[Eddard_Stark]

First, let me thank everyone for their contributions. I have been lurking a few weeks and am still amazed and the depth of knowledge shared on the site.

I am a residential contractor of sorts and have been contemplating a CNC wood router build that will be used primarily for sheet goods. Mostly .75" melamine and plywood. (Cabinets, closets, etc.) As such I need to be able to have a y-axis working envelope 60" or greater to accommodate both 5 x 5 and 4 x 8 stock. While I don't expect to replicate an Onsrud, this is not a hobby machine. It may only run a few hours a week, but it needs to be capable of processing a decent number of sheets daily if need be. (30ish, maybe 40??) Based on my research thus far I have the following requirements: (Which are probably all wrong!)

1.) It must be able to consistently and accurately make clean cuts in a nested routing application. (more on this later)
2.) Vacuum hold down system
3.) Pneumatic pins for positioning
4.) Air drill for boring as I anticipate having a single spindle w/o ATC capabilities
5.) AORN I am planning on a 3kw chinese spindle (water cooled?) with a good VFD (Hitachi?)
6.) My understanding is cutting speed affects tool life and it certainly affects the bottom line, so I want something that can cut decently fast and clean.
7.) Budget: Under 10k is the goal, but I have an additional 5k in the budget. (That I would rather not spend!)

As for my skills they are as follows: I have zero hands on CNC experience. My fabrication abilities are OK, but I have some friends who are very capable and I am not afraid to pay someone to do something I am not capable of. Anything beyond installing a ceiling fan or light switch is uncharted territory for me, so I will definitely need the Wiring for Dummies manual on this project! While I don't have any specific relevant software experience, I am not worried about learning whatever software is necessary.

I have done a fair amount of research on the Mechmate builds, as well as CNCRouterparts builds, and read the entire Widgetmaster Big Red thread. (And the MadVac build as well) Due to my inexperience I am leaning towards a known design, preferably one with several active users.

As such, the obvious candidates (to me!) are the Mechmate and CNCRouterparts builds. I have concerns about cut quality on the Mechmate (fuzzy edges on MDF) and I have concerns about the rigidity of an 8020 build (for my purposes). Also, the CNCRP kits are a maximum 4x8, so I was thinking more of adopting the linear motion, motors, electronics, etc. of a CNCRP to a steel frame.

So other than asking what should I build? I have three specific questions:

1.) Does anyone have experience with the Mechmate specifically wrt sheet goods for cabinets? If so, what kind of feed rate and cut quality are you experiencing? (I realize that build quality of MM differ, thus leading to differing outcomes, but I am curious if anyone has hands on experience with one. I plan on posting on the MM forum as well)
a.) Are some of the issues with MM cut quality (that I could totally have misinterpreted btw) due to direct drive? I see some people have geared their MM and the quality seems good to me, but I am curious about consistency with nested routing.

2.) Is my conclusion that steel, as opposed to 8020, will be better suited for my application correct?

3.) What kind of precision and life could one expect, ceteris parabus, between a vee bearing liner motion sytem, CNCRP milled steel with skate assembly, and linear rail w/bearing (Hiwin?).

Thanks all! Btw I hope I posted this in the right place.

[ger21]

1) I think the info you've received at the Mechmate forum should cover this.

2) Yes, I'd agree that steel is a better choice for a machine that large.

3) Hiwin type rails and bearings should last a lifetime in your application, and are an order of magnitude stiffer than the others. I think I've seen Mechmates using linear rails before, which is a great upgrade. However, getting a straight, perfectly flat surface to mount them to might be difficult.
The V bearings used on the Mechmate work well, and have been used on ShopBots for years. It's a proven, low cost system that works pretty well.




Since you really want a proven design, the Mechmate is probably your best bet. It's also the least expensive option, but it's a bit labor intensive. With belt reduction drives, I don't think there are really any cut quality issues to be worried about.

I'd recommend at least a 4Kw spindle, as spindle power is usually what limits the feedrates you can cut at.

A really good vacuum system is also important when nesting full sheets. You might want to read through the threads on the ShopBot forum from the links here.
ShopBot Vacuum Motors

[Eddard_Stark]

3) Hiwin type rails and bearings should last a lifetime in your application, and are an order of magnitude stiffer than the others. I think I've seen Mechmates using linear rails before, which is a great upgrade. However, getting a straight, perfectly flat surface to mount them to might be difficult.
The V bearings used on the Mechmate work well, and have been used on ShopBots for years. It's a proven, low cost system that works pretty well.

Re: linear rails; that it would be difficult/expensive to get a straight perfectly flat surface to mount to was a contributing factor that led me to lean towards a cheaper, less robust linear motion system. Being a proven low cost system that works pretty well is right up my alley. I hope the machine I am planning with be version 1.0 and I can shoot for more performance in a subsequent machine with the aid of experience.

Since you really want a proven design, the Mechmate is probably your best bet. It's also the least expensive option, but it's a bit labor intensive. With belt reduction drives, I don't think there are really any cut quality issues to be worried about.

This is good to hear and is pretty much inline with the feedback thus far on the MM forum.

I'd recommend at least a 4Kw spindle, as spindle power is usually what limits the feedrates you can cut at.

More power. I like it.

A really good vacuum system is also important when nesting full sheets. You might want to read through the threads on the ShopBot forum from the links here.
ShopBot Vacuum Motors

Will do. Thx for the advice.

[Eddard_Stark]

Ok, I received the plans for the MM and have begun narrowing down specific components. I don't intend for this to be a duplicate of my MM build thread, but I do have more general CNC related questions that I think would be useful to other people so I will keep those here.

Additionally, I can do a abridged version of my build here on the Zone if there is any interest. You must be registered to view MM threads. I haven't started ordering parts yet, but I am on the cusp. This project is ~98 percent to go forward.

The questions:

1.) I am planning on using an ethernet SmoothStepper motion controller on my MM. It is my understanding I can get away from the parallel port completely if I do this; is this correct?

2.) Since my motion control processing will be offloaded to a dedicated piece of hardware does that mean I can upgrade to a newer PC and OS w/o fear of it interfering with Mach3? It's important b/c it seems that the long awaited Mach 4 arrival is not far off. So there is a possibility I may be able to use Mach4 if not from the beginning, then soon after; and it is my understanding Mach 4 will require a much more capable machine than the typical dumpster PC's w/XP people have been using out of necessity due to Mach 3 and the Parallel Port.

IOW I don't want to set up my machine on rapidly obsolete hardware and software (PP and XP), especially if I may need more modern hardware very soon to run Mach 4.

[ger21]

Yes, and yes.

[wizard]

Ok, I received the plans for the MM and have begun narrowing down specific components. I don't intend for this to be a duplicate of my MM build thread, but I do have more general CNC related questions that I think would be useful to other people so I will keep those here.

Additionally, I can do a abridged version of my build here on the Zone if there is any interest. You must be registered to view MM threads. I haven't started ordering parts yet, but I am on the cusp. This project is ~98 percent to go forward.

The questions:

1.) I am planning on using an ethernet SmoothStepper motion controller on my MM. It is my understanding I can get away from the parallel port completely if I do this; is this correct?

Yes and No. Why? Because it depends upon how much I/O you need. Do understand though that there are different ways to get more I/O. I only mention this because you indicated the possibility of features either requiring more I/O or manual operation.

2.) Since my motion control processing will be offloaded to a dedicated piece of hardware does that mean I can upgrade to a newer PC and OS w/o fear of it interfering with Mach3?

It is possible but it is also possible that new hardware will create issues for Mach 3/4. This is especially the case where manufactures are going after low power in a system. In other words just because hardware is new don't expect miracles.

It's important b/c it seems that the long awaited Mach 4 arrival is not far off. So there is a possibility I may be able to use Mach4 if not from the beginning, then soon after; and it is my understanding Mach 4 will require a much more capable machine than the typical dumpster PC's w/XP people have been using out of necessity due to Mach 3 and the Parallel Port.

Well Windows certainly requires better hardware as you leave XP behind. I do wonder how stuck you are on Mach, there are other options out there such as KFlop, LinuxCNC and others. The thing here is the ease with which they are configured for more I/O.

IOW I don't want to set up my machine on rapidly obsolete hardware and software (PP and XP), especially if I may need more modern hardware very soon to run Mach 4.

[hanermo]

The right solutions for you, in my experience, are;

Forget the SS. Get a CSMIO-S.
There is no comparison, none.

Use steel. Its cheap and easy to work with.
Have the steel surfaced, cheaply, anywhere local with a largish mill.

Using real steel, on the critical surfaces, it wont twist and distor.
Ie Dont use structural steel, ie mild steel, get tool steel.
Its still cheap, about 1€/kg.
Now you can always make it stiffer and bigger later, easily and cheaply.
For maybe 200-500$ you can get all your beams milled flat, and pre drilled for mounting linear rails and corner brackets.
Get the corners reamed for locating pins.

This eliminates 95% of all work and headaches, as the alignments now come out right.
The above will save 50-200 hours work on a large machine.
Alignments and twist are the real problems on bigger machines.

Forget anything other than profiled linear rails like Hiwin etc.
Its cheap in comparison to what you need.

I built a mill in steel, 1600 mm x 500 mm table.
A table of 300 kg still moves effortlessly, even with a tiny nema 23 stepper.
Mass will not hurt, only help.

Larger sections will scale up very easily for a router.
Just keep the rations the same.

The mechmate is the best guide, with above caveats.

I would also suggest brushless ac servos.
Again, cheap for what you need.
400€/axis, for 400W 80V servo motors, drives and everything.

5x more accurate than steppers (2000 / rev vs 400 realisticwith 4000 cpr encoders), 5x more accceleration, and 3x more top speed (3000 rpm vs 500-800).
And noiseless (well the better new stepper drivers with sinusoidal stuff are also very good, like the 542 series).

You can get an auto tool change spindle for your budget.
This is your nr 1 most impotrtant moneymaker and productivity enchancer, bar none.

Likely, you can try it with anything else, and then get an auto tool changer spindle.

[Eddard_Stark]

wizard,

Thanks for the detailed reply. It took me a while to reply b/c you and hanermo gave me a lot to think about and research.

AORN = As of right now. It was superfluous. I was using it to indicate that I was unsure of the spindle sizing.

As for the rest you make several good points, some of which have been in the back of my mind.

1.) Profiled linear rails; I am beginning to think I will regret using a lower cost system such as the vee bearing. While it is a proven low cost system, I think the benefits of upgrading to profiled linear rails outweigh the marginal costs. Especially at the prices I see on 25mm Hiwin rails at Automation Overstock. (I am still looking into milling prices/options for the gantry and main beams)

2.) Of course the upgraded linear rails will be for naught if I don't have a frame adequately matched to performance of the machine. Which has me leaning away from the MM design some. Not b/c there is anything wrong with it, but if I am going to make significant changes to beef it up (and switch to profiled linear rails) then I might as well venture off onto another design that allows side loading. The single worst feature of the MM for me is the necessity of end-loading sheets. Decisions...decisions.

3.) You keep stressing that I need to be mindful of my I/O. My understanding of electronics is lacking to say the least. Can you recommend some reading or tutorials specifically relevant to CNC control applications? I don't have an understanding of the fundamentals at all. Consequently it is difficult for me to understand why things are done, need to be changed, etc. I have an Arduino Uno on order with some simple outputs (like LED's, small steppers, etc) so I can start getting some hands on controller experience at a basic level.

4.) I am not married to Mach, it's just that's what most people use so I know support will be readily available.

[Eddard_Stark]

hanermo,

Thanks for the reply.

As I wrote above your point about profiled linear rails is well taken and I think that is the way I am going to go. Ditto on the machining if I do indeed use the rails.

As for the 400 W AC servos - I was surprised to see that I can buy Leadshine AC servos w/drives @ $400.00 per. OM PK299-F4.5A 880oz/in (~6.21N/m) steppers are ~$350.00 w/ Gecko 503v drives. If I were to go with profiled linear rails and beefed up gantry (relative to the MM) [4:1 belt reduction running on rack and pinion] the performance characteristics of the servos compared to the steppers is not even close imo.

The question that kept me reading into the wee hours of the night was the issue of reduction on the 400 W servos. While 4 : 1 seems to be sufficient for the aforementioned steppers, I can afford to reduce the servos lower to increase the resolution and torque.

Looking at the PK299-F4.5A speed/torque graph, if wired 48V DC I would be getting ~200oz/in @ 1000 rpm which would give me ~800oz/in(5.6N/m) torque @ 785 ipm (.33 m/s). This would be jogging. It looks like they can go a little higher rpm, but from what I understand that is pushing them pretty hard and I am rapidly losing torque with every increased rpm.

Looking at the servos the same torque and speed is childsplay. They are rated 181oz/in torque and are rated for 3000 rpm. Holy Cow. Looking at the same 4 : 1 we have a little less torque @ 724oz/in (5.11N/m) but we can do that @ 3000 rpm which would increase the speed by a factor of 3! A blazing 2355 ipm (~1m/s). Ouch. If I had to venture a guess those speeds are going to require a larger frame than I want/need to build.

So this leaves me with two questions:

1.) Assuming I have a 4KW spindle (~5hp), chosen due to availability of moderately priced Chinese spindles, and my main application is going to be .75" sheet goods, what feed rate would practical to design around? (I am assuming the spindle will control here) Based on my reading I don't think 4-500ipm feed rate (jog rate of 1k+) is out of the question, but please let me know if I am in outer-space here. (Again this is based on moving away from the MM design in favor of something side loading and more rigid.)

2.) Once a feed rate is established I can determine a gearing ratio. And then decide on belt vs. planetary. It would be nice from a cost standpoint if I could stay with a one stage belt reduction, but I am not sure of what is advisable here or what the upper limit of one stage is; belt or gear? I have read that Bayside Planetary occasionally has surplus gearing on ebay, and that may be the way to go. Not sure.

Lastly, I forsee myself starting with a single spindle w/drill but probably upgrading to ATC later as you mentioned.

[ger21]

I believe those Leadshine servos have a peak torque of over 400oz in, so they actually have much more torque than you think.
Without very quick acceleration, you'll never get to 2200ipm, let alone be able to use it. I think you can cut full depth 3/4" sheets with a 3/8" compression bit at around 400-500ipm with a 5HP spindle, and you should be able to get 1000ipm rapids (Maybe more) out of those servos. Depends on how much weight your moving.
I plan on using the same servos on the machine I'm building. They'll probably cost closer to $500 when you add in the longer encoder and power cables you'll need. They do occasionally have them on Ebay for about $20 less, and you can save even more importing them from China.

I'd spend some time looking for planetary gearboxes on Ebay. Somewhere around 10:1 would be ideal, but you can probably go up to ~20:1 and use a larger pinion, which will give smoother motion.
Something like these.
Apex Dynamics PE070 Planetary Gearbox Ratio 010 1 | eBay

Watch out for backlash, though. Go to Apex Dynamics website and find the correct model numbers you need, and find the numbers with low backlash. Also look at Harmonic Drives. I picked up a pair of these with <1arc min of backlash for $225 each a few months ago..
HPG 20A 21 F0 New Harmonic Drive System Reducer Ratio 1 21 for 200W Servo Motor | eBay

A single belt reduction is limited to around 4:1, maybe 5:1. If you can get a good deal on gearboxes, they won't cost that much more than a well built belt reduction will cost.
You can also do a 2 stage belt reduction, but they'll still cost quite a bit, and add a bit of complexity.

Be aware that moving away from the mechmate will result in basically a custom designed machine. But if you know what you're doing, you can get a much better machine and still stay within your budget.

[wizard]

wizard,

Thanks for the detailed reply. It took me a while to reply b/c you and hanermo gave me a lot to think about and research.

AORN = As of right now. It was superfluous. I was using it to indicate that I was unsure of the spindle sizing.

Never saw AORN before!

As for the rest you make several good points, some of which have been in the back of my mind.

1.) Profiled linear rails; I am beginning to think I will regret using a lower cost system such as the vee bearing. While it is a proven low cost system, I think the benefits of upgrading to profiled linear rails outweigh the marginal costs. Especially at the prices I see on 25mm Hiwin rails at Automation Overstock. (I am still looking into milling prices/options for the gantry and main beams)

It really depends upon what you want to accomplish. In general though I think you would be happier with profile rails though you are building a rather large machine. If you can't get the frame and gantry machine locally so that you can properly mount the profile rails they could be seen as a waste of money. In any event if your goal is a machine capable of precise work I'd go with the profile rails.

2.) Of course the upgraded linear rails will be for naught if I don't have a frame adequately matched to performance of the machine. Which has me leaning away from the MM design some. Not b/c there is anything wrong with it, but if I am going to make significant changes to beef it up (and switch to profiled linear rails) then I might as well venture off onto another design that allows side loading. The single worst feature of the MM for me is the necessity of end-loading sheets. Decisions...decisions.

Well consideration for things like usability in your context I can't address but if end loading is out another design would be needed. The problem with building a machine this size is one of access to a machine shop that can handle the large parts of the machine. So it probably wouldn't hurt to review a design with your local shops to verify that they can indeed handle any machining you may need.

3.) You keep stressing that I need to be mindful of my I/O. My understanding of electronics is lacking to say the least. Can you recommend some reading or tutorials specifically relevant to CNC control applications? I don't have an understanding of the fundamentals at all. Consequently it is difficult for me to understand why things are done, need to be changed, etc. I have an Arduino Uno on order with some simple outputs (like LED's, small steppers, etc) so I can start getting some hands on controller experience at a basic level.

Many of the very low cost controllers on the market have limited I/O which may be a problem if you want to do a lot of stuff under the control of the G-Code. What I mean by limited is the number of inputs for example switches and outputs for example to switch a vacuum pump on and off. This isn't a huge problem as most CNC controller software is fairly easily expanded via additional I/O cards.

4.) I am not married to Mach, it's just that's what most people use so I know support will be readily available.

Well that is debatable. For more advanced systems I see people leaning towards other hardware and software. Mach 4 is coming soon but that will apparently come in different versions. I only mentioned the alternatives because you seemed to have implied that you want more out of the controller than the basic XYZ. Maybe I read too much into what you posted.

[Eddard_Stark]

Be aware that moving away from the mechmate will result in basically a custom designed machine. But if you know what you're doing, you can get a much better machine and still stay within your budget.

Well, it remains to be seen if I know what I am doing, but the above sentence encapsulates my thoughts exactly.

- - - Updated - - -

[Eddard_Stark]

Well that is debatable. For more advanced systems I see people leaning towards other hardware and software. Mach 4 is coming soon but that will apparently come in different versions. I only mentioned the alternatives because you seemed to have implied that you want more out of the controller than the basic XYZ. Maybe I read too much into what you posted.

You are correct I am looking for more than basic XYZ. I don't know who has better support or what software will be best suited for my needs, I am just going on what I have read, and can deduce from others experiences. Having said that, I will probably try out both Linux CNC and Mach3 to see which one I feel suits my needs the best; at the very least I will be doing extensive research into each.

To that end, and given the fact I have zero Linux experience, today I loaded Ubuntu onto an older Dell Inspiron that was running Vista at such a pace my brother- in-law gave me a laptop that he thought was useless. Well, I after loading Ubuntu it is a new machine. I am so impressed by what I have seen so far, going forward I won't be installing Windows on any new machines, save any that will be running Windows specific software.