[dvirb]

Hi everyone!

As the title explains, I was wondering about the metal machining capabilities of the Taig CNC machine which was recommended to me a couple of times.

I am looking to machine small precision parts (precision of ~0.0001 of an inch) in brass.

I was thinking about matching German silver, would the Taig be capable of machining the more-harder-then-brass type of material (I know some people machine mild steel with the Taig, but I am not sure how well does it go)

A little of topic - do you know of an affordable and available quick tool change spindle/chuck for the Taig ?


Any help will be so much appreciated!

Thank you



Sent from my iPhone using Tapatalk

[Eldon_Joh]

you might have to replace the leadscrews with 1204 ballscrews to get that kind of precision. or use a spring to provide constant load on the leadscrews.

yes it can handle the harder material, depends on how sharp your cutting tool is.

[awerby]

I'm a Taig reseller, but I wouldn't claim that the Taig would hold those tolerances. You'd be hard-pressed to achieve that kind of accuracy with a mill costing 20 times as much. Even measuring that precisely is problematic - metal objects will expand or contract more than that if you take them into a warm room from a cold one.

German "silver" (actually a copper/nickel alloy) shouldn't be a problem to machine on the Taig; it's softer than mild steel.

[dvirb]

Thank you awerby for your straightforward honesty!
I can do with a tolerance close to that as well (i'd say I need approx. \ around 0.0001 of an inch).
BTW - As a Taig reseller, do you know of any standard quick-change (manual) tool chuck for the Taig ?


@Eldon_Joh thanks for your tip! does such a conversion for the Taig (leadscrews to ballscrews) is a straightforward remove and replace type of conversion ? or does it require more ?


Thanks in advance!

[awerby]

With care, you can achieve tolerances of +/- .001" on the Taig. If you need ten times better, be prepared to spend a whole lot more. But often people think they need much higher tolerances than they really do - engineers often drive machinists nuts by over-tolerancing simple parts, just because it's easy for them to do.

I don't know of any quick-change system for the Taig. It doesn't actually take long to change tools using a couple of wrenches (the ER system is a lot easier than changing tools on an R-8 spindle); I don't think a "quick-change" system would really save much time for most of us. An ATC is different, but there's nothing like that on the market that would fit the Taig.

People often talk about switching from the stock screws to fancy ballscrews, but it's rare that they actually succeed in doing it. A set of ballscrews as precise as the stock screws would cost more than the whole mill, and it would take a lot of excavation to make room for the nuts, which are much bulkier than the nuts it comes with.

[robertk925]

You can see a really cool ATC for the Taig here:

Log in

I don't think it is for sale yet, and it looks like it might eat up a bit of valuable Z height, but it is impressive.

[Tkamsker]

as awerby wrote are this to,erances really needed ? it is possible to do that on an good stable mill but you need servos and really the high end (grinded ,.., ) spindles then temperature is important ,.. i would also be verry interested how you measure that tolerances and if you are doing one offs or series ,..


Gesendet von iPad mit Tapatalk

[bhurts]

If your looking to truly hold that tolerance you are looking at the wrong machine. If you spend 100k on a serious vmc you are still not going to get that every time. Grinding is really the closest tolerance machining your ever likely to see. Even then your very unlikely to see anyone calling that out on a regular basis if ever. If you would have said .0003-.0005 I would think you could achieve it on a high end machine. If your looking at the lower end hobby machines that people use here your never going to get what your looking for. For what they are they are really nice machines but .001 is about the best you can ask. For .0001a single screw that's capable of that will cost more then a taig and likely not far from the price of a tormach.

Ben

[dvirb]

Thank you all for the replies and tips!

So let's assume I am getting a Taig cnc mill, what can be done to improve its accuracy and reduce backlash?

I understand that replacing the stock lead screws to ball screws is a difficult task, so how about replacing the Nema stepper motors to servos ? Would that help with accuracy and less backlash ? And what type of servos/stepper motors are better?

How about keeping the lead screws and replacing the nuts to upgraded nuts ? (I believe I've read something about it somewhere)

Does Taig offer customized machines or does everything has to be done by my self?

Thanks in advance


Sent from my iPhone using Tapatalk

[dvirb]

Regarding to tolerance I think .001 would be enough


Sent from my iPhone using Tapatalk

[awerby]

Servo motors aren't any more accurate than stepper motors, although you can put high-count encoders on them that give you more resolution. But these tend to bog down your controller and don't have any practical effect on accuracy. The steppers supplied with the Taig are as accurate as any, and Taig's half-stepping system reduces inaccuracy due to microstepping. Taig mills have adjustable bronze nuts that can be tightened to reduce backlash, but they won't eliminate it entirely. In Mach3, you can use backlash compensation to zero out measured backlash.

Taig doesn't offer any customization services for their mill; if you want to tinker with it you're on your own. If you're getting one, I'd suggest trying it out before launching into any modification projects; you might find that it works well enough as-is.

[tony978]

I am running at servo (dc motor with encoders )driven taig and I use hsm pathing . I would not run that pathing with steppers . As for holding a tolerance with the machine you need to know your speeds \ feeds and how the machine thermal expands over the run of your part . After you know how the machine is cutting the part you can mod your program to cut the part you need . Even with a industrial cnc you need to make adjustments as you run to hold +-.001 " . I have production and mold making experience.

Tony

[Tkamsker]

i converted an bridgeport 412 i to linuxcnc where i learned it is all about the accuracy of all parts errors sum up. so linear rails and a good spindle is a must i would then in your case use an closed lopp system on the machines we produce i can get normal 0.015 mm accuracy only on the z axis i need the feedback for that accuracy because of the verry heavvy head ( which is good for steel ) so maybe for that accuracy if you live in usa look at tormach or novakon if you live in eu look at our machines


Gesendet von iPad mit Tapatalk

[lsces]

dvirb I think most points have been made, but just to summarise ...
The original Taig manual machine was supplied with a simple split brass nut, and the 'improved' nut came as part of the 'CNC ready' machine, but nowadays every machine is supplied with the improved nut, which has set screws to adjust the fit on the leadscrew. I understand that this is even 'lapped in' which has some pluses and a couple of minuses. Switching to the Kress spindle with a 25mm long mount and spacer on the Y axis gives a useful extra work envelope, but any step in the leadscrew prevents that,
The comparison between the 20tpi simple screw and a ballscrew is referring to 'backlash', but the simple fact is that the raw accuracy of a Taig mill is lost if you fit ballscrews. Just work out the step size on 20tpi over say 3mm pitch ballscrew.
Yes keeping the backlash tight is more important on the screw, but to get accuracy back on the ballscrew one has to microstep on a stepper system, which is why a switch to a servo is almost essential, while the stepper approach with simple quarter stepping gives a 1/16th thou step size. Which is a lot more accurate than we can hold the backlash, but something which can be coped with. Raster profiling one simple switches the cutting pass to the one direction, but profiling is affected by backlash, and as has been said, one gets what one pays for. The Taig IS however very capable even cutting harder materials such as steel, and as long as the tolerances are not a problem a stock machine can do very competent work such as required by jewellers and the like, but not as tight as you are looking for.
Even throwing ballscrews at the taig with a high accuracy servo control system, I think you will find the 'slack' in the gib strips will prevent holding the sort of accuracy you are asking for. I think you will need to add pre-tensioned ball race slides into the mix, but I suspect even tool deflection and ware will thwart getting down to 1/10th thou sustainable accuracy?

[Tkamsker]

Hi isces i fully agree with your comment i think that a linear rail system with some rigidity and pre loaded ball bearings are a must. Probably you can do a 1:3 or 1:5 reduction from Steppers to Spindle to get some more accuracy. I would also prefer the servo solution
But maybe it is an interesting learning project that endeavor might be worth watching to see if it is possible ,..

[109jb]

Tony - I don't disagree with most of the stuff you said, but I see absolutely no reason a stepper or servo driven machine makes any difference in regard to HSM machining. If a stepper system is set up properly it can run an HSM toolpath without any problem. Although it isn't a Taig, my stepper driven G0704 has no problem running HSM toolpaths. I don't understand your reasoning for that statement.

[tony978]

My history with steppers with a flashcut system is not good


Tony

Sent from my iPhone using Tapatalk

[Eldon_Joh]

The comparison between the 20tpi simple screw and a ballscrew is referring to 'backlash', but the simple fact is that the raw accuracy of a Taig mill is lost if you fit ballscrews. Just work out the step size on 20tpi over say 3mm pitch ballscrew.
Yes keeping the backlash tight is more important on the screw, but to get accuracy back on the ballscrew one has to microstep on a stepper system[...]
Even throwing ballscrews at the taig with a high accuracy servo control system, I think you will find the 'slack' in the gib strips will prevent holding the sort of accuracy you are asking for. I think you will need to add pre-tensioned ball race slides into the mix, but I suspect even tool deflection and ware will thwart getting down to 1/10th thou sustainable accuracy?

yes if you fit 1204 ballscrews which is a major modifiction, you may need to belt drive the ball screws. 200 steps per 4mm is 8 tenths. 10x microstepping would be enough with a big enough (stiff enough) stepper motor to get to tenths, 16x microstepping would be good enough imo.

If you buy cheap ballscrews you will need two nuts spring preloaded or you will need to reball the nuts.

with mach3 driving my steppers at ~.01 inches per minute, you can count the 40 microsteps per thousandth of an inch on a dial indicator like watching the second hand tick by on a clock. however, backlash with 1/2-20 60degree V thread is just too unpredictable. and the screw thread on mine is off as well, by as much as 1.003/1. leadscrew mapping will be required anyway..

you might be better off replacing the leadscrews with 1/2-10 or 16tpi precision acme thread and spending a bunch of time lapping it.

another problem with the taig for tenthousanths work is you will need to scrap in the y axis casting to 20-40 points per inch.. because otherwise the oil film thickness will vary each time you move the machine. you can't scrap in the aluminum annodized table either, so you're going to be stuck with whatever warpage it already has.


no point in throwing ball races at it unless you intend to do something like this:
frankenTaig - Album on Imgur

only problem is now its not difficult to generate enough cutting force to break the 608 bearings that hold the leadscrews..

[dvirb]

thank you all for the participation

I plan on machining small amount of parts per day, not a commercial production line. the parts are small precision parts (as small as 10-15mm).


How about if I slow down the machining process (slower and more torque), would that help with avoiding the backlash for some percentage ?

How about using springs to keep the tension on the X-Y axis (as much as I've heard the Z does not need to be preloaded since the heavy motor weight does that already ?)

What other methods can be used to preload the axis ? (other then the weight pulley system)

Can I use specialised screw nuts to keep the tension on the axis ? or should I keep the stock nuts ?


@Regarding the 8-ball ATC mentioned - it looks good. hopefully it would be sold for an affordable price too


thanks alot!

[Tkamsker]

Hi,
i think you should try it probably it works to some degree but maybe you learn on the way which precision you really need. but if you run a part a day every day it is about production so maybe you then look for a Tormach or Novakon depending where you live thomas