[Oce]

Hi everyone, I'm in the process of building my first machine...It has all taken a good while so far, and I haven't started building anything yet...Like to do a good bit of research before I do.
Sounds like a hardcore machine for a first-timer, I know, but I like to do things big when I first take the time to do it...

So far, I've ordered 25mm bearing blocks and support rails, breakout board from PMDX, 4 1200 oz-in steppers from Keling Inc, and are about to order drivers, probably Gecko or some with same specs.

My target machine has a working area of about 2.5x2.2x0.6 m (XYZ)
(Open for discussion and reduction on this point)
I bought 4 steppers to ensure that they can move the gentry properly, so I will have one on each side of it.

The router will be used to process mostly wood, and the occasional aluminium / steel pieces...

I will use mostly steel and aluminium for construction, as I'm quite familiar with working metals, and have a fairly good place to do the necessary work.

But then comes the tricky part; what should I use to drive it?
I've looked at a lot of different solutions, and the Nexen RPS system really got my attention. I've made contact with a local distributor, and they will give me a quote as soon as they get the necessary data.
I know, ballscrews are an option also, but I've read that the RPS can, in many cases, match or have a lower price than that of ballscrews, while at the same time be faster and more precise...
(Sounds like I'm taking the factorys word for it, right?) (nuts)

And this is were you guys come into the picture; any idea what a "reasonable" speed, acceleration etc would be for a machine this size?

The information requested;
Total system weight kg
Max speed m/sec
Acceleration time sec
Max acceleration m/sec2
Guide System Friction µ
Degrees from Horizontal
Desired run length m
Cycles per day

I assume the weight will exceed 200+ lbs, no doubt about that, so let's just say 400+ to be on the safe side.
Guide system friction; I didn't find anything at the dealers site (marchant dice in the UK)
Max speed, acceleration time, etc, I'm open for suggestions?

Just give me YOUR honest opinion, so I have something to work with.
Any videos from youtube etc demonstration estimated speeds would be great!

[DocTanner]

Oce,
You might want to look at www.mechmate.com <--- Fantastic Machine Free Plans and Gerald is willing to "coach" you through it. I'm 99% complete with mine

EZ-Router.com and Shopbottoolscom for specs on machines similar to what you want

DocTanner

[Oce]

Yes, I have seen the plans, but I like the idea of building something on my own... maybe I will post my own plans here in the future as well
Thanks for the tip, I'll look into it!

[Zumba]

First things first. You need to draw your machine using a CAD program and really figure out exactly what you're going to be making, right down to the last bolt. You should really consider building a small machine since this is your first. Large machines are exponentially heavier and exponentially more expensive. Anyone who says otherwise hasn't done the math.

So what are the exact components you've purchased so far? What are these "25mm bearing blocks" you speak of?

I noticed that you want 0.6m of Z-axis travel. That much travel will require a Z-axis that's around 1.5m tall. For your first machine, try something more along lines of 0.15m of Z travel.

Nexen RPS... don't think you'll be saving any money with these. They claim to be cheaper than ballscrews. But they're comparing them to precision ground ballscrews, not the rolled screws that most of us use.

The questions you ask regarding friction, speed, etc. are impossible to answer. Kind of like asking, "how fast will my project car go? I've purchased a 400bhp motor so far."

[Oce]

Hehe, I realise that they are difficult to answer, I just want opinions, not definite answers (thats pretty impossible I understand.. )

I've done some small drawings so far, but nothing big.
I use AutoCAD at school, so I think I wil draw a simple 2D drawing.
Haven't started drawing in 3D yet, and I've heard it's pretty hard, vs that of Solidworks... Learing another program isn't the favorable option, although I might have to...

The price & weight doesn't bother me too much, I plan on using a lot of time and money into this, making it a great machine


Parts-list:
http://www.kelinginc.net/KL34H2120-42-8A.pdf 4x
http://www.pmdx.com/PMDX-122/index.html 1x
http://www.marchantdice.com/linear/SBR/index.htm 12x + a lot of rails in 100mm sections.

About the ballscrew VS RPS; I haven't looked too much into the advantages/disadvantages, but I like the idea of Nexen having 0,5m and 1,0m racks.
However, IF I had some estimates for the "recommended" speed etc, I could get a price quote, and maybe adjust accordingly...
(or just build-as-I-go.. )

Thanks,
Anders

[ger21]

The price & weight doesn't bother me too much, I plan on using a lot of time and money into this, making it a great machine

If you're going to spend the money, I'd shoot for 400-500 ipm cutting speeds and at least 1000 ipm rapids. You'll need a good spindle to handle those cutting speeds, though.

[Oce]

Thanks Gerry, sounds like a pretty high speed machine..
My vision is to start with something extremely good, fast, reliable etc...So the dissappointment of the final performance isn't too big...

Any suggestions on a price for ballscrews for a machine this size?

(It's not like price doesn't matter, it always does, but if I have to pay a little extra to get better overall performance, I probably will.)

[JerryFlyGuy]

Well I can offer some opinions and some of my experiances.

The steppers seem large enough for the job at hand, however you'll never run those steppers through the Nexen Rack's and keep the pinion's from blowing up [check the torque spec's]. I also doubt that they are more accurate than Ballscrews. I guess depending on the grade of ballscrew they might be but.. don't matter as they'll just bust on ya if you use them. Note also that this is assuming your using some type of reduction as direct drive would be too course of a resolution [distance per step] to be of use. For a table that size Ballscrews are the order of the day. Check out some websites like Nook or Misuimi [spelling] and see what they recommend for sizes to counter the whip issues that are going to crop up. You'll need to figure out what reduction you'll use as well as the needed power [at stepper rpm] to cut at the speed you want. Then know what lead to use to get that speed, then you can figure out screw RPM and figure out what size will not whip. I'd think a 1.5 to 2" ballscrew is going to be needed [if not larger].

I think that you could do it in 400 lb's but that would be lite. I'd shoot for more like 1000. If you really want to cut Alum/steel, if its just wood you'd be fine at 4-500.

I'm building a fairly large mill and I'm over a couple thousand already. I'm using rack and pinion for the X and Y axis, but then they are 20 and 10 ft respectivly.

As Zumba said, get it designed as its a complex pc of equipment and you'll having nothing but headaches if you don't know how its going to go together before you start. Either that, or it will look like it has warts all over it from just 'hanging' things wherever they 'look about right'.

Ger21 has pointed out that larger rapids and feeds will make it a real workhorse, also he mentioned that it will take a pretty good spindle, at these speeds I don't think a 3hp plunge router is up to the task, do you have a spindle picked out? I would first decide on what I want to spend, then find a spindle in the range and design towards that, that way you won't have a machine w/ a 2 hp spindle that could [ in theory] at 400ipm, but effectivly can only cut at 100 because of the spindle. 7-10hp is probably going to be required, that is unless you only want a DOC of 1/16" then the 2-3 hp would probably do it!

Keep at it! It's a terribly addicting past time, as well as expensive but I think it has its rewards for sure.. nothing like building something on a machine you've created and built. [ I've not felt that feeling yet, but in the next few weeks I hope to! ]

Jerry

[Zumba]

Thanks Gerry, sounds like a pretty high speed machine..
My vision is to start with something extremely good, fast, reliable etc...So the dissappointment of the final performance isn't too big...

Any suggestions on a price for ballscrews for a machine this size?

(It's not like price doesn't matter, it always does, but if I have to pay a little extra to get better overall performance, I probably will.)

You need to realize that "thinking big" is not necessarily a good thing here. If you really have the money and the time to do things right, take my advice and build a small 0.5m x 1.0 m prototype first.

Those linear rails... they're not bad. But they're not what you'd see on a machine that's "extremely good". Not even close. Those types of rails were never meant to carry heavy loads. They were meant to be very accurate, used to carry moderate loads in a clean environment, e.g. medical. I believe MRI machines use those in 50mm sizes. Then China came along and made junk versions of them that aren't very accurate. For a heavy duty CNC application, the only rails you should even consider are the square rails/blocks from THK/Hiwin. You'll noticed that the spec charts for all linear motion components have a "Dynamic Load" capacity column. If the dynamic load is 900lbs, that doesn't mean you're "safe" with anything under 900lbs. This number is used to directly calculate the life expectancy of your bearing blocks. The round rails you linked have a dynamic load capacity somewhere in the neighborhood of 10-25% that of THKs/Hiwins... far inferior. Another problem is that they need to be mounted upright to retain their load capacity. Once you turn them sideways or upside down, the load capacity goes from mediocre to pitiful. They're also more prone to dust contamition, as the seals that come with those bearing blocks are much thinner than those found on THKs/Hiwins. If you already bought em, sell em on ebay and recoup some of your money.

As for what size ballscrew... a prototype .5m x 1.0m would only require 20mm and even 15mm screws. But the machine you described would require 35-50mm screws on the long axes unless you put the screws under tension. Take a look at the prices for the big screws and compare.

The 1200oz-in motors have lots of torque, but the high inductance due to the length of the rotor/stator limits the top speed. Big, fast machines cannot be powered by steppers... servos only. There is a reason why machines such as the Shopbot use steppers.... they're very lightweight. But mass = rigidity. If you want a heavy, accurate machine that flies, you'll need the right drive system.

Don't let any of this discourage you. But I can tell you that jumping into a large machine and wanting the best will lead you to disappointment.

[Oce]

Thanks for the advice Jerry, I have also figured that there will be a need for some sort of reduction due to the torque...But I will take a look at the Nexen RPS none the less.
Might come up with some useful info for others as well...
(And that I know is a huge plus to other users!)

Ballscrew size; I imagine something along the lines you said, therefore I'm a bit worried about it getting awfully expensive.
But then again, IF and WHEN if turns out great, I'm sure it will be worth it...

Spindle; I have prepared a space in my garage / workshop, with 3-phase power socket for the router.
Still haven't thought about which one yet, suggestions are welcome

Zumba; wow, you really know how to put it to a layman
I think I will survive, better with the cold hard truth, than working my ass off and getting crap in the end.
I will consider the possibility to buy hiwin or THK, I just haven't found any decent shops here in Norway, that's one terrible part about all of this, customer service and availability in Norway SUCKS.
I found some used ones in Germany though, what do you think about that?
Is this type of rails generaly the best choice, or just THK / Hiwin?
Here are a selection of rails availible in Germany:
http://search.ebay.de/search/search....e=search&fgtp=


Thanks again, all of your opinions mean a lot to me and my work, I will never accomplish this without you!

[JerryFlyGuy]

Zuma said it VERY well. I didn't actually look at the rails you'd linked to, but he's right, they aren't what you want. You want a THK or Hiwin or INA or SKF or.. you name it, there's lots. I'd think not less than 20mm and probably 25mm rails and blocks to suit, if you want really good precision you should get a larger, pre-loaded block over the 'standard'. I personally stear clear of Ebay rails as I don't know their history, they could have been run on a grinding machine and I wouldn't know and won't be able to tell until they are set in place and checked for wear. Unless they are NIB on Ebay, I'd pass [again the precision thing].

Cost, I found it very suprising [as have others] but again as Zuma stated.. large machines cost more money [exponentially] than small. In reality I doubt you'd be able to make the machine your thinking for less than $20,000 USD, if you do you've skipped a step someplace in the 'precision' or the 'large' part of the equation.

You won't need 3 phz as you can get VFD's which take single and switch it for you. However, since you've got it so much the better. I'd keep an eye on ebay as higher HP Columbo spindles do come around from time to time. You should be able to pick up one for $2500-3000USD if your lucky, sometimes less. If you want to rebuild one, then less than even that.. The only reason I'm ok w/ Ebay for things like this is you CAN get them rebuilt, rails can't be. New spindles are lots of cash, I priced a 2hp HKS [ I think it was] spindle w/ pneumatic tool changer and it came out to $9kUSD for one the size your looking for, I'd expect $10K+USD [new].

You'll be able to get a VFD off ebay as well [Hitachi is a good name], keep in mind that you should buy your spindle first, as a VFD is NOT a one size fits all, you've got to spec it and the reactors to match the spindle.

There is a great bunch of guy's on here, who make the 'zone' what it is [ I'm not talking about me ] I know I wouldn't be where I am in my project w/out their help, thats a certainty..
HTH
Jerry

[Oce]

Uhm, you 're making me nervous about the whole project by saying this...But I've given it some thoughts together with a friend of mine...
Let's say we DO use the rails mentioned, and take the chance of them breaking relativly soon...What do you think about changing rails later on? Very much work, or can it work out?

20.000 sounds like a lot of money, where do you think most of it will go? Sounds like an awful lot right now, but I guess costs add up quickly enough...
(Looks like I won't be changing cars during this year after all...)

Thanks again!

[ger21]

I didn't look at your choice of rails, either. Go with what they said. To add to what Jerry said, I believe you can run up to a ~5HP spindle with single phase VFD, from what I've read.

[JerryFlyGuy]

There is no doubt that it's a pile of money. I don't think its worth investing it [ your situation may differ] unless your going to get some type of income back out of it. If its a toy then.. well either you can afford expensive toys :d or.. you need to find a cheaper toy, or change the parameters so it comes out cheaper [translates into lower expectations]

You could use the rails mentioned, and they won't 'break' per say, they just won't deliver as good of cut's as the higher performance unit's. As the get loaded, they will flex more, and eventually wear out before higher priced units would. If you want to be able to machine and hit tenth's then you need the higher performance rails. If your happy w/ +/-0.005 or so, then these would probably do the job. If your convinced that you need to cut Alum & steel, then they probably won't, either that or you'll have to take VERY slow, shallow cuts to get the job done. If your mainly interested in cutting wood, and occasionly [ once every couple months] want to cut Alum/steel then this may be ok. If your planning to daily cut Alum/steel and wood on occasion, then the higher performance rails are pretty much a must.

Same goes for the weight of the machine. If your going to be doing heavy cuts all day long in metal then it needs to be heavier, if its a once in a while thing, then thats less of a concern. My gantry is just slightly longer than yours at 10ft [~3m] and the gantry beam weldment is somewhere around 1200-1500 lb's alone and I'm still not sure that it will cut steel, I'm pretty sure it will work ok for Alum and some of the softer metals but time will tell. Steel wasn't really a priority. For me, cutting Alum was as was wood and foam.

One thing that effects ALL design's is compromise. If you want the best, your gonna have to pay for it, there is NO EASY WAY AROUND IT! In the machine world, you get what you pay for. There are guy's out there w/ little wood routers who regularly do Alum, but I don't think they are shooting for super high tolerances, they can't as their construction medium swells and shrinks several thousand's just from the change in humidity. You have to decide what your happy with [ sure it might not be the Ferrari you had hoped, but a nice 'M3' might fit the bill better and a price you can afford [or be willing to pay] ]

As Zumba has mentioned, it would be worth while to take the components you've purchased and build a little wood/MDF unit to get some experiance w/ it all. I wish I had, but then.. if I had, I might not have started the project I did. Building this stuff really takes dedication, the only reason I've stuck w/ it is because I've got so much cash into it that I HAVE to finish it.. [I wish I only had $20,000 invested ]
If you look up the downloads section there is a little wood router called a JGRO or something like that.. it look's like a good little unit, build that first, then use it to make the parts you need for your larger unit [like motor mounts etc] in the interm you'll get some great experiance from it!

So where does $20k all go? Well its suprising how much you spend on the little stuff which is hard to define. The big stuff takes money too, but thats easily definable.

Rails are going to cost ya..~30-50cents per mm [USD] so if you want 2.5 x 2.0 x 0.6 thats 10200mm [ two rails per axis] so thats $4590.00USD [guessing 45cents/mm]
Your spindle is say$3000, the VFD is $1000, plus the reactors which you could get on ebay for ~$200[you should really get two as its a fairly large spindle].
Then you needs trucks/carriages for the rails. If you go for the cheap ones, they are ~$75-100 each but the higher precision ones are ~$250-300 [and higher] so you'll need min of 12 for a 3 axis machine [ two on each rail] so thats $1200 to $3600 depending how much pre-load you want. So far thats what..$11390[ using an average carriage price]

Now you've got to buy drives and a BOB and a control program. A Mach 3 BOB from Campbell's is $245 for the combo board. Four G202's are $137 each [$548]

Mach3 is $159

Now we're up to $12342.

You still need to buy all your mat'l, say you end up with a 800lb unit. I'm not sure what steel is worth over there but a good rule of thumb number is ~$1/lb. with scrap and cutoff's your going to spend $1000.

You still need wire which can cost ~$2-3/ft for a 16ga x 4 cond shielded flex wire [you'll need ~100ft] possibly $300 w/ all the limit switch wiring as well.

You'll need connectors which can be as much as $10 each depending what you buy.
Paint is going to be ~$150 if done properly w/ primers etc.
Tools will eat up a HUGE amount of money, precision level's and sqr's and calipers and micrometers and dial indicators and straight edges most likely more than $1000 all told for decent [ read not super expensive] stuff.

Then there are your consumables, things like drill bits, tap's maybe a welder, grinder disks. It wouldn't take very much cash and the last $5-6000 would be gone.

I would seriously sit down and think about it. Have you designed heavy industrial equipment before? Is fabricating metal 'stuff' new to you? If yes to either of these, there will be a price to pay to aquire these skills through trial and error.

Oce, I'm not trying to discourage you, to the contrary, I'm TRYING to incourage you but in a realistic light. Built that wood router and have fun w/ it, then take another look at what you REALLY want and build it w/ the experiance you've gained w/ the first go-around.

Regardless what you decide, remember that there are people on here who will help you as much as they can to succeed. I know they've helped me,[ mainly by telling me I was crazy when I started (chair)] I wouldn't be here still, nor as far along in my project if it wasn't for them.

All the best

Jerry [sorry this got sooo long.. ]

[Oce]

About your thread being long, I would say the contrary; long threads and posts ARE the best when it comes to learning...The longer the thread, the better the information...
Thanks for giving me a little perspective on what to expect and what to consider, I think I have a much better background now than ever.

I'm not so discouraged now as I was yesterday Hehe...but then again, the more I get told now, the less the dissappointment it will be.
I guess we have a different opinion as for the precision, if I can get 0.005 +/-, I would say that is pretty good for a homebuilt machine.

I like to consider myself having a fairly okay experience with welding and fabrication, but yeah, some trial and error would be expected.

Like mentioned earlier on, I will mainly cut wood. I might try going with aluminium, but like you said, it would be very slow...

But let's get this thread back on track, I thought I might send a mail with some data to the Nexen dealer, to get some information on the prices, just for the fun of it. Any suggestions for the data?
(Let's use YOUR machine with ballscrews etc, just to have something to compare pricing with.)

Thanks again to G, J & Z for helping out !

[Zumba]

For the Nexen quote, tell them 600 IPM cutting, 1500 IPM rapids. Total gantry weight = 200 lbs. Y-axis car + Z-axis weight = 50 lbs. Z-axis = 30 lbs.

As for the accuracy goals, .005" over the entire table would be absolutely spectacular. That would destroy most commercial routers on the market under $30K. Even premium machine builders like Komo ($150K for their lowest priced machine) will not admit to extremely high accuracy. Repeatability, yes. Accuracy, no. Shopbot only claims a .015" accuracy and I don't think it's over the entire 8 foot table.

Someone else mentioned a budget of about $20K. That's a bit on the high side I think. Careful sourcing of your parts will get it down a couple grand.

[Oce]

Alright Zumba, I've submitted the info, hopefully they are fast and reply during the next day or so... I will post info as soon as I get it !

[JerryFlyGuy]

I'm glad I didn't discourage you! That wasn't my intent at all.

Did you mention your torque spec to them? I would think 1000oz-in would be 'ok', you could put 1200 but that is only at stall.

I went back to Nexens site to refresh my memory of why they wouldn't work for my system. I'm using a 640oz stepper [2 on the x axis] through a 10:1 planetary reducer. In theory the 20 series unit's would work for my system as I'd be putting out 400in-lbs [worst case]. However they are fairly large in having a 2.5" pitch dia. Also while designing my system, I had in the back of my mind that I was going to switch to a servo system at some point, hence the large reduction in the reducers. I didn't want to change to much when making the switch, other than maybe a pinion or something. After ordering my steppers I realized that 640 wasn't going to be enough but I continued w/ the system keeping in the back of my mind that I was going to need 1000oz or more servo's.

So why didn't I use the Nexen's?? Well because they were to large in pitch dia to get a usable resolution and force at the tool. I would have needed to increase the reduction. I got into a situation [in aviation its called 'coffin corner'] where I needed to increase the resolution [and decrease the speed], so I upped the reduction, but this increased the torque to much, so I upped the pinion size, which decreased the resolution [and increased the speed] so I upped the reducer size, which.. well.. you can see the 'corner' or 'circle' I was going in..

Your situation might differ,but w/ 1200oz steppers I'd question if you could get a workable solution. If your just cutting wood, it might be worth getting smaller steppers to do the job as those 1200's are WAY overkill [in my opinion anyway].Your going to need a fair amount of reduction as those steppers are going to have pretty good cogging issues if they aren't run fast enough [read 100rpm or more<--guess]

Building a system which can hit 10 thou is fairly attainable. I guess when we hear 'high precision" and "not concerned about expense" we each see a different machine. I see a machine which can hit a thou or two over the length of the machine [ or less] this could even be taken as far as to have temperature controls or sensoring to ensure that the machine stays at the same temp while its operating to limit thermal expansion issues. I see high dollar, heavily preloaded carriages and ballscrews. Glass scales for the length of the machine etc, etc. To someone else, a high tech machine might be made from metal and thats it, because their cnc experiance has been a machine made from wood. It's all perspective.

Looking at the precision you've listed, I still think that proper linear rails would be needed w/ just the standard carriages [no preload]. Ballscews or your Nexen rack's would still be needed, however you MIGHT get away w/ a precision Acme thread as well. Your overall machine weight would be ok at 500ish lb's.

So, can a person build all this for less than $20k, maybe but probably not less than $15k. It would all depend on how 'professional' you want it to look and how frugal/good you are at finding "deal's". All those little things like cable chains and proper wire and Control consoles etc.. they still add up.

As far as some of your original questions. For acceleration I'd think 1-2 seconds to feed speed and 2-4 seconds to rapid speed would be ok, less than that and you'll just have longer machine times, higher than that and you'll get lots of 'banging' due to acceleration, it will also require a heavier structure to withstand the forces of acceleration. That being said the new version of Mach3 coming out sure looks like it has some nice s-curving features in the accel and decel side of things, this should limit the 'banging' to an extent, it won't be out for a while yet though [as I understand it].

If your looking for a price reduction, another place would be the spindle, however if you go down to a 2-3hp I think your cut speeds will suffer. It all depends on DOC and bit size, but I would think that cutting much over 200ipm in wood [1/4" cutter and say 3/8" DOC] might be the limit. I don't KNOW this from experiance, only what I've read and gleaned from others coversations. Someone who is running a system of this sort could tell you better than I could for sure.

Be sure to keep us posted on the Nexen stuff, I might be in the market for more of that type of thing in the near future... when I start my next machine.. (chair) <--my wife... that is if the price is right.. I'm kinda a ballscrew guy when it comes to precision stuff.. we'll see

Jerry

[Oce]

when I start my next machine.. (chair) <--my wife...
Jerry

Hehe, I'm hearing the same thing...talking about taking the missis away for holidays etc, and all I talk about is how me and my friend are going about to build it and how much we will spend.. No wonder they get mad...

You put it exactly as I thought yesterday, we all have a different opinion about resolution / repeatability and general performance...

Anyways, I'm looking forward to hear from the Nexen rep, and will keep you informed on how it all goes.. (but that will happen in another part of this board...)

Anders

[Oce]

Alright, I've gotten a lot of mails from the Nexen rep. lately, and the prices was pretty high.
I got to see a couple of pictures from a plasma cutting table useing the RPS, looked VERY nice... But still;
A little rough calculation tells us that the RPS25 sprocket costs about 600$ for each piece (probably a lot cheaper in the us due to VAT etc, but still)
For each 1m (3ft) of rails you would have to pay a little less, but in the same 600-ish area...
The torque part mentioned was not an issue with this model, as it is rated up to 48Nm...

What do you think about that relative to ballscrews or other solutions?

I know it's too much for me to invest in my machine, although I could just...hm...keep my old car and...
Oh well, maybe not.