What are Reasonable Goals for Home CNC?

[redneckalbertan]

I think it would be neat to knurl stuff that isn't flat.

Why couldn't you? You would have to use an appropriate cutter to get the desired profile, then make the machine move in a manner to make the knurl. A standard diamond knurl is a multi start right hand thread on top of a multi start left hand thread... Or is it a left hand thread on top of a right hand thread... We will have to let the experts weigh in on that.

[SteveHGraham]

It would mess with people's minds to show them a curved part with knurling on it.

[Jaxian]

This is why its best to start with a CNC machine designed this way versus converting a manual machine. A CNC machine has been designed and engineered to work together with all the components.

This is the gospel truth that really can not be stressed enough. When I jumped into machining with both feet to make car and motorcycle accessories I bought a Shopmaster Patriot and figured I was good to go. What followed was the longest 2-1/2 months of my machining life. I tried to learn using it as a manual machine. When that didn't work very well I tried to jump right in doing CNC. Yeah, try getting your feeds and speeds right by just typing them in and hoping without a lot of experience. Broke lots of things doing that.

When I started pinging all the local machinist and the machinery places I got a serious reality check. First was you can't make big parts on home machines. Not only for the simple fact that they won't fit, but the machines are not set up for that type of duty cycle. If you try to make a part that takes ten hours to cut you will kill a home grade machine. It will be so far out of spec by the end you might was well scrap the part. They are just not designed to run that long without a break.

Hence what people are saying above about knowing what you want to do. Don't try to bite off more than you can chew job size wise. This takes some careful planning to make the job fit the machine.

The biggest thing to remember, and I had to have this demonstrated to me a number of times before it sunk in is this. You CANNOT have a machine that is both a very good manual machine and a good CNC machine. You get one or the other. The converted or compromised ones do not work well at all for more than the most basic tasks.

The first and most important factor is simply this. The good ballscrew attributes of a CNC machine are directly contrary to the leadscrew good attributes of a manual machine. I have seen people take industrial level CNC machines with large tables 50"x25" like you would find in a 40x20 machine and with the servos not energized be able to move the table around just by placing their hand on it and pushing. This ease of movement and lack of slop is what the servos and their encoders need to maintain accuracy.

Conversely, a manual machine needs a leadscrew that will resist the cutting forces while still allowing the machine to be easily manipulated. Acme leadscrews have backlash and resistance. This is a requirement for good accurate manual machines. If you have ballscrews in a manual machine you better have VERY good table locks and never forget to use them. Even when cutting the ballscrews can allow chatter or fluctuation as they have no drag to dampen it.

That is the tip of the iceberg difference between the two but it pretty much stops serious dual use machines right off the bat.

[SteveHGraham]

Let me make things even worse. I blogged about this, and I stated the tentative conclusions I had reached. One was that old CNC machines were often sold because the bulky electronics were obsolete, and it was possible to hook them up to PCs and get good results. I got that impression from reading various forums and sites. Here is what a friend of mine said:

CNC controls are purpose built, and probably always will be. Only hobbyists use pc’s to control cnc, despite the hue and cry among the hobbyists. Fanuc is the gold standard, and again, probably always will be- the Fanuc 0 control has been in place and in use, unchanged except for processor speed, for twenty years. Several companies have attempted to use PC’s to operate commercial machines, and those companies are all bankrupt now. Those heavy useless built in electronics are still in demand and still in place making pretty much everything everywhere. Fanuc controls, Haas controls, Mitsubishi controls, and to a lesser extent Dynapath and even NUM are bulletproof and reliable, and are what everyone wants when running cnc. Many manufacturers use embedded PC’s to act as operator consoles for these controls, but the control is still a purpose built piece of equipment, and more than likely always will be.

You don’t have the power to run any but the most rudimentary real cnc controls, and the PC based ones won’t have the oomph to do anything but light work, but for most people the light work is OK. If you ever get to the point where you move to a place with a better shop with better power, you want a Makino KE55 mill, and a Harrison Alpha lathe. Both can be used as excellent manual machines with extensive and very easy to use features, and both will accept nc code to make complex parts, but neither will run on anything less than about 20 kva of 3 phase power.

The things you’re learning will lend themselves to plenty of other technologies, and that’s when the stuff gets the geeks excited. Keep experimenting, it’s rewarding and lots of fun when it works.

If he is to be believed, home CNC is a joke. If I understand him correctly, he's saying that you can learn a lot from hobby machines, but they are inferior to the real thing in every way, including the computer end. I don't want to invest three thousand dollars in a Grizzly mill and then watch it disintegrate over the course of a year.

I don't know what the truth is. I see people doing neat stuff with Grizzly G0704s and so on, but I don't think they spend much time publicizing their failures and limitations.

"Small" machining centers are gigantic. If I put one in my garage, first off, it would cost a minimum of seven grand, and apart from that, I would have to give up maybe 25 square feet of floor, making it necessary to get rid of a lot of my tools. Then I would have to go through it fixing and replacing expensive parts that were worn out.

He also added this:

The Fanuc Robodrill is probably the most useful piece of equipment you can buy for the money. Very difficult to kill, good toolchanger, simple to operate. Common as dirt and easy to repair. Not anywhere near as nice or useful as a Makino KE66, but easy to find cheap. From a stability and reliability standpoint much nicer than a knee mil of any kind.

I looked that machine up. It has a tiny table, but it would take up half of my garage.

It sort of looks like the only option I have is converting a small mill. If I buy something used, it will cost a screaming fortune, and it will probably be very screwed up, because if it were in good shape, it would be even more expensive. But if I get a small mill, I'll have lots of problems and limitations.

[Jaxian]

He is right about the things he says and you appear to be drawing the correct conclusions. As far as failures I owned my home CNC for 10 weeks, of which around 9 of them where spent just trying to make it work and fixing things. There are a number of threads over on the CNC zone following my failed exploits.

The Makino's are one of the few machines that I have hear being able to be used as semi manual and CNC. They are open machines though. But more to the point think of a used KEV-55 in the same way as a used classic Ferrari. You don't just hit up Craigslist and grab one. To illustrate this the guy who runs Practical Machinist is a machine dealer, has been for a long time. Finds tons of rare machines and goes to all the auctions and knows everyone. He has a sticky on the Commerce section of his forum trying to buy a Makino KEV-55, it's been there for quite a while. They are apparently VERY hard to come by if even he can't get the one he wants. Also the prices on those in good condition are enough to buy you a 5 year old VMC so really, there isn't a point unless you really can NOT find the space to fit a full size CNC any other way.

I have been shopping for a CNC VMC now for around 9 months in the $30k range without finding anything that meets my needs I truly feel your pain. All the older stuff the computer hardware is proprietary and expensive. Also feeble by modern PC based CPU power. Your PC would die in an industrial environment and there are technical issues to sort through but the bottom line is you smart phone has more raw processing power than a ten year old CNC machine...even a good one.

To illustrate this go hit up Haas website, they let you build machines like Dell does computers. Take say a TL1 series CNC lathe. They come with 1mb as standard and you can upgrade to 750mb for....$1695. I mean I can pick up a solid state 16GB SD card at the checkout stand at Staples for $30. The electronics are really a joke....BUT they are embedded and single purpose and that makes all the difference. It has really thrown me off trying to find machines that can do Helical Interpolation or High Speed Machining which basically perform functions you would expect any normal computer to be able to do, and PC's can, and find out they were all $5k, to $10k options and most don't have them. Very frustrating. They are very basic, very feeble, but very robust.

But to what your friend says. I don't see any way around it unless you are a electronics genius to convert a machine and have it perform remotely acceptably. So you sit there and wince as someone tries to upsell you a industrial CNC machine by bragging that it has 64mb of onboard memory so you don't have to trickle the program to it. Right after you put a $15 16GB SD card in your GoPro to film your latest machining trick. It's kind of sad....but what can you do?

[ctwo]

If you guys are just going to compare specs on paper, then there is no point because of course the commercial machines are going to have it.

Industry is not going to experiment with PC when purpose built controls have already been built, optimized, and taken to feature. There is nothing to gain on that path other than risk. I cannot comment about "all" of industry that has gone the PC route and are now "all" bankrupt. They may be, but it would be a hard sell to point at the PC as the sole cause of their demise.

If the goal is to home build a CNC simply to compete with a commercial industrial machine, then start with a multi-million dollar machine building factory.

My kitchen does not fair to well against a commercial kitchen, but it turns out some pretty good meals. Because my kitchen won't compete with those industrial kitchens with 150k BTU burners is no reason to toss in the towel.

Steve, you didn't really expect that your Big Dawg lathe was going to compete with any commercial CNC, did you?

[hobgobbln]

Having three small converted machines at the moment I figured I'd chime in.

Though I have never run a "real" machine I will admit right up front I have no doubt they are superior. After all, that is why they were built. Having said that though, my machines run quite well for my needs. Are there quirks and moments of frustration? Yes. Do I sometimes wish I could afford a $200,000 machine? Absolutely. But I also have realistic expectations from $3,000 self built machines.

For the last year and a half I have been using those machines to produce parts that are all over the world. I have made close to 12,000 parts between the three of them. They're small parts, mostly aluminum or delrin. The most popular parts are threaded aluminum. I have checked every dimension of every part I have made before they went out the door and I have been able to maintain +-0.0005 on every dimension on every one of them. Don't get me wrong, I have scrapped plenty of parts, but most of them were operator (my) error, not the machines. I know I could hold better tolerances and make the parts faster with real cnc machines but I also would have to guarantee I could earn enough every month to cover the cost of it, never mind always needing new tooling for the next project.

So, to me, whether you buy or build comes down to your own requirements and a simple question..... What do you NEED to make your parts?

Griz

[Marty_Escarcega]

Having three small converted machines at the moment I figured I'd chime in.

So, to me, whether you buy or build comes down to your own requirements and a simple question..... What do you NEED to make your parts?

Griz

This is a true statement. What are you trying to do? What do you want to make?

A home shop class machine can make nice parts. IF the machine is assembled properly with matched components.
Anyone that says a PC has no place in a CNC machine is kidding you. They are there. Not cheap components mind you. There is plenty of computing power in a PC to drive a CNC machine. Solid state drives make them more reliable.

For the home hobbyist its usually about learning about cnc and making something with your hands. It is for me anyway. But I did learn early on that finding a CNC machine with a dead control is much better than refitting a machine with ballscrews, thrust bearings, timing pulleys, motor mounts etc. At least I know those components are matched. Though I understand I have to live with the wear. Fortunately, most of those machines were overbuilt for production and with automatic lubrication systems so the iron tends to last. Some have led a hard abused life. Those are apparent simply by looking at them, and the environment they came from.

That said, for many that embark on the CNC ride, its about the journey as well. They enjoy the mechanics and fitting of precision components, and electronics, and that's cool if that's your goal start to finish.

I do know it can become expensive, especially when you don't buy components that work well or efficiently. Then you are buying more components to make it work. My suggestion is do LOTS of reading on the internet and learn what seemed to work for others. Let someone else make the learning mistakes.

My .02
Marty

[magic9r]

If he is to be believed, home CNC is a joke.

He's not.
It isn't.
The technologies used in home CNC are not now unusual in commercial scenarios.
Mach3 has crept into commercial applications where conventional controllers aren't flexible enough and the cost of having custom built controllers manufactured would be prohibitive.
A good read around the Mach3 support forum to see some of the commercial applications is an real eye-opener.
Using Mach3 doesn't limit you to parallel port or stepper motors either, there are excellent Ethernet motion controllers and servo controllers which can be used.
Don't try to inform or educate him, he doesn't need it and won't appreciate it

- Nick

[Mid Day Machining]

I have a small CNC milling machine in my garage and I do just fine with it.

I have a Tormach PCNC 1100 Series II. While I do mostly hobby work, I have one customer who makes fiber optic connectors and enclosures and I do about 90% of their short run and prototype work.

My business card states "ONE TO TEN PIECES IS OUR SPECIALTY". I will take a 25 piece job, but I will turn down 26 pieces. I don't like big production jobs. They bore me.

I owned a real CNC shop for about the last 10 years I worked full time. In my shop, I had a FADAL 3016 and a Haas TM1. My Tormach will do ANYTHING those machines would do, it just takes a little longer. My FADAL had 22 horsepower and the Haas had 7.5. The Tormach has 1.5. The FADAL had a 10,000 RPM spindle, the Haas had 4,000. The FADAL would rapid at 400 IPM, the Haas 200 and the Tormach rapids at a rip snortin' 90 IPM.

On the FADAL or the Haas, I could work within .0005 and My Tormach will hold +/- .001 all day long.

So to answer your question, ARE HOBBY CNC'S ACCURATE, my answer would be yes, IF YOU BUY THE RIGHT MACHINE.

[greenbuggy]

I feel like there's a wide spectrum that isn't really being talked about in this thread. Certainly most people don't have the space in their garage, probably not the power to feed and definitely don't want to spend the money on a big enclosed vertical machining center. On the far opposite side of the spectrum we've got tiny Harbor-freight grade machines that are converted into home-shop CNC keeping the least-accurate parts of them and no closed loop feedback to even let the controller know if a step has been missed.

There needs to be some sort of reasonable expectations for what you've got into the machines you make stuff with and what they could possibly do. What will work for a guy making parts for model RC or a hobby is a far cry from what will work for a machine shop doing lights-out high-precision production runs. Similarly the cost for a machine that would suit the production shop is going to be orders of magnitudes higher than a grizzly tabletop mill, some steppers and a Chinese breakout board and motor controllers.

In my shop I started with some crappy old non-CNC machines and I quickly realized that converting a clapped out acme-screw machine into something worthwhile (and especially doing it right) would cost more than starting with something worthwhile in need of an upgrade. So I hocked the manual mill I had and found a relatively "old" cnc with good iron that needed a new controller and drives, and I have <$3000 into a real heavy 3 HP Lagun mill with a (relatively) big table, running LinuxCNC and Mesa FPGA and motor controllers. I'm relatively happy with this, it gets a lot of work done in my shop. My only regret is that it took a few months to get completely up and functional, but the time and money it took me to get it all put together was considerably less than it would have been to buy an off the shelf solution with that much power and table size. Money I'd rather have been spending (and did spend) on measuring equipment, cutting tools, vises and other accessories.

[tornitore45]

If you are making jet engine blades or submarine propellers then a CNC is practically indispensable, although that stuff was done before CNC.

For HSM let's be practical, conversion needs ball screws, yes you can use all-thread rods and loaded nuts but that would be a kludge, then you got a project to motorize all axis plus expensive stepper drivers, then you have to invest in the software to make the drawing and the software to run the CNC.

Then you have to get through the learning curve of using both software. followed by few crashes along the way. Then you must run a PC in a shop environment.

At this point it takes you as much time to program and make the FIRST part as it would take you to make it manually. The next 100 parts are practically making themselves.

How many of your projects require multiple copies of the same part?
How many of your projects requires complex curves/shapes you can not make with angles and RT.? A guy on Model Engine Builder magazine made a gearbox that look like a casting on a manual machine.
How fat is your wallet? I do not mean how rich are you, but how much of your wealth you want to dedicate to your hobby or desire to be self-sufficient in this field.

While I admire, with a bit of envy, what my friends can do with full blown commercial CNC and home made conversions, I have decided that for my need and satisfaction an import manual mill and lathe provide me with all the capabilities I need to pursue my hobby: building functional model engines. I usually strive and achieve 0.001 tolerance for any non critical dimension; better for piston and shaft fits.

This is a subject where opinions and the weights each pro/con carries is highly personal, although previous poster made some very appropriate and rational observations, disagreement can be raised at any point I made, but you asked for opinions.