CNC Router Question

As I was poking around the the internet the other day I found a CNC router called the Workbee. It comes as a kit you put together and for the 1500mm x 1500mm version it’s only around $1500, to me seems too good to be true. Has anyone ever heard of this or better yet had personal experience with it?

Doesn’t look anywhere close to powerful enough to cut aluminum.

As for prototyping, it may have some uses but overall doesn’t seem very powerful.

What wouldn’t be powerful enough, the spindle?

Did not see axis stepper specifications listed.
Look like 150 oz in NEMA 23 steppers.

So it is like a bridge mill version of a MaxNC milling machine with a much more powerful, much higher RPM router for a spindle.

Deflection due to low rigidity will likely be an issue.

The 16k-20k+ spindle speed will make things less than idle for speeds and feeds even with an O-flute cutter.

The rigidity and non-ideal spindle speed will slow your IPM of cutting down…but steppers have more torque at low RPM anyway.

The steppers do not have encoders, so if you miss a step or have microstepping issues you won’t notice till you measure probably after removing the work piece.

It does not say if it supports microstepping, so what is the repeatable accuracy and the travel per rotation via the belts?

Update 1
Dug around more, I was very close:
Motors - NEMA23 175oz Stepper Motors (X, Y & Z)
Accuracy - 0.05 - 0.10mm (Screw Driven) Or 0.10 - 0.20mm (Belt Driven)

I have many 1.8 degree steppers exactly like those - I question the validity of that 175oz value. In practice it will be closer to the 150oz I mentioned above.

Update 2
If you plan on pushing that router in aluminum plan on using the screw version of the gantry, not the GT belt version. At first I thought they meant they put belts on the stepper to act like gearing, no, they use the belts like a rack and pinion (think 3D printer).

I notice they don’t offer a completely screw driven version at 1500mm x 1500mm. Just at 1000mm x 1000mm.

Update 3
The lead nuts are designed so one can adjust the backlash, but they are not spring loaded anti-backlash nuts. I would expect some adjustments as things wear in.

Update 4
The 1000mm x 1000mm is $1,664.77 at 1 British pound = 1.39 U.S. dollar and do not forget VAT which could get you close to $2k.

Update 5
A MaxNC 15, Sherline, Taig mill will cost $3k+ CNC without 4th axis or tapping unit.
So even at $2k that is not horrible…but you are in the barely adequate spot for general aluminum work.
Sure you can make it work, but for a little more, you can get something with less tinkering required.
With a small spindle make an adequate engraver.

For wood and softer it would be fine.
It would have trouble with aluminum, as the design lacks some basic rigidity. Motors can deliver plenty of power, as can the spindle, but with unpredictable deflections of several thousandths - aluminum is tough - you might be disappointed.

As a comparison, find an image of a KRMx02. I have one and it is just rigid enough for aluminum, but a lot of it or very fast.

832 has an EZ-Router that’s about 30 x 50, it cost $20k, and is sufficiently rigid.

I’m still hoping to get into a little deeper discussion (video series) this summer on CNC routers, the basics behind them, and how to “cut aluminum”. It seems to be a common question that is not well answered by manufacturers, primarily because what we ask these machines to do for FRC really isn’t what the average industrial user asks of them.

First, even your cheapest CNC router will “cut aluminum”. Stick a sharp bit on it, send it at the material, and it will start to cut. You might even find a few Youtube videos of the router “cutting aluminum”. What we want to know is whether it can do it with a reasonable consumable lifetime, rate of cutting, and accuracy that is acceptable for our application.

For most FRC applications, cutting aluminum refers to cutting something between a solid sheet of ~1/16" to 1/4" thick aluminum sheet, or with the correct jig, possibly working on aluminum tubing. We really don’t need to make parts particularly fast. I could probably make a robot’s worth of parts on our mill in 2-3 hours, and if I could pay 10x the cost to make them 10x faster, it really doesn’t help me. The gantry isn’t the limiting factor in our design process. I also wouldn’t want a gantry mill that is 10x cheaper and takes 20-30 hours of babysitting to make parts. Finally, my absolute accuracy needed can range from +/- 0.020" or so for most of our gussets and things like that, but need to be tightened for some applications considerably. This is different than “positioning accuracy” many manufacturers will quote, which is basically how accurately the machine can position the bit in free space, not necessarily the accuracy of the part from a machine cutting under load.

When you cut aluminum, the material pushes back. This push back is proportional (roughly) to the power you are putting into the cut, which is a function of the material itself, and how much material you are removing (how big of a bit, deep of a cut, etc). This pushback causes EVERYTHING to deflect. Your bit bends slightly, your connection from the spindle to the Z axis bends, the Z axis to the Y axis flexes on its bearings/guides, the gantry might twist slightly versus the X axis, etc.

This is the first effect you will see when you say a machine “can cut aluminum”. For a given bit, you will be taking a feed of something like 0.001" to 0.010" per tooth on the cutting edge. If your deflection at the bit, for the cut you are trying to take is relatively large compared to that number, your bit might not actually cut, but instead rub and get pushed back. Then on the next go around, the machine has moved it forward, and it has to take “two cuts” of material in a single bite. This action is called chatter, and will rapidly destroy the bit due to the heat generated, even if it cuts at first.

Some times people try to reduce this cutting forces by taking an exceptionally light cut (not very deep). This can work, but with a few caveats. We usually use something called a “single flute, upcut bit” for CNC routers, because they solve several problems with our spindles/routers spinning fast and the machine having low rigidity, as well as how to get chips out of a slot (generally not recommended for machining, but common in our applications). However, these bits generate an upward force on the material, and unless it is held down very well, it will vibrate against your bit, cause heat buildup and fail. This problem is worse with thin material (sheet aluminum). Additionally, your bed might be warped, warp over time with cutting… your material might not be lying perfectly flat on it, etc. If you depend on taking an extremely light pass to make the cut, and the bed is warped, you might be cutting air in some areas and overcutting others.

So usually when we say something can “cut aluminum well”, we mean first that the machine has enough rigidity (which is usually proportional to weight for a well designed machine), for its given work envelope, to take a reasonably deep cut, with a reasonable cutting rate, without chattering.

As a comparison point, I’ve been running the CNC Router Parts Pro Series (~$6k all in for a 48"x48") for several years now. It will run a 3/16" bit at 20-30IPM through 0.08" aluminum in a single pass, provided the material is well secured and I run mist coolant. This is about as cheap as I’ve seen that can do this, and it certainly didn’t work out of the box as well as I would have hoped. I am purchasing another one of these for my home shop this summer, as I still think it is the best value for my application. If I had a few more bucks to spend or wasn’t willing to mess with dialing it in, I’d go with the Velox.

More to come :wink:

As an aside:

One day I want to revive my project to make an FRC router for aluminum that is made with mostly FRC parts and driven by CIM motors. Would need a car battery to really get it working but would be a fun hybrid.

FRC robot building an FRC robot.

Looks almost identical to a Shapeoko 3. S3s can cut aluminum consistently, if slowly. They’re also cheaper than whatever knockoff this is.
That being said, just spend 3k and get an Omio X8. I don’t think I’ve met anyone that regrets that decision yet. Save up if you need to.

Looks almost identical to a Shapeoko 3. S3s can cut aluminum consistently, if slowly. They’re also cheaper than whatever knockoff this is.

I haven’t had a chance to work with a Shapeoko and would be happy to get some more information on them. As a comparison though, I looked for a couple videos of them “cutting aluminum”.

In this one, you can see in the first minute or so, the chips that are being formed are really just slivers and not actual chips. At ~1:20 or so, you can see the quality of the cut on the side walls the bit was traveling in, and how much play is in the machine just by all the facets in the cut surface.

Here is another by the manufacturer claiming to “cut aluminum like a boss”.
The bit is squeeling the entire time, and while granted it is a poor bit choice and recipe, you can hear the motor rpms dropping as it trying to ramp into the part, clearly underpowered for the cut.

Here is a shot of our mill doing a ~20IPM recipe, 3/16" bit, 0.08 aluminum.
The sound of the compressed air and spindle motor is actually louder than the cut. The cut edge is a mirror finish. That in addition to the sound indicates extremely little vibration in the cut. The chips formed are a true chip.

Whether or not the Shapeoko’s quality of cut is sufficient for an FRC team is a whole different question, but just a little basis of comparison from some quick Googling and personal experience.

Large parts of this robot (excepting the bellypan and claw) were made on a Shapeoko 3.

Yes,it should be able to mill aluminum. We built a CNC router similar to this. Ours is built stronger and does have NEMA 24 steppers, and uses belts for the x and y axis. Still playing around with feeds and speeds. Was doing 0.4mm DOC @ 1200mm/min. Lately have been slowing router RPM and have been pushing and 1/8 bit 1.3mm DOC @ 500 mm/min. All of our gear box plates and various mounting bracket so we’re cut on it. Gearbox turned out so smooth, feels like butter in motion. Link to video milling a gearbox plate

The issue is really not that you can’t cut aluminum.

The issue is if you pull on belts like that long enough they will expand and you’ll loose accuracy…eventually they’ll snap. In a continuous loop driven perpendicular to the cogged pulley the expansion of the belt simply risks skipping on the cog…not elongating the rack surface. I have a steel rack and steel pinion on one of my custom gantry mills…the steel rack bolted to 2" steel tube is not going to get longer pushing on aluminum, even 7075 aluminum.

The wear and tear on the cutter using it against good practice is also being ignored when you do this. A factor most people care about because these O-flute up-cut bits can cost $10-$25 each. These machines rarely have collets like R8/MT tools have; so at every manual tool change you have to reset your tool depth and the wear and tear will force more tool changes.

Will it cut is secondary to will it cut well or wisely. Perhaps saving a few thousand now is worth investing in more belts, bits and extra material (cause you won’t notice the failures till you ruin the work) later. Putting in the quality time to keep things working nicely. Just remember though: I have the experience to work the problems till they are not issues…how many people on your team will have the patience to get experienced like that?

It is a magic triangle issue: cost/time/quality.
I am of the opinion that a Tormach/Haas is probably overkill for just some aluminum FRC parts and I was FRC11’s CNC mentor on the Haas TM1P.

However the capacity of a ShopBot or Velox setup like this is quite a bit more than this tool that would be a very nice engraver. To put some of this in perspective - why would you use belts when you can now get cheap (<$30) Acme lead screw/lead nuts like this:
Even a few years ago an Acme screw like this might cost $100 but not now:

Okay maybe your design will travel more than 915mm/36" (mount on lead nut reduces overall travel). How often will you do that and would maybe putting some dowel pin holes into your design, then sliding the work, be even more effective? Cause a 48" by 48" table is nice but also remember you just lost 16 sq ft of your shop to it. Are you going to spend 98% of that machine use in 3 sq ft of that 16 sq ft? If so, is it worth it? Of course the space in a professional shop where you pay for square footage may not be relevant if your school gives you a lot of room you don’t pay for. You can always rough cut large pieces of aluminum on a table saw or band saw to size if the concern is buying 48" x 48" pieces of aluminim. At least the carbide blade or band will cost less in the end and you then end up with 2 tools in that square footage you may be able to run at the same time. Plus pushing that sheet through a circular saw will be way faster than dragging a light pass of the router to simply sever the sheet.

I know that this thread is about a cheaper machine, but my team has a Shopbot Desktop D2418 and it is a great machine. I recommend it to anyone who is searching for a CNC machine for FRC.

Thank you everyone for the replies, now I have a bit better idea of what I’m looking for.