What's new
What's new

Need CNC advice for new business doing product dev, prototyping and small batch prod

stev

Cast Iron
Joined
Aug 19, 2004
Location
Oregon
Hello Gents,

I am in the beginning stages of building a new low volume business that will take small product ideas from conception to prototypes to small batch production (less than 100 parts), if the demand for a product is sufficient I would probably farm out the process to a real production shop or possibly sell the rights to another manufacturer.

I have several ideas for small products costing less than a few hundred dollars and not larger than a 12x12x12 envelope, generally out of plastic, aluminum, stainless, steel. These products will not generally have more than 10 separate machined pieces.

I want to start small until I can prove out the viability of my concepts. Initially, it will be just myself, I have a wide variety of design, computer, electronics, cad and manufacturing skills, but will eventually grow the business if demand requires it. I've owned a small home machine shop with high end, self refurbished manual machine tools, monarch, deckle etc for a decade or so and intend to farm out any work I cannot do myself.

I would love to have some input on the following thoughts/questions:

1. I'm looking for a small used CNC like a haas TM2p or equivalent to produce most of the parts? So far I like the Haas, but I've also considered Fryer, Trak, Southwestern ind., Tormach, should I be looking at any other manufacturers?

2. I am also considering an inexpensive 3d printer for the prototyping. I am new to 3d printing and am not sure how useful this would be, have you found them useful for this purpose and do you have a suggestion on a good model to start with?

3. If you've done something similar and have any advice to share I would welcome it greatly, either PM or here!

Any ideas are welcome, Thank you!
Steve
 
Last edited:
The Haas control is very nice to work with.

Definitely buy an enclosed machine so you can run flood for aluminum. Fastest spindle you can get for that, too.

Largest work envelope you can fit and afford. You WILL fill it up with multiple vises, etc.

Making one at a time on a CNC is slow and really sucks because you are probably stuck at the machine vs. loading it up and working on something else while it runs.

Probably should decide now if you want to get into 4 or 5 axis. 3 gets a bit limiting, but it totally depends on what you are making.
Just a heads up doing the flip-flop routine between multiple vises can get parts out of tolerance quickly (ask me how I know!).

I'd say the decision would be 5-axis or not, since you are not after high volume production and your work size is smaller. Ours is higher volume, and variety, but we can't deal with one at a time, so 5-axis is out. We have a 3 and a 4. Although I'd sure like to just grab a part one time, and have it almost finished, it would drive me nuts for what we do, and we'd have to hire a full time person to load blanks and push the green button.

Hope that helps...
 
Depending on what you are doing... If you know what you need, it be much cheaper to have the first bits done by someone else. Basically, if you can draw it, have it made, redraw and be done, it would be much cheaper to send that work out. Or if you have little to no experience, it ain't worth it. All depends though. If you have many things or fosisticated parts requiring lots of time to develop, it may pay off to have your own setup.

Trak is southwestern... And probably the easiest to learn and use. But don't expect to get too crazy with that as for putting out pieces. And to the other extreme, if you know what you are doing, and if your ideas require big time setup and may take days to setup and get going..... If it were me, I like Okuma. No homing, no worries, turn it on and go, it knows where it is. You would not have to worry about picking back up on zeros. And a better machine that any you mentioned. But my views aren't always agreeable..... Well, hardly ever.... Teehee
 
If your part envelope is 12 x 12 x 12 you don't need the travel available on a TM2P, possibly you should consider a SuperMiniMill or SuperMiniMill 2 with much faster rapids, toolchange and spindle speed. The TM2P is an economical way to get into doing longish parts but the 6000 rpm spindle speed and the tediously long toolchanges get tiring aftera while.
 
The answer to your question is entirely dependent on the kind of parts you'll be doing.

Prismatic parts? Turned parts? Complex surfaces? High precision? Consumer 'faux' precision?

I keep toying with the idea of getting a mill (a used Robodrill or Brother) and I constantly have trouble penciling it out for the simple reason that my part variance is very very high - I do prismatic parts, I do heavily contoured parts, I do mill/turned parts. It's become obvious to me that a simple 3 axis mill could work, but I would be doing a lot of trickery on a bunch of my parts (faking lathe work on a mill, for example, or using rotary indexers). As much as I would love to justify having a mill just to fuck around and invent stuff, it still makes more sense for me to rapidly iterate using 3D printers and do a $1000, 20 part beta test run through my machine shop.

Also, as you get into the lower-end tools (Tormach, Haas TM), the space and rigidity limitations just make pushing out weird parts even more of a time suck. Go on YouTube and watch the home machinist crowd working, and you quickly see that they spend a LOT of their time and effort working around the limitations of their hardware (and shitty free/cheap CAM software).
 
Love this forum, I can always depend on great advice and insight from the people here.

I almost looked up 'fosisticated" as well, but then saw the other posts!


CTAC - Check, large work envelope, enclosure, 20 ATC, definitely would like 5th axis, my in house volume would stay low enough to make that work. Once I have a design dialed in and tested, I'll out source for sure.

annoying - I've been wondering where I need to be on the out-source vs in-house, good insights. Eventually, I'll get to the point where I'm just doing the design and sending it out, for now I feel better going through the process myself as I work it out. Some of the parts are gonna take several iterations and require reverse engineering complex fits, mounts, etc. I'll also look into Okuma, do you have a TM2 equiv recommendation?

Hdpg - I started looking at the sminimills, but moved to the toolmills when I realized I could get a 20 pc atc and longer X. I do have some parts that will require the longer X, but not sure I'm gonna make them yet. I'll have to reconsider if the length is worth it. It would be nice to have that with the faster rapids and spindle, but don't think I can afford it.

gkoenig -
1. A little bit of everything.
2. This is why I mentioned 3d printing, while I'd still purchase a mill, I can see building much of my prototyping with 3d and then out sourcing the beta test parts. Can you recommend a decent 3d printer?
3. Yeah, I know the trade off on machine quality. I'll buy the best machine that I can justify. I don't mind upgrading once money is coming in the door. I'm not sure that the Trak and Tormach are really up there, but my understanding is that the Haas is pretty decent if not a little slow. I could be wrong though.
 
Love this forum, I can always depend on great advice and insight from the people here.

I almost looked up 'fosisticated" as well, but then saw the other posts!


CTAC - Check, large work envelope, enclosure, 20 ATC, definitely would like 5th axis, my in house volume would stay low enough to make that work. Once I have a design dialed in and tested, I'll out source for sure.

annoying - I've been wondering where I need to be on the out-source vs in-house, good insights. Eventually, I'll get to the point where I'm just doing the design and sending it out, for now I feel better going through the process myself as I work it out. Some of the parts are gonna take several iterations and require reverse engineering complex fits, mounts, etc. I'll also look into Okuma, do you have a TM2 equiv recommendation?

Hdpg - I started looking at the sminimills, but moved to the toolmills when I realized I could get a 20 pc atc and longer X. I do have some parts that will require the longer X, but not sure I'm gonna make them yet. I'll have to reconsider if the length is worth it. It would be nice to have that with the faster rapids and spindle, but don't think I can afford it.

gkoenig -
1. A little bit of everything.
2. This is why I mentioned 3d printing, while I'd still purchase a mill, I can see building much of my prototyping with 3d and then out sourcing the beta test parts. Can you recommend a decent 3d printer?
3. Yeah, I know the trade off on machine quality. I'll buy the best machine that I can justify. I don't mind upgrading once money is coming in the door. I'm not sure that the Trak and Tormach are really up there, but my understanding is that the Haas is pretty decent if not a little slow. I could be wrong though.

I just tossed okuma out there. I never did feel like trusting the way most machines "home" out. Okuma does not, absolute encoders so it knows where it is. Bigger budget and nothing that small, so it isn't likely a choice you would consider. But, some of the older vmcs could be had for decent price.
 
I've been doing something vaguely similar, started as a hobby and kept getting more serious until I found myself renting an industrial space a year ago and buying bigger toys. I've tooled up for my own product development goals (mostly robotics-related prototyping) but work has found me. I'll offer a few observations based on my experience.

- No ATC gets old when doing production, even batches of 10. Cycle time isn't that much slower, but it means you can't do anything that requires concentration while the machine is running.

- I have 4-axis capability but so far have only needed 2.5D. 4th is useful for creative multi-part fixturing though. YMMV though. Your machine will in part define the jobs you can take. There are comparatively fewer shops that can do 4/5-axis prototype parts, so if you can, you'll have a better shot at those jobs. I would look carefully at the downsides of a 5-axis machine, and not just cost, and think about how those would affect your 2.5/3D parts. Depending on the mix (which is probably going to be a guess), it might or might not be a good play.

- Someone once wrote here that a good prototype machinist thinks the opposite way that a good production machinist does. In production you might spend days to shave a few minutes off cycle time. In prototyping, an hour of setup and an hour of spindle time are often equal, so using stock that's oversize (but fits a fixture you made for another job), tools that are undersize (but set in toolholders and ready to run), and default CAM routines that have plenty of room for optimization are all money-makers. Lately I've been thinking that a bar-fed mill-turn machine would be pretty interesting for a lot of parts--the stock utilization would be poor, but aluminum is cheap!

- I've been somewhat lucky in that Boston has a lot of brainiacs (around half of my clients have a Ph.D.) who have ideas and a few bucks in their pockets. Many also have at least some familiarity with the business of making things so their expectations are not too ridiculous. They are also more likely to build prototypes that turn out to need more than one. My first client came looking for 5 each of 5 parts and in the eight months since then he's started selling them and I've made about a hundred sets. He's now looking for someone who could make batches of 100, because I can't and don't really want to. That's due partly to me not having a large enough machine or ATC to run that many, and partly to not actually wanting that much work.

- What's really going to determine your success is not your ability to make parts (though that matters!), but your ability to find customers, quote well, and sell those quotes. The transition from doing this as a hobby to thinking about it as a profitable venture has been slow and filled with under-estimates. I've also been nowhere near break-even, but I haven't been marketing aggressively because I basically look at paid jobs as a side dish rather than the meal. Before investing a lot of money, I would try selling the capabilities you have today. First, you'll figure out if the business side of the machining business isn't for you, and second, you'll get a better idea of where you should invest.
 
I hav done pretty much what you are proposing. For aluminum you want a high spindle speed, 8K minimum, and that means an enclosure, which you also need for flood. I would consider a tool changer mandatory for more than two or three parts at a time, hell its useful even for one offs. I would also consider electronic probing to be a must have. In both prototyping and short production runs you do a lot of setups compared to machining time, electronic probing will speed this up immensely. When I bought the VMC I use, it was going to cost as much to move the conveyor as the mill, and I almost decided to leave it behind - how many chips could I possibly make anyway? I did bring it along and have blessed that thing every day I use the mill - unless you do tiny little parts, you will spend half the day shoveling chips and cleaning up the resulting mess.

When considering the in-house vs. buyout, it may depend on how complicated and accurate the parts need to be. On my product, I was revising it very often, sometimes as I machined the parts. Some of this was a result of realizing what would make them easy to machine, from the experience of trying to do it. This will drive a job shop crazy, and if they are smart they will charge you for it. With my in house machine, I could change it at will as I refined the design and its manufacture. As the volume increased from one at a time to 5 or 10 at a time I changed things again. Once this settles down, then it makes sense to look outside. On accuracy, I found that the short run job shops in my area really did not want to talk about the accuracy I could routinely hold in house. That may or may not be an issue in your area, with your parts - but consider it.
 
Low production and prototype, Knee mills are nice. Drawback is the mess. For that reason alone I'd be itching for an enclosed machine...Haas let you open the side windows for added length...AKA...part hangs over the table and out the door. Not ideal, but works.

Haas is not what I would consider slow...but would consider it not a very rigid machine, but best in the class you mention. Dirt simple machine, user friendly.

Big deciding factor that needs to be taken into account, parts and service...make sure whatever you choose you can get them at a reasonable price and time. You do not want to have your one and only machine down with an interested customer and you cannot produce for 2-3 weeks as your power supply board went down...or monitor. Or you cannot figure out what is wrong and need a tech to diagnose and best they can offer is two weeks or you have to fly somebody in.

The rest has been said...good luck.
 
Terryz - I'll check out Mazak, any recommendations there?

Doug - I agree with starting with what I need to make money and upgrading when the need and opportunity presents itself. 3,4,5 axis is definitely something that I need to think about… not sure where I'll land, but want to make sure that I can add the capability in the future, if I need it. Agreed, ideas are a dime a dozen, getting them to market is the hard part! As for 3D printing, I've started researching these and I'm a little confused about their usefulness. Ideally, I could print the initial parts in plastic and get all of the rough changes out of the way, before I turn on the mill. But I'm just not clear on the quality and resolution of these printed parts. Sounds like sending the printing out would be the best way to start.


sansbury - Nice, all good advice. This will be the third business that I've built over the years, there are always twists and turns on the the way to finding the path. There are quite a few startups in the area, while this business is primarily for my own products, keeping the equipment busy with other parts from neighboring startups is an interesting idea. I love making things and love machining as much as I enjoy creating ideas for products.

swarf_rat - Agreed, electronic probing is definitely a must have. I would love to have a higher speed spindle for aluminum, that's what I did like about the Trak machines. But I don't want to drop below the level of the Haas to get it. That's why I started with the minimills first, but they're more expensive and a smaller envelope. Even though I need a 12x12x12 envelope I don't want to limit myself in that area either. So, far the Toolroom mills catch the most "needed" features for the least price. Do you have another mill recommendation?
Your second paragraph captures the exact reasons that I want to stay in house to start! thank you.

SIM - Thank you, when you say the Haas isn't very rigid, could you compare that to a manual machine for me? Like a bridgeport vs a big horizontal? I'm leaning towards the Hass for all of the reasons you mention, it seems to cover a lot, aside from rigidity and moderate speeds. Any thing seems fast compared to manual. :-)
 
I have been prototyping for medical devices for quite a while and have found extremely limited usefulness to 3D printed parts. The tolerance and finish are not very good compared to a machined part. The strength of the parts is not very good either. Couple all of that together and a printed part is really not representative of fit, finish or function. Sometimes it is nice to be able to put a "concept" part in someone's hand, but that never seemed to justify the cost of the machine.
 
There may be some Sharp 2412 type mills out there. For a long time you would never see a used one for sale, got snapped up too quick I guess. But I have seen some lately. The people that have them seem to like them.

Another consideration for prototyping and development though it access to the table. An open mill like a Deckel FPNC is ideal for this, really good access to setup, measure, ponder. You usually sacrifice the enclosure, spindle speed, and tool changer for this though. Looking at enclosed mills, some of them have very restricted access, others pretty good. If you have a choice, that might make the difference.
 








 
Back
Top