This machine, AMN-6040-M CNC Router: Mach 3 or Mach 4? (etc)

Hi All,

First post, newb to manufacturing engineering, etc. so be gentile...

I'm starting a business (think very, very large drones) and need to build cost-effective prototypes and maybe progress to initial light production. I'm a design engineer and new to the manufacturing side, but since I can't seem to get parts made at a cost I can afford I've decided to make them myself. So I'm about to order my first CNC machine. Sorry, this 'new machine', but 'buy' not 'build' - I simply don't have time no matter how much fun that would be.

I need to mill semi-complex parts out of 6061-T6 and I think I've settled on the AMN-6040-M CNC Router from Alpha CNC with a few options. This is the best (cheapest) off-the-shelf unit I can find that looks to meet my needs at a price I can afford. My budget is ~$5K to start for this item; this machine starts at $3200 and a nice Delta B2 servo upgrade runs an additional $1500. Looks pretty sweet: beefy frame, is designed to cut hard metals, and hopefully can grow with my business for a few years.

By the way, you won't find that machine on their website but when I inquired their sales folks and gave them my requirements, this is what they suggested.

The plan is to use Fusion-360 to translate my CAD models into G-code and run that on the mill. Since they don't offer it, it's looking like I'm going to need to buy or build an A/B-axis table to mount on the XY table to cut some of the parts but I'm starting out with the 3-axis plus a rotary A-axis they do offer that looks more targeted at lathing so it probably won't work for what I need but the price is right so I'm throwing it in.

I'm asking them about getting a controller that supports 5-axises, but they're hedging and recommending I stick with the Mach-3 4-axis unit that comes default with the machine. They warn to stay away from Mach-4 because they say it's not reliable but I doubt this is true, especially nowadays with a recent version of Mach-4 and driving servos not steppers. Wives tales die hard, especially when functional pirated software is involved.

They're willing to supply it (with some resistance & $$) with either a 'Mach-3' 6-axis or 'Mach-4' controller. I put that in air-quotes because it's hardware, not software. I think they're offering it with this (or some knockoff):
Products ??????????? (in the 4-axis variant)
but I'm wanting to go with something like this:
Products ??????????? (the 6-axis version)

I have absolutely no problem purchasing a licensed copy of Mach-4. Probably hobby version until I actually sell something. My only concern is, what motion controller is going to work best for my application? And should I go with Mach3 (free) or v4 software? Order with one of these CDX HTC Tech products, or something else? Or should I stick with the bare-bones (which they know and support) and plan on upgrading to get the additional axises I need?

So that's my first post question, hope it's not too stupid. But to make sure its a lame 1st-post I need to mention I need to place this order ASAP (EOW at the latest) so a fast reply on at least the controller question would be ????.

And yeah, I'll attach the AMN-6040-M PDF if this board will allow little ole newb me to attach files in case anyone is interested. It looks like a pretty nice setup.

Any feedback/thoughts/ideas are welcome, especially anything on A/B tables that can run 6061-T6 reasonably quickly. Finally, I'm looking forward to joining the community here. Looks like a lot of knowledgeable folks hang out here.

Hi Dwg - I'd be careful starting a business to make money and buying that type of machinery to do do it. You end up being a poor machinist vs a product designer and seller. Once you have this machinery its difficult to place a commercial value on the parts as your time and machine payback is a rubbery value. You are subsidising the machine and the work. If you can't subcontract the parts to proper machine shops that will make parts 10x faster then you can then how do you scale up? So find better machine shops and when you get quotes find out the cost of 10 units or 100 units to see what scaling up does for you. It's your call but I've been there done that and unless its thought out you don't win. Peter

Welcome to the Zone.

3D Printing to Prototype/Testing for your birds come to mind. There maybe a MakersSpace or some group with Machines in your area who might be willing to work with you or you can join, also.

Good on your mission,
DJ

Hi Peteeng, thanks for the thoughtful response. I think you're 100% right based on the info I gave. It's a little more complicated than I let on.
Maybe the following is 'TMI' but:
You definitely hit the nail on the head. I've been spinning on this exact issue for too long.
Problem is I need 48 parts for my first full-sized prototype & the lowest quotation I've been able to find is $1000/ea. Which is way more than I expected.
A shop is willing to work with me on some redesign that might get them down to $500 but still that works out to a $24K bill for just these parts for one vehicle.

My fail was not anticipating the costs - the motor/prop/esc's I've selected cost less than that per unit. And they're pretty freakin big, 32" props.
I've already got $50K+ sunk into a patent, $20K into parts (props, motors, etc) and expect this to go up.
Trying to minimize costs where possible. Would rather not seek investors until I can demo it lifting cars, etc. Call it 'sweat capital'. ;-)
Also FWIW although I'm not wealthy I am semi-retired and have some (limited) means, this is kinda my dream to start my own business and I have time to spend if not waste.

Actually my friend/biz-partner on this project owns a sheet-metal shop, we plan on reworking the into bent & welded plate when if/when we scale.
But he has a business to run & family stuff (as do I) and his paying customers come first. My guess is it'll take many months going that route.

I'm taking your words to heart though and rethinking because you're right its a steep learning curve, etc.

Actually my background is as a software engineer evolved to a design engineer evolved to a mechanical engineer. All self taught but I've done pretty well.
Core issue is (I'm told) I don't have a Ph.D. in Aerospace Engineering, so I have 0 credibility.
Weird, I say, my scaled-down prototype seems to work pretty well. Not perfect but it works as expected.

Realization was I need to build & demo a kick-ass prototype to get some credibility. Lift cars & stuff. (Yeah, its big) Or sell out to VCs for cents on the dollar.
Anyway, I need not only these parts but the ability to make revs and additional prototype parts and push them out on the fly. At reasonable cost.

FWIW these parts so far are only for the 'wing' sections (think props laid out on a space frame, 36 coaxial props @ 50lbf safe lift/ea).
I still have to do detailed design the fuselage. Which will be more machined $parts$.
And its hard for me to swallow $25K for a partial bill on a 1st prototype vs $5K for a machine + materials then having future low cost capability.

But still think your advice is good. But unless I can find a low-cost alternative, the ability to turn around parts cheap & fast, I don't see a better path.
Thanks a lot for your input. Ideas & advice always welcome.

And sorry about the brain-dump. Just thinking with my fingers I guess.

Hi DJ,
Yeah, thanks for that. The thing is, some of these parts need to take serous loads. Max bar-force calculation is almost 5000 lbf thru-part in the worst case.
I've FEA'ed the heck out of the design I came up with and learned a lot along the way (slip-critical vs. bearing fasteners was a big surprise).
Bottom line is the quotations I got back for 3D printed parts in an appropriate metal were ~$3K/ea
Multiply this by 48 parts and I just killed my dream project unless I find a cheaper alternative.
Outsourcing to a machine shop took costs down to $500-$1K but still a serious stretch compared to a $5K machine and time to learn the ins and outs of machining.
Especially considering I have more parts to build. It kinda sounds like you're going to have to skim the rants of yet another newb.
But don't worry, I won't be that bad. 30 years ago I actually programmed steppers via a parallel port (raw, no drivers). So I'm not totally green.
If you think of any ideas on how to outsource my problems ;-), I'd love to hear about it.
DWG

Back to the original question at hand: Does anyone have an opinion on Mach3 vs. Mach4? Both the hardware motion controller & software.

To recap, I've been warned away from Mach4 from the tool manufacturer. But although Mach3 looks like it would work and has a large following, I'd prefer to get support on (and to support the developers of) Mach4.

So what's the bottom line? Get the 'Mach 4' motion control hardware & Mach-4 software? Or go with Mach 3 hardware & software? Or some mix?

From what I gather, the motion controller hardware just takes G-code and translates it into stepper pulse-trains more or less in real-time while various monitoring inputs, possibly triggering stop events, etc. So it seems to boil down to smoothness of motion, event handling, and bells & whistles 4 has over 3. Am I missing anything? Or am I exposing just what a newbie I am to it?

I've skimmed a lot of forum posts and the answer still isn't clear. If anyone could point out killer features or deal killers of either, I'd greatly appreciate it.

If your only option is the Chines controller, I'd probably have to say Mach3?
I haven't heard of ANY Chinese controller working properly with Mach4. The vast majority of them don't fully support Mach3, but work well enough for normal milling. Expect issues if you need to do any probing.
And Mach3 isn't free. The Chinese give you a pirated copy.

Hard to say without seeing the machine, but for complex aluminum parts, expect to spend MANY hours per part, and a marginal surface finish, from a cheap Chinese router.

Hi DWG - Being involved with prototyping and various machinery design for a long time and working with people on their projects like you vs working with companies on projects I understand your position. Its a tenuous one as your pockets are not deep enough. In the car industry we usually have a low function prototype before we jump to the production spec. There is even a "pre-production" spec. So its an appearance model, a low function model, a minimal function model, a pre-production model and then the production model. This could be a scaled down model to cut costs and it sounds like you have done that. Instead of jumping to your aluminium full function model what about cutting plywood parts full scale in the knowledge that it won't lift full operational weight but it will be fully operational to flight test systems out? It seems your vision or biz model is blurry. You are starting with billet machined aluminium, then when scaling you are going to sheet metal. This is contrary to all types of aircraft at the moment. You need to go composites. Billet aluminium and sheet metal will not scale and will not achieve your current or future performance or cost requirements. A smarter approach is to use the router to make timber moulds and make composite parts. This scales beautifully and produces the best parts. Another learning curve but I assure you, you can make excellent carbon fibre parts in your garage that will outperform metal, be cost competitive etc etc.. Do a search for infusion processing of composites and use epoxy only.

Plus with a moulded part you have geometry freedom and size freedom. You can make the entire drone in 3 or 4 pieces vs the many you are talking about... To start you could make plywood parts and over laminate with glass or CF to improve its stiffness...all very doable. aluminium 2700kg/m3 lightweight plywood 400kg/m3 CF 1500kg/m3 well ahead of AL once you get your head around it...So even if you have a small router you break the big object down into small parts then glue them all together into the big part, then overlaminate Good luck, where there's a will there's a way..Peter

Thanks, I'm coming to the same conclusion, at least on the motion controller. I'm more or less assuming I'll have to upgrade the controller at some point. So it's probably fine to just go with the crap controller they ship it with and suffer thru it until upgrading to something better once I get my feet wet.

Huh... I just realized my attachment with the PDF of the machine spec didn't make it into my original post. I'll try to attach it again here.

If it makes it thru and you take a look you might (or might not) agree that this is actually not bad. It has a pretty beefy cast iron fixed-gantry / sliding table design that's (supposedly) well suited to cutting metals. It comes with Taiwan-style Hiwin rails & ball screws, so it looks to be pretty stiff. But who knows until I receive it?

I'm ordering with Delta B2 750W servo XYZ motors ($1500, ouch!), my understanding is this should kick butt for cutting even steel.

Anyway, if the quotation makes it thru take a look and let me know what you think!

Hey Peter, thanks again for the feedback. You may be right, but right now I'm seeing things from a different, possibly naive angle. Let me explain.

The 'wings' of the vehicle are basically a triangular lattice structure that supports lots of rotors (36 in this 'small' but full-scale prototype) with aerodynamics taken into account. The basic idea is lots of rotors aggregate lift and add redundancy. There are a ton of other benefits I won't bore you with. But a key feature I'm trying to achieve is no single-points-of-failure. Anywhere. Safety thru redundancy. So, although there is a cantilever strut involved to support the high-off-axis loads, I want the wing to support max load even if the support strut breaks. And remember, I want to scale this thing up to lifting thousands of pounds.

This thing also generates really high stresses along the frame due to the leverages involved. Many thousands of pounds. It's a long story but another nice angle is the frame design also heavily replicates a few parts, I.e. Many instances of a few designs. Think tinker-toys and you're not far off.

I thought about this a long time and tried a lot of different approaches and basically came up with using warren trusses to link the tringular matrix nodes together. I'm trying to machine these nodes that link the trusses together. That's my initial goal. They look a lot like a hub of a wheel or a star, but optimized for various reasons.

Here, I'll attach a jpeg of the current design for your amusement. This version was intended to be 3D printed, I need to redesign it to be more machinable since printing is too expensive.

I also want this to be cheap. Safe but cheap. And when your vehicle is capable of lifting thousands of pounds, shaving weight becomes a lower priority. Right now the frame I designed weights in (calculated weight that is) at about 500 lbs but is designed max lift is 2500. Warning, I'm throwing around rough numbers, I realize it's a lot more complicated than that, its just been a long day and I don't have the brain cell capacity to lay it all out better right now.

So while I think you're right from the conventional sense, it's also true that large aircraft are made of metal including a lot of sheet metal. I may have issues with buckling (Euler's column law) I have plans to deal with that if/when if fails static worst-case stress testing.

Then there's the difficulty in validating carbon fiber. It's pretty easy to mess up a layup and bury the issue under additional layup. I'm no expert on CF, i've only watched videos on laying up CF props, etc. You're right it doesn't look so hard. But I'm also trying to think low (or reasonable) cost manfacturability. CF layup is a highly skilled task with a lot of toxic chemicals involved. So I'm not a big fan of CF these days.

I'm no Elon Musk (probably qualify as a wanna-be) but he dumped CF for Starship. When you go big the rules change. And now his Cybertruck, damn, that's genius: monocoque stainless steel exoskeleton. Badass. I really like how they bend the unbendable stuff by milling (sawing?) the inside of the bend line to thin the metal. Genius.

So hey you might be right. You certainly are for smaller vehicles where efficiency is paramount due to battery limitations. But when you have a bigger vehicle other power supply options become available, it opens up a lot of options. The rules change.

Sorry for the word salad there. Again, thanks for the feedback and keep the comments coming...
DWG

If anyone is wondering about this machine, the PDF didn't attach to the post for some reason. May be a first-post issue.

I'm attaching it here in case anyone is interested. I think it looks not half bad for the price, all things considered.

But I'm a newb to milling so theres that.

Hi Dwg
1) Metals fail catastrophically via fracture composites fail safe as they delaminate but do not loss much strength or stiffness in this process and they certainly don't fall apart
2) 6061-T6 has a YS=240MPa from memory a good CF laminate has a flexural strength of 1200MPa (and does not fatigue) so its a no brainer for me
3) saving weight on any vehicle is anyways a priority. No one has ever asked me to design a heavy vehicle (boat, plane, car or 1000T trailer that I'm doing at the moment ) they always have payload to gross weight issues
4) The largest commercial aircraft, the airbus thingy is mainly made from a laminate of aluminium and fibreglass so its a composite (its called GLARE look it up)
5) You can validate CF quite easily I've been doing it for 30 years plus. The suppliers now have a good idea of whats needed. When I started on these things I'd ask ?? and the suppliers would say "I don't know" now they have answers for nearly everything
6) In your case its not easy to muck up a laminate stack. Using infusion is the safest process around as the resin is either in a bucket or a tube or in a plastic bag.
7) Many issues with CF at the moment revolves around you can't buy CF in big quantities. Airbus and Boeing and other large companies have forward orders in for the next 15 years on all the good stuff!! CF companies are doubling production at the moment but it will take a few years to get all the new plants running
8) Sorry the rules don't change, I think you have painted yourself into a corner.
9 ) Keep at it... Peter


https://en.wikipedia.org/wiki/6061_aluminium_alloy