PDA

View Full Version : Arvid's first CNC machine!



arvidb
10-30-2003, 07:50 AM
Hi all!

This is my first message to this list and I'd like to start with saying that this really is a nice place! I've been checking the board for new messages something like three times a day or so since I found it :) (Thanks balsaman for the link!)

I've had the thought of building some kind of CNC router or milling machine for some years now, but haven't had neither the space nor knowledge enough to feel ready to begin until now. During the last year or so I've been doing some construction (CAD) work at work, and also some manufacturing of parts on a small desktop manual mill/drill, and I've learnt a lot. Enough to feel ready to begin building!

This is what I had in mind:

* work area 600 mm x 420 mm x 100 (150?) mm (This is the same size as an A2 paper, nice if I want to plot large drawings or stuff like that. Also big enough for loudspeaker building :)
* repeatability better than 0.1 mm (preferrably <0.05 mm)
* ability to cut aluminium, wood, delrin etc

I've done quite some research and this is what I have found out (feel free to comment on this, i's all "hearsay", I haven't tried it):

The demand to cut aluminium is the controlling factor. It requires a machine with high rigidity and no backlash, even for lighter cuts. Wood and aluminium also requires different types of spindles, where the aluminium needs lower speed/higher torque unless the machine has the power and rigidity to sustain a really high feed rate.

Ok, so the machine has to be rigid. That doesn't really help :-/ ... I decided (by guessing) to aim for a machine that can take cutting forces of about 1000 N. This is what the linear guides, leadscrews and motors have to be able to handle. I have no idea of how much the MDF machine frame will flex under this load, or what kind of cuts I will be able to make, but we'll see :)

I don't have access to a machine shop, so the machine will be built out of MDF. To get the machine as rigid as possible, it will be a fixed bridge moving table machine.
After looking at different choices of linear guides, leadscrews, and motors, I think a reasonable price to aim for would be $1000 for these parts and the MDF.

Parts:

* X axis (600 mm) linear guides: I will use 20 mm dia, 1100 mm long guide shafts and linear ball bearings from INA for this one. The guide shafts are unsupported which is not good, but since I already have them I'll use them. I will support them at the middle in addition to the ends, which will limit my travel to about 500 mm, but I can live with that for now.
* Y axis (420 mm) linear guides: I bought real linear rails and slides (THK HSR25) from ebay for this one. Quite a lot of money, but very much fun :cool: I got them just a few days ago, and they're beautiful!
* Z axis (100 mm) linear guides: haven't decided yet, want to compare the linear shafts and the linear rails first.
* Leadscrews: I first thought I would use ACME lead screws with anti-backlash Acetal nuts from Ball Screws and Actuators (BS&A). The 16 mm version have a design load of 175 lbs (about 775 N), and the 20 mm ones are rated for 250 lbs (about 1110 N). I can buy these locally. However, it would be nice to find a used ball screw to play around with, at least for one of the axes. Probably the Y one, since it'll also have the nice linear rails :) I will use axial bearings and radial bushings (both from FAG) with these. This is the cheapest solution I've found that's backlash free. I found a local source for these as well.
* Motors: one design goal is 1000 N of cutting force. I want to be able to cut at 0.01 m/s at this force (guessing again). Lets also say I want a top speed of 0.05 m/s (about 120 IPM). 1000 N and 0.01 m/s gives 10 W of power at the table (P [W] = F [N] * v [m/s]).
To figure out what speed and torque I need, I used the following equations:

* n = v/s, where

n = motor speed [r/s, revs per second]
v = feed rate [m/s]
s = lead of screw [m/r, meter per rev]

and

* M = F * s/(2*pi) * 1/Eff, where

M = motor torque [Nm]
F = force at table [N]
Eff = efficiency of screw, about 0.4-0.5 for the BS&A ACME screws, and about 0.9 for a ball screw.

The lead of the BS&A screws are 4 mm/turn, or 0.004 meters/turn. This gives: n = 0.05/0.004 = 12.5 r/s = 750 RPM, and M = 1000 * 0.004/(2*pi) * 1/0.4 = about 1.6 Nm. Now, with a belt reduction of 2.5:1, this becomes 1875 RPM and 0.64 Nm.
(Hmm... 1875 RPM = angular velocity of 1875/60 * 2*pi = about 196 [rad/s]. 196 [rad/s] * 0.64 [Nm] = almost 126 watts! (Angular velocity [rad/s] * Torque [Nm] = Power [W].) But only 10 watts at the table! Where did all the power go? The answer is that I don't need maximum speed and maximum cutting force at the same time. I do however need a motor that can supply 0.64 Nm at low speeds, and have a top speed of at least 1875 RPM while still supplying some torque.)
After looking around a bit I found a source of used, 200 watt Yaskawa servo motors with encoders. These are WAY overkill (stepper motors would be better suited to this project, I think), but I have fallen in love with them :) and they're not too expensive. They will be very difficult to drive however since they are 3 phase AC motors. I will have to build a servo driver board for them, but I think I can do this (and if I succeed, perhaps I can start selling the drivers and actually make a profit out of this! :-D). They have a rated/peak torque of 0.637/1.91 Nm, and a rated/peak speed of 3000/4500 RPM. I ordered one motor a few days ago, and it'll be here in about a week if everything works out fine. I long to hold it, caress it ... *ahem* ... let's just say I'm eagerly awaiting it :)

I think that's it for now (wow, that's a lot of text! :)) Attached is a CAD drawing of the machine frame. It's preliminary since I don't have all the parts yet, but should give you an idea. I used Pro/DESKTOP to draw it, which can be downloaded for FREE from www.ptc.com. You have to register (free) to get the activation key, though. The program has some annoying bugs and quirks but is really easy to work with. And did I mention it's free? :-D

Please feel free to comment.

Arvid

arvidb
10-30-2003, 09:58 AM
... and here's a pic of the linear guides. Aren't they beautiful? :)

Arvid

Mr.Chips
10-30-2003, 11:10 AM
Arvidb,
Welcome to the CNC Zone. I'm a newbie too, and in the process of building mine. I'll let the resident experts give you advise, just wanted to say you have a good start and it looks pretty rigid, what with the box construction and bracing.
I purchased Turbocad but haven’t begun to use it. Will build mine from sketches.
Regards,
Hager

machine ed
10-30-2003, 11:38 AM
Hi,
I'm a newbie to CNC also, some of the machines for "first timers" are REALLY making me humble!
ARVIDB, looks good!
I plan on starting as simply and cheaply as I can. I'm collecting linear bearings, shafting, motors, etc. but have a question. Has anyone used UHMWPE or Delrin for the linear bearing? (drilled and reamed for class 2 running fit?)
I'm drafting a 8-1/2" x 11" cutting area machine for my first and will post for critiquing when ready.
Thanks for the input in advance.
This is a wonderful place to learn, I'm thankful I found the Zone!
Take care,
Ed

Mr.Ed
10-30-2003, 11:45 AM
Hi there, welcome and join the club.

What i was wondering is, you desided to use Linear slides on your Y-axis. Great idea, i'm still working on my first machine (without slides) and planning on using them on my second machine.

So why not using them on your x-axis as well?

Ed. (It's crowded with Ed's overhere) :D

arvidb
10-30-2003, 04:39 PM
Thanks for the welcomes, folks!


Originally posted by Mr.Ed
Hi there, welcome and join the club.

What i was wondering is, you desided to use Linear slides on your Y-axis. Great idea, i'm still working on my first machine (without slides) and planning on using them on my second machine.

So why not using them on your x-axis as well?

Ed. (It's crowded with Ed's overhere) :D

The reason is cost - linear slides are expensive! I bought mine on eBay, and the shipping to Sweden was 30% of the total cost. Then add duty (some 4%) and VAT (25% in Sweden) and they get REALLY expensive! (Actually I haven't had to pay any duty or VAT yet, but I expect an invoice will show up soner or later...)

I have tried to find these kinds of things (ballscrews, linear slides, servo motors) here in Sweden, but it's not very easy, unfortunately. Actually, no luck in the whole of EU! Perhaps one would need to search in German, which I don't speak... Anyone's got any good European sources?

Arvid

Mr.Ed
10-30-2003, 04:48 PM
Thats's my point !

I just bought one set of linear slides on ebay as well, and i have to pay 95 us dollars for shipment alone. :eek: :eek:

I opted for quality so there is nobody to blame but me, but still...

I have allso searched everywhere, and keep on ending up at the other side of the ocean. :mad:

I think this will encourage me even more to perform better and build to the best of my ability, trying to make something out of this.

Anyway, balsaman made me addicted to these slides, i will just blaim him. :D

Ed.

balsaman
10-30-2003, 08:41 PM
Ahh, it's all my fault again is it? I was just minding my own business....

:)

Eric

arvidb
12-21-2003, 03:22 PM
After two weeks in transit, two used ballscrews landed on my doorstep (well, actually at my postal office, where I collected them :))

I will use these for my Y and Z axes. Still have to order ACME lead screw for the X.

arvidb
12-21-2003, 03:24 PM
Closeup of the Z axis one (about 130 mm of travel):

arvidb
12-21-2003, 03:29 PM
... and a closeup of the angular contact ball bearings of the Y axis one. The nut that keeps them in place have been loosened and the bearings separated somewhat. I'm actually not really sure how to mount these, so if anyone have an idea...

cncadmin
12-21-2003, 10:10 PM
You are not playing around with this bullet proof set-up are you? :)

arvidb
12-22-2003, 06:00 AM
Bullet proof? What's bullet proof? :)

No, seriously, I paid less for these used ballscrews than I would have had to pay for new ACME screws with AB nuts. And they came complete with machined ends and bearings, too! One of those rare win-win situations :banana: .

// Arvid

Mr.Chips
12-22-2003, 09:45 AM
How would the resolution on these screws be, since the TPI is fairly low?

Bloy2004
12-22-2003, 10:58 AM
If you have a high ratio from motor to shaft would that keep the motor rpms up and increase resolution?

balsaman
12-22-2003, 01:04 PM
Yep.

E

arvidb
12-22-2003, 02:19 PM
Originally posted by Mr.Chips
How would the resolution on these screws be, since the TPI is fairly low?

Well, the Z axis one is 10 mm/turn and the Y axis is 15 mm/turn. The manual for the motors I hope I'll be able to use specifies the encoders as "2048 P/R", which should mean that there's either 4096 or 8192 counts/rev. And I'm going to use a timing belt with a ratio of between 2:1 and 3:1. So worst case resolution for the Y axis is below 2 microns (15 mm/(2*4096)), which should be enough, I think :). Of course, with servos, I have to take into account that they might be off a certain number of counts, but hopefully they'll stay within +/- 10 counts.

Actually I had much rather used 5 mm/turn screws, not so much for resolution, but because of the higher cutting force. The servos would still have plenty of speed. But I guess I'll take what I can find...

// Arvid

balsaman
12-22-2003, 07:34 PM
Your speed will be limited to what your pc can pump out. I think 35 khz is max for mach 2. Turbocnc maxes out around 25khz. 250-500 cpr encoders would be better. I think the G340 drive has a pulse devider for high count encoders.

Eric

turmite
12-23-2003, 01:27 PM
Eric Mach2 can output 45k hrtz if the computer is up to it. I have finally got mine running and I cannot believe how much smoother the machine is than what it used to be with the full /half step system.

turmite

arvidb
12-25-2003, 07:42 PM
Hi everybody! I hope you all are having a nice christmas!

I need some advice about timing pulleys. My motors have 14 mm axles, the screws have 10 mm axles. I want about a 2.5:1 reduction. Also, I need a type with low backlash.

Do you guys drill/turn your own bores in the pulleys? If so, how to do that without a lathe? :) And will a setscrew be enough to keep the pulleys from skidding on the axles? The axles are all smooth, no flats.

I haven't been able to find anything that would fit my axle dimensions and reduction ratio :( .

// Arvid

arvidb
12-26-2003, 11:28 AM
Originally posted by balsaman
Your speed will be limited to what your pc can pump out. I think 35 khz is max for mach 2. Turbocnc maxes out around 25khz. 250-500 cpr encoders would be better. I think the G340 drive has a pulse devider for high count encoders.

These motors are 3-ph AC, and thus won't run with Gecko drivers. The drivers I can get that will run these motors lack a position servo loop (i.e. they have only the torque and speed servo loops). So I will have to build my own position PID controller, and could implement a count multiplication/division in this as needed.

This way I could set things up so that 1 step from the PC gives a movement of, say, 0.005 mm (=> 0.1 m/s @20 kHz), but still use the full encoder resolution for positioning accuracy.

BTW, a 500 counts/rev encoder with a 10 mm/turn screw, and 2.5:1 motor-to-screw reduction gives a speed of 0.16 m/s or 375 IPM @20 kHz! Or 755 IPM for a 250 counts/rev encoder! A bit fast perhaps? :)

// Arvid

HuFlungDung
12-26-2003, 01:39 PM
Originally posted by arvidb
Hi everybody! I hope you all are having a nice christmas!

I need some advice about timing pulleys. My motors have 14 mm axles, the screws have 10 mm axles. I want about a 2.5:1 reduction. Also, I need a type with low backlash.

Do you guys drill/turn your own bores in the pulleys? If so, how to do that without a lathe? :) And will a setscrew be enough to keep the pulleys from skidding on the axles? The axles are all smooth, no flats.

I haven't been able to find anything that would fit my axle dimensions and reduction ratio :( .

// Arvid

Hi Arvid,

Have you checked with a supplier like Browning or Gates? They manufacture two piece timing pulleys with taperlock bushings which I think you can buy to match standard shaft sizes, both inch and metric. There is no backlash in a timing belt drive, unless your slides were extremely stiff for some reason. I have used timing belts on many servo drives and never had any backlash attributable to the belt and pulleys.

arvidb
12-29-2003, 07:18 PM
Thanks for the tip! Those taperlock bushings look really nice, just what I was looking for! :banana:

I haven't found anything that fits my 10 mm screw journals yet, but now at least I know what to look for.

// Arvid

arvidb
01-19-2004, 07:35 AM
Hi everybody,

I thought I would report on the situation:

I'm still looking around for a timing transmission. I'm having a thorough email exchange with a local manufacturer of timing belts and pulleys, and they gave me some valuable suggestions.
First, they recommend the AT tooth profile with polyurethane belts to get minimum backlash. If I understand correctly, HTD is worst in this respect, followed by T, GT, and AT, in that order.
Second, to get good grip on smooth journals, one can use either taperlock bushings or another kind of bushing that sits between the bore of the pulley and the journal. For an example, see "Heavy Duty Keyless Bushings" on the McMaster-Carr website. Taperlock would be nice, but I don't think there exists taperlock pulleys for my small journal dimensions.

The only problem so far is that a complete, new timing belt transmission of this kind is 2-3 times what I can afford. Maybe I can find something at ebay.

Also, I have finally found some time to do CAD drawings of my ball screws and some other components. The timing pulleys are very simply drawn, since I don't know exactly what kind they will be. They are the same size as AT5 pulleys with 22 and 60 teeth, resp. Here is a more detailed look of the "bridge bar" (without the Z slide and guides because I don't have them yet!):

It's a bit incomplete; for example the only ball bearing that's got a complete suspension is the far Y axis one.

ger21
01-19-2004, 10:21 AM
From what I understand, GT2 belts have the least amount of backlash.

Gerry

Hobbiest
01-19-2004, 03:27 PM
Hey Arvid and Mr. ED,
www.Igus.com, the company is based in Stuutgard Germany (did I spell that right?) Might have distribution from that home location also!
Arvid, What CAD program do you use, and how user friendly would it be to a moron? I have tried a couple, but they are very difficult to get the hang of, and make really cheesy looking drawings.
Igus' stuff is REALLY nice!!

Mike F
01-19-2004, 04:15 PM
Hobbiest

Take a look at ProdeskTop Express from PTC.com for a CAD package, it's free!!!! It is a cut down version of ProEngineer and I wouldn't mind betting it will do all you want it to.

This software has been made available to all UK schools, complete with support, and is an excellent package for nothing, well worth a look.

Mike

Mike F
01-19-2004, 04:37 PM
Sorry! Meant to attach a drawing done with ProdeskTop Express.

arvidb
01-19-2004, 05:46 PM
Originally posted by Mike F
Hobbiest

Take a look at ProdeskTop Express from PTC.com for a CAD package, it's free!!!! It is a cut down version of ProEngineer and I wouldn't mind betting it will do all you want it to.

Mike

Unfortunately, as of the new year, PTC has discontinued the free version of Pro/DESKTOP Express. You can no longer get the activation code needed to run the program. And no use asking anyone else for it either, since the code is unique for each installation :-(

I'm using Pro/DESKTOP Express, but if (when!) I ever need to re-install my computer, well then that's it!

// Arvid

arvidb
01-19-2004, 06:04 PM
Originally posted by ger21
From what I understand, GT2 belts have the least amount of backlash.

Gerry

Actually, I haven't found any real information on this! Do you have any comparison data?

This is from the "Gates Poly Chain GT2 Drive Design Manual":
"PowerGrip®HTD® Drives have improved torque carrying capability and resist ratcheting, but have a significant amount of backlash. Polychain GT2 Drives have considerably improved torque carrying capability, and backlash characteristics in between that of PowerGrip HTD and PowerGrip Timing Drives."

This seems to be a comparison between the Polychain GT2 and Powergrip GT/GT2 products. It's a bit unclear, but indicates that the "registration" (or backlash) is similar between the GT and the GT2 (Powergrip).

From Stock Drive Products (SDP) "THE WORLD OF TIMING BELTS":
"Trapezoidal Timing Belt Drives are known for having relatively little backlash. PowerGrip HTD Drives have improved torque carrying capability and resist ratcheting, but have a significant amount of backlash. PowerGrip GT Drives have even further improved torque carrying capability, and have as little or less backlash than trapezoidal timing belt drives."

The T belt profile is a trapezoidal timing belt. The AT is an improved version designed especially for low backlash (the AT is curvilinear, but close to trapezoidal, I think). So I draw the conclusion that AT is the best there is in this regard. The GT2 would be better in all other areas, though, I guess :)

If anyone's got better info regarding this, I'd very much like to share it!

// Arvid

Hobbiest
01-19-2004, 06:16 PM
Mike F,
Thanks for the info...I'll still check into it. What is that drawing? Some kind of one way check valve?

HuFlungDung
01-19-2004, 06:34 PM
Here is another belt to check on: Eagle PD synchronous belts by Goodyear. This is a "herringbone" style of belt which is supposed to be much quieter, perhaps 15 dba. The pulleys do not need flanges because of this belt tooth pattern. For the "white" belt series (second to narrowest belt), the backlash is .8 degrees on 22 tooth sprocket.

Mike F
01-19-2004, 06:46 PM
Spot on Hobbiest. It's a non-return valve I made for a vacuum bagging, prepreg oven here in my workshop. I build competition model aircraft (gliders), machine the patterns with CNC then make moulds and production items using prepreg composites in the oven. Keeps me off the streets.

I intend to make a much larger, approx 2000mm CNC for making larger wings. Looking on ebay at this moment for bits.

Mike

ger21
01-19-2004, 11:15 PM
IN my SDP/SI Handbook of Timing Belts and Pulleys, "On drives using a low installation tension, small pulleys, and light loads, the backlash of the Powergrip GT2 system will be slightly better than the trapezoidal timing belt system. Hpwever, with increased tension and/ or loads and/or pulley sizes, the performance of the Powergrip GT2 system becomes significantly better than the trapezoidal timing belt system. "

This is the paragraph right before the one you quoted. The only problem with this info is that they have NO info on any other types of belt. Your guess is as good as mine.

Gerry

trent2530
02-02-2004, 10:34 AM
Say Balsaman, what happened to your roller skate bearings?

Trent

PS: I just bought some drawer slides from Home Depot mayself. Experimenting.

Hobbiest
02-02-2004, 02:11 PM
Arvid...use a file to cut flats on you shafts. Don't use a grinder...to much heat.

arvidb
02-02-2004, 03:39 PM
Hobbiest,

thanks for the tip. I'm now looking into using tapered bushings. These clamp down very well on plain shafts, but may be a bit too expensive. If I cannot get these for a reasonable price, I may go with your idea. I guess you mean to use a set screw that "sets" on the flat?

Arvid

Hobbiest
02-15-2004, 02:54 PM
Call Igus and talk to a sales rep. Have them send you a sample of some of their products. They sent me a sample of their bushings for round rod, and I tried it out on zinc plated rod from the hardware store. Also tried it on polished Thomson rod. The zinc rod worked a lot better! They say that their bushings work better with rougher shaft, and this means that you don't have to spend an arm on shaft, and a leg on bearings! I just got my set of bushings, 8 3/4" open, and 4 1/2" closed, and with shipping it was only like $30! Good deal so far.

arvidb
02-15-2004, 04:03 PM
Hobbiest,

I think we might be talking about different things here. I'm talking about bushings for fastening timing pulleys to screw journals, while I think you are talking about linear guides?

I actually did contact Igus a few months ago, asked them a few questions and got their catalogue. I think their stuff looks very good for things like balsa cutting - very tolerant of dust and misalignment, cheap, good precision under light loads, and not least, looks good. It would be my #1 choice for a light machine.

Unfortunately they explicitly states that you should not use DryLin:
* when you need a high precision, preloaded linear system
* for heavy load applications

so I went for used THK style guides instead. With a lot of patience you can find those used for the same amount of money you would pay for new DryLin guides.

Arvid

balsaman
02-15-2004, 04:16 PM
Originally posted by trent2530
Say Balsaman, what happened to your roller skate bearings?

What rollerskate bearings? I have never used any on a cnc.

E

Hobbiest
02-15-2004, 04:19 PM
Definately talking about different things! I am going to put a steel frame machine together...eventually, that will use these bushings, then see what they can do! As far as your bushings, why not tap a commercial bushing for the thread?

arvidb
02-15-2004, 05:56 PM
It would be interesting to hear your judgement on the Igus slides when you are done!

I'm not sure what you mean by the "tapping a bushing" comment? The thing is, I want to rigidly attach a timing pulley to my leadscrew journal, for driving the screw. But since the journal is so small (only 10 mm dia) and also smooth, it's difficult to find a pulley that would grip the axle well enough.

There are pulleys with something called taperlock bushings that would work well, except these have a too big dia for my journal. So now I've found a "loose" taperlock-style bushing that you can place between the hub of the pulley and the axle, that expands and therefore holds the pulley in place on the journal. Only they're a bit expensive :(

Arvid

Hobbiest
02-15-2004, 06:39 PM
I thought you meant that you wanted to attatch a bushing to the thread of the rod to enlarge it for a given pulley ID. Sounds like what you need is to have an extended hub on the pulley, with a set screw to tighten against a flat on the lead screw journal. If the ID of the pulley is too big for that, talk to a machine shop about turning up a bushing that will prss fit on the lead screw journal. Shouldn't cost too much. What time is it in Sweden right now?
Jens

arvidb
02-16-2004, 04:16 AM
It's also a question of torque handling capability; peak torque will be somewhere around 5.75 Nm or about 810 oz-in. But I'll keep it in mind as an option!

When I wrote my previous message, the time was about 11:30 PM here. It's now about 9 in the morning. (GMT+1)

Arvid

arvidb
02-21-2004, 11:52 AM
I found a pair of 280 mm THK HSR25 rails with slides on ebay. I *am* a lucky one! They are perfect for Z guides, especially considering I already bought THK HSR25 for my Y. They are even preloaded!

(I hope one of the administrators tells me to stop if I post too many pictures...)

balsaman
02-21-2004, 11:57 AM
Nice!

No need to stop posting pictures. Keep going.

E

arvidb
02-21-2004, 02:20 PM
Thank you balsaman! I will :).

... and I recently got a Servo Pack (Yaskawa term for servo amplifier) for my motor. I'm having some trouble getting it to work at the moment... the Servo Amplifier thinks the motor encoder is 1024 pulses per rev, while it is, in fact, 2048... and as of now, I have no way to set this. At least I *think* this is the problem...

The Servo Pack and Servo Motor were bought used, and all the connector cables had been cut right off when they were dismounted from their previous machine. So there were some tricky soldering of micro-DB20 connectors and splicing of cables to get this far.

In the picture below, you can see one of the connectors (that luckily were delivered with the Servo Pack) lying on the table, and the Servo Pack and Motor connected. The motor has two cables coming out of it, one for the encoder and one for power. The grey one is power (U, V, W phases + Frame Ground), and connects to the cable shoe terminal on the left of the Servo Pack. The black one is the connector cable, connected to the Servo Pack with the micro-DB connector.

The white cable connected to the Servo Pack is for connecting it to a computer for programming. But I don't have the correct program for this (yet) :(.

In the background, wrapped in plastic, are the different linear guides.

Arvid

arvidb
02-21-2004, 02:21 PM
Closeup of the motor... crudely clamped for testing :)

arvidb
02-21-2004, 02:22 PM
... and the Servo Pack

arvidb
02-21-2004, 02:29 PM
And a warning label on the outside of the Servo Pack cable shoe terminal lid.

There's no transformer in the Servo Pack, the mains is simply rectified and smoothed with two big capacitors.

This is lethal in two ways: first, there's about 20 uF of >300 VDC waiting to kill you, and secondly, there's no isolation from the mains. So I obey the warning labels, believe me!

Arvid

kevincnc
02-21-2004, 03:03 PM
Here's another way to attach pulleys on a shaft with at lease one flat. I's better than setscrews or a keyway bacause it won't slip. With no flats I'd go for a taper-lock if you can find one small enough.

kevincnc
02-21-2004, 05:14 PM
Sorry I obviously need an automatic spellchecker.

balsaman
02-21-2004, 05:29 PM
Avrid, do those drives use step and dir signals? You should be able to fool the drives by setting the resolution incorrectly in the software. Two wrongs making a right in this case....?

E

ballendo
02-21-2004, 06:09 PM
Kevin,

That's a good idea!

How is the vibration? How fast--rpm's-- have you run pulleys modified this way?

Thank you in advance,

Ballendo


Originally posted by kevincnc
Here's another way to attach pulleys on a shaft with at lease one flat. I's better than setscrews or a keyway bacause it won't slip. With no flats I'd go for a taper-lock if you can find one small enough.

arvidb
02-21-2004, 06:33 PM
Originally posted by kevincnc
Here's another way to attach pulleys on a shaft with at lease one flat. I's better than setscrews or a keyway bacause it won't slip. With no flats I'd go for a taper-lock if you can find one small enough.

Yeah, I'm looking for taperlock, either pulleys with a taperlock bushing, or a separate taperlock to insert between journal and pulley bore.

Thanks for the nice pic! It's a great idea, if I can't find taperlocks for the right price, I could very well file flats on my journals and go with your idea.

I had kinda postponed the timing reduction problems for now, focusing a bit on the motors. But this is a good solution!

Arvid

arvidb
02-21-2004, 06:39 PM
Originally posted by balsaman
Avrid, do those drives use step and dir signals? You should be able to fool the drives by setting the resolution incorrectly in the software. Two wrongs making a right in this case....?

E

They use an analogue +/-10 V signal for speed input. There's no position servo loop built-in, I'll have to build electronics for that part myself.

And unfortunately, since they are 3-ph AC servos, the amplifier needs to know the position of the rotor to run the motors. No way to get around this, I'm afraid. Well unless I build a hardware divide-by-two circuit between the motor and amplifier, something I probably could but would rather not do.

As it is now, the commutation loses sync with the rotor magnets, so the motor just jumps and hums when I jog it.

Arvid

kevincnc
02-21-2004, 06:43 PM
I thought it was a good idea too, but it wasn't mine. The pulleys came on some brushless motors I bought a while back. I wouldn't think vibration would be a problem unless the motor is on a very light assembly. I haven't tried it yet but will on my router I'm building.

Another pretty good method (better than setscrews on flats in my opinion,) is to drill a "drill point" for a cup-piont setscrew to tighten into. That keeps the shaft from scarring and works even better if you use a short one and tighten another on top of it to keep the first one from backing out. Sorry if someone already suggested that.

HuFlungDung
02-28-2004, 11:58 PM
Originally posted by HuFlungDung
Here is another belt to check on: Eagle PD synchronous belts by Goodyear. This is a "herringbone" style of belt which is supposed to be much quieter, perhaps 15 dba. The pulleys do not need flanges because of this belt tooth pattern. For the "white" belt series (second to narrowest belt), the backlash is .8 degrees on 22 tooth sprocket.


I resurrected this post just to update: when I retrofitted my knee mills, of course I threw out the variable speed pulleys and went to a VFD drive.

On one machine, to connect the motor to the spindle, I used a PolyV belt drive, which is a really fine ribbed belt and pulley. This is nice and quiet, but maybe not really dependable for high torque transmission at low rpm (for tapping).

The other mill, at first I used a Gates PolyChain. This was all right (very positive drive) up to about 2400 rpm, but when it got up to 3600 rpm, it was damn painfully noisy to be near. And, I have to be near it :( I was running 6" dia pulleys.

Then, I went and got the EaglePD belt and pulley system. This is a herringbone belt and pulley. I dropped the pulley diameter to 4.5", because I noticed the momentum of the larger pulleys on the Polychain drive created kind of a start up "lurch".

Anyways, the Eagle PD herringbone belt is really nice. They say 15 db quieter than regular gear belts, but it is likely 40db quieter than that PolyChain at full rpm.

Stevie
02-29-2004, 03:45 PM
Originally posted by Mr.Chips
How would the resolution on these screws be, since the TPI is fairly low?

did you look closely at them; they must be .500" pitch

arvidb
02-29-2004, 06:21 PM
I want to thank everybody for all the ideas on timing reductions. I still haven't decided, but I will keep the different solutions in mind when it comes to chosing. All ideas are appreciated!

Arvid

arvidb
03-07-2004, 09:13 AM
I'm beginning to think I've gotten far enough on the planning to actually start building!

However, I think it is really ugly! Or perhaps it's only because my CAD program doesn't do perspective transforms properly? Or something? *LOL*

So please comment the overall design, it'll soon be too late to change.

Arvid

arvidb
03-07-2004, 09:15 AM
From the front with the Z axis lowered:

arvidb
03-07-2004, 09:20 AM
Close-up on the Z axis with the front plate removed. Not much space left here! Maybe I should have chosen smaller components (THK/ballscrews) to make things less clumsy?

Arvid

Hobbiest
03-07-2004, 03:48 PM
I think it looks really good Arvid. Don't hink the horizontal parts of the table frame need to be so tall, but then that may be from your software like you said. Start building it!

arvidb
03-07-2004, 05:43 PM
Hobbiest,

Thanks for the encouragement!

But what do you think about this alternate design? (Idea from Pistachio's machine, http://www.cnczone.com/showthread.php?threadid=3211).

Arvid

Hobbiest
03-07-2004, 06:27 PM
One thing that I liked about your origional design, was that the carriage actually moves past the edge of the uprights. This is handy, in that it keeps the table size closer to the width of most of the machine. That is one thing that I think makes Cranky's designs look even smaller. I also like the stiffeners on the gantry uprights. I like the table height better on the second design. Overall it does look cleaner. I would rather build something that looks like a Sherman though, instead of a Miata, if that means that the tank will be better engineered. The extra mass, etc. all makes the product nicer.

arvidb
03-07-2004, 06:51 PM
Hobbiest,

thanks for the really good input! I'm happy you identified my thoughts with the first design (less floor space and rigidity).

But since I first came up with the original design, after reading lots of the threads here, I have realised two drawbacks:

1) Dust. It would end up everywhere. In the second design it has a natural "sand box" to fill first.

2) Enclosure. It would be a lot easier to build a sound dampening enclosure for the second design.

It wouldn't be very difficult at all to add stiffeners to the gantry uprights on the second design, either. I could even cut them on the machine! :)

I'll have to think some more about this, but right now I'm leaning towards the second design. Strange, just a few hours ago I was convinced I would build according to my first idea. See what happens when you spend too much time here at CNCzone :D

Arvid

Hobbiest
03-08-2004, 12:24 AM
Is a "natural sand box" really something desireable? Seems to me that any dust would best be kept away from the machine.

arvidb
03-30-2004, 05:20 AM
*aahrg* I just paid $395 for 3 servo drivers for my nice 3-ph servos. They are very professional and all, and doesn't even require a power supply (they are just hooked up to the mains)...

BUT I'm not completely sure they'll work with my motors! And I don't have a clue where to get the correct motors for them if the ones I've got doesn't work :eek:

So I'm really nervous here. Hopefully they'll be here in a week or so... :drowning:


Hobbiest,

I've thought about the "sand box" (chip box?) issue, and I think I would prefer to have the chips stay in the box below the machine, than going everywhere. I have very limited space and it won't be a dedicated machining room. I will have to make sure to clean it between uses though...


Arvid

mhackney
03-30-2004, 09:09 AM
arvidb and hobbiest - the sandbox idea is fine if you hook up a vacuum system to remove the collected dust/chips. That is one of the first steps in outfitting a woodworking tool like a table saw for dust collection - enclose the cutting area and provide a "trap" (aka sandbox) for the dust chips to accumulate and be sucked out.

Graham S
03-30-2004, 08:49 PM
What made you go for 3 phase servos? Given the cost of geckdrives and DC servos and the fact that you don't have to buy expensive motion control cards or develope your own control system it seems a difficult way to do things. I hope you can get it working, it will be awesome when it does!

Graham

balsaman
03-31-2004, 12:35 AM
If you think your design is ugly one thing that can be done is to taper the 4 upright supports towards the top (half as wide at the top and just a little wider at the bottom?), and maybe even eliminate the ones at the back... I think that will change the look a bit and it will still be rigid. Try it and have a look.

Eric

arvidb
03-31-2004, 07:33 AM
Originally posted by Graham S
What made you go for 3 phase servos? Given the cost of geckdrives and DC servos and the fact that you don't have to buy expensive motion control cards or develope your own control system it seems a difficult way to do things. I hope you can get it working, it will be awesome when it does!

Graham

Heh, maybe I just like to make things difficult for myself?

Actually, I think when I started looking around for servos (that was before I found this excellent board), most of the info I found was about industry servo systems (I searched google). I didn't know about the geckos, xylotex or rutex drivers.

And when I found a guy that was actually selling these nice used motors and drivers at a very reasonable price ($40 for the motors and $80 for the drives) there was really no question about it.

Later, when it turned out the drivers he sold were actually not dismantled from the same machines as the motors, and did not work with the motors ($80 + $30 shipping down the drain - or maybe not completely, he might compensate me somehow)... well, it was only after that I began to wonder if I had made the right decision. 'Cause by then I had found this place and learnt about the availability of DC servos.

But I don't know... even if it turns out these new drivers won't work, I'm still reluctant to switch to DC servos. I think I'm a bit of a perfectionist when it comes to ... I don't know; sleekness and simplicity, beauty of physical systems. The 3-ph servos are simply a so much better mechanical device than the DC servo, and it would bug me to change them out.

Have you compared the construction and specs of 3-ph and DC servos? 3-ph servos have the windings in the stator, and suddenly you don't have to rely on air flow (which is *very* inefficient compared to conduction through metal) though the motor to cool them. The motor can be sealed; no need for air holes that let dust and chips in. There's also no windings in the rotor, so lower moment of inertia and a smaller motor too. No brushes, which means no brush wear, and an electrically much more quiet motor.

It would be a very fun project to DIY a 3-ph servo driver. But it's a big project, and I already have too many unfinished projects running... two of them actually needing the router for their completion :rolleyes: . But perhaps I'd go that way, instead of chosing DC servos. I'm a very stubborn person :D .

Arvid

arvidb
03-31-2004, 07:58 AM
Originally posted by balsaman
If you think your design is ugly one thing that can be done is to taper the 4 upright supports towards the top (half as wide at the top and just a little wider at the bottom?), and maybe even eliminate the ones at the back... I think that will change the look a bit and it will still be rigid. Try it and have a look.

Eric

Ok, I tried it, and it did make the machine look better, especially removing the supports at the back (tapering while retaining the back supports made it look *really* weird :)). There's not much space available for tapering though, and I will have to figure out a new way to mount the Y axis motor if I don't have the back supports.

I'm still leaning towards the "new" design since it allows me to easily enclose the machine and collect (at least some of) the dust.

Arvid

Hobbiest
04-13-2004, 01:28 AM
Will you have to build a phase converter for you industrial strength servos?

arvidb
04-13-2004, 05:13 AM
Hobbiest,

I'm not sure what a phase converter is?

The motors are 3-ph, and to drive them the 1-ph 50 Hz mains will have to be "converted" into 3 variable frequency and current controlled phases. This is done with a servo driver (like a Gecko, but for 3-ph motors). Is this what you mean by a phase converter?

I have bought three servo drivers (or Servopacks as the manufacturer Yaskawa calls them). They are on the way, but delivery seems to have been delayed now over the easter holiday :-/ I hope to get them within a few days.

Arvid

arvidb
04-14-2004, 11:14 AM
*long sigh of relief...*

the Servopacks seems to work just fine with my motor, even though they are different type (200W motor and 400W drive). I'll just have to limit the current accordingly... I hope. Still not sure they won't fault out once I put a load on the motor, but why be pessimistic? :)

If this actually turns out OK, I think it's really great:

$330 + $65 S/H for 3 servo drivers, including built-in power supply
$120 + $100 S/H for 3 servo motors with 2048 pulses/rev differential encoders

= $615 for a complete 3-axis 200W servo system with analogue speed input (+/-10V) and scalable (1-4096 counts/rev settable in driver) quadrature output.

I will need hardware for the position loop (to convert parallell port step/dir to +/-10V speed signals), but I think I can build that myself. It would be lots of fun!

*whew* - it's been a nightmare laying out all that money without knowing if it would work though. But worth it now :)

Arvid

Hobbiest
04-14-2004, 11:01 PM
I was being funny about the phase converter. A phase converter is used to run 3 phase equipment (lathe, mill, etc.) with a single phase circuit.

arvidb
04-15-2004, 05:51 AM
Okay, maybe it's time for an explanation... I'm aware that I don't *need* all these fancy things... like 3 ph servos, THK linear guides etc. I will never be able to build a machine that comes close to their tolerances anyway.

I put down all this effort simply because I think it's fun. I like these nice components, and they will make me happier just by the way they look and the fact that they could have been sitting in a $50,000 machine. Even if they don't make my machine better than it had been with DC servos and skate bearings on supported shaft.

And all that math is also because I'm fascinated by it; this is a perfect opportunity to actually test how well all that one has learnt in school actually fit in the real world - and I believe it's a useful tool that answers a lot of questions.

I'm not trying to show off here (well, not more than anybody else ;)). We all think CNC machines are cool, right?


Arvid

arvidb
04-15-2004, 05:54 AM
... and to more directly answer your phase converter question: they may be industrial quality, but they are not very large. The mains 230V @ 10A should be plenty for 3 axes (axises? :))

Arvid

arvidb
04-16-2004, 04:24 AM
More Good News,

the seller of the first Servopack (the one that didn't work) has been extremely nice to deal with, and will compensate me in full. If you read this - thanks!

Arvid

NEATman
04-16-2004, 10:48 AM
arvidb- Welcome! I'm a newbie to this board too. You are going to build a converter for step and direction to ±10vdc? If you can accomplish that, I'll trade you some linear bearings and a ballscrew for your lower axis. When the lower axis rail is fully supported, you are less likely to have vibration vertically. I am retrofitting a bridgeport CNC machine, and the controller is dead, but the servo's and drives are 100%. Plus the machine is only 10 years old. The amplifiers that I have are just like yours. The only low cost controllers out there are designed for stepper (step and direction), not the other industry standard of servo (±10vdc input). I work at a company that tosses out lots of linear ways, ballscrews, even motors and drives. Let me know if you are interested.

By the way, the machine design looks pretty good. One thing that I would suggest is the way you are mounting the duplex angular contact hearings on the end of the ballscrews. Instead of clamping a thin sheet metal looking piece between the bearings, usually they are put into a metal block with a flat bottomed bore, and clamped that way. Usually the bearings are designed to have the inner races touch when you tighten the nut on the ballscrew. The outer races are then squeezed axialy by a clamp plate with a hole in it. If I can figure out how to attach a picture, I'll show you what I mean. For reference, take a look at our catalog. It has lots of tips about servo vs. stepper, and it might help to take a look at some of our standard products. www.neat.com
[url=http://www.cnczone.com/gallery/showphoto.php/photo/329][img]

Take a look if this works!

Neatman

arvidb
04-16-2004, 04:33 PM
Neatman,

I'm very interested!

I have some questions though:

1) How long are you ready to wait for me to finish this thing? (As you can see from my title I'm a Time Optimist :D)
2) What's the resolution of your encoders, and how fast do you need to spin your motors?
3) Since you work with these things I guess you're aware that you'll not get the same performance and features as with an industrial fully closed loop servo system? (The position loop will only be closed to the "converter" card.) That is, no following error graphs, no graphing utilities to help set PI parameters, no feed-forward etc.

The graphing utilities could be implemented I guess, but I will try to keep it as simple as possible to start with.

What I have in mind is an FPGA that counts pulses from the parallell port and encoders and compares them through a PI controller. The result would be fed to DACs for +/-10V output. I would probably also include a binary step multiplier so that higher resolution encoders can be used while still getting ok feed rate.

What do you (all) think about this? What other features would be required to make it work? For instance, how is homing normally done? It would be neat to just move until a limit switch is reached, and then back off until the encoder index pulse is hit - an easy way to get a repeatable home position. But how would the communication between g-code interpreter and hardware look? Guess I have some reading to do :)

Arvid

arvidb
04-16-2004, 04:36 PM
... and thanks for the tips on how to mount the bearings! Unfortunately I don't have access to either a lathe or a mill, so I have tried to design it so that it can be made with normal hand tools. Do you think it could work as I drew it?

Arvid

arvidb
05-09-2004, 11:32 AM
Hi again,

I've been doing some work on my controller card. But I wonder a bit, what inputs and outputs are useful?

The card will be set up and controlled via RS232. It will be run with step/dir signals from the parallell port.

In normal "run" mode it will listen for step/dir signals and encoder signals and make sure they agree by telling the motors how fast to spin. It will also monitor for positive and negative overtravel and disable the servos in case those signals are hit. It will send out error count over RS232 so that one can see how well the actual position coincides with commanded position, and it will map overtravel and home signals to the parallell port somehow - please help me out here to what's practical, I mean, there's four (five?) inputs on the port and three signals per axis that want out... :confused:

There will also be a "home" mode where one can home and reset the counters for the individual axes via commands over RS232. This mode will use limit switches and the encoder index pulse to find a reliable home for each axis. Will this be compatible with the different controller programs (i.e. is it possible to just zero their counter after an "external" homing has been performed?) :confused:

I will also put in a "set-up" mode that will hopefully help set the parameters for the PID controller - here the motors will be run open loop so that one can watch their step response.

So all in all, the inputs and outputs to the card will be:
* A, B, Z encoder inputs for each axis
* pos, neg overtravel inputs for each axis
* a home switch input for each axis (if one would like to home with the parallell port controller program without the aid of index pulses - or for Z maybe this could be used for a 0 plate?)
* +/-10V motor speed outputs for each axis
* One Servo Enable output that will be active as long as there is no fault condition (will be tripped by encoder errors, excessive position errors, overtravel - anything else? :confused: )

Sorry for another one of my long posts... and please comment! What would you be missing if this were your controller card?

Arvid