation, please don't....

What did you get from Zapp?


Fudge me, that's a big motor!! I hope my Z-axis can cope with 16mm rails and ballscrew.

It will if done right - do post the drawing first.

[Tenson]

Mine has an ER16 clamping nut. I guess I can't use ER20 as the part the collet goes into (spindle shaft?) will be too small?

[Jonathan]

Mine has an ER16 clamping nut. I guess I can't use ER20 as the part the collet goes into (spindle shaft?) will be too small?

Yes the dimensions of the taper for ER20 are completely different so you'll have to use ER16, so upto 10mm:

http://www.ebay.co.uk/itm/FULL-ER16-SPRING-COLLET-SET-10-PCS-CNC-MILLING-LATHE-TOOL-WORKHOLDING-NEW-A03-/150784725024?pt=LH_DefaultDomain_3&hash=item231b785020

[Tenson]

Has anyone done this before?

Linear bearing - £6

http://i608.photobucket.com/albums/tt169/tenson_uk/LinearBearing.jpg


Bearing taken out, and Ballnut put in:

http://i608.photobucket.com/albums/tt169/tenson_uk/NutBracket.jpg


http://i608.photobucket.com/albums/tt169/tenson_uk/NutBracket-1.jpg


Now it's not quite a perfect fit because the top of the nut is slightly ~0.5mm) taller than the top of the bracket housing. So one could either fit a thin sheet on top of the bracket, or trim the top off the nut if you can figure out how to do it. Or just account for it on the plate you are mounting.

Anyway, seems better than paying ~£35 for a ballnut housing? Isn't it?

[Lee Roberts]

Hi Tenson,

I see your cnc endeavours are coming along nicely!

I have toyed about a little with ball nut mounts myself using the "off the shelf" components, what I found with using the SMA units is the mounting holes for the Nut. If memory serves me right only two of the fixing holes could be used.

Your drive nut is properly one of if not the most critical part of an axis, personally I wouldn’t be happy with such a fixing, however I have seen machine’s up and running and only using the same two fixing points. I am guilty of being a perfectionist though with a little OCD topping to add to the flavour :)

In the end I decided to design my own drive nut mounts, I wasn’t able to machine them myself but another member on the forum was able to and did a great job for me. The price wasn’t a lot however they did stress that clamping was an issue due to the shape/design.

I have attached the finished picks for you to view, if you want the part details I am happy to post them for you.

.Me

[Tenson]

Thanks Lee! Four of the six screw holes can be used, so I thought that is enough. I also considered putting a bit of lock-tight around the nut before inserting it. Not too much, since it should come out again with a whack.

P.S. your mounts look like a super-charger!

[Jonathan]

Thanks Lee! Four of the six screw holes can be used, so I thought that is enough. I also considered putting a bit of lock-tight around the nut before inserting it. Not too much, since it should come out again with a whack.

I wouldn't put lock-tight round it. You'll need a bit of room for adjustment when you first fix everything together to ensure the nut can be set precisely concentric to the screw.

I've made a few ballnut mounts for myself and these for people on the forum:
55815582558355845585\

I hope you're not intending to use those unsupported rails on this machine?

[Shinobiwan]

Hi Ant,

I knew you got a CNC machine but only after reading about it on here as I haven't been reading DIYA much lately.When I read about your machine I thought 'oh, you too!'.

Aye. I'd been looking longingly(as you do) at them for years but things fell into place the middle of last year and I suddenly found myself with a decent size workshop.


But I did get a wife and a bigger listening room in exchange.

Simon, it seems like we're in the twilight zone here. We've both bought a CNC, we build speakers and now we've both saddled ourselves with partners who moan about all the MDF dust we walk into the house or that the dinning room isn't an assembly area.


I'm going for 2ftx4ft as sheets commonly come in that size, and also they fit in my car so I can collect myself. I thought about vertical mounting too, but I'm worried about the practicality of clamping work pieces in place, and I also want to make a dust/sound cover.

You can build in a cover and have it vertical. Its more work in the design stage but can make for a really neat machine when space is an issue(when isn't it?). If I'd DIY'd I'd have likely done something very similar to this:

http://www.denfordata.com/pdfs/product-literature/vertical-router.pdf

Very tidy package that would easily allow for a 4x4ft bed. All the crap falling into a pan at the bottom rather than all over the workpiece or into you rails etc. is another plus. Really there's not many downsides aside from increased build complexity and more awkward clamping. You could add vacuum bed later to solve the latter.

If your spending the sort of money I think you are then its worth considering.


I got the spindle today actually :) I was unsure if I should get the 1.5Kw or 2.2Kw motor, but decided to go for 2.2Kw firstly because it was available from CNC4U at the same price as Chinese sellers, and also because I don't want to limit my cutting speed for the sake of £100.

Good stuff but the ER16 collet might be a bit limiting. Can I ask why you choose that over the ER20? I looked around long and hard at cutters and most of the extra long series cutter are 12mm shank. These cutters allow thick(ie. 100mm) baffles to be safely machined.


So what are you doing for a water-pump, and does it need a radiator to cool the water? If so, I thought I might run the water through one of the alu extrusions.

Absolutely no need for a radiator/fan setup in this country, maybe the Sahara. I'm going for simple so that means a bucket with about 25ltrs of water, a dash of anti freeze and a regular submersible pump in the bottom. That's all you need and anything else is additional expense and not warranted. The only thing you need to watch out for is the pump head height capacity something like 3m is fine.


By the way, Mr. ZappAutomation, please don't use the shredded paper for packing my floor is all messy now. At least get a paper shredder that makes long lines of it rather than flakes. :) Other than that thank very much for your great and fast service!

Yep can't say I'm a fan either. No matter how carefully you try to remove things from the box that stuff still goes everywhere. I'm sure Gary is laughing whilst he's packing these things.

[Lee Roberts]

Now it's not quite a perfect fit because the top of the nut is slightly ~0.5mm) taller than the top of the bracket housing. So one could either fit a thin sheet on top of the bracket, or trim the top off the nut if you can figure out how to do it.

You could use shimming steel to do this, a bench grinder can also be used to take the nut down a little.


Thanks Lee! Four of the six screw holes can be used, so I thought that is enough. I also considered putting a bit of lock-tight around the nut before inserting it.

Humm are you sure?, the top two holes look as though they would pass through the mounting holes for the SMA unit, the middle two dont look like they would have much material for a thread because of the bore in the unit?

As Jonathan said you may need to jig the nut about to get it concentric to the screw, all sounds a little to "make do" for me when a little creativity is all that is needed. Our right honorable gentleman Jonathan could properly fix you up with some custom mounts for a reasonable price.

Somthing to think about also is machine maintenance in the future, maybe unlikly but what if a ball gets jammed or stuck in the nut and you need to take it apart....smash it out with a hammer ?!?!? Obserred, how dare you :tongue:

:joker:

.Me

[Tenson]

I wouldn't put lock-tight round it. You'll need a bit of room for adjustment when you first fix everything together to ensure the nut can be set precisely concentric to the screw.

Do the mounts usually have much room to move? This fits absolutely perfectly and tightly so no room to move at all, although it could rotate a little.



I've made a few ballnut mounts for myself and these for people on the forum:

*Pictures*

I hope you're not intending to use those unsupported rails on this machine?

Very nice mounts. If I use your super rotating ballnut systems on the Y and X then I only need a nut-bracket for the Z right?

The unsupported rails are for the 25cm Z axis.

[Tenson]

Simon, it seems like we're in the twilight zone here. We've both bought a CNC, we build speakers and now we've both saddled ourselves with partners who moan about all the MDF dust we walk into the house or that the dinning room isn't an assembly area.

Eunji doesn't mind too much, she just wants me to sell lots! She is very understanding about hi-fi stuff too since she works in music. Actually she is very helpful in blind testing stuff since she did 3 years of critical listening training.

Just for fun, a quick photo of the table in the living room right now, lol.

http://i608.photobucket.com/albums/tt169/tenson_uk/LivingRoom.jpg



You can build in a cover and have it vertical. Its more work in the design stage but can make for a really neat machine when space is an issue(when isn't it?). If I'd DIY'd I'd have likely done something very similar to this:

http://www.denfordata.com/pdfs/product-literature/vertical-router.pdf

Very tidy package that would easily allow for a 4x4ft bed. All the crap falling into a pan at the bottom rather than all over the workpiece or into you rails etc. is another plus. Really there's not many downsides aside from increased build complexity and more awkward clamping. You could add vacuum bed later to solve the latter.

If your spending the sort of money I think you are then its worth considering.



Ooooh that is nice! I don't think it will save alot of space for me though because the gantry height, plus z-axis motor and machine frame end up about 65cm, and the width is only about 85cm. I think I'd prefer easy clamping. That said, I could easily turn the machine on it's side if I want.




Good stuff but the ER16 collet might be a bit limiting. Can I ask why you choose that over the ER20? I looked around long and hard at cutters and most of the extra long series cutter are 12mm shank. These cutters allow thick(ie. 100mm) baffles to be safely machined.

All the router bits I have are 6mm shank, and I have a fair bit of money in that. I also thought it was possible to just buy an ER20 collet... nevermind!

So are you doing custom built speakers professionally now?

I still need to have a go with Acourate!

[itsmillertime]

Tenson that looks like my front room, my wife reckons it won't be long till I've got it looking like delboys flat in only fools and horses. Minus the inflatable dolls of course ;-)

I wonder if we should start a thread on "How diverse are the things that we make?" That could be an eye opener! Got a few good ideas of my own.

[Shinobiwan]

Just for fun, a quick photo of the table in the living room right now, lol.

Looks familiar :)


Ooooh that is nice! I don't think it will save alot of space for me though because the gantry height, plus z-axis motor and machine frame end up about 65cm, and the width is only about 85cm. I think I'd prefer easy clamping. That said, I could easily turn the machine on it's side if I want.

The gantry is offset by you being able to go upto 4x4ft. Generally a 4x4 will stick out at least 1400mm from whatever wall its pushed up against. Vertical would be less than half that.


All the router bits I have are 6mm shank, and I have a fair bit of money in that. I also thought it was possible to just buy an ER20 collet... nevermind!

When you say router bits do you mean the regular 1/4" collet types with straight flutes? They'll work fine but you might want to consider getting yourself some end mills. These are generally better quality, plunge much better, remove material better and last longer.

Use up what you've got and once they're starting to dull get yourself something like these:

http://www.ebay.co.uk/itm/140548881261?ssPageName=STRK:MEWNX:IT&_trksid=p3984.m1439.l2649#ht_3253wt_1163

They're really cheap for what they are and were recommended to me by folks like Jonathan and Jazz who've used them to cut aluminium. They'll eat MDF and ply for breakfast and last ages.


So are you doing custom built speakers professionally now?

Pretty much but nothing official. Its mostly word of mouth but I hope to bring a small lineup of speakers eventually. There's a lot of planning involved before that can happen not to mention prototypes, listening tests, crossover tuning, auditioning. I had hoped to have a floor standing speaker available before the end of this year but I've recently taken on a custom project for Arthur(Ackcheng on DIYA) over in Hong Kong and its scale is very ambitious. A total of 8x 10" drivers, 4x 8", 4x 5" and a pair of ribbons all housed in 2100mm x 320mm x 750mm cabinets. Its the audio equivalent of the Deathstar.

And last year I designed(not built) a huge OB dipole line array with seperate ripole sub stacks for a guy in New York. 8x 21", 18x 8", 16x BG Neo10's and lots of Mundorf AMT's. I believe he's now taken the CAD files to various machine shops in the US and is in the process of having them manufactured. Materials are alu and acrylic and the project is about the cost of a new mid range Benz.

[JAZZCNC]

The unsupported rails are for the 25cm Z axis.

CAN NOT STRESS ENOUGH. . . . . .DONT DO IT. . . . ABSOLUTLY DONT DO IT.

You will regret using unsupported rail and even more so on the Z Axis.!!

[Shinobiwan]

CAN NOT STRESS ENOUGH. . . . . .DONT DO IT. . . . ABSOLUTLY DONT DO IT.

You will regret using unsupported rail and even more so on the Z Axis.!!

Simon, I'd really listen to this advice. I know you think you know what your letting yourself into here but honestly, you have no idea. I mean that in the nicest way possible.

I had to have my Z-axis rebuilt because it was using unsupported rails and was shit. About +/-1mm of flex at the bit and that means you'll get a crap finish even in MDF with anything like a decent DOC and feedrate. Your parts will be undersized, oversized and generally a mess. And you can forget it if your thinking of aluminium. Plus your using a heavy spindle with heavy 16mm alu plate. That's a bunch of weight right there before you've even got to the cutting forces.

Do your self a massive favour, ditch it and go with supported round rail. It'll add nothing to build complexity and hardly anything to the price. Cannot stress enough how shit round rail is, it should be banned on anything calling itself a decent CNC router.

Listen to the knowledgeable folks on here and save yourself from a world of pain, you'll get you a much better finish, more accurate parts and will let you move up to harder materials like alu, Corian and so on.

I'm trying to stress this because I fell in to the same trap your about to and its costs me a lot of money to put it right. Do it right the first time.

[Jonathan]

You could use shimming steel to do this, a bench grinder can also be used to take the nut down a little.

I'd be very wary of grinding a ballnut since any dust from grinding in the ballnut is a recipy for disaster. You could try and cover it up but I don't think its worthe the risk.



CAN NOT STRESS ENOUGH. . . . . .DONT DO IT. . . . ABSOLUTLY DONT DO IT.

+1

When I said earlier that 16mm rail would be fine 'if done right' I was assuming supported rails. You'll be much better off getting some SBR20 (supported) rails.

The 6mm single flute cutters are brilliant...get lots! Although they are much better than most at plunging you shouldn't need to plunge - use spiral toolpaths or ramping. Router cutters with a 1/2Q shank are much much stronger than the 1/4" ones, so well worth the extra.



Very nice mounts. If I use your super rotating ballnut systems on the Y and X then I only need a nut-bracket for the Z right?

Yes... but how long is the Y-axis screw going to be? It's a bit sledgehammer to crack a nut using a rotating nut on the gantry, althought it would work well.

[JAZZCNC]

Don't under estimate the job the ballnut mount does.!!

Ballscrew mounts, alignment and nut holding are crucial to a good working machine, if any one is off the slightest amount the result is friction.
This friction as to be overcome by the motors, the motors lose torque as the speed increase. Has the speed increase's friction heats things up which inturn creates more friction sapping more power end result stalling motors or missed steps at best.!!. . . . Worse case things lock and prematurely wear away often quickly in tight ballnut's case and die suddenly.

Saving a fiver can cost hundred.!!.. . . Is it really worth it.????

[Tenson]

I had to have my Z-axis rebuilt because it was using unsupported rails and was shit. About +/-1mm of flex at the bit

I find that quite surprising. I rather thought with the very short length and least weight of any axis, the Z would be okay running unsupported. If it really is that bad I'll get some supported rails. Building a Z-axis assembly to have opposing supported rails seem quite tricky though at first thought. Any tips / pictures?

Anyone want to buy 2x 30cm unsupported rails? ;)

[Shinobiwan]

I find that quite surprising. I rather thought with the very short length and least weight of any axis, the Z would be okay running unsupported. If it really is that bad I'll get some supported rails.

Anyone want to buy 2x 30cm unsupported rails? ;)

Ask Jazz about it, he came around to help me out with the machine and saw it. Its the cutting forces that will really show this up but even pulling back and forth on the z axis by hand will tell you how sturdy it is. The more rigid the design, the better the finish and accuracy you'll get.

If you do it right you'll be cutting MDF with very tight tolerances and even aluminium won't be an issue. Just keep asking loads of questions on here and tell folks about every design decision your making as they'll be able to advise if its a good one or a bad one. Building a good machine doesn't mean its more expensive but building a bad one will cost you to fix it.

[Tenson]

Building a good machine doesn't mean its more expensive.

Bloody seems like it! :tongue:

[Shinobiwan]

Bloody seems like it! :tongue:

lol, ok well not that much more expensive. You'll have a really nice machine if you do it right and it'll last you years.

[Tenson]

Is there a big advantage in build quality terms using profile rails for the Z-Axis rather than supported round rail?

I must admit I can't see an easy way to use supported rail so that the bearings oppose each other on the Z axis. I have seen Jazz used HiWin rails facing out, not opposing, so is this only okay with profile rail?

[JAZZCNC]

Is there a big advantage in build quality terms using profile rails for the Z-Axis rather than supported round rail?

I must admit I can't see an easy way to use supported rail so that the bearings oppose each other on the Z axis. I have seen Jazz used HiWin rails facing out, not opposing, so is this only okay with profile rail?

The advantage of profiled rails are much tighter bearing tolerences with a smooth action helping to reduce resonance and flex they also last much longer, esp in harsh conditions like a Z axis where it's constantly in the firing line and open to crap from the cutter. They don't really add much more strength than supported round rail if any.!

You can use supported rails just the same as profiled like you have seen me do. Absolutely no difference other than they stick off the Z axis rear plate (which you would call running across X axis, me Y axis) slightly more due to height of supported rail.

[Jonathan]

The main advantage with profile rails is that they are preloaded. This means there is negative clearance between the bearings and the rail, so there will not be any play. It also makes any deflection linear and crucially, with rails with 4 rows of balls which are the type you should get, the load ratings are equal in all directions. Conversely supported round rails have little if any preload and the load rating varies significantly depending on the direction the force is applied since the open side of the bearing clearly offers less support. Ideally we would mount supported rails so they 'oppose each other' on Z, however as you say it's quite difficult to accomplish rigidly without a somewhat radical gantry design. Hence most people mount them both the same way round. If I measure the deflection when a force is applied to my Z-axis parallel to X the reading varies quite a bit, due to the uneven load rating.
However this is of course all relative - compared to unsupported rails the deflection is tiny. You wont regret using supported or profile rails on the Z-axis - ultimately depends on what you can afford.

[Tenson]

Thanks I rather thought it might be the case, though good to know open bearings on supported round rail are not a no-no. The HiWin rail is okay priced but the bearings seem a bit steep. How about other brands? RS Components sell one called IGUS DryLin T (http://uk.rs-online.com/web/p/linear-guides-rails-carriages/4885079/) which for 300mm is quite affordable (GBP14 rail, GBP16 Carriage, I can't do the pound sign on this Korean computer) but doesn't offer longer rails. There are many other brands too.

[Tenson]

Hi Guys,

I got some SBR rail from CNC-4-U with open bearings. The bearings look like they are normal ones that have been cut open. Is this normal? They also have little grub screws in the top that hold the bearing in and when tightened up they close the bearing up on the rail a little. The motion was a little jerky at first but after a quick sanding of the cut endge of the bearing thye are smooth now. They seem a bit home made. Should I be sending them back or are they okay?

http://i608.photobucket.com/albums/tt169/tenson_uk/20120416082.jpg

http://i608.photobucket.com/albums/tt169/tenson_uk/20120416080.jpg

[Jonathan]

They seem a bit home made. Should I be sending them back or are they okay?

That's perfectly normal. Same as you'd get from linearmotionbearings2008 on eBay. I ordered some recently from china (not for myself I hasten to add!) and they came in less than a week. The're essentially the same except the grubscrew is now on the side of the bearings, not on the base.

[Tenson]

Thanks for that. I like CNC-4-U. They are UK based so I'm supporting local trade and the prices are not too far from the eBay ones. I got the 2.2KW VFD spindle and 4Nm steppers from them too.

Jonathan, can you link me to some more pictures of your machine? I saw it in the video for the single flute cutters but I'd like to see more detail.

[JAZZCNC]

Personally I hate these bearings they are rubbish.!
There hardly ever flat or square making mounting without binding difficult they go sloppy on the rail in no time, they wear like crazy if you operate them in any reasonably dusty or abrasive conditions.

Do your self a favor and save up for some real profiled rails or at least buy some decent bearings and blocks, to me it's just false economy because if you build a decent machine frame these will quickly become the weak link and you'll want to change them and it's not often a quick swap over job to change to profiled rails.!! . . Spending on profiled rails is one area where I promise you won't regret or begrudge the extra expense.

[Tenson]

Hi Dean, I thought about profile and besides the cost, it seems like it would be more fussy about the surface they are mounted on. I don't have the ability to machine the mounting surface 100% flat, so I think the stiff supported rail will help there. I have to say I'm not very impressed by these bearings though. Can you recommend some higher quality open bearing for this type of supported rail?

[Tenson]

I must say I'm starting to wonder if I will just use the unsupported rail. The bearings seem much smoother and tighter tolerance! It's all very well having stiff rail, but I can find some places on the supported rail where the bearing will wobble slightly side to side.

[Shinobiwan]

lol, you is crazy.

Pardon my French but it sounds like you bought shite.

I await your post showing pictures of your newly purchased profile rails

[JAZZCNC]

Hi Dean, Can you recommend some higher quality open bearing for this type of supported rail?

Sorry didn't see this post. Other than Gary at Zapp then no not really because I won't touch them personaly and I only ever come into contact with them when others have all ready bought them. I hate the bloody things because I'm always having to machine them flat and true even then there still left with shite bearings.!!

Said it before many times but If you want real accurecy and longevity then get rid and save for some profiled rails and don't even go the unsupported route because you will regret that path promise you that.!!

[Swarfing]

Mmmmm sower grapes me thinks????:stung:....I love my unsupported rail and never had a problem, maybe that is just me :-)

Reality is you get what you pay for in most cases, if you can afford profile rails and bearings then grab them first. If you can not then anything else needs to be as thick as you can afford to give yourself a chance.

[Tenson]

After I looked more carefully at the parts I found one rail was strange, it seemed like it was slightly more narrow at one end because regardless of the bearings used, it was tight at one end and sloppy at the other. So it isn't the bearings so much as a strange rail.

I sent a short video to the guys at CNC 4 U and they called me back to say they are checking their stock and if others are okay they will send me a replacement, without me even needing to return the duff one.

[Tenson]

I await your post showing pictures of your newly purchased profile rails

I'd love to, but although the HiWin rail is not too badly priced, the bearing blocks are very costly. Are there no alternative brands who make them to fit the same rail?

[Robin Hewitt]

If you want real accurecy and longevity then get rid and save for some profiled rails and don't even go the unsupported route because you will regret that path promise you that.!!

Ah! Someone with experience of Hiwin rails :beer:

I'm drawing up a plasma cutter with a 2 meter axis on the gantry and I want to use one profile rail, probably the 20mm.
My problem is the spec says they come in two flavours...
1: with slop to allow for slight misalignment between 2 rails.
2: with oversize balls to get a tight fit.
Zapp seems to sell only the sloppy variety and I can't afford twist around the long axial because the head sits quite far forwards, about 65mm ahead of the races.
My question is, how much axial twist do I expect?
Should I add a spring roller running on the rail behind to keep it in tension? That wouldn't be hard because I have to screw the rail from behind so there are no visible holes that might carry abrasive crut passed the wipers.

[m.marino]

Robin,

You can also get the flush plugs for rails mounting from the front and SBC rail might offer a few different options, then there is going over to traceparts and looking up the French/Swiss rails (not much different in price from Hiwin from what I have seen) and seeing what they offer as well. I need to look up th Italian firm that sells different types of rail on the per project basis as most want you to buy way to much. Good luck

Michael

[Tenson]

Hi Jonathan you PM box is full.

I'm about to buy the ballscrews from Chai. Just to be sure, I need to get the ballnuts from him right? They are not part of the mount. Also, if he sends me the twin-start thread like he did to Jazz recently, will it work okay for the nut-mounts?

[Jonathan]

Hi Jonathan you PM box is full.

Sorry I didn't realise - sorted now.



I'm about to buy the ballscrews from Chai. Just to be sure, I need to get the ballnuts from him right? They are not part of the mount. Also, if he sends me the twin-start thread like he did to Jazz recently, will it work okay for the nut-mounts?

Yes you do need to get the ballnuts. I recently got some of the twin start from him and I think they will be more rigid than the single start. I asked for twin-start and the price was the same...rotating ballnut mounts will work just fine with both.

[Tenson]

I cut a few bits of 16mm aluminium to size today for the Z axis.
5815

[Jonathan]

Is the watering can your coolant system? :distracted:

[Tenson]

Yes it is. I have never needed it before to cut wood (obviously) but I found my scissors got quite hot cutting the 16mm aluminium. I made a coolant drip with a thin bit of tube and siphon effect.

[Tenson]

I thought I'd share a couple pictures of what my CNC3040 was doing today. This was making a pocket in 6061-T6 (heat-treated and artificially aged) aluminium.

Initial go-around.
5837


Pocketing
5838