[Hyrum R]

Hi again everyone. Once again, the CNC bug has bitten me, and I'm looking to upgrade my current machine to a full size 4x8 (maybe 5x8 if I can squeeze it in the space I have.)

I would probably consider a MechMate, but the footprint is just too large. I'm VERY limited on the space I have, and I want a machine that can cut a full 4x8 sheet for cabinet parts (and any other ideas I come up with. )

I've pretty well decided what I want to do for the mechanical. I've looked through CarveOne's thread, and he's got it pretty well like what I'm thinking. It's going to be a welded steel table, and I'm (most likely)going to be using Ahren's bearing blocks and Nema 34 R&P system. I've considered using the more expensive linear rail systems, but I don't know that the extra cost would be worth it for me. Anyone have any experience this way?

The main thing I need help with is deciding on what motors to use. I've been through hundreds of posts over on the Mechmate forums about what motors to use. I've decided to go with Nema 34 at least for the X&Y, and to use Gecko 203 drivers. I'm having a hard time deciding what motors to get though. According to the guys over at Mechmate, it's best not to go much over about 800 oz*in. I've been looking at Keling's motors, and they have a 960 ozin motor (http://www.kelinginc.net/KL34H295-43-8A.jpg) that looks pretty good to me. How well would a 48v unregulated power supply work with this motor? Should I try to get closer to a 60v power supply? (according to Gecko's formula, this motor's optimal voltage would be 58v). Should I be looking at a different motor? Also, can I hook up 3 Nema 34s for the X&Y(2 for x, one for y) and use a Nema 23 for the Z? I ask because I have a really nice Z carriage assembly from K2 that has a Nema 23 mount. Or should I just redo the motor mount for a Nema 34? I want to put at least a PC 7518 router on this thing, and possibly a spindle later on.

I just have so many questions going through my head about this, and I want to do it right with this machine. I emailed Ahren earlier asking about his Nema 34 R&P system, and he said that with a G201 driver and a 960 oz*in motor, it was pretty well unstoppable at anything under 1000ipm.

Thanks in advance. I can't wait to just start building, but that stupid problem of $$$$ keeps getting in the way.

[Hyrum R]

No replies? If some of you experts could chime in about the motors, I'd be most grateful.

Is anyone using Ahren's Nema 34 R&P system? If so, how is it working out? I've been slightly concerned that the 2:1 gearing might not be enough for smooth chatter-free cuts, but I calculated that my resolution should be at approx. .0007 if I have a 10 microstep driver. How that translates to reality might be a slightly different story.

[CarveOne]

If you count the NEMA 23 motors that stick outward, my machine has a footprint of 12 feet long by almost 7 feet wide. The actual cutting size is 50 inches by 11 feet 2 inches.

I'm using Ahrens short and long carriages and his NEMA 23 pinion drives, Gecko G203V driver modules, CNC4PC C11G breakout board, and four Keling 495 oz-in steppers and their 72vdc 20 amp PSU. There is lots of power available.

If you consider using a very heavy gantry with a vertical rail like I am, note that all of the weight is on the upper carriages. The lower carriages bear little or no weight load. So I have ~175 to 189 pounds riding on just two ABEC7 bearings on each side of the X axis. That's ok with me and I'll just have to expect to replace these bearings more often than with the horizontal rail scheme that Ahren recommends. I don't expect to replace them as often with hobby uses. If it were a production machine I would keep plenty of the bearings on hand. Removing a carriage for replacing bearings is not hard.

I would recommend using 3" square box tube for the legs and 1-1/2" x 3" box tube for the other tubing if you can. 3/16" thick tubing walls if this is to be used in a production machine.

The G203V drives can't exceed 80vdc input. 72vdc is pushing it a little but has been fine so far and others are using this. If you want a little more margin at slightly less torque, use a Keling 65vdc PSU. The Gecko documentation on their site can help you determine if the Keling 960 oz-in motors are ok with either of those voltages. I haven't given any consideration to them so far.

The dollars kept me from buying the steel I needed too, but in the end I spent more than enough to correct that issue. Buy what you really want for the table construction. The rest will come in time. While it's fun to build, rebuild, and rebuild again it just wastes time and you'll end up spending much more than you should.

CarveOne
(aka "General Overbuild")

[hemsworthlad]

Have you concidered servo motor's.?

Personally I wouldnt go the stepper route when I could have these 900W AC brushless servo's from DMM Motion Control Products - DMM Technology or the even the 400w.

Servo's are far more powerful than steppers esp when the speeds rise so even twin 400W would probably do what you need.? Twin 900W would be awsome.
Only problem I could see with the 900W and R&P's need for gearing is the low 1050rpm, but that said Nema 34 steppers wont give you much more than 1000rpm that will be of any use.
Unlike steppers servo's produce power across there whole speed range and full power is max when flatout so with twin 900W you would have more than enough power to gear them down if extra speed is required.
Twin 400W that spin at 3000rpm geared 3/1 would still give you the same rapid speeds as nema 34 steppers but with extra resolution and probably more torque esp at higher speeds. (say probably because would have to do the calc's to be sure)
Then there's the bennifit of the speed/accurecy of servo encoders which will help with twin driven gantry.

To me the little extra that it cost's would be worth it and you would have a very very accurate machine.

It's the route I would take if starting again from scratch, infact i'm going to upgrade my machine with twin 400W turning my 5mm ballscrews geared 2/1.
At present I use 1 x 1200oz/in nema34 linking the screws with a long timing belt with gecko 203V's @ 74V.
I want rid of the long belts but dont like slaved steppers because of the abilty to miss steps and potential racking esp at higher cutting speeds, Servo's I think will do a better job of keeping things in check and allow high accurate cutting feeds. (It's a personal thing so dont you slaved stepper peep's jump on me. .Lol)

Cheers
Dean.

[Hyrum R]

I've considered servos. I just need a lot more knowledge of motors, encoders, drivers, etc. before I'd feel comfortable. I'm sure I could get a lot of help here on the Zone, though. So far, I'm not committed on any design yet. I just wanted to post here to get some ideas of what would be recommended. I've just seen almost everyone using steppers on their woodworking routers, and so that's all I have experience with. I'm sure I'd never regret using servos. So I'll seriously consider it.

I don't want to go bankrupt building this machine, but I do want a good reliable machine that I can use day after day. If that means spending a little more to do it right the first time, I'm willing to do it.

[hemsworthlad]

I've considered servos. I just need a lot more knowledge of motors, encoders, drivers, etc. before I'd feel comfortable. I'm sure I could get a lot of help here on the Zone, though. So far, I'm not committed on any design yet. I just wanted to post here to get some ideas of what would be recommended. I've just seen almost everyone using steppers on their woodworking routers, and so that's all I have experience with. I'm sure I'd never regret using servos. So I'll seriously consider it.

I don't want to go bankrupt building this machine, but I do want a good reliable machine that I can use day after day. If that means spending a little more to do it right the first time, I'm willing to do it.

Yes I understand the need to fully understand before commitment and a wise policy.
Really the answer to why you mainly see steppers on DIY cnc machines is mostly because in the past the price diffferance between servo's n steppers was too big, but i think like everything as technology advances prices drop and thats whats happening regards servo technology.
If servo's where the same price as steppers nobody would choose steppers over servo's and to me with these DMM servo's it's getting very marginal if planning on buying hi-end stepper components like Gecko's etc.

Another way to think or justify the extra cost other than a superior quality, accurecy, speed of the machine is the extra price and desireabilty it will add to it's value if you ever come to sell it.

[Hyrum R]

Ok, regarding servos/steppers. Is it recommended to gear servos down more than steppers? Do I want them spinning faster rather than slower? I just have so very little knowledge about this sort of thing. I'm willing to learn though.

Looks like things have come up that may put my plans on hold for a little while, at least for making the table. But if I can use that time to figure out the electronics, design, etc. that would be time well spent, I think.

[hemsworthlad]

Ok Well i'm no servo expert but heres my understanding.
The reason for the gearing is because servo's are best used and most efficent when cutting running at or around there rated rpm which tends to be around 3 or 4000rpm. This tends to be too fast for most cnc machines also the other advantage to gearing is the torque increase's with the gearing. IE 400oz/in would give around 800oz/in if geared 2/1 may not be exactly double but you get the picture. This means smaller servo's can be used, also the other thing with servo's they have 2 ratings a continous and peak rating. When you see 400w continous this what it will run at all day long 24/7 and not be stressed the peak rating is the max power it will give for a limited period when pushed hard.
The differance between steppers and servo's is stepper torque ratings tend to be given when stood still and has you may or may not know the torque drops proportional as speed rise's to the point that it dissappears completely and the motor stalls. Around 1000rpm there's not much torque left in a stepper and even lower rpm's with larger nema34 motors.
Servo continous ratings on the other hand tends to be across the whole speed range but tend to be at their most effiecent when running at the rated rpm.
Thats why when using servo's it best to gear and setup for having the servo's running around the rated speed with regards to the desired cutting feeds for best performance. This will give far more power than a stepper can give at higher cutting feeds. (Stepper fans will be all over me know for this comment. . Lol )

Like i say i'm no expert and i'm sure one will be along soon to rebuff me and tell me i'm talking ****e.?

[Hyrum R]

Well, it sounds like you're talking sense to me. I've done some limited research into servos, but there are just sooooo many of the diy machines using steppers, that I couldn't find a whole lot of good info out there. So, let's say I get some servos like you're recommending. They have a continuous rating of 408 oz-in. Now, if you're saying that a servo likes to run at 3-4k rpm, that means that with no gearing at all, with a 1" pinion gear, we're talking approx. 9425 ipm?

Ok, that's totally unrealistic and would probably destroy more than just my machine. So, I think I'd want to gear down at least 8:1. Right? Maybe even more?

The other thing I'm not totally understanding is how servos move. I mean, with a stepper, the faster your computer can send pulses, the faster you can run your stepper. How exactly do servos work? I can't imagine them working quite the same way. Anyway, just lots of stuff to process in my mind.

[hemsworthlad]

Now, if you're saying that a servo likes to run at 3-4k rpm, that means that with no gearing at all, with a 1" pinion gear, we're talking approx. 9425 ipm?

Ok, that's totally unrealistic and would probably destroy more than just my machine. So, I think I'd want to gear down at least 8:1. Right? Maybe even more?

Yes thats right but you control the speed and dont have to run at that speed. That said you would never run 1/1 any way due to the poor resolution it would give. So running a higher ratio gives the best of both worlds speed and resolution with more torque. A realistic ratio of 4/1 or 5/1 would give good speed and resolution with loads of torque.

The other thing I'm not totally understanding is how servos move. I mean, with a stepper, the faster your computer can send pulses, the faster you can run your stepper. How exactly do servos work? I can't imagine them working quite the same way. Anyway, just lots of stuff to process in my mind.

They work just the same with the servo drives take step & Dir signals from the controller just like steppers, but your wrong thinking the faster comp sends pulse's faster the steppers go. Yes faster pulse rate helps some what but ultimately it's the stepper motor it's self and it's make up plus the drive combo that restricts speed. Even the best stepper motors and drives cant match the speed/torque servo's.
For instance I run a smooth stepper which takes the parallel port away and gives me the abilty potentialy run at 4Mhz.! Even the fastest computers with super fast PP can only manage 100Khz. But my nema 34 motors and Gecko 203v's combo running at 74V still give up the ghost at 1000rpm or so and the actual usable range is less than 800rpm.

Again I'm no expert on the way servo's work but they do work differant to steppers and it's the combo of this differance and the differant way the drive's work that allow themto work at these speeds.
It's the encoders that or the resolution of them that allow the precision with servo's and the higher the encoder resolution better the precision.

[hemsworthlad]

If you want to check out these youtube vids of DMM servo's on a machine cutting. Also shows his control box.

[nomedia="http://www.youtube.com/watch?v=ZfXR7iXUn6U&feature=mfu_in_order&list=UL"]YouTube - Rod's cnc Tiger Vcarve[/nomedia]

[nomedia="http://www.youtube.com/watch?v=S5zlySq2WYA&feature=related"]YouTube - My first attempt building a AC Servo controller the "home shop machinist" way.[/nomedia]

This is the guys build thread. http://www.cnczone.com/forums/diy-cn...router-11.html

[Hyrum R]

Thanks. I took a look at those, and that looks like a sweet machine. I didn't realize that the DMM servos have a max speed of 1080 rpm. That's actually a lot more realistic for me. So that means if I gear it down 3:1, I would have a max speed of about 1130 ipm. If I use Ahren's Nema 34 mounts, he's gearing them 2:1 though, so that would give me closer to 1700 ipm. Personally, I think I'd rather go 3:1. Lots of thinking to do.

[hemsworthlad]

Thanks. I took a look at those, and that looks like a sweet machine. I didn't realize that the DMM servos have a max speed of 1080 rpm. That's actually a lot more realistic for me. So that means if I gear it down 3:1, I would have a max speed of about 1130 ipm. If I use Ahren's Nema 34 mounts, he's gearing them 2:1 though, so that would give me closer to 1700 ipm. Personally, I think I'd rather go 3:1. Lots of thinking to do.

Only the 900W have this low 1080rpm the 400W are 3000rpm. I think the 400W would be a better option personally and gear 5/1 and still endup with 1885ipm but with higher resolution and still have lots of torque, esp if using twin slaved 400W. Obviously this will largely depend on how you build the machine and how heavy the gantry etc but I would want at least 3/1 with a 1" pinion. Again this will depend on what you intend to do with the machine.

[ahren]

Having worked with servo motors, stepper motors and linear motors connected to various different mechanical systems, I can say they all have their place. In the case of our Nema 34 R&P system, the product is specifically designed to work with stepper motors, as it is geared aggressively to take advantage of their high torque at lower RPMs. At cutting speeds, with a 960 oz-in stepper, this system can develop over 400 lbs of linear force in a dual drive system, and I have been able to achieve rapids of around 1600 IPM, so I would not say that performance is terribly compromised.

While servo motors do have the ability to develop more power in a smaller package size, they require significant gearing to develop this power. A ballscrew system with a greater mechanical advantage is a great candidate for a servo drive system, and we used these extensively on chip placement robots for this reason. However, for our R&P system, without adding an expensive gearhead to the system, you won't be taking full advantage of the servo's power. A stepper-based system will be a signficantly better bang for your buck.

Best regards,

Ahren
CNCRouterParts

[Hyrum R]

Decisions, decisions! At present, I'm probably leaning a little more towards the stepper side. Mainly because I'm familiar with them, and also so I can take advantage of Ahren's r&p system. The next reason is cost. I figure I can get my motors and drives for about $650. I already have a BOB I can use. I'm thinking of going with the Gecko 201x drives (on sale for $84 right now), as I'm comfortable enough with stepper wiring to not do anything too stupid.

But having the feedback of a servo does sound enticing. I just don't know that it would actually be necessary though. I've never had my current CNC miss steps unless I did something stupid with it.

[Hyrum R]

Well, it's time to bring this thread back. I've decided to use the Motionking 34HS9801 stepper motor. 34HS Stepper Motor 86mm 1.8degree, MotionKing Stepper Motor

That one should work absolutely beautifully with a 65v power supply from Keling and some Gecko 201x drivers. The cost of the motors including shipping is just over $300. Several people have used them on their Mechmates, so I think they're a wise choice. Also, I have a 305 oz-in motor from CNCRouter parts, and it has the exact same inductance, so I plan on using 3 of the Nema 34 motors for the x&y, and the Nema 23 for the Z. This way, I can reuse my K2 Z carriage system.

What I'm wondering now is should I spend the extra $$$ on some industrial linear rails or do you think that the bearing blocks from CNCRouter parts and the steel rails will be fine? I'm looking at something like the Hiwin linear rail system that Keling (and a few other places)sell. One of the biggest problems I currently have with the steel rails and bearing blocks I currently have is the dust build up. I know, I know, I should put something on there to brush it clean and I'm going to get around to it sometime. I know I'm looking at a lot more $$ to do it, but I want this machine to have everything done right from the first. I'm looking very hard at this system:
LGW20HA1TxxxxZ1H/E1, Combo, 20mm T Rail, Long Flange Block, Specify Length, Hiwin LG 20 Combos - automation products from Automation 4 Less

I figure about $650-$700 for the rails and blocks. Does that seem like money well spent to you guys?

[hemsworthlad]

I figure about $650-$700 for the rails and blocks. Does that seem like money well spent to you guys?

Yes it will be money very well spent, Hassle free and will handle everything you throw at them and last much much longer. IMO Makes the differance between an OK machine and a Great machine.

[lgalla]

HyrumR
Those rails are T rails.The mounting bolts come up from the underside.There is no through holes for mounting.Could prove difficult to use and mount if you do not design for this type of rail
Larry

[Hyrum R]

Thanks for the heads up there, Igalla. I wasn't even paying attention to that. These are the older "legacy" style rails I was looking at. I can't find any on their site with the mounting from the top. Possibly they have sold out?

So, what's the difference between square type, flange type, etc. Do I need anything bigger than 15mm for this machine? What model would any of you recommend for this machine? Thanks again.

[Hyrum R]

Ok, I found some 15mm rails in stock from Automation-overstock:
automation products from Automation Overstock: AGR15RxxxxH, AG Linear Rail, 15mm, *specify length, Hiwin AG 15 Rail

Do you think 15mm is sufficient? If it will work, that will save me a bundle over going with the newer type of Hiwin guides. I figure with the blocks and lengths of rails, I'm at about $715.