[bean7795]

correction 304s.s. not 340

[Graham S]

Am I being daft but aren't those blades very thin indeed, how do you machine them without deformation?

[DHK]

I've used a Taig CNC mill extensively in the past and currently have another one I'm rigging (using a different CNC setup than the factory models). As much depends on the endmill and feedrates used as what is being cut (in your case inconel?). Harvey tool has some nice miniature crystalline CVD diamond and amorphous diamond endmills which will fit the Taig mill (www.harveytool.com). These can substantially cut down on temperatures when cutting so you won't work harden stainless. One of the main reasons for selecting a smaller desktop is that they can spin at 10,000 RPM all day. In fact, Taig runs their spindles in at 10,000 RPM for a whole day before they ship them (their bearings are rated for 15,000 continuous). A larger machine won't be happy at those speeds for extended periods of time. A contractor friend used to run his large Fadal at 10,000 for small parts and he ended up replacing his spindle. While it was running at 10,000 RPM it made an unhappy "screaming" screech. The Taig mill, on the other hand, was very quiet -- just a fan-like humming sound. I ran the Taig very hard -- often at 10,000 RPM for 24 hour periods.

Backlash is the primary limitation on accuracies on these small machines. One way to zero out all mechanical backlash is to use simple pulleys and weights to preload the X and Y axes. You'll need about 7 pounds on the X and Y. I'll be rigging up a Taig with weights soon -- perhaps I can post pictures on this 'board when I'm done. The Z axis has zero backlash because it is already "weighted" by the weight of the spindle and motor. I never had any problems with the CNC drive on the Z axis. With a 20:1 feedscrew, 7 pounds of weight adds a trivial amount of back-torque. The stepper motors used on the factory Taig CNC mill are very strong (they'll bend a 5/16 endmill right over). They do run warm and I never had problems with lost steps. I'll be equipping the Taig that I'm currently rigging with a DC servo drive system from Imsrv (www.imsrv.com). This will result in greater accuracies and much cooler (and quieter) running. Steppers can be irritatingly noisy. I've already got the imsrv servo system up and running. The DeskCNC CAM software used with the imsrv system uses a serial port and has no problems running in Win2K. My biggest gripe about the desktop mill parallel port stepper software was the need to run in DOS mode. A few such software packages claimed to run in Windows, but these often generated errors and were happiest running in DOS mode.

A larger mill will carve out stainless faster and with common tooling. But, there's something to be said for a nice and quiet desktop mill which you can run all day. I'd caution that when cutting stainless always wear eye protection, especially with a non-enclosed mill. Some varieties of stainless are non-magnetic -- and a few slivers/flakes in the eye can take a lot of flushing to remove. I unfortunately know this from direct experience...

-D

[mgamber]

See Mach2 for Windows based control software.

[Meltdown]

DHK

I also have the servo's from imsrv. They are nice but the problem i'm having with them is the backlash in the gears. I thought about buying some ball-screws but didn't because with this gear backlash in the servo's it would defeat their purpose. I'm looking into how to improve that without shelling out for expensive planetary gears. Any thoughts on this would be most welcome

[metlmunchr]

Re the cutting of Inconel.........I don't know which Inco alloy you're looking at using, but 718 is a common one for use in jet engine parts. The guy who proposes to sell you a chunk of inconel with the advice that it machines "just like any other stainless" is being just a tiny bit less than honest IMO.

Milling 718 is similar to SS in that it needs slow speeds and heavy feeds, but my limited experience with it also says it takes about twice the power to drive either the spindle or the axis as compared to 304, and using light feedrates to compensate for lack of power is an option that simply does not exist. It only makes a bad situation worse. Razor sharp tools help more than anything else. T-15 HSS endmills often work better than carbide on these materials. T-15 has good red heat hardness, and can take and hold an edge that's next to impossible to get on carbide. Avoidance of heat generation, or the use of flood coolants will have little effect on whether the stock work hardens. Work hardening occurs due to the rubbing action of the tool against the stock, so that's what has to be avoided. It might be worthwhile to check with Carpenter Technology and find out whether they produce any relatively free cutting high temperature stainless alloys that could be suitable for your parts. Materials like this are generally available in small quantities, since no one can afford large quantities of them

[DHK]

Meltdown,

The Pittman DC gearmotor servos used in the Imsrv system (part# GM9236C534-R2) do have a bit of backlash in them. Using DeskCNC, you can select the "Abrupt" style backlash compensation to consistently remove this. I think this is done in the 2.067 and 2.070 versions by unselecting the "Metered" backlash compensation box in one of the configuration menus. The Taig Z axis, despite the weight of the motor and spindle, doesn't have any static back-torque on the feedscrew (which would eliminate the slight rotational backlash in the servomotor gearbox). I think even if one used a much better lubricant (Taig recommends ATF -- auto trans fluid -- as the lubricant) you probably couldn't achieve static back-torque with a 20:1 feedscrew with any reasonable amount of weight loading. Ballscrews may well be different. Preloaded axes with ballscrew drives may have enough static back-torque to eliminate the servomotor gearbox backlash. But, this might also continually stress the DC servomotors to remain in any static position (my experience with ballscrews is limited). I think the important part about servomotor gearbox backlash is that it remains consistent irrespective of axis position. Software backlash compensation, of the simple abrupt type, should render this backlash negligible.

By the way, you can get these servos as N.O.S. surplus from www.servosystems.com for around $49 each. They're the exact same Pittman part (GM9236C534-R2) used by Imserv in their system. The only difference is that the "flat" on the output shaft hasn't been machined/ground on them. I figured that Imsrv did this as Pittman makes no mention of a flat on the output shaft for this part number. I picked up a set of these units myself -- nice to have as spares as the brushes on these motors do not look to be user serviceable.

-D

[Meltdown]

Thanks DHK for the servo link! I will get me some extra servo's just in case. You're right on the constant backlash in the gears. The problem i'm having is that with with the screws i use the backlash is constantly changing. The only cure will be ball-screws i guess. After that the gear backlash should be constant and manageable in the software . Now how to find affordable 12mm ballscrews is another matter