Buy commercial machine for 6k or build your own?

[Menatep]

I've been lurking around on this site for about a week now and it has helped me tremendously to learn about cnc. I am in the market for a CNC router, but i am not exactly sure whether it's best to buy one commercially or build one yourself.

I've narrowed down some of my requirements and they are listed below:

- 24" by 24" or 24" x 36" (preferably) working area with Z-axis travel of 5"
- Would like to use servo motors but might consider steppers depending on size. Anything in 200-500 oz range would be ok.
- Would like to have ballscrews
- Would like to have limit switches

Commercial offerings that i looked at are the following:

- K2-3925-G model for $6450.95 (with 3 axis controller, servos, electronics, aluminum t-slotted table top, standard ABBA brand rails, and a router mount)
- K2-2539 model for $5810 for the same configuration as above.
- DynaCNC TT2436ST 24x36 model for $4695 without ballscrews and without servo motors. If it's upgraded to servo motors, limit switches and ballscrews price shoots up to $6,069.90
- DynaCNC TT2436B - This is a barebones 24x36 table and it's listed at $1495 without any motors or electronics. They have a ballscrew upgrade for $499 and servo motor and electronics upgrade for $1995. The price comes up to $3989 but i'm not sure what else is missing and why there is a difference of almost $2000 between TT2436ST and TT2436B model with servo motors and ballscrews upgrades. I guess i'll have to call the company up to clear things up.
- A model i found on ebay... looks like it was home-made
http://cgi.ebay.com/ws/eBayISAPI.dll...m=220071348278
It looks like this model is using unsupported rails and acme lead screws and stepper motors. Not sure about how sturdy the frame is. I thought of just buying the table without any motors or electronics or lead screws but the seller does not want to sell just the frame. He's asking around $3200 for the 20x30 setup.

Looking at the DynaCNC it seems that their 24x36 table-top models has unsupported rails so that might cause some deflection problems if i'm cutting aluminum plates for example. I have read a lot of positive posts about DynaCNC tables, but i'm not sure if any of them came from users of their table-top machines. Does anyone here have their table top machines? If yes, how do they perform? Please post your experiences...

K2 CNC has nice offerings, but i noticed some people had issues with build quality and their customer support. Have all the problems been solved?

I actually almost bought a 25x25 k2 cnc router on ebay yesterday but someone overbid me and it ended up selling for $3700 or so.

So right now i have the following options in front of me...

1 - Buy a new K2 CNC machine for around 6k (either 25x25 or 25x39)
2 - Buy a Dyna CNC TT2436ST that is similarly priced
3 - Buy a Dyna CNC table and try to upgrade it somehow
4 - Build a machine out of 8020 extrusions in the same way that STEVESPO and CNCFOAM did. If i'm able to get precisely cut extrusions, brackets, all fasteners, linear rails and ballscrews it would be fairly easy to assemble the router table. I would buy the router mount from K2 CNC but everything else shouldn't be that hard. The thing that I don't know yet is which servo motor/controller/power supply combination would work best in this case.
5 - Try to find a used K2 machine somewhere - I actually almost bought a 25x25 K2 machine yesterday on ebay for $3600 but someone overbid me by $130 or so.

Please let me know what you would recommend or what your opinions are about any of the commercial offerings.... I have no problems spending 6k on a k2 machine but I am a DIY guy at heart and i'm mechanically at heart so if i can build myself for considerably less than 6k i'd do it.

All your replies would be greatly appreciated. Thanks in advance!

[sploo]

I would suspect you could source parts quite cheaply, and if you're not costing your own time then yes, you could make a comparable machine for vastly less than $6k.

However... I'd note two things:

1) I recall seeing someone (probably on this site) noting that he'd spent much more overall than the actual cost of his machine, because of the trial and error process with designing bits and finding they don't work (this of course is less of an issue if you build to an existing design).

This is a bigger machine, and I've no personal experience of it, but what comments I've seen have been positive: http://mechmate.com

2) Do you want to be building a machine, or making stuff with it? That was the consideration that finally made me decide to buy not build.

[Menatep]

Thanks for your reply... the mechmate looks like an interesting setup but i prefer working with 8020 - it's easier to be more precise with that stuff, considering all the rais can be joined with prefabricated mounts. It seems there's less room for error.

I want to build things with the machine, but building it might not be a bad idea. I guess i will have to do a comparison of what's available and what the prices would be to see what's worth it.

Does anyone know where i can get ABBA or THK rails and ballscrews? Any other places other than ebay?

[Konstantin]

If I was in your shoes I would buy a commercial router, start making money with it. Learn about the machine I bought, decide about the design and what I would like to have different and then build the second machine.
I am a mechanical at heart too, building one for hobby is one thing, but if it is to be a primary source of income you may want to get the money income rolling before venturing on building it.

Also if you have a lathe and a mill and tools it speeds up the building process a lot.

I wonder if cutting 8020 precisely will you cut all the pieces in the same day ?
What is the coefficient of expansion of 8020 extrusion? If the temperature varies how length varies?



Konstantin.

[steve323]

I just purchased a DynaCNC TT2436B and am very happy with the quality. I am still assembling it so I cannot comment on cutting accuracy. The fit and finish of the components looks great. It appears to be constructed with very heavy material and tight tolerances.

The long rail (they call it the Y axis, I consider it to be the X axis) uses 1" X ~42" stainless rods with a 5/8 X 10TPI Acme thread. There are 2 motor mounts and 2 Acme threads, so you either need 2 steppers or build a chain drive to connect both sides. I am using 2 steppers and a 4 axis controller.

The footprint is 36" by 48" and the base is very heavy. I don't expect too much deflection on the rails or the base. I will have to report back later with real data.

You need to budget extra money for a 4 axis controller and a router. I am using a 4 axis HobbyCNC controller with 200oz-in steppers. Cost is around $300 plus another $100 for a transformer, case, and connectors, etc. A decent router will run another $100. Shipping cost for the main unit was nearly $500. The total price will end up around $2500. I don't consider it worthwhile to nearly double the price to add servos and ballscrews, but that is just my opinion.

[Menatep]

It's good to see some input and i'm very grateful for your replies guys...

Konstantin: When cutting 8020 extrusions i will be cutting them in pairs for making the frame. That way two pieces should be the same length and everything should be square. I'd also cut them in the same day with a quality dropsaw.

If i use 100 pound weight, LOAD-CENTERED deflection properties of a single 8020 extrusion are as follows:

model 1530 (1.5 x 3 in) 25 inch length: 0.001 IN / 0.029 mm to 0.004 IN / 0.110 mm depending on which side the weight is applied from.

model 1530 (1.5 x 3 in) 39 inch length: Deflection x: 0.005 IN / 0.115 mm to 0.017 IN / 0.430 mm depending on which side the weight is applied from.

K2 3925 model is made out of 1530 8020 material and it has 2 extrusions running for the lengths and 5 extrusions used as width braces,

Steve: Could you please post some pictures of your machine as you assemble it and once it's assembled maybe. I just got a quotation from DynaCNC 24x36 machine with ballscrews and servos for 3044$. Their electronics kit is $1395. If i go with their electronics i'd be at $4439, which is better than the website price for a similarly equipped machine. what worries me is that those unsupported rails might deflect under pressure, but then again Heiz High-Z model uses the same design and i've seen videos of that machine and it's not bad.

[steve323]

I will post some pictures of my partially built DynaCNC system early next week. It will probably be a few weeks before it will be completed.

[drmosh]

My opinion is that the idea of building your own machine, while seductive, ends up being way more expensive and time consuming than most imagine. If your hobby is CNC, then build it. If you're going to use CNC for your hobby, then buy the machine and get back to your hobby.

[Les the Luthier]

Hi,
I was able to find a Techno Isel 067 router for $2500 w/o controller. The 32"X36"X8" work envelope was what I had wanted. The controller I chose is a Camtronics 2000 servo controller which came to $1500. So at $4000 I've got the basic system. It already had ballscrews and servomotors. I was fortunate enough to talk to an engineer at Geco who patiently explained the issues confronting servomotor users. To be sure there were some incompatabilities such as the Z axis brake on the servo had to go because of the 1/2 second delay before the servos engage. The thought of starting from scratch with a pile of extrusions and a blueprint was simply daunting to me. I still have to connect up the home limit switches for X,Y and Z. Using Mach 3 I can set soft limits for end of travel on the machine. I wouldn't dream of building a machine until I have run this one for a long while. Most of us don't have the lathe and mill to fabricate precision parts.

[Menatep]

Yes indeed... the idea of building a cnc machine is seductive. This CNC thing is just a hobby for me but i do have some niche products in mind that can be made with a cnc router.

So for around 4k i could get a dynacnc 24x36 table.... it sounds like a more reasonable choice, but i'm not sure about those unsupported rails. decisions decisions decisions.

I never imagined it would be so tough to make a decision like this.

[NC Cams]

First you need to determine if this is a BUSINESS or HOBBY.

If it is a business, buy the machine because the lost time involved in tuning/developing/repaing/crafting on a home brew device cuts into your ability to earn a living. Period, Paragraph.

If it is a hobby endeavor, you can take the luxury or time to play with and learn how to do stuff and create/recreate and do-over to your heart's content.

However, having taken what was essentially a hobby and turned it into a small cottage industry, I can say from experience that the hobby can evolve into a business.

BUT, if you have the business going and need to deliver product to the client, they expect parts to be delivered on time and to-print. They could care less if the parts come off of your home brew machine or a commercial one that you bought and simply run. Besides, when the chips are down and parts are due, you do NOT need the added stress of trying to figure out why some home brewed machine that worked perfectly 20 minutes ago now refuses to work for goodness knows what reason.

Granted, professionaly built machines sometimes do that but at least they can be professionaly serviced from parts that SHOULD be available off the shelf as opposed to spot buys/special E-bay/never to be found again pricing/availabilities.

Only you know what you're REALLY trying to do/achieve. If it is a business venture, buy something that can be serviced with minimal hassle and/or crafting required.

If it is a hobby, let your imagination and creativity and tolerance for fiasco's be your guide. Even so, life can be too short to mess with some projects......

[RandK]

To be sure there were some incompatabilities such as the Z axis brake on the servo had to go because of the 1/2 second delay before the servos engage.

The Z brake is your friend it's a solonoid dead-man thing that keeps the spindle from crashing into the stock should something bad like a power failure happen. In the Techno controller the brake is disengaged (by applying juice)whenever their controller app is running on the PC and re-engaged (no juice) only when you exit the controller program; so it does nothing (no delay) while you are running a program. The controller keeps power to the motor to keep it in position while it is running.