freejoth turret mill to cnc

[joeavaerage]

Hi,

i think i will go with the 1.8kw dmm ac servo motors and drives

They are great and powerful servos.

Hi rodger my reason for the 1.8 kw ac motors is because ac motors are less efficient than a their dc counterparts

That is incorrect, if anything the torque density of an AC servo exceeds that of a DC one.
I must say I agree with Roger, I think 1.8kW as axis servos is over the top.
We have a customer who has an aging but huge double column Okuma mill and its axis servos are 750W and it rattles through its work with
500kg on the table.

ac motors is less maintenance and no interference from the contact of brushes and commutator

Yes that is certainly true. What you may not be aware of is that the modern AC servo drives have control options which are light years, I mean
LIGHT YEARS ahead of all the old DC stuff. You are in for a pleasant surprise with those DMM's

Craig

[joeavaerage]

Hi,
maybe I should explain my assertion:

if anything the torque density of an AC servo exceeds that of a DC one

These may be called AC servos but they are not AC in the traditional sense.

AC induction motors are somewhat less efficient than DC motors of the same size, because some of the applied current is required to induce
the magnetic field in the rotor and therefore unavailable to generate torque..

An AC servo motor however is a synchronous machine. Its rotor is permanently magnetized, and by rare earth magnets with fields of
close to 1 Telsa. Thus NONE of the applied current is required to magnetize the rotor. Additionally the algorithm for applying the current,
called field oriented control, applies the current in such a manner that is always at 90⁰ to the rotor field and therefore always
maximally producing torque from the lowest possible current. Thus the torque density of a modern AC servo is as good or better than a
DC servo. Additionally the heat loss in a DC servo is in the rotor (difficult to cool) whereas its in the stator (easy to cool) in an AC servo.

There are a couple of YouTube videos posted by Texas Instruments about Field Oriented Control, truly fascinating, it gives any control engineer
a sweaty palms in anticipation.

Craig

[Sterob]

I got nuthin......



lol

[mechhamo]

hi

i think i need to a bit more research on the bob board, but i would like to thank everyone that has replied to my posts and as always i have learnt a lot, and your help is very much appreciated.

regards wayne

[Pearo]

I can now give info on the cnc technics break out board - that is, don't buy it. The board might be ok, but the documentation is atrocious. Unless they fix that up, steer clear.

[Sterob]

I can now give info on the cnc technics break out board - that is, don't buy it. The board might be ok, but the documentation is atrocious. Unless they fix that up, steer clear.

Are you going to persist or 'bail out' and buy something else?

[Pearo]

Are you going to persist or 'bail out' and buy something else?

I'll persist for now since I own it. I have managed to get it working, but anyone with out a solid experience in electronics I think would struggle.

[mechhamo]

Thanks for the advice pearo

After spending a lot of time reading on this forum i am amazed at what is possible, and after i convert my mill i would then like to use it to build a cnc router plasma combo, which brings me to my next question ? is there such a thing as a working reliable combo or would you have to build a dedicated one for each for the best performance.

regards wayne

[joeavaerage]

Hi,
a router and a CNC plasma table are similar, essentially 2.5 D mills.

The difference comes in the required rigidity and accuracy. A router has to contend with cutting forces that a plasma does not.
Plasma tables often have to travel fast but accuracy smaller than 0.1 mm is seldom required.

By the time you've built an accurate (0.01mm) router with the rigidity to withstand cutting forces using exotic and expensive components I doubt
you will want to spray the whole thing with cutting dross of a plasma!

Craig

[Pearo]

My next build is going to be a combined router and plasma, but I wont be chasing accuracy better than what is required for a plasma. I cant see myself doing it for a little while yet, but I am currently researching how I am going to do it. Plan is to be able to cut full 1200*2400 sheets of timber.

CNC Router parts do a combo which is worth a look at.

[mechhamo]

i can see what you mean craig ,i was thinking of making one end the router end and the other the plasma with the router part higher,
craig would it be less messy if i made a water bath.

like pearo i would make it 3 meters 3 meters so i could cut a 2400x1200 sheet of timber one end and the other end a sheet of steel 2400x1200 would be awesome

thanks wayne

[joeavaerage]

Hi,
that makes it worse not better. The bigger the table then to achieve a given level of rigidity the cross sections of materials
you have to use go up exponentially and weight likewise. With increased weight comes increased horsepower from your servos
and on it goes.

If your table had 1200 mm travel in X then you could use a longish rolled ballscrew, maybe $200 or so. At 3m you'll have to use
rack and pinion, with backlash adjustments $500 or more.

When it comes to CNC big means very VERY expensive.

I would guess that you could build two 1200 X 2400 tables for less than you could build one 3m X 3m.

Craig

[hanermo]

Yes.
No.

ALL posts re weight and mass grossly overestimate the effect of very heavy tables and beams.
They are practically immaterial for hobby use.
"10x heavier" in steel is usually just about right.

A tiny 40$ 3Nm nema 23 stepper effortlessly moves my old 1600x500 mm steel milling machine table with 100, 200, or 300 kg of stuff on it.
The table is 200 kg in mass and 3 cm thick tool steel.

The commercial guys, I used to be one, have to use bigger servos to get silly rapids (never used to effect) - due to advertising and competitors propaganda.
So the same 1000 kg mill table, 15 cm thick cast iron honeycomb, today has 2 kW servos instead of 1 kW like 5-10 years ago.

3m screws and linear guides are std stock items and not at all expensive.
I am happy to sell and source original hiwin equipment in europe, if someone needs some.
Just use about 25mm++ guides and 32 mm screws.
I use 32 mm screws on my lathe and VMC (35 mm guides).

The VMC z mount without the assy is == 150 kg and has == 8kg stiction and under 1kg friction.
Strength is theoretically about 16 metric tons.
The tiny stepper moves it fine, I switched to 400W ac servos, 290€ all-in, for better accuracy and publicity.

Hi,
that makes it worse not better. The bigger the table then to achieve a given level of rigidity the cross sections of materials
you have to use go up exponentially and weight likewise. With increased weight comes increased horsepower from your servos
and on it goes.

If your table had 1200 mm travel in X then you could use a longish rolled ballscrew, maybe $200 or so. At 3m you'll have to use
rack and pinion, with backlash adjustments $500 or more.

When it comes to CNC big means very VERY expensive.

I would guess that you could build two 1200 X 2400 tables for less than you could build one 3m X 3m.

Craig

[mechhamo]

I would guess that you could build two 1200 X 2400 tables for less than you could build one 3m X 3m.

That is probably correct craig ,so you think keep the size down and use ball screws rather than rack and pinion on both x and y and drive ball screws directly with servo motors,what would be the maximum length of a ball screw without getting whip and obviously the dia of the screw changes that.
You don't seem to be a fan of rack and pinion,another thing i would like to know is how good are the linear rails and carriages or are there some poor quality ones that i should stay away

thanks wayne

[joeavaerage]

Hi,

You don't seem to be a fan of rack and pinion,

No thats not quite correct. Rack and pinion are good and the only option beyond 2m-3m because a ballscrew will whip.
The problem is that all rack and pinions have backlash. There are some very clever (and expensive) pinion drives that can
improve or eliminate backlash.

The cost of the rack and pinions aren't bad but when you add the low-lash reduction box....that gets expensive.

I use C5 ground ballscrews on my mill, they are short, only 500mm overall and I got them second hand, they are way too expensive
for me to buy new. They are preloaded...so no backlash and have a manufacturer guaranteed 18um per 300 mm accuracy and 8um within
one turn. In short they are very good, with such ballscrews you can make parts to 0.01mm with confidence.

You will not achieve that with rack and pinion.

I deal with plasma tables at work and even with water tables everything ends up with black gritty grime over everything. If you want to pay
big dollars for nice components like rails and ballscrews and then use them in a plasma table where they WILL get covered in rubbish that's your
call.

Craig

[mechhamo]

Hi craig thanks for the advice.

What dia ball screw would be needed at 3 meters to get no whip, and because i want to drive the ball screws 1,1 what pitch ball screw would i be looking at needing 10 mm pitch or 20mm pitch,
you said that you use ground ball screws does this mean that rolled ball screws are not great but are they useable.

If i used ball screws is it any advantage using double ball nuts.

thanks wayne

[joeavaerage]

Hi,
hanermo previously responded that 32mm diameter screws would be adequate at that length. I have no experience with long screws, in
fact I can't find any screws over 2.5m in 32mm size. As to the pitch.....I would take what I can get. The high pitch types allow high axis
speeds at low rotational speeds, ie good for avoiding whip, but then require high torque servos or gear reductions to get the required torque.
5mm pitch screws will provide greater thrust with direct coupled servos but you may be speed limited before the screw starts to whip.

Rolled screws are commonly C7 grade, 50um per 300mm accuracy but up to 30um in one turn. They are adequate but not for highly accurate
machines. You should avoid the 'T' grades, they are for transport purposes and not accurate enough for CNC axes.

A 1200mm X 20mm C7 rolled screw with preload ballnut without bearing blocks might cost around $75. A ground C5 of the same dimensions
might cost $1500. Unless you want to spend mega-bucks rolled C7's are your choice.

Double nutted ballscrews are the original backlash elimination technique. The flexibility with which they can be applied makes them still a good choice
despite the extra expense and loss of travel.

Craig

[mechhamo]

Hi craig since your last post i have been looking for ball screws that are 25 mm or bigger they are very few and far between and costly
a 32 mm ball screw and nut 3000 long are about $600 each 2 for the y axis and 1 shorter one for the x axis,now i can see if you want it to be very accurate and at a lower cost make it smaller and stronger,but i would be kicking myself if i was limited by size at what i can do after i built it i think i will make it 3000 x 1500 which will give me room around a full sheet and because i will be making most of the parts in the build myself i can afford to put good components in the rest, even with the cost of the ball nuts linear guides and rails ac servo motors and drives i recond i can build a 3000x1500 super strong table for $5000au or less which is still a lot cheaper than buying one of the same quality, i have never been one to do a half a job it's not me.

are machined ball screws any good and do you rate them is it (ground > machined and then rolled

regards wayne

[joeavaerage]

Hi,
yes long ballscews are expensive. I would not go less than 32mm at that length or the rotational speed of the screw will be limited
by its potential to whip.

For your budget, assuming you use ball screws at all, then rolled screws are your only choice. You could spend your entire budget
on one ground ballscrew alone.

I would rate rolled screws as suitable for machines with accuracy of 0.05mm to 0.1mm whereas ground screws are required for
0.01mm accuracy.

You should consider rack and pinion:
discounted automation products: Motion Control, Rack and Pinion
Cost effective depending on the sophistication of the pinion drive.

Craig

[mechhamo]

I think i could save a bit by making the y axis 1200 and the x axis 2400 that way that way the two ball screws for the y axis will be cheaper and use on long screw on the x.

i did want to machine aluminium parts for engines with high precision, but after what i have read and been told i think it might be a good idea to build a plasma table than i can learn from that build and it will make me much more aware of things that might need changing in the router build.

I think i need to rethink the size of the router build because any reduction in size is going to be a lot cheaper to build, and i also have to ask myself how much more work am i going to get by machining objects larger than a meter x meter which would be a lot cheaper to build and you could make it precise and strong.

!!! lots to think about

regards wayne