Shop OBEX P1 Docs P2 Docs Learn Events
Looking for any experiences or recommendations of a desktop CNC milling machine — Parallax Forums

Looking for any experiences or recommendations of a desktop CNC milling machine

I would use it mainly for milling on extruded aluminum enclosures. In particular for cutting out area in cover for LCD display and slits on ends for switch or connectors. Aluminum is approx 1.4 to 1.8 mm thick.
I was looking at some inexpensive models on Amazon (< $500) . They say they will mill aluminum but I wonder how many passes it would take to cut through. I'm not against spending up to $2k if it would buy something worthwhile.
The work area I may need could be 12"x12"x4" for example. So I would like to hear from those that may have something like this, use it, recommendations and any pros and cons.

Thanks. Don

Comments

  • Ken GraceyKen Gracey Posts: 7,386
    edited 2020-02-24 06:08
    We have three Haas CNC mills at Parallax and prior to the purchase of these full-sized machines, we owned several desktop CNC machines. These smaller desktop machines were always a compromise - hardly useful for much other than prototypes. No coolant, no tool changer, junky tool holding systems, no support. I can provide many cons to the idea. I hate to discourage you, so hopefully, somebody in favor can tell you the pros side of this argument.

    Unless you really want to get into this as a hobby, I wouldn't buy any CNC machines for $2,000 or less. I wouldn't get started with this kind of machining for less than $10K. But spend $25K for a good used one and the machine capability increases 10-50x over a $10K CNC milling machine - value ratio goes up really fast the more you spend, far from a linear relationship. Flood coolant, tool changers and enclosures are absolutely necessary for cutting aluminum. If you don't like anything I write and MUST proceed, start here https://www.tormach.com/ with $15K in hand.

    .060 enclosure aluminum is really too thin to mill effectively and I think waterjets and CNC punches would be a better way to go.

    The software startup has a learning curve, from idea to 2.5D CAD, to CAM, to the part. Before I brought CNC into Parallax I bought a Haas CNC Mini Mill of eBay and moved it to my garage, where I ran it for eight months. When I finally understood how to run the machine (this was prior to MakerSpace-type training) I started making our 12V Motor Mount and Wheel Kit products. Eventually, we moved the Haas to Parallax and put it into production, buying two more machines.

    Here's what I would do after reading about your goal. Identify the main commonalities you want in your enclosure every time: LCD cutouts, mounts for the board, and USB cable, power supply, etc. Make this "flattened" design (let them figure out the box folding and radii required) to a CNC shop locally with a punch, brake, and have them make 20 of your design. They may prefer to waterjet the cutouts. They'll charge you a setup cost of $300 and maybe $20/unit in these volumes. It'll seem like a big investment but it's so much less expensive than trying to do it on your own. Then, further modify your standard enclosure with a drill press each time you want to add something: 1/4" button mounts, potentiometer holes, etc.

    There's an economic argument of efficiency in favor of the shop (their big powerful tools work fast and they know how to set up processes and fixtures) and an opportunity cost you'll incur if you were to do this yourself (your expertise is in the products, coding, customers, design). When we all do what we're good at doing, things get done as effectively as possible and the highest level of success may be found. It may not be as fun as doing it yourself.

    One more reason to consider outsourcing. When it comes to buying the bigger tools, like the Haas machines, I think that 99% of these are sold on leases. Hmmmmm. Their prices are somewhere between $50-200K for a CNC. Many of the purchases are made by small shops or individuals who have a fantasy of making parts - like Orange County Choppers. Often, these people are more creative than business-minded. The machines are sold with little collateral or credit check. They hold their value and they're easy to repossess. New shops have challenges valuing their time in the beginning. Their machines might often sit idle, too, costing (vs making) money. My personal/professional view is that starting up a CNC shop can have an unfavorable ratio of fantasy/reality for the owner, leaving "spindle time" open for people like yourself. Help out one of the MANY shops in your Midwest location and save yourself a lot of hassle along the way.

    Now, there is also a reason or ten to buy a CNC machine and do this on your own. If you start to obtain some volume in your CNC parts you might want more control over the outcomes. You may want to tweak the design frequently, between iterations. Our Gripper is an example - the first two iterations (going back 10+ yrs) were made outside and some parts came from China. Well, you know how that works out now - tariffs, coronavirus, high minimum order quantities, etc. These all add up to time and money that we don't want to spend. So, now, we're making our first 100 Grippers inside Parallax. And the next 100, and the next 100. And, as long as I work at Parallax we shall NEVER outsource another Gripper. We'll make them on our own CNC machines.

    Party pooper *out*.....

    Ken Gracey

  • When your Haas machines run, do they run unattended? I ask because I don't know what the industry practice is, if this is proper.

    If these machines need to be attended anyways, then you might as well do lubrication and chip overload removal manually, obviating the need for coolant. (There should be zero heat from milling the enclosure window)

    I would spend just enough money to get something that doesn't use a 'Dremel Moto tool' for a milling head, because I assume the bit wouldn't last for even one enclosure. Hopefully would only cost a few hundred $

    Does such a thing exist? So far all I've seen for these cheap CNC machines online are lame 'Moto tool' setups, even going so far as securing it with hose clamps.

  • Have you thought about using a plasma cutter?

    Mike
  • The_Master wrote: »
    When your Haas machines run, do they run unattended? I ask because I don't know what the industry practice is, if this is proper.

    It's an excellent question. Once you've optimized tool paths and made a few parts, these machines run unattended. The operator is usually close by, listening for tool changes and has the CNC's heartbeat in their head - but it's not necessary to intervene unless there's a tool break, in which case the spindle feedback sensor detects the absence of a tool and it shuts down.

    On the low-cost machines, the operator is certainly standing by, removing chips, blowing coolant, hoping it all works out. It can be a rough ride!

    Ken Gracey
  • Don MDon M Posts: 1,647
    @"Ken Gracey" Thanks for some great insight. I'm wasn't thinking of doing any machine work other than some quick prototypes for myself. I've been using enclosures from Hammond Manufacturing and Context Engineering in some of my designs. I like the durability and feel of these enclosures. I've had some of this work done locally in a very nice machine shop but unfortunately he has retired. I've looked at Front Panel Express to farm some of this work out to and Context also offers services to do the same with their products. I was looking at the cost of sending it out and trying to amortize it into owning something small on my own to do when needed. I had already thought about the learning curve and would agree with you that my time is better spent in the design and programming.

    Here are some photos to show some of the enclosures I've had made. I also had a local firm do laser marking. They use a YAG laser and the results are great with nice sharp white letters. I've also had another local firm try laser marking with their CO2 laser and the results were disappointing as the letters were not crisp and bright. This process is not cheap either.

    So in moving forward I'll probably stick to farming out the work as it makes the most sense.
    500 x 329 - 250K
    4032 x 2268 - 3M
  • Don MDon M Posts: 1,647
    @"Ken Gracey" Btw... these devices do use the Propeller. I have another device on the drawing board now.
  • You mentioned using Hammond enclosures and I remembered they do custom machining

    Here's an extract from page 3 of the catalog

    Some years ago we did some tests with laser marking of these, from memory they used an Epilog engraving laser. The contrast achieved depended on the anodizing color





    924 x 426 - 250K
  • Tubular wrote: »
    You mentioned using Hammond enclosures and I remembered they do custom machining

    Here's an extract from page 3 of the catalog

    Some years ago we did some tests with laser marking of these, from memory they used an Epilog engraving laser. The contrast achieved depended on the anodizing color

    If this is an option I'd take it over ANY of the alternatives mentioned above.

    Ken

  • Ken

    Very nice input :)
  • MicksterMickster Posts: 2,588
    edited 2020-02-25 07:16
    @"Ken Gracey"

    Excellent advice!

    IIRC, I had a HL-4 and a VM10(?). At the time, my VM10 (or was it VMC10?) was the biggest in MI and we'd load it up and let it run all night.

    Do it right or farm the work out.

    My Android tablet enclosures are carved out of billet aluminum and turn out like a piece of jewellery.

    IMG-20190526-WA0002.jpg
    1280 x 700 - 111K
  • When I started out making parts I had zero experience with Cnc and nobody to explain anything. Just searching online I found the MaxNC 10. It was around 2500. A closed loop stepper system that I replaced with my own stepper drivers and tosses the encoders. I used Mach3 and loved it. Bobcad was and still is the CADCAM, cheap, but works. I used to cut 1” thick aluminum on this thing. Ran it for years before changing to a new 24”x36” machine. Cutting .06” aluminum is trivial. Even with no lubricant just run slow RPM and make a few passes around .02” per pass at 20IPS and it won’t overheat and bind up. You can make a jig with holes to mount your part, and run these little machines all day long. You’ll wear out the Delrin acme thread nuts often but who cares. I use to make my own and just swap out often once the backlash gets bad. I have a $10k CamMaster Stinger now with WinCnc. Very reliable. not really made for aluminum but I run 1/8 and 1/4 aluminum regularly at slow speeds. A real work horse. Just hit the material with WD40 white lithium and leave it alone for sometimes 1+ hour runs. never an issue.


    Pic is the old MaxNC. Other table tops are Taig etc.

    PS. It doesn’t matter if you can run it in once pass unless you are running big production and time is a factor. Use a spiral cut 1/8” aluminum router bit. easy stuff. But $500? Hard to do for something new.
  • W9GFOW9GFO Posts: 4,010
    Ken Gracey wrote: »

    .060 enclosure aluminum is really too thin to mill effectively and I think waterjets and CNC punches would be a better way to go.

    This is my only point of disagreement with the "Party pooper". :-) In the context of milling cutouts in a small piece of 1.8mm aluminum a mini CNC mill can do a great job. The edges will be crisp and clean and have none of the distortion or striations that you get with punches or water jets. However if the idea is to put in a large sheet of aluminum and then mill out a dozen panels at a time, then yeah, milling is not the way to go. But for small runs every so often, a (good) mini mill would do a great job. Making a good jig is the key.

    I bought a Sherline mill 20 years ago (desktop). I've made all kinds of things on it, like an injector nozzle for a 800 lb thrust Lox/kerosene rocket engine, a Sterling cycle engine, pulleys for toothed belts and gears (using a rotary table that I converted to CNC) and lots of prototype things. I've milled steel, stainless steel, brass, acrylic, Delrin, plywood, plastic, fiberglass, graphite and a whole lot of aluminum; from thin sheets to 3" thick billets. Occasionally I have done small production runs, I don't remember how many exactly at a time, twenty to fifty probably.

    The nice thing about it is that it is small, it takes up about a quarter of the space on top of my tool box, yet with some creativity you can work on parts that are larger than you might expect possible. I think it would be ideal for milling small aluminum enclosures. You can also get an accessory kit to swap out the pulleys so that the spindle rpm is much higher. I actually have that kit but have not yet installed it. I got it for milling PCBs, it would also work well for milling cutouts in thin aluminum using small diameter end mills, 1.8mm in one pass easily. The X,Y range is only about 5" x 8" though, that may eliminate it as a choice for you - but it looks like it would work for the enclosures you have shown.

    With the exception of engraving some vector graphics (and PCB traces) absolutely everything ever cut using this mill has been hand coded (or cut manually by spinning the knobs). G-code is pretty darned simple, and for cutting out shapes like for enclosures it makes no sense to me to use a CAD program. In other words, the learning curve is quite shallow. Even when using the fourth axis I will enter the G-code manually, and I am a mediocre coder - at best.

    If I had to do it over, I don't think that I would get a bigger mill. Certainly it would be really nice to have a large mill but space is limited, for what I do I would rather use the space for a CNC router/plasma table. A well built small router table will easily cut aluminum, might need two passes but it would work well and would have a lot of room.

  • W9GFO wrote:
    I bought a Sherline mill 20 years ago (desktop).

    I bought one of those in 1979 (with a Sears Craftsman label on it). I took off the cranks and added steppers with gearing. The objective was to mill molds for lead fishing jigs (Point Wilson DartTM). It was driven by a TRS-80 whose printer port connected to an array of 12 Opto22 solid state relays that drove the steppers. Every little stepper motor step was programmed in RS (Microsoft) BASIC and Z-80 assembler. No CAD, no CAM, no G-code. Frankly, the mill was a total PITA. Between the slide gibs and the drive backlash (which varied, depending upon where along the 1/4-20 lead screws you were milling), it was a constant source of maladjustment. But I was able to make the molds and help bring a still thriving tackle company to life (although I lasted only two years in it).

    My only bit of advice is to opt for a mill with closed-loop servos and ball-screw drives. And go for recirculating linear ball bearings on a round slide. No steppers, acme screws or dove-tail slides.

    Subsequent to retiring the Sherline mill long, long ago, I purchased a used proLight PL2000, which met my above criteria and uses G-codes. I was still contracting to the lure company for molds, so I had to convert my RS BASIC programs to QuickBasic on a PC and have them output G-codes instead of stepper commands.

    But you know what? I hardly ever use it since getting a laser cutter. Granted the laser cutter can't cut metal, but it's been a thousand times more useful than the mill. (And I don't make fishing lure molds anymore.)

    -Phil
  • Backlash PhiPi! Forgot about that little demon that rears his ugly head in every desktop CNC.

    I'm glad Rich provided a counter-perspective to the self-proclaimed party-pooper. I'll add to Rich's valuable comments the fact that he's very skilled, diligent, methodical and experienced with machines and materials.

    Ken Gracey
  • W9GFOW9GFO Posts: 4,010
    I bought one of those in 1979 (with a Sears Craftsman label on it).
    Yeah, backlash... I guess I'm used to it, and maybe in the two decades between our purchases some improvement was made. Ball screws would be awesome, I don't think closed loop would help much - but sure wouldn't hurt. And I do use the mill a lot less since getting the laser and 3d printers.

    Is there a mini mill available that uses ball screws and linear ball bearings?

    Ken, you're too kind...


  • I'm not a mechanical engineer but one of my clients asked me to come up with a custom enclosure they could get in small runs. At the suggestion of a friend, we went semi-custom with www.polycase.com. We found a case that worked and I sent them drawings for milling each side, and separate drawings for pad printing. It turned out really well.
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2020-02-26 23:55
    I've always maintained that you should design boards around stock enclosures and not the other way around. It's easy to make a custom board fit an existing enclosure. It's expensive to make a custom enclosure fit an already-designed board. So start with the enclosure, then design a board to fit -- with the mounting holes in the right places.

    -Phil

  • I've always maintained that you should design boards around stock enclosures and not the other way around. It's easy to make a custom board fit an existing enclosure. It's expensive to make a custom enclosure fit an already-designed board. So start with the enclosure, then design a board to fit -- with the mounting holes in the right places.

    -Phil

    An idea I learned from you years ago that has paid off for me. :)
  • Hi
    Perhaps a bit off topic but-
    For home use- not professional- a very serviceable enclosure can be made from cheap electrical conduit square section type with the snap on lid. It comes by the meter from diy stores. It cuts and drills easily and the ends are glued with plumbers waste pipe glue.
    Its amazing what can be pressed into service if absolute professional finish is not required.
    Dave
  • Great post Ken.

    .06 Alum can also be routed.
  • Tracy AllenTracy Allen Posts: 6,656
    edited 2020-02-27 19:17
    I too have had a Sherline mill for a little over 5 years, and use it mainly for slots, cutouts and arrays of holes in plastic (ABS or PC) off the shelf enclosures. That is for prototyping or making small numbers or modifications. It's been very useful for that. In one case, I used it to correct an error in the ground plane of a run of 200 circuit boards, cutting away a couple of areas of copper. Saved my butt and the time and expense of a re-run. For prototyping is a great help in aligning things like board-mounted connectors and SD cards with the cutouts in the box. That can save a lot of back and forth when sending the job out for production.

    I bought the baseline model and added aftermarket steppers and power drivers, àla carte. It was a long but educational process. Especially dealing with z-axis slip. Luckily at the time I had a young employee who was hugely interested in it and all the fine tuning and liked to program directly in G-code. I use it with Mach 4 software. The caveat is that the accessories and special tooling, should you need them or crave them can run into bucks and more and more bucks. A lot you can improvise yourself, but setup is time consuming, rewarding if you have the inclination. Mach 4 is okay, but there is a constant back and forth of issues between their patches and Windows versions. I hardly touch its capabilities. It is all simple stuff that can stilll be written directly in G-codes.

    I do agree that production is best left to the oems, who have the right equipment and skills. I've been happy with quality jobs done by Serpac, Phoenix Mechano , and Vynco. I have a job going with Serpac now, using their H-67 ABS enclosure with 9 custom cutouts and a UV printed logo, and @qty100 the box itself is $5 each, and the customization adds $10 each.



  • W9GFOW9GFO Posts: 4,010
    Regarding Mach 4, I have been using Mach 2 (Could be version 3 but I don’t think so) on an old dedicated, isolated computer. For what I do I have never exceeded the limits of the demo version, and it never gets updated so those related issues are avoided.

    Initially I used TurboCNC, which worked just fine but I prefer Mach.
  • W9GFO wrote: »
    I bought one of those in 1979 (with a Sears Craftsman label on it).
    Yeah, backlash... I guess I'm used to it, and maybe in the two decades between our purchases some improvement was made. Ball screws would be awesome, I don't think closed loop would help much - but sure wouldn't hurt. And I do use the mill a lot less since getting the laser and 3d printers.

    Is there a mini mill available that uses ball screws and linear ball bearings?

    Ken, you're too kind...


    Dual-Loop Feedback
  • As an integrator, my problem is that my end-products need to look like "bought ones". I am ripping out Siemens and Beckhoff boxes, it wouldn't look too good if I replaced them with project boxes that scream "some guy made this in his shed".

    For this reason I've gone with the:
    Tibbo Project System

    This is an awesome concept IMO. One starts with a motherboard and adds peripheral modules to suit. The only thing I don't care for is their programming languages as they leave a lot to be desired.

    Tibbo are happy to simply supply just the enclosure, along with packaging.

    I am not a board designer or a Prop-based motherboard would exist.

    Maybe I could contract this out.

    Some of us want to simply write code and ship product.
Sign In or Register to comment.