Fail of the Week: Hackaday Writer’s First CNC Adventure

This Fail of the Week post focuses on a project from [Limkpin] aka [Mathieu Stephan], one of the Hackaday contributors. He wanted a CNC mill of his very own and decided to go with a kit that you assemble yourself. If it had been clear sailing we wouldn’t be talking about it here. Unfortunately he was met with a multitude of fails during his adventure. We’ll cover the highlights below.

At the top of his post he lists the features he was looking for in a mill. He wanted something that had good community support, could be used for milling aluminum, and worked with tight precision. He settled on the Probotix Fireball V90. The kit is shipped from the US, and an international (shipping) incident caused the first fail. The front panel on the controller was bashed in, which in turn damaged the electronics on the circuit board. What’s worse is that it took several weeks to get replacement parts because of the insurance claim process with the shipping agent.

Next came problems with connecting parts — to us it sounds like the issue was with Probotix’s stock of screws. One was too large for the milled hole in a connector, and another was too short to use with the three others that made up the set. Once replacement parts were shipped [Limpkin] thought he was finally in business. He was wrong.

With the rig finally assembled he realized the Z-axis had some play in it due to the machined holes for the X-axis bearings. The video after the break shows a close-up of the issue. After receiving a replacement he started testing and realized there was an alignment issue with the Z-axis ball screw.  Another clip below shows it misbehaving, resulting in a screech as the sled passes a certain point on the screw. It seems there was a bend somewhere that caused this. [Limpkin] did get around the issue by using a helicoidal coupler. This small success finally got his mill up and running.

So this Fail post actually ends in a successful CNC build running the sound-dampening box you see above. We thought about featuring this some time ago when it wasn’t working but [Limpkin] wanted to wait for the company to respond to his support requests. We’re glad we waited. The Fail of the Week series isn’t about trashing on creators of failed projects, so we don’t want to hear you bad-mouth Probotix. Instead we’d like to hear your stories about problems with your first CNC machine project. This example should drive home the message that whether you buy a kit or build from scratch, you need to be ready to deal with adversity the first time around.


2013-09-05-Hackaday-Fail-tips-tileFail of the Week is a Hackaday column which runs every Wednesday. Help keep the fun rolling by writing about your past failures and sending us a link to the story – or sending in links to fail write ups you find in your Internet travels.

39 thoughts on “Fail of the Week: Hackaday Writer’s First CNC Adventure

  1. so we don’t want to hear you bad-mouth Probotix.

    There’s really not much more we can add.

    To be fair, even the $500,000 machine we just got came with problems (the motors cut out, and the mfgr can’t figure out why)

  2. I build my first CNC from scratched with a design based on the largest linear rails I could find on ebay. It is a large gantry style setup. I originally wanted to mill wood, aluminum and maybe even steel. Unfortunately, there is way too much flex in my design, and twist in the gantry.

    There’s a reason ‘real’ CNC mills look like they do. Eventually I got a CNC Fusion Kit for my X2 clone, and am quite happy with it. I only use the big one (BFCNC3K) for wood, plastic or other materials that don’t need as much precision and rigidity.

    A common lesson for any CNC enthusiast.

  3. Not owning a CNC mill is my biggest fail, a failure of finances for fun toys like CNC mills and 3D printers, yes even the cheaper DIY ones. Even though I have a full time job with an employer everyone automatically assumes must pay well…

    However, I thought fail blog was supposed to be about personal fubars that cause your project to fail, something you did that in hindsight was a mistake that caused the fail? This is clearly a fail by probotix and the shipping company… But now I guess I know who to stay away from if I ever do get the funds for such a thing. Looks really nice though.

    I will say that using foam acoustic insulation like that is a poor idea that is going to cause a fail in the form of a fire some day… Even if it is the flame “retardant” stuff, it’s going to collect all kinds of nasty dust. I would suggest a box within a box with that stuff sandwiched in between, or some super fine steel filter mesh or someother thing that dust won’t be as prone to stick to for the inner layer if you think something hard like plywood or plexi will defeat the purpose of acoustic foam.

  4. I ordered a ShapeOko2 and I hope that the kit will not have the problems seen in Limkpin’s situation.

    I agree that with these failures you at least learn how NOT to do it, which is valuable but cost frustration and time.

    I also find it funny with the last quote on the page …”you will have 3 cnc’s”… 1) The learner 2) Unbalanced 3) Perfect

    I hope to merge 1) and 2) with the upgrades (spindle, motor and beefing up the frame) and have the next one be THE ONE.

    1. The ShapeOko2 will have it’s own set of problems. Still it works. It’s darn good fun. (I scratch built one 48″ x 24″).

      How about a software workflow Fail? Once you have the CNC you still need to process cad to cam. There are so many “it almost works but…” software options out there.

  5. Any company (and in this case – that means Probotix) that lets their customer eat the delays in Shipping claims is a FAIL all by itself.

    Customer paid money, Vendor shipped product, shipper damages product – THEN THE VENDOR SHIPS OUT A REPLACEMENT ASAP and worries about the shipping/insurance claim on their own time.

  6. I’m in the middle of a two year journey building my own CNC router and the entire process has it’s up and downs. For a few hundred more and far less time I could have just started with a kit such as the Fireball and had fun setting up the electronics and the software.

    TL;DR: Building precision machinery is hard.

    1. You need precise machinery(and knowledge to use it) to build the precise parts you need to build a precise machine. I have super respect for the early watch builders.

  7. “an international (shipping) incident”

    *sigh* Was it coming from China? I’ve received a lot of parts from them wrapped in nothing but bubble wrap, some of which break apart in traveling halfway across the world (the parts AND the packaging). Some items have shown up DOA, some with broken pins/plastic/etc.

    You get what you pay for, I guess.

      1. Yes, really.

        In an older life I sent similarly-sized items internationally to a large number of places (Spain, India, Korea, Brunei, even here to the US) from Australia.

        Despite explicit warnings and even premiums paid for “special handling” on items sent via Air Freight, things got damaged, went missing (India was always a fun place to send things) and G-shock stickers visible, nothing was safe. I never had a shipment without a claim.

        You have to wonder when a 100G sticker comes back-not just “activated’, but SOLID RED on a cardboard box INSIDE a wooden crate and surrounded by foam packing…

  8. How normal is it for these kits to use COTS tools? Looking at their $3,200 fireball comet, it uses some sort of DeWalt tool to do the actual cutting. And are any of these probotix CNC routers really rated for aluminum? All the pictures on their site show it working wood.

    1. A commodity router has a lot of advantages. The mass production of that part greatly lowers the cost, and you can choose a router that suits your specific needs. And when you are ready you can upgade to a more purpose built spindle, maybe even liquid cooled.

      You can easily spend a thousand on a used spindle, and thousands on a good spingle. A good spindle will allow you to run off the shelf collets.

      So a router for $60-$200 is a great deal. The Bosch units are a decent choice.

    2. I recently purchased a Fireball Comet. I’m running the Porter Cable 892 router, with low TIR collets and collet nut purchased separately. I had been looking into building my own, and this machine is so close to exactly what I was designing it’s amazing. It’s a good solid machine. I’m using it right now for wood and acrylic, but will be running aluminum and copper on it eventually. I have no worry about it not being capable of these materials. I merely need to get the proper tooling and fixturing in place so I can do it.

      1. Copper will be a “fun” metal to mill. It’s very soft and grabby. It work hardens from deformation and gets soft when heated and cooled rapidly.

  9. Look at it this way every problem is an opportunity to learn something new. One thing I hope this person learns is there are a lot of CNC machine designs but no one CNC machine that does everything.

  10. First of all, I have had the Probotix Fireball V90 for about two years now and love it. Probotix, on several occasions, has gone out of their way to support their product–most recently, they put together a special refurb kit special for me to replace the aging frame; and they gave it to me at their cost. I have pushed this machine way beyond its intended use of wood and plastic. I use it almost every day for plastic, graphite, ceramics, aluminum, copper, tin, nickel and even limited 304ss (under 10mil). I have been generally surprised how well this hobby mill with a Bosch colt router has done.

    Though I will have to say, when I first got it I had several fails of my own. First, wiring up limit switches was a bit of a headache. They seem pretty straight forward, but I had issues with my chosen limit switch of falling apart while being used and the machine not stopping. So, if you wire your own limit switches, be aware that most of them separate.

    Second, learning to zero Z got me for awhile. While first using the machine, I gouged my spoil board way more times that I can remember.

    Speeds were also a fail at first (and I’ll throw in cooling, when dealing with metal). I went through hundreds of bits learning about feed rates. You can read up on the subject all you want, but there is nothing like experience when it comes to feed rates. Two years later, I almost never break a bit.

    I bought a kit, but even if you build it yourself or get a higher end unit, just as in any kind of machining there is a lot to learn and no substitute for experience. CNC is almost as much of an art as science.

    1. I ran into the same learning experiences. My spoil board has many “battle wounds” from a wonky spindle that thinks it’s higher than it actually is.

      That being said, I ordered their monster cnc electronics kit and have had a excellent experience with them. Not knowing everything, I asked questions on voltage and amperages and they were super helpful! Even swapped out a driver board that failed on me under warranty. I’d recommend them! (I live in the states though…)

  11. I have the excellent opportunity that I got to learn on the big toys. I have a Haas VF4 and a G&L600 bore very large machines. Both are older machines and have massive capabilities. The fail of my machines is maintenance. The machines are older units and have a lot of bells and whistles, however, people didn’t use them properly and never did the maintenance on them. So I have to replace all the oils, the greases, the ball screws and things that you never change unless the machines have been abused.

    Luckily, performing the maintenance activities and replacing and rebuilding the broken parts means I have two very high end machines costing less than what a new or totally refurbished unit would cost.

  12. I know the guy who actually developed the V90 and he sold the rights to it since he got too swamped. I own the guys Microcarve A4 machine and love it. The V90 is a great general purpose machine. If I listed all my CNC fails (and some success) you’d have like 4 HAD articles on it.

  13. Very useful post. I am in the market for a CNC that can engrave small aluminum end panels and make a few DB9 cut outs. I haven’t seen this machine before and, despite the troubles, I would certainly consider this one. Are there are any suggestions for alternate inexpensive-ish machines?

      1. The cheap chinese(3040 etc) models are great for aluminium, I’m not a fan of the shapeoko designs. The chinese machine really work well, only the electronics let it down.

        1. Thanks Phil, that’s a good idea. Go through some portion of the learning curve on the el-cheapo, upgrade components after I start to get a clu.

  14. I own a Probotix Meteor and love it. It came fully assembled; just plug it up and go. Only problem I had was a linear bearing went bad after about 4 months. I called them and had 2 new bearings setting on my doorstep the next day. My first CNC was scratch built and had several problems. Some of them were alignment issues, sagging linear rails, Stepper driver problems, and speed issues. Glad to see [Limkpin] got his working.

  15. Those foam panels are not designed to dampen sound, but to reduce echo, that’s why they’re used on recording studios. They often add real sound dampening panels on the back to insulate sound. There are some panels that are a mix of both, but using them on a cnc machine has no point, since you don’t need to reduce the echo there.

  16. I have a 23 year old Acra knee mill with a 10×50 inch table. I’m working on doing a CNC refit. Not a retrofit, it was originally CNC. The 1990 vintage system was partially broken and partially missing so I took it all off.

    The big old 140 volt brush type DC servos are being replaced by smaller, lighter, more powerful stepper motors and an ordinary PC will replace the proprietary Anilam Crusader M control system. Sold all the good parts, still have two of the servo amp boards that for some reason I can’t even give away for $50 each. Yes, that’s a give away price, refurbs on these are nearly $500 and you don’t want to know what the manufacturer charges to blow the dust off their antique design to build a new one.

    The only original electric components that will remain are the 3HP Baldor spindle motor, the glass scales on X and y and the rotary encoder on the Z axis.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s