3 Attachment(s)
New (to me) Build - Mini CNC
Hello all! Long time lurker, first time poster. Today I picked up the guts of someone elses's cnc project, and I'm looking for a few pointers from the experienced builders. The person I bought this from was planning to build a machine out of it, but didn't get the time. I don't know the history before that. I believe it's part of someone's DIY build. Everything is put together nicely.
What I've got is a machine that looks like it is designed to operate like a knee mill, the X/Y axis would ride on the Z with a fixed head. I don't have a column or a head that came with this. Building a large column seems like it would be a bit inefficient compared to disconnecting the table from the Z and mounting a spindle to the Z axis like a traditional VMC. Open to suggestions here??
What I have is not super strong, but it seems like robust enough that I should be able to mill aluminum. The entire frame is made from aluminum. Parts list is:
Nema 24 on X/Y axis with 7/16" ball screws - Sanyo Denki 103H7823-1740 stepper
Nema 34 on Z with lead screw - Sanyo Denki SM2862-5255 stepper
15mm rails on X/Y
20mm rails on Z
I'm not sure what size spindle would be a good match for these rails/steppers. My mill is BT30, having shared tool holders would be awesome, but that just seems way to big for this little machine.
For a control system, I'm thinking Raspberry Pi with LinuxCNC. Mesa for the breakout/control boards?
This is all new to me. I wanted to build a little machine like this to learn the ins and outs of doing it. And if I can make aluminum parts along the way, that would be great. In the future, when I have more space, I would like to get a fullsize VMC to retrofit, so I thought learning some of what to do and what not to do on a little machine would be smart.
Thanks for any and all input!
Re: New (to me) Build - Mini CNC
On second thought, reading more about using a Raspberry Pi, it seems problematic. Barebones PC seems like a better bet now. Plenty on ebay for the price of a raspberry pi. I was reading on the linuxcnc website that built in graphics may not work great and an external card was recommended. At the same time, it sounds like that data was from 2008...
For a Mesa card, it looks like the 7i96 is the way to go.
Re: New (to me) Build - Mini CNC
I figured I could edit my above post again, but I don't seem to see the button. Anyway... Been doing my reading on spindles. It seems like the Chinese 1.5 and 2.2kw spindles common on eBay are suited for wood/plastic only. A BT30 spindle is about $260, ER16 is about $130. Alternatively Taig makes an ER16 spindle for about the same price. With all these options, I could either use a 110v motor and pulleys or I have seen some have adapted a servo motor. I don't know how to do that just yet, but that seems like it could make for nice controls!
Re: New (to me) Build - Mini CNC
Hi,
I made my own mini-mill about seven years ago. It was not entirely dissimilar in size to what you have.
I fitted a 800W 24000rpm ER16 Mechatron spindle and VFD....its been brilliant. I don't know how many thousands of hours use I have had but lots and lots.
It works great on circuit boards, my main use, but also on plastics, aluminum and brass.
What it will not do is mill steel or at least very well. Steel, in fact all ferrous metals, require a spindle of much higher torque but much lower speed so about the same power, just
much much slower.
In order to do steel I made a second spindle using a secondhand 1.8kW 3500rpm 6Nm (cont) Allen Bradley AC servo, a Rego-fix cylindrical ER25 toolholder and P4 angular
contact bearings. It too has been great. The only real problem is that it has enough torque to cause my mini-mill to flex...so I have to take it easy to
accommodate the modest rigidity of my mini-mill.
Over the last few weeks I have commissioned my new build mill. It has 115 kg cast iron axis beds, twin "L" 32mm med-tensile steel frame, 32mm doublenut C5 THK ballscrews
and THK HSR 20mm linear rails and cars. All-in-all a much MUCH bigger and more rigid machine. My 1.8kW Allen Bradley based spindle has now come into its own,
it's a much better match than my mini-mill.
The important point here is that the spindle must match the machine. Having a high torque slow revving spindle is great for steel but what's the bet it will twist your little
machine into a pretzel. I personally have found that the low torque high rpm spindle to be much more useful. Addmittedly its not much good on steel but on everything else it
has been fine.
May I suggest that a 1.5kW 24000rpm spindle would be a good match. I would recommend water cooling as it means you can use the VFD to slow the spindle
to 6000-9000 rpm without huge risk of overheating it. An air-cooled spindle would very likely overheat by comparison.
My experience is that such a spindle will give you lots of service and will make lots of small to medium parts for you, and in short you'll have a lot of fun.
Having a BT30 toolholder would be nice, but you pay one hell of a premium for an ATC type spindle. A straight manual ER20 or ER25 spindle is less convenient to use
but likely 1/3-1/2 the price.
Having said that my next spindle will be ATC, BT30 probably. I have a second hand remanufactured 3kW 3000rpm (rated) 12Nm (cont) AC servo from circa '95. I want
to use it for another high(er) torque low speed spindle, again for steel and stainless.
Craig
Re: New (to me) Build - Mini CNC
Thanks Craig, it must have been your post I saw with the Servo/ER spindle you built. Do you have any pictures of that spindle build? I tried to find if you had a thread or anything on that, but I didn't find one.
Something like this in combination with an ER toolholder build like yours might be a good in-between. Smaller servo than yours, but looks like 5000 RPM rated.
https://www.ebay.com/itm/23409492325...QAAOSwp9Vg9M0K
I have not worked with anything other than metals in my shop, with the majority being aluminum, so I don't have much need for the high speed end of a router type spindle. So a servo like the above would be enough RPM for aluminum even in a 1:1 drive while still being able to drop down to a slow enough speed for steel. Price-wise, it looks like I could build a spindle like that for a similar price to a 1.5kw water cooled 24,000 rpm one.
Reading more about driving the servo... It seems like a matched servo drive and servo motor are needed. A VFD works in some cases but not ideal? Did you find an Allen Bradley drive to go with your motor?
Re: New (to me) Build - Mini CNC
Quote:
Originally Posted by
joeavaerage
Hi,
I made my own mini-mill about seven years ago. It was not entirely dissimilar in size to what you have.
I fitted a 800W 24000rpm ER16 Mechatron spindle and VFD....its been brilliant. I don't know how many thousands of hours use I have had but lots and lots.
It works great on circuit boards, my main use, but also on plastics, aluminum and brass.
What it will not do is mill steel or at least very well. Steel, in fact all ferrous metals, require a spindle of much higher torque but much lower speed so about the same power, just
much much slower.
In order to do steel I made a second spindle using a secondhand 1.8kW 3500rpm 6Nm (cont) Allen Bradley AC servo, a Rego-fix cylindrical ER25 toolholder and P4 angular
contact bearings. It too has been great. The only real problem is that it has enough torque to cause my mini-mill to flex...so I have to take it easy to
accommodate the modest rigidity of my mini-mill.
Over the last few weeks I have commissioned my new build mill. It has 115 kg cast iron axis beds, twin "L" 32mm med-tensile steel frame, 32mm doublenut C5 THK ballscrews
and THK HSR 20mm linear rails and cars. All-in-all a much MUCH bigger and more rigid machine. My 1.8kW Allen Bradley based spindle has now come into its own,
it's a much better match than my mini-mill.
The important point here is that the spindle must match the machine. Having a high torque slow revving spindle is great for steel but what's the bet it will twist your little
machine into a pretzel. I personally have found that the low torque high rpm spindle to be much more useful. Addmittedly its not much good on steel but on everything else it
has been fine.
May I suggest that a 1.5kW 24000rpm spindle would be a good match. I would recommend water cooling as it means you can use the VFD to slow the spindle
to 6000-9000 rpm without huge risk of overheating it. An air-cooled spindle would very likely overheat by comparison.
My experience is that such a spindle will give you lots of service and will make lots of small to medium parts for you, and in short you'll have a lot of fun.
Having a BT30 toolholder would be nice, but you pay one hell of a premium for an ATC type spindle. A straight manual ER20 or ER25 spindle is less convenient to use
but likely 1/3-1/2 the price.
Having said that my next spindle will be ATC, BT30 probably. I have a second hand remanufactured 3kW 3000rpm (rated) 12Nm (cont) AC servo from circa '95. I want
to use it for another high(er) torque low speed spindle, again for steel and stainless.
Craig
Are you not able to mill steel with a smaller bit? Even with 24krpm you should be able to match the SFM and CPT up with your spindle power for the material and not flex your mini-mill.
Re: New (to me) Build - Mini CNC
Hi,
Quote:
Thanks Craig, it must have been your post I saw with the Servo/ER spindle you built. Do you have any pictures of that spindle build?
No I don't have any photos on hand. I will post some in a day or two, we have just entered a new lock down and it may be several days before I can get in to work.
Of the two spindles I have, the 24000rpm Mechatron and the 3500rpm Allen Bradley the Mechatron sees 95% of all use. Excepting steel the high rpm spindle is brilliant.
Quote:
Are you not able to mill steel with a smaller bit? Even with 24krpm you should be able to match the SFM and CPT up with your spindle power for the material and not flex your mini-mill.
Yes, I can, and have, and still occasionally do, run my spindle down to 9000rpm with a 3mm tool and cut steel. The problem is that it has so little torque (0.3Nm) and consequently
any decent chip load will stall the tool....so you have to take very light cuts.....but it does work.
The drawbacks are:
1) Very long cycle times
2) Light cuts means that very little of the tool in engaged and so when that part of the tool which is engaged is dull 99% of the rest of the tool flanks are untouched.
This is a poor use of tools.
3) My spindle is air-cooled and running at 9000rpm is a sure way to overheating. I make sure that if I do so that I allow plenty of time for cooling or even run the spindle up to 24000rpm
at no load every few minutes or so for 30-60 secs just to ensure the spindle stays cool.
As a consequence of these problems you might guess that I consider the high speed spindle not a practical proposition for cutting steel. Yes it can be done, and for a very small tasks
when I just can't be bothered to swap the spindle over that's what I do. When I do swap over when I have a larger or more complex job on hand the extra torque and cutting forces
have to be considered....but that's normal. Even the great big Cincinnatti mill I've used, many tons of cast iron, you still have to use commonsense when using high torque/low speed with
large diameter tools, and surprise surprise I still have to use commonsense with either my first mini-mill or my new build mill.
As far as making your own spindle......it is not quite as straight forward as it sounds. For instance I used a Rego-fix ER25 toolholder, but Rego-fix being Swiss is eye-wateringly expensive.
I paid $450NZD (about $320USD) for it.....I was in 'perfectionist mode' at the time. Matched P4 angular contact bearings are not cheap either. Then when it came to making
it I realised that really my skills and the lathe I had were not really up to the task and just about all the bits I would have to send out for grinding....all beyond my budget at that time.
In the event I just had to accept the fact that I would have to make it myself with whatever level of skill and equipment I had or budget would preclude me from having anything to show for
the money I had already spent. So thats what I did....and I have to say am quite proud of the result....but its far from perfect. The bearing lands should have been ground etc. None-the less
I have a usable spindle and I've had many hours use from it. All up it cost me about $2000NZD ($1400USD) which is competitive with what was available at the time. Given that this is a hobby
I'm quite happy, if however I was doing it as a business the equation does not look so good.
As far as powering an AC servo with a VFD....been there....tried that....don't waste your time or your money.
If you buy a servo buy the MATCHING servo drive....especially if this is your first servo. Once you have fiddled around with servos for a while you soon realise that despite the plethora
of manufacturers all servos are very similar, if you can operate one then you can probably operate all of them. The downside is the learning curve is VERY steep and you do not
want to make it any harder by non matching servo and drive.
There are plenty of very cheap Chinese brands of AC servos, and while I think they are cost effective and probably good enough, the documentation and support is crap. Avoid like
the plague. All decent servos have manufacturer specific set-up and tuning software, take advantage of it. While often you can program a drive by pushing buttons like a microwave
its very tedious and error prone....don't try it. For my new build mill I bought three 750W Delta B2 series (160,000 count per rev) servos which have the aforementioned set-up software,
they are brilliant. If buying new do not stoop any lower than Delta (Taiwanese made in China) or DMM (Canadian made in China), both good quality, well supported at fair prices.
When I bought my Allen Bradley servo I found it on a NZ auction site and I bought it for $900NZD (650USD) for the servo and drive including shipping. I still had to find/buy the cables,
in the event I was able to make one, but the plug of the other required that I buy a new cable for the other. I also paid $200NZD (140USD) for the set-up software, the 3000 series
drives CANNOT be programmed WITHOUT the software. Still the 1.8kW servo, drive, cables and software cost me $1400NZD (980USD) which I consider was fair value.
Craig
Re: New (to me) Build - Mini CNC
I appreciate all the input!
I will do a bit of digging into a matching driver, cables, and software to the AB servo I found. Part of me wants to go servo to do my learning on this machine, when I opt to build a bigger machine in the future, I'll have a stronger background. I see the Chinese servos you mention, the price is tempting. Ive dealt with the cheap Chinese VFDs before, I know what you mean about documentation...
On the ER collet you built, I see what you mean about the difficulty building it. That's where I think buying the offthe shelf spindle from that taig would be a good bet. Only Er16, but I think that's fine.
Re: New (to me) Build - Mini CNC
Hi,
Quote:
I will do a bit of digging into a matching driver, cables, and software to the AB servo I found. Part of me wants to go servo to do my learning on this machine, when I opt to build a bigger machine in the future, I'll have a stronger background. I see the Chinese servos you mention, the price is tempting. Ive dealt with the cheap Chinese VFDs before, I know what you mean about documentation...
Yes, that is a valid approach. Now that I reflect on my experience with my first mini-mill I've learnt heaps and that has placed well to make all new sorts of mistakes with my new mill!!
While I take issue with documentation and support of the cheapest Chinese made servos I can personally recommend Delta. Addmittedly not the very cheapest....but close.
For instance this is a 750W B2 series, complete with servo, driver and cables for $438USD plus shipping. Note that you will need a IEEE1394 cable to program the drive, maybe another
$30 or so.
https://www.fasttobuy.com/flange-80m...er_p28084.html
Buying secondhand requires patience. It seems there are plenty of secondhand servos but few good drives to suit. Probably not surprising, after all the electronics are more likely to fail than the servo.
Replacement drives for old servos is a losing battle, you'd be better off buying a new matched pair. I got lucky with my Allen Bradley servo. As it turns out the 2098-DSD series of Allen Bradley drives
are still in good supply at reasonable prices and drive all the 3000 series servos nicely. Note that various encoders were used throughout the 3000 series production run requiring a matching drive,
your research is required.
This one is single phase input but capable of driving a 1.8kW servo:
https://www.ebay.com/itm/18174907121...QAAOSwrklVWfAq
You could, at a stretch, use it to drive this 2.2kW 5000 rpm 6Nm (cont) servo:
https://www.ebay.com/itm/18484721802...IAAOSw~55gpkVI
Although the somewhat harder to find 2098-DSD-030 drives would be better:
https://www.ebay.com/itm/25484025724...YAAOSwEvZgCQ9B
Note that all the 2098-DSD series drives are single phase input and the 2098-DSD-030 drive is the largest output (15A cont) of any single phase drive I've ever found.
It is marginal whether you are saving any money buying secondhand however. If you want a spindle motor then this is hard to beat:
https://www.fasttobuy.com/delta-2kw-...nc_p33936.html
A complete kit, 2kW, 6Nm at 3000rpm (rated) for $895 plus shipping is pretty good value.
What you may not be aware of is a mode of operation called 'field weakening mode' which is available to AC servos. It allows the controller to, in very clever manner, reduce the effective
armature magnetisim and therefore exceed its rated speed by in effect, reduction in BackEMF constant. As I say very clever indeed. For example my 750W B2 series servo is rated at 3000rpm
but is capable of 5000rpm in field weakening mode which allows blindly fast G0 rapid traverses. I would guess you could push the 3000rpm (rated) of the 2kW servo to which I've linked
out to 5000 rpm using the same principle. Not cheap as spindle motors go but you'd have up to 18Nm at overload, very impressive!!
Craig
Re: New (to me) Build - Mini CNC
Hi,
my Delta servos are rated at 3000rpm but perform more than adequately at 5000 rpm. Note that I only use that 'extra' on G0 moves, ie rapid traverses which occupy such a small percentage of operating
time that I've never encountered any propensity to overheat. Above the rated rpm the power is near enough constant, and therefore the torque decreases approximately linearly from rated (2.4Nm)
at rated speed to about 1.45Nm at 5000rpm.
If you were to use such a servo as a spindle motor continuously in the field weakened zone then yes it would be wise to consider some derating. I have not tried.
Noticeably the Allen Bradley servo I use as a spindle motor does not appear to offer field weakened operation, it runs up to rated speed but no higher. It's probable that if I dug around in the multiple
screens of tuning data that I might find it, but it works perfectly well as is and have little interest in trying to expand its envelope in this way.
As I posted earlier I have a 3kW servo that I wish to use as a spindle motor. It is equipped with a resolver rather than an encoder which very much restricts the choice of drives. Given that electronics is my
thing I have elected to make my own Field Oriented Control drive. Its proving to be a learning challenge but I am extremely impressed by Texas Instruments C2000 MCUs, with all the
'fruity bits' for motor control. The servo is rated to 3000 rpm but the manufacturers nameplate specifies 4000rpm max. It rated line-to-line voltage is 180V, which suggests that with a supply
derived from 230VAC, ie 320VDC, I should be able to achieve 4000rpm with little field weakening provided I limit current to prevent overheating.
I am thinking that I will have to use power factor correction in order to tame the input supply current demand, in which case the DC link voltage will be closer to 400VDC, and I might get my 4000rpm
without any field weakening whatever.
Its a good project but I have other things to do first....like way covers for my new mill, new high volume coolant pump, chip filtration/conveyer, more lighting, line reactors for all servos/VFD and
existing servo based spindle motor.....and the list goes on.
Craig
Re: New (to me) Build - Mini CNC
This is a lot of info to take in. I've got a lot of research to do! I appreciate all the info.
The drives/motors you link to above are all in the 2kw region. I was thinking for my machine, about 500w to 750w would be plenty for a spindle motor. The rails are HG series 15mm for X and Y, so not too big. But I also have no experience with them to know how much of a cut they can take.
Re: New (to me) Build - Mini CNC
Hi,
well by all means make the spindle out of a 750W B2 series Delta servo, $438USD plus shipping for servo, drive and cables.
The other alternative is to make one in the region of 2kW so that when you do build your bigger machine it will swap over to it. After all there is no rule
that says you have to use all the power and torque of a given spindle. Like my Allen Bradley based spindle I had to use the power and torque judiciously
in my mini-mill....but its a good match for my new mill.
Craig
3 Attachment(s)
Re: New (to me) Build - Mini CNC
Couple updates on my build. Found a lathe bed locally for sale, it's from an Emco Maximat. 6" wide, about 40" long. Perfect for a base and column in this build. Probably overkill. I could always get bigger rails and balls screws in the future with this frame.
Sliced it in half with my horizontal band saw. Cleaned up the ends with a 6" long end mill. I had a hole box full of these end mills I bought at auction, not realizing how big they were at the time of purchase... Haha. Unfortunately, my Wells Index has a fixed head, so I can't easily rotate it and hang this table off the edge of the table. The good news is that it's more stiff than an average bridgeport, so it actually handled this 6" cut fairly well.
Still haven't bought a servo. It looks like the Allen Bradley software is free now and the Ultra 3000 drives are popular to use, even with other brand servos. So it's looks like that's my route.
Re: New (to me) Build - Mini CNC
Hi,
Quote:
It looks like the Allen Bradley software is free now and the Ultra 3000 drives are popular to use, even with other brand servos.
When I tried about four years ago to download the software I couldn't get it anywhere, hence I bought it.
Be aware that the Ultra 3000 drives were made to accommodate maybe ten models of servos, differing mainly in the type of encoder. Download the manuals
and read them carefully so that you can identify the part number variations.....there is a lot of them.
That is a good use of the lathe bed.
When I made my mini-mill seven/eight years ago I got some great big rectangular cast iron elevator weights from the scrap yard.
The cast iron was not particularly good but I did end up with some good axis beds once I milled the rime off the outside.
Craig
