[Fatamorgana]

Hello!

In cad stage of cnc diy build, seem to stagnate somehow because I cant seem to find the right ballscrews (grounded 25mm 5mm pitch, preload nuts) at dy-global.

Is bst automation a good choice for ballscrews?

Lars


Sent from my iPhone using Tapatalk

[TMToronto]

I recently posted here about a very good experience I had with my purchase from dy-global.
As you mention however, I had a long wait to try to find something that worked for my build - not exactly what I wanted but with respect to length of rails but the ball screw assembly worked well for my needs.
The variety of components and pricing are excellent which kept me a frequent visitor to the store over many, many months.

Had I not found something, I would have tried BST or Maxtenso Motion.

[joeavaerage]

Hi,
I waited until the right ballscrews came up on dy-global, and designed my machine around what I could get. They were 3 x C5 BNFN double nut 32mm x 5 mm screws by THK with six FK25 support bearings.
I paid less for all three than I would have paid for just one, and that does not count the support bearings, had I bought new.

The downside is that the selection is not maybe what you want.....but that is what it is in the secondhand/new old stock market.

Craig

[catahoula]

i ordered custom length TBI C5 double nut preloaded (OFUR) ballscrews from BST motion. they can do whatever end machining you require. parts arrived quickly and price was good, i'm happy.

[joeavaerage]

Hi,
beware that there are Chinese companies whom claim to sell ground C5 ballscrews and even more outrageous ground C3 ballscrews at little more than
a C7 rolled screw. I call BS. There is genuine ground C5's and C3's........and then there is the rest.

i ordered custom length TBI C5 double nut preloaded (OFUR) ballscrews from BST motion

TBI is a known brand and BST would seem to have a good reputation which suggests in turn that they would not sell counterfeit parts.

To get a feel for the market try pricing THK and NSK C5 screws....they are eye-wateringly expensive.....then ask yourself is what you are being offered genuine relative to the two
market leaders?

Craig

[peteeng]

Hi Lars - I have found Fred at BST to be very happy to supply parts that are not on the BST site. Just need a good specification for what you want and he will try to get it. So far he has been able to supply everything I have asked for in machine parts. World Peace that seems to be a difficult ask. Peter

[Fatamorgana]

Thanks for the good replies, really helpful!

Im having a conversation going with dy-global over the last weeks, and he can offer me some 32mm ballscrews that suits my build in terms of everything I need (except im looking for 25mm screws) So now I am trying to calculate how this will match up with my 750w 3.18Nm servos.

Can someone please assist me at this calculation sheet, is linear speed the same as feed rate? 500mm seems like an overkill, but it doesnt make sense with the required motor torque, and one more thing, machine force have I entered 1000nm, it seems like an standard value for these calculations?

[pippin88]

Max feed rate 500mm/sec
Screw length 500mm
Acceleration time 5 sec

You will never get close to max feed with that acceleration time

Acceleration is a major determinant of torque required

I'm an amateur, so I used multiple tools to see if I got consistent believable results.

https://www.kollmorgen.com/en-us/ser...neering-online
Is a good tool

[Fatamorgana]

Max feed rate 500mm/sec
Screw length 500mm
Acceleration time 5 sec

You will never get close to max feed with that acceleration time

Acceleration is a major determinant of torque required

I'm an amateur, so I used multiple tools to see if I got consistent believable results.

https://www.kollmorgen.com/en-us/ser...neering-online
Is a good tool

Thanks, I will try that, what is typical acceleration values?


Sent from my iPhone using Tapatalk

[peteeng]

Hi Lars - I've done a quick calculation here. If you are moving 120kg horizontally over 500mm then the mass can accel over 250mm and decel over 250mm. This is the least accel. This takes 2sec. I've used a 25mm ballscrew for the calcs. At 3Nm torque the screw can exert 366kgf. The 25mm screw buckles at over 31 tonne so no issues there (using a hiwin formula). This accel is 0.125m/s/s which is not too shabby for a hobby machine. Given those conditions the motor needs to provide 0.06Nm to do this and it has 3Nm available so all things being equal the 3/0.06= 50x so you could accel considerably more if you want. I'd say that combo is a very powerful combo and will be very serviceable. Typically you need to specify what max accel or cycle time you need to achieve then calculate from there. Although the screw can exert 366kgf it only needs to push against the inertia and friction. Friction is tiny with balled cars (say 0.01x120=1.2kgf) so all you need to move the 120kg mass is 120*0.125=15kg of inertial force...

So you could use a much smaller motor unless you want to push this thing around at high levels of g's. Is this what you are looking for? Peter

There are a couple of machining conditions to look at:
1) when machining the tool is travelling at a constant feed so there are no major inertial forces so the calculation shows that the machine can push up to 366kgf which is very considerable, way more than you need for most milling. You may need this for drilling with large drills
2) For rapids you need to move masses quickly otherwise the cycle time can be very large if the air time is large. The attached calculating is more about rapids then cutting... If you describe what you want the machine to do I can temper the calculations and advice if needed...

typical accels? 0.1g 0.2g maker level , commercial machines 1g, some 2g some even faster 1g is really scary stuff even light 0.2g machines dance.... have to nail them to the floor.

[peteeng]

here's the calc for 0.5g - the motor doesn't quite have enough torque to do 0.5g but a servo can be overdriven and it would achieve it briefly. But seems 0.4g is fine... Peter 0.4g is not a slouch movement. 0.1-0.2g is fine for a general machine. Peter

edit - Opps in my prior calcs I was out by a factor of 10 on trans torque so the D25 screw at 0.125m/s/s needs 0.17Nm not 0.06Nm... always check my numbers please... also please note that the D32 screw requires a lot of torque vs the D25 screw.. If you want to downsize the motor downsize to D25 screws...

[Fatamorgana]

First en foremost - humble thanks for the time and calculations for my question, really really appreciate that level of help. To sum it up, this calculations are above my head, but I can understand your conclusions so that it seems that 25mm 5mm pitch screws willl fit my build the best. Dy-global could set me up with 32mm ballscrews, but as I intrepetid the calculations 750w servo motors would be a good fit to 25mm screws. I have 3 x 750w servos from an earlier build, so thats why thats Im basing the setup after this hardware. Im currently waiting for a quotation from BST automation for 25mm screws 5mm pitch / grounded.

Im concidering starting a thread as I will need further advide for my build with cad-drawing etc. But im unsure of where it would be natural to start, here in the DIY or in general metallwork mills. Any suggestions?

Thanks again to community for all replies.

[joeavaerage]

Hi,
I have posted calculations for momentum equations several times, this post fairly recently:

the size of the motor requires that you calculate the momentum equation.

the angular acceleration of a rotating assembly is:

dw/dt= T/Jtotal

where:
dw/dt is the angular acceleration in radians/ sec2
Jtotal is the total first moment of inertia of the assembly in kg.m2
T is the torque in N.m

In your circumstance the total first moment of inertia is the sum of smaller factors, the first moment of the ballscrew, the first moment of the armature of the servo and the first moment of the linearly accelerationg axis mass
refered to the rotating ballscrew

Jtotal=Jballscrew + Jservo + Jlinear

Jballscrew= 1/2. rho. r4. PI. l

where:
r=radius of ballscrew in m
rho=density of steel (aprrox 8000kg/m3)
l= length of ballscrew in m

Jservo is published by the servo manufacturer. For example my 750W Delta B2 series servos have a first moment of 1.16 x 10-4kg.m2

Jlinear= 1/40 p2. m

where:
m= axis mass in kg
p is the pitch of the ballscrew in m

I will start by reviewing the calaculation I made for my machine. I have 32mm 5mm pitch C5 ground ballscrews. They have a travel of 350mm but are 700mm long. The double nut reduces the travel somewhat.
These screws would be similar in size to what you would require for a 300mm travel. I have 750W delta B2 servos, as above. I assumed a linear axis mass of 150kg. The cast iron bed is 115 kg, and the balance is the
weight of the vice and the part.

Jballscrew= 0.5 x 3.141 x 0.0164 x .7
=5.76 x 10-4kg.m2

Jservo=1.16 x 10-4kg.m2

Jlinear=1/40 x 0.0052 x 150
=0.94 x 10-4kg.m2

Jtotal= (5.76 + 1.16 + 0.94) =7.86 x 10-4kg.m2

This is an exceptionally important result, and one you must understand, the rotational momentum is dominated(72%) by the ballscrew....not the linear axis mass(12%).

If you choose to use a 32mm diameter screw then it is likely that the momentum will be dominated by the screw, not the 900kg you pile on the top!!! In this regard pippin's comment about a 32mm screw being
overkill warrants consideration. The momentum of a screw is proportional to the fourth power of its diameter. Thus if you were to decrease the size of the ballscrew to 20mm (rather than 32mm) the
ballscrew momentum would diminish by 85%!!!!

The first moment of the linear axis mass is proportional to the second power of pitch...so pitch is an important design parameter also but much less so than the ballscrew diameter.

Lets say you use the same servo and ballcrew as in my machine (32mm diameter, 5mm pitch,700mm long and 750W Delta B2) and recalculate for 200kg axis mass and 900kg axis mass respectively.

Jtotal,200kg axis= (5.76 + 1.16 + 1.25) x 10-4kg.m2 ......so the linear axis is only 15% of the total momentum.
Jtotal,900kg axis=(5.76 + 1.16 + 5.52 ) x 10-4kg.m2........so the linear axis mass is now 44% of the total momentum.

Despite a heavy axis mass the rotating parts (the ballscrew and servo) still dominate the momentum. Is this what you were expecting?

All you have to decide now is what you want your acceleration to be and that will determine the torque required and therefore the servo required. Lets imagine you chose the Delta B2 as I have....what
would the acceleration be with the 900kg axis, the worst case?

dw/dt= 2.4 / 12.44 x 10-4.....where the rated servo torque is 2.4Nm
=1930 radians/sec2

This corresponds to a linear acceleration (determined by the pitch) of:
accel=1.53 m/sec2 or .15g
Note that this is the acceleration achieved with just the rated torque of the servo....not its overload rating which would be 0.45g!!!

My machine is somewhat lighter than yours and I get 0.27g at rated servo torque, but that is too scary fast for me. I dial it back to 0.15g and it's still scary bloody fast. Thus depending on what you want
to do even a 900kg axis can be accelerated by a 750W servo to 0.15g with a 32mm diameter screw and 5mm pitch. Is that enough?

The most important takeaway is that the momentum of the ballscrew is to the fourth power of diameter. Thus going from 25mm to 32mm implies and increase in angular momentum of the ballscrew
of 2.7 times! The contribution of the momentum of the linear axis mass is proportional to the square of the pitch. In many circumstances the rotational momentum dominates and the linear momentum is often less
the 25% of the whole.

In the post above I used my machine as an example. I used 32mm ballscrews (from dy-global) and 750W Delta servos of 2.4Nm (rated) torque. Even these large, you might say out-sized screws, still result
in 0.27g acceleration at rated torque. Your servos are somewhat beefier albeit at a lower speed, so in the same machine you might expect 0.36g at rated torque. Personal experience suggests that
is way WAY WAY enough. I detune my machine to 0.15g, and still find it scary fast.

Give me some numbers from your machine (diameter and pitch of the screw, the screw length, the linear axis mass and the first moment of inertia of your servo) and I'll do the calculation.
I think the conclusion will be that while 25mm is best suited, if you can get 32mm they too will be adequate. The acceleration will be somewhat reduced, but your servo will still spin up a 32mm
screw beyond your wildest expectation.

Craig

[peteeng]

Hi Lars - At D32mm the motor is basically turning the screw and not pushing the payload. So a D25mm screw is a much better match. Keep at it. In regard to starting a new thread just keep this one rolling... Peter

There is a possible advantage to D32 ballscrews. 1) if the cost is attractive (vs D25 screws cost) ie its surplus stock or fire sale and 2) the D32 bearings are higher and depending on your rail size the nut height may match the car height better, if you're mounting the rails and screw on thew same surface.

So check the geometry out before the final decision. In fact its good advice to not buy ANYTHING until you have resolved the design in CAD. Cost and source everything but sort the design first.... This will save you money in the long run. Most of us have gotten excited about something and bought it but its still sitting on the garage shelf as something else changed. So get serious in CAD... Peter

[joeavaerage]

Hi peteeng,

At D32mm the motor is basically turning the screw and not the payload. So a D25mm screw is a much better match

That is true.........BUT........I have 32mm C5 screws in my machine with 750W servos and they still accelerate wildly fast....so much so that I deliberately de-tune them ........down to 0.15g
and that is still way fast enough.

25mm screws might be a better match, but C5 ground screws are going to cost a fortune ($1000 each at a guess), whereas 32mm C5 ground screws from dy-global are like one third or less the cost,
and they will perform very VERY well despite being somewhat out-sized.

Craig

[joeavaerage]

Hi,
Just as an example, this is for an NSK 25mm ground C5 ballscrew 1m long:

https://www.acorn-ind.co.uk/p/nsk/st...05n1d0999-nsk/

That is nearly 1000GBP!!!!

New ground ballscrews cost a bloody fortune....you think servos are expensive...BS....ground ballscrews are way WAY WAY more expensive.

Craig

[joeavaerage]

Hi,

Most of us have gotten excited about something and bought it but its still sitting on the garage shelf as something else changed. So get serious in CAD... Peter

In the case of ground (C5 & C3) ballscrews I disagree. They are so damned expensive that they are effectively precluded for hobby use....so in order to get C5 or C3 screws you have to buy second hand
or new old stock....or something, otherwise you can never afford to get them. When it comes to ballscrews....find the right size (approx), quality and cost....and get them. Design your machine around
what ballscrews you can get...not the other way around.

C5 NSK and THK ballscrews of 20mm and 25mm are in excess of $1000USD per meter....and that's without the support bearings. I bought three THK ground C5 32mmx5mm pitch BNFN (double nut) screws including six THK FK25
support bearings second hand for $1000USD for all three including free shipping to New Zealand. They would have cost over $8000USD had I bought new. When I got them I'd swear that they had never been used
and still had the numbers matching manufactures test certificate. My machine would NEVER have been built if I had to buy new ballscrews. The dimensions, travels and even style of construction have all been determined
to get the best possible use from the ballscrews I could get.....not some ballscrews that I could never afford.

Craig

[catahoula]

yes not cheap, still almost $1000usd for the set of 3, but drastically less than THK, NSK etc
from my invoice (this was 3 years ago)

TBI Ballscrew
OFUR1605-DGC5-470-P1 1 Machining US$204.00 US$204.00 overall length
OFUR2005-DGC5-780-P1 1 Machining US$288.00 US$288.00 overall length
OFUR2005-DGC5-880-P1 1 Machining US$308.00 US$308.00 overall length

[joeavaerage]

Hi catahoula,
according to the part number they are rolled screws which would describe why they are that much more affordable, and while they might be rolled they are C5 (JIS B 1192).

BoschRexroth do the same thing, they roll the screw and then give it a light grind to get C5 spec, and a little cheaper. I think they must have had a cockup though, instead of taking
a zero off the price, some b****d added a zero to the price!!

Craig

[peteeng]

Hi Lars - What do you intend to use this Mill for and what accuracy are you aiming for? I say this because if you go with C5 (or better) you have to be able to make the rest of the machine commensurate with the ballscrews. Otherwise some other area of the machine will not meet that spec and let the machine down. A C5 screw has a V300 of 0.018mm. So it could be +/-0.009mm out every 300mm. This is simplified to +/-0.01mm accuracy. To make the rest of the machine to that spec requires machining and setting up to +/-0.001mm which is a tall task unless done on very good machines. You will achieve very good repeatability, better repeatability then using C7 screws V300 = 0.050mm so what do you intend to do with this machine? Peter