[kanankeban]

If you had to start a machining business and had 50K to get a used CNC Lathe and a CNC Horizontal Machining Center which ones you will look for?
And what do I spect to receive for this amount of money?
Regards...
Hector

[motordude]

Hello,
I bought some older CNC`s I while ago, and looking back this is what I will do next time around if buying a lathe and a horizontal (or anything else CNC):

1. Machines preferred to be of same make and identical controller. Takes less time to learn the programming and postprocessor setup.
If the controller/servo on one machine goes bad you can interchange parts and find the error easier.
2. Find out sparepart prices on selected hardware and electronics and compare. I had quotes on my Cincinnati for $2000 for a memory board. A cheap machine is not always cheap when you have to fix it! Even availability may be an issue if the machine is old.
3. Stay away from oddball machines, go for the recognised brands. If you have to close down they are easier to sell.
4. Never buy anything that is not in working order even if it is "just a contact breaker costing $10". See the machines working. Even a newer stored machine will give you trouble if standing idle for some time.
5. Buying a machine from a professional seller will cost you more, but you can go back and kick ass if something goes wrong. (within a relatively short period).

Good luck with your machines!

regards
John

[ynneb]

Not that I am that experienced, but from what I have heard, a PC controlled machine is easier and cheaper to maintain. Get some quotes on retrofitting a machine and compare the cost. I would guess your dollar would go a lot further with a retro fitted machine too. Worsted case scenario, lets imagine every servo, and servo driver failed and the computer blew up. It would cost no more than $3000 to replace the lot, unlike having to spend $2000 just for a memory card on a proprietary machine.

[metlmunchr]

In a horizontal machining center, I'd look for either a Makino MC-40 with a Fanuc 11M or a Mazak H-400 with a Maza-cam control. These are both 400 mm nominal work envelope machines, and either of them can be had for less than $10K with a bit of shopping for machines made in the late 80's.

For turning centers, I've had such good luck with Okuma that I wouldn't look for anything else unless it was awfully nice and awfully cheap. Avoid any turning center that doesn't include the full complement of ID and OD toolblocks. About all lathes use bolt-on ID tool blocks, and some also use OD bolt ons as well. You DONT want to have to buy these from the manufacturer after you thought you got a good buy on the lathe. There is no other source than the original manufacturer, and a full set of these items can cost thousands. My own preference on Okuma lathes is the LC series with a 5000 or 5020 control. LC's are built like a tank. One of these from the last half of the 80's can be bought for $15 to $20K in good shape, and sometimes for quite a bit less than that.

Once you're looking at auto toolchanger machines, the so called PC-based controls are largely a PC front end with proprietary motion control behind the panel. Machining centers in particular have so many auxiliary functions which must be controlled that any home-brew PC retrofit is out of the question unless you're a control wizard. The true all-in-software systems for retrofit such as OpenCNC are not cheap by any means.

In any machine purchase, make sure to get all the manuals. If they're absent, figure $1000 per machine to purchase them. Also, a lathe or mill with noisy spindle bearings or worn out ballscrews might not be a good deal even if it's free. Either of these items can be a mega-dollar repair for both parts and labor.

[kanankeban]

What about brands...? Which ones are good options? I've hear good things about mazak and really bad things about HAAS...I now its hard to say which one is better that other, but I know that ford is hardier for work than chrysler...so you know what I mean...
Regards...
Hector

[Scott_bob]

Hector,

Where in the USA are you?
By the way, PC based controls have far more flexibility than proprietary controls and are less than half the cost of say Fanuc for example. Old used controls are not worth repairing or in my opinion upgrading. Money is better spent on retrofitting the entire control system instead. Then you get a service organization to work on your machine when you need help. Very important...

You can buy a used CNC machine, retrofit the control and get better performance than buying a new machine, if you choose the right control retrofit...

[metlmunchr]

Hector,

Where in the USA are you?
By the way, PC based controls have far more flexibility than proprietary controls and are less than half the cost of say Fanuc for example. Old used controls are not worth repairing or in my opinion upgrading. Money is better spent on retrofitting the entire control system instead. Then you get a service organization to work on your machine when you need help. Very important...

You can buy a used CNC machine, retrofit the control and get better performance than buying a new machine, if you choose the right control retrofit...

I'd agree with this.......for a mill......for a specialized application like 3D surfacing or mold work. However, that work represents less than 15% of the machining currently being done in the US, and the percentage is shrinking every day. However, for general machine work, a full blown turnkey retrofit on a 15 year old VMC purchased in good mechanical condition will push the investment close to the cost of a new mill. And, having a usable machine is predicated upon finding a good retrofitter. From what I've seen, the bad results outnumber the good by an easy 3 or 4 to 1 margin. For a turning center, I'd have a hard time imagining a situation where spending the money for a retrofit would be worthwhile on a normal size machine with as much as a 10 or 12 inch chuck. Most any of them will spin faster and feed faster than the tooling can stand as long as the right size machine is being used for the job. On turned parts with any substantial cycle time, it's dirt simple to take a 20 yr old 4 axis machine and leave the latest $70 or $80K offerings with fast rapids and servo driven turrets gasping for air. Everything depends on the specific application, but the only thing that really matters is when a machine finishes paying for itself and starts putting money in your pocket.

[Scott_bob]

However, for general machine work, a full blown turnkey retrofit on a 15 year old VMC purchased in good mechanical condition will push the investment close to the cost of a new mill. From what I've seen, the bad results outnumber the good by an easy 3 or 4 to 1 margin.

So not true!
IMO, cost of the decent CNC mill: 20k + 30k (for an excellent PC based Retrofit)=50k for a CNC that is more accurate, less expensive to maintain, faster and therefore more productive than a new CNC machine you choose which one. I'd compare the performance of any new machine tool builder with the performance of an excellent PC based control retrofit any day. Price to performance ratio is unsurpassed. I don't know who you're talking about when you say:

From what I've seen, the bad results outnumber the good by an easy 3 or 4 to 1 margin..

We have a retrofit PC based control retrofitted to a 10 year old VMC. This machine has old DC current drives (old school). If one axis motor goes bad, we could replace just one axis servomotor with a new AC drive and the control can be configured for this odd motor and the CNC would be just as good. Try this with a proprietary control (mix and match DC / AC servo motors). Add control memory in a proprietary control at today’s PC hard drive prices:

Fanuc memory .5 Meg = $1,000
Old Fadal control memory 8 Meg = $1,000
PC hard drive 20 Gigabytes = $200
Which would you like to pay?

[motordude]

Hello all,
I am considering retrofitting my Cincinnati 5VC75. I have checked out Siemens, Fanuc and other types of controller. The best price/performance is so far Fagor. They are the worlds no. 3 (or 4) in CNC controls.
The packet includes the following:

3x axis DC motors and drives.
1x AC spindle motor
wiring etc.
Fagor 8040 controller.

Price:14300EUR.

If you can keep the existing servo and drives the price would be half that.

I feel that when starting a new business you want something that works, and to buy a machine to have it retrofitted, will cost you more than the price of the retrofit itself. You need to be making money fast, not to wait for the installation of a retrofit.

I hear good things about Okuma, Mazak and other japanese brands. My next machine will be japanese. And new!

regards
John

[metlmunchr]

Scott_bob, it's interesting to note that your Fadal was 8 yrs old about 6 months ago, but now it's 10. Since wear varies roughly as the square of speed, if you've doubled the average feedrates, you've probably put 2 years of wear on the machine in the past 6 months. Assuming a commodity VMC has a useful life of 12 years before it's ready for a complete mechanical rebuild, would you assume your doubled feedrates will generate enough additional revenue to overhaul that new machine in 3 years instead of 12? What happens if you double your 30K retrofit expenditure and spend it on a second VMC, running both at the feedrates they were designed to run? Surely the productivity will be greater than with your one high speed machine, given the dead time for toolchanges, part loading, etc that tend to be constants in the process. And both machines would have a projected twelve year useful life. From a tax perspective, the retrofit is a capitalized repair, whereas the 2nd machine is a possibly expensable investment. That said, I'd be the last person you'd ever hear singing the praises of proprietary controls and the rip-off prices attached to their parts. True PC based controls will take over soon, because there's too much money to be made, even at a fraction of the price of proprietary controls, for it not to happen. IMO, the sooner the better. However, as with all processes developed with an eye toward profit, there's a balance point that maximizes profit. Not feedrate, or cycle time, or any other factor, but profit over the long term. I'm afraid some of your statements about how a machine can be run faster and yet cheaper, simply because a cheaper to maintain control is able to drive it at higher speeds, are based solely on next month's bottom line. Somewhat like high school science problems which begin with the phrase "ignoring friction".

[Scott_bob]

metlmunchr,

Ok, this CNC I'm talking about, was built in 1996 making it 8 years old. Where did you get the principle of: "wear varies roughly as the spuare of speed"?
Is it possible to reduce wear on mechanical components even though greater speed of motion is occuring?
If a ballscrew and nut assy is accelerated and decelerated smoothly at a higher rotational speed, will it wear slower or faster than a ballscrew assy who's acc/dec is not smooth but is rotating at a slower rotational speed? In the case of the control that we removed, the acc/dec was really jerky. Other users of this control technology have reported reduced maintenance costs, increased tool life, greater speed, greater accuracy, all beneficiaries of smooth motion control.

Which motion path below represents less wear on a CNC?

[metlmunchr]

Where did you get the principle of: "wear varies roughly as the spuare of speed"?

For sliding motion, friction losses will increase as a square function of sliding speed, and wear will increase in fairly direct proportion to friction. Based on these two principles, its common to make the above statement.

Is it possible to reduce wear on mechanical components even though greater speed of motion is occuring?

In simple terms....No its not possible if the materials in question remain constant. The exception to this might be if you were pumping lube at a rate that allowed a hydrodynamic wedge to form and separate the surfaces. Not a likely scenario on a machining center. A car engine is a good example of another machine composed of rotary and sliding motion with variable loads. At 60 mph and roughly 2000 rpm, a large percentage of today's engines will go 200,000 miles with no problem. If doubling the speed reduces the life by a factor of 4, then we should be able to project that about the same percentage of engines would run at 4000 rpm and 120 mph for 50,000 miles. In actuality, probably less than 10 percent of them would survive that. Why? Because the little problems that inevitably occur in machines of all types suddenly become major problems when the machine is running at twice its design speed.

If a ballscrew and nut assy is accelerated and decelerated smoothly at a higher rotational speed, will it wear slower or faster than a ballscrew assy who's acc/dec is not smooth but is rotating at a slower rotational speed?

The wear rate will still be a function of speed. "Smooth" however is a rather subjective term. Unless you're cutting a straight line, the faster you run, the more steeply you must accelerate with change in direction. Otherwise, one axis must slow to allow for the other to come up to speed. If the jerky motion of the original control is such that screw loading is kept below the fatigue limit, chances are it will have little effect on the screw life. Another consideration to keep in mind regarding the screws is heat generation. Once again, this is a square function. I assume your Fadal has the dowtherm cooled screws, but doubling the average screw speed means the dowtherm has to remove 4 times the heat, or remove heat 4 times as fast, either way you want to say it. The nut's only means for transferring heat to the screw is via the balls. Not a good heat tranfer path, particularly since their rolling friction is whats generating the heat in the first place. Even though the screw is mechanically cooled, the nut is not. Assuming the machine is properly designed, at projected normal speeds the screw would be able to carry heat away from the nut sufficiently fast to protect it. Whether it can run twice as fast and perform this function adequately over time is something only time, or detailed analysis, will tell.

Which motion path below represents less wear on a CNC?

This is a good graph, because it demonstrates how closely you have to examine things which are thrown at you to help make the sale. If the lower average speed line did indeed represent average feedrate, then the area below the line and bounded by the various triangles and the average line would be equal to the area bounded by the same factors above the line. It doesn't take any calculation to see there's definitely more area above the line than below, indicating the "average" line is actually further down than it should be for an accurate representation of the average. The first logical question, once you notice that little faux pas, would be about whether some filtering was applied to the other curve to make it so nice and pretty.

I don't doubt for one second that your machine runs faster and smoother than it did before, and I didn't indicate I doubted it in the previous posts. My entire point is whether, over the long haul, it's possible to "soup up" a commodity machine without paying a big price in terms of useful life. If this can successfully be done, it says the machine was grossly overbuilt initially. The competitive structure of the commodity VMC market today would make it highly unlikely they're overbuilt to that degree. No doubt there's more and more pressure every day for shops to make it faster and cheaper. I just feel its important that we all look very closely at some of the long term costs of making it faster before deciding a price reduction is in order. We all know the typical customer could give a rat's ass if we scrap an entire machine, so long as we can cut another nickel off the cost of his part. Some shops feel this pressure because the customer's alternative is to source the parts to the pacific rim, but far more every day are getting it purely as a result of greed on the part of the customer. They "need" a 3 percent cost reduction from all suppliers, but when they get it, we see their price edge up by another 3 percent. We've gotta look out for ourselves, cause it's for sure no one else will.

To touch on a statement I made in a previous post about the percentage of "good" retrofits......obviously you got a good retrofit job. You may not realize how fortunate (lucky) you were in that respect, because good retrofitters are the exception rather than the rule. Half-assed cobbled up crap is far more common. I own a couple glaring examples myself One of my customers spent close to $200K with one of the largest rebuild/retrofit houses in the country on an upgrade to a large King VTL. It was at the rebuilder's facility for 6 months and returned "ready to run". 6 weeks later, their men were still at my customer's plant trying to get the machine to run, and it had yet to cut a chip. Once it did run, I made close to $10K of various parts to rebuild the supposedly rebuilt turret. The "shims" they had installed around the turret rotation bearings still had "Coke" clearly visible on a couple pieces of "shim stock". We decided the Coke was there to accompany the cabinet full of spaghetti they called wiring

[HuFlungDung]

Remind me never to argue with Metlmunchr

[IJ.]

Remind me never to argue with Metlmunchr

He's very informative though !

You always been a Mod Hu or is this a new suit for you?

I can see MM's point and have to agree with the "2 x machines to double throughput" while keeping within the "comfort zone" of the machine.

You see the hot up mentality in many different places Car's PC's CNC's and so on and while it's great to get something for nothing in the way of extra performance there's usually a price to pay sooner or later...

Being a CNC Newbie I'm not sure how much head room is built into current machines but using PC's as an example in the old days you could over clock many components and get a great performing system at a budget price.

These days as the head room shrinks people are finding more and more ways to Clock higher and usually this involves some exotic cooling solution (Can we see the similarity to MM's post?).

Same goes with Cars, You drop the killer Turbo system in and if the cooling side of things can't cope the lifespan decreases.

I guess my point here is the "Performance Costs" "How fast do you want to go" another that springs to mind that is applicable here would be
"Cheap
Fast
Reliable"

Pick 2.

I've followed quite a few of the HSM threads and have always enjoyed the informative nature of them, And while passions run high at times they have always managed to stay "friendly" I hope this is always the case!

[Scott_bob]

I'd like to ask again: "Where did you get the principle of: "wear varies roughly as the square of speed"?

Now I have to ask: Where did you get the principle of:

For sliding motion, friction losses will increase as a square function of sliding speed, and wear will increase in fairly direct proportion to friction. Based on these two principles, its common to make the above statement.

I don't come to the same conclusions as you have because of the variables in these equations. I don't believe one can define the "wear" in a mechanical system as "friction or speed x friction or speed" or as you say "square of speed". I think you're being way too abstract here.
Is it possible to reduce wear on mechanical components even though greater speed of motion is occurring?

In simple terms....No it’s not possible if the materials in question remain constant. The exception to this might be if you were pumping lube at a rate that allowed a hydrodynamic wedge to form and separate the surfaces. Not a likely scenario on a machining center.

Pumping way lube at a rate that allows a hydrodynamic wedge is exactly what machining centers do... That is what way lube is for.

A car engine is a good example of another machine

(This is just too abstract to be of any value in a comparison to high speed motion)... But I'm thinking right now of an example of increased life in a racecar when a driver can be easy or hard on his tires. Usually the winner of the race is the team that stays out of trouble, goes fast and manages tire wear. The driver who does not care about tire wear and just goes out there and tears them up will not win...

The wear rate will still be a function of speed. "Smooth" however is a rather subjective term. Unless you're cutting a straight line, the faster you run, the more steeply you must accelerate with change in direction. Otherwise, one axis must slow to allow for the other to come up to speed. If the jerky motion of the original control is such that screw loading is kept below the fatigue limit, chances are it will have little effect on the screw life.

You have not been a witness to the quality of or the benefits of outstanding motion control. Improvements to accuracy, machine life, tool life, and increased feed rates. All these improve with better motion out of a control that has replaced one of the worst controls on CNC machines today. This is part of what makes this so dramatic. We went from really bad to really good.

Another consideration to keep in mind regarding the screws is heat generation. Once again, this is a square function. I assume your Fadal has the dowtherm cooled screws, but doubling the average screw speed means the dowtherm has to remove 4 times the heat, or remove heat 4 times as fast, either way you want to say it.
...
Whether it can run twice as fast and perform this function adequately over time is something only time, or detailed analysis, will tell.

You like to use that "square function" and now you bring out "4 times the heat"... Where do you come up with these figures?

Let me just clarify a couple of details:
This retrofit CNC is not turning the ballscews any faster than the CNC was originally designed to operate at.
Max Rapid was 900 IPM, still is, only now the control is capable of feeding accurately within our tolerances of +/-.002 at feeds 2 to 3 times what it used to with the original control.
Most controls cannot perform up to their own designed specifications; this is exactly why many people have removed their control in hopes of addressing the problem. THE CONTROL...
So far, F500. is as fast as we have fed. The control is now capable of smooth (check out the high speed machining video) with dynamic feed compensation like I have never seen before...

[IndHobby]

I don’t ask this to sound rhetorical but I think your question got lost in the “discussion”, also where are you in the country?

Long and short, if you are looking to open a job shop you might want to look around to see how the other job shops are doing and plan accordingly.

In the Northeast (Maine) I placed an ad for a “Retired Machinist” for part time work. The next day I had LOTS of replies and some were job shop owners looking for anything.

Some parts of the country are dieing from a manufacturing point of view and these small job shops are being forced under.

Sad but true

[Scott_bob]

For anyone interested in trying to understand wear, friction, and heat in a ballscrew and nut assembly:

http://www.designnews.com/index.asp?...cleID=CA435593

Bottom line ref:
Identify the Heat Load

Average output power divided by the efficiency is input power, and the difference between powers is the loss that heats the ball screw. Preload torque times mean speed must be added to the loss. A usage factor (duty cycle) must be applied to reflect the ball screw dwell times (stopped conditions) that are not included in the duty cycle.

In order to go from heating (power) to temperature rise, the rate of heat dissipation is needed. But due to the variety of factors, it is almost impossible to approach the problem from an analytical standpoint. Instead it is necessary to use data obtained from actual applications as a guideline for the expected temperature range.

Kind of like I said earlier...

[metlmunchr]

Scott_bob,

I guess the fact that power, as referenced in the design news article, is an exponentially derived term, flew right over your head too as it swelled with satisfaction at having found something you thought might refute what I said. Allow me to give you a bit of the education you obviously don't have. Most every motion related task that performs work (defined as power over some specific time) reduces to a kinetic energy problem. The reason I "like to use those squares" is real simple. Kinetic energy is half the mass times the velocity squared. It's not that I like to use those square functions, its just that's what actual engineers tend to do since those are the basic scientific functions by which the calculations are made. Did I dream up the fact that twice the speed will create 4 times the heat as you seemed to suggest before? Uhhhh....hardly. If we move the speed from 1 to 2, then the energy moves from 1x1=1 to 2x2=4. Did I pick that as a trick pair of numbers to prove an invalid point? Don't think so. Move the speed from 3 to 6, and the energy moves from 3x3=9 to 6x6=36, once again 4 times as much. Even someone like you might be able to see a pattern developing here. Use any pair of numbers you like, and the results will always be the same. It's beneficial that you brought up the difficulty of empirically calculating the heat rejection from a screw, even if you had no idea why it's difficult. Modes of heat transfer and their associated calculations are much more complex than the simple energy relationships that allow estimating how much heat might be created. In this particular case, with a ballscrew spinning in air, contacting the balls in the nut and the bearings at the ends, you have convective (due to air) transfer, conductive (due to direct contact with other unheated objects) transfer, and radiant (due to close proximity to other unheated surfaces) transfer. The boundary layers are difficult to define, as are the contact conditions and temperature gradients within the components contacting the screw and nut. The very lube oil that reduces friction and helps prevent overheating also creates a barrier to certain modes of heat transfer. Lube obviously helps far more than it hurts, but its impeding effect is just another of the variables that make pencil and paper calculations inaccurate for complex systems. Calculations would generally assume conditions more ideal than would actually exist, and, over time, these small inaccuracies would cause the screw to run hotter than the numbers would suggest. On the other hand, I wouldn't tend to be nearly as concerned about the screws with above-design operating speeds as I would be concerned about the effect of these speeds on way life. Screws aren't cheap, but they're easily replaceable. On the other hand, worn out ways and the cost to redo them usually signals the end of useful life for the typical commodity VMC. All the "bogus" graphs and "jerk factors" in the world will not overcome the basic laws of physics, and those basic laws show unequivocally that heat generation and wear are exponentially related to speed. Your claim of "extended machine life" is ludicrous. If wear was only linear (the best we could hope for), and we double the speed, then its easy to predict the life of the machine in running hours would be halved. To claim extended life at accelerated speed should extrapolate to a conclusion that, if we run the machine fast enough, it would last forever. That's getting too close to perpetual motion and frictionless bearings for me to even comment on any further.

[motordude]

This thread is becomming really boring.

John

[BIG AL]

Strange, but I found it to be rather enjoyable. Thanks folks for the fight, boys it opened up new thoughts for me! :cheers: