[efinkg]

I'm working on getting a Taig 2019 Manual mill. I "grew up" working on a couple of Bridgeports with DROs, and I'm a complete slave to having one. I've seen that the "normal" way of doing it is cheap chinese calipers modded to work as linear slides. I'm considering eventually CNCing this, and I have two questions that are connected.

Does the output from the chinese slides work similarly enough to a quadrature encoder that they could be used as encoders for a closed loop system? If not, is the backlash in the Taig low enough that you could reasonably use a rotary encoder on the shaft connected to a Shumatech 350 DRO and get good results? I don't want to spend a couple hundred on slides that I can't use since the Shumatech can take quadrature encoder inputs according to the website.

At home I have a big closed loop CNC that I have converted from original control software to use EMC2, DC Servos on ballscrews, and it is rock solid. I've used a Taig on steppers and it's a solid machine, I just love the direct input of working on a manual mill. The second option is to use Mesa Electronics boards as if I were doing a CNC conversion and just skip the DC motors for now, since a closed loop system is considered optimal while driving all the axes though CNC, I expect that the encoders will be good enough for use with a DRO.

Please feel free to tell me why I am wrong.

[hub]

I'm working on getting a Taig 2019 Manual mill. I "grew up" working on a couple of Bridgeports with DROs, and I'm a complete slave to having one. I've seen that the "normal" way of doing it is cheap chinese calipers modded to work as linear slides. I'm considering eventually CNCing this, and I have two questions that are connected.

Does the output from the chinese slides work similarly enough to a quadrature encoder that they could be used as encoders for a closed loop system? If not, is the backlash in the Taig low enough that you could reasonably use a rotary encoder on the shaft connected to a Shumatech 350 DRO and get good results? I don't want to spend a couple hundred on slides that I can't use since the Shumatech can take quadrature encoder inputs according to the website.

At home I have a big closed loop CNC that I have converted from original control software to use EMC2, DC Servos on ballscrews, and it is rock solid. I've used a Taig on steppers and it's a solid machine, I just love the direct input of working on a manual mill. The second option is to use Mesa Electronics boards as if I were doing a CNC conversion and just skip the DC motors for now, since a closed loop system is considered optimal while driving all the axes though CNC, I expect that the encoders will be good enough for use with a DRO.

Please feel free to tell me why I am wrong.

Just my opinion..
Linear measuring would be best, but I'm not so sure "chinese converted calipers" would be it.. The "proper" glass ones are expensive though.
I would go for the encoders if price matters.
And if there is backlash, I would try to fix it mechanically if possible.
Just my thoughts
Hub

[awetmore]

There isn't a lot of room on the Taig for glass encoders.

If you are eventually going to go to CNC then I don't see much reason to put a DRO on first. There isn't much equipment which can be shared between them. Few Taig CNC conversions (or small benchtop machines in general) are closed loop. I also like manual machines, but if I have to have one big machine and one small machine I'd prefer my small one to be the CNC, not the large one. My manual machine (6x26 knee mill) does have a DRO.

[efinkg]

So I was looking at throwing a set of the DigiKey Encoders on, which have a variety of hole mounts. I am looking at a Taig since I am currently attending college as a mechanical engineer and there is no machine shop, which I consider ridiculous, but I don't think there is the impetus to get fullsize machines at the moment. I am unfortunately a ridiculous distance from my rescued Hurco KM3, and miss having such tools available to me.

I can't really put the time into a CNC conversion right now, and I'm not thrilled with the prices that Taig has for full CNC'd mills out of the box, I think I can do much better for much less, especially since I have the experience of converting my KM3 already under my belt, I just need a month or two that I can work on it seriously, and I don't have that time.

[cameraguy]

A closed loop system is good for servos (required, actually), but not so good with steppers. A Taig-sized machine is best suited to steppers. I would avoid attempting to set up a closed loop system on one. The only commercial closed-loop system for them that I know of is godawful and completely pointless.

As for manual DRO's, the Chinese slides use a bizarre coding system that is simple, but entirely their own. I do recall the someone made a simple converter that plugged into them and output standard quadrature, however. Sure a search will turn it up.

[efinkg]

Would standard quadrature encoders be an acceptable way of doing a manual DRO? It would be cheaper than linear slides and seems like it should be as accurate, especially since I would get as much accuracy as using stepper driven axes.

[cameraguy]

You asked about backlash, and on the Taig you should get about .0015 or so on X and Y, zero on Z of course, on a totally stock mill. There are very easy (and cheap) mods to get that to a true zero backlash on all axes if it is required, but it is often not necessary to completely eliminate it for most purposes. Trying to use rotary shaft encoders would be one of those times, however!

[efinkg]

Okay, so in your opinion what is my best approach?

[lsces]

A closed loop system is good for servos (required, actually), but not so good with steppers. A Taig-sized machine is best suited to steppers. I would avoid attempting to set up a closed loop system on one. The only commercial closed-loop system for them that I know of is godawful and completely pointless.

As for manual DRO's, the Chinese slides use a bizarre coding system that is simple, but entirely their own. I do recall the someone made a simple converter that plugged into them and output standard quadrature, however. Sure a search will turn it up.

The simple answer here is that one has to manage the physical backlash properly and keep things as 'tight' as you can. The latency in getting a position from a Chinese scale means that it is no use for controlling the actual position, only confirming a position once movement has ceased. Even the quadrature scale inputs into Mach3 can only be used to confirm position and produces an error if a final position is not reached.

'Closed loop' will only work by directly driving the motor and controlling the MOTOR position and since the Taig backlash is after the motor, even switching to servo's will not do anything for that. Putting the feedback sensor on the bed rather than the motor shaft will just create an unstable mess on this type of closed loop controller.

Mach3's backlash compensation works reasonably well as long as the backlash is being managed well by regular maintenance.

As for CNC versions of the Taig mill ... Buy the CNC ready basic mill and three stepper motor mounts, and then the only 'conversion' work is to add your own stepper motors and drivers. The cost of scales would go some way to buying the CNC kit instead?

[hub]

Would standard quadrature encoders be an acceptable way of doing a manual DRO? It would be cheaper than linear slides and seems like it should be as accurate, especially since I would get as much accuracy as using stepper driven axes.

Depends on where you would mount the encoders. If on the leadscrew your reading would be a reference to the leadscrew revs. It there is backlash between leadscrew & nut, the DRO reading would be off (by backlash amount), at least in one direction.
But that might not be a problem, you could always calculate off the backlash amount in one direction..

Hub

[efinkg]

Iisces, I understand this, I just don't see A) where my benefit is driving the axes with steppers off the leadscrews over counting rotations on the leadscrews, unless there is no backlash compensation in the shumatech, and B) I really would prefer to be able to do a DRO with by-hand machining, and AFAIK, you cannot use back driven stepper motors for this. I've been trying to get information on the improvements ill get for purchasing a computer ready model, since you are advocating one, I'm all ears. from what I can tell all I get are motor mounts and a bigger motor, both of which I can purchase seperately.

[cameraguy]

Bigger motor, motor mounts, and better leadscrew nuts that allow backlash adjustment. All of which can indeed be purchased separately.
If you want to go manual, I would just purchase the CR kit and a set of handcranks, and put the stepper mounts on the shelf until later. The motor upgrade and the split nuts are nice to have even just using it in manual mode.

[efinkg]

I have looked at prices of all the added benefits, and it looks like unless the motor mounts are going to run me over $30 per I come out ahead upgrading the motor and lead screw nuts for a separate cost. I'm not trying to be difficult and I appreciate all the help, I just want to ensure I am getting the best deal I can. I should probably email Taig to make sure I am not missing out on anything I should be worrying about.

[efinkg]

Hub,

I understand this also, I should take a look at the shumatech manual and see if I can do backlash compensation through that.

[hub]

Hub,

I understand this also, I should take a look at the shumatech manual and see if I can do backlash compensation through that.

That was for manual. For NC, backlash compensation works fine when doing straight left-right or forth-back moves, but doing circles & arcs for example you might not get what you want.

[lsces]

Hub,

I understand this also, I should take a look at the shumatech manual and see if I can do backlash compensation through that.

The DRO350 does not have anything other than straight display of the reading head. Using quadrature headers via the QCC100 just does a straight count and numeric output.

It would be possible to implement backlash compensation on the DRO550 and all the code for it is open source (opendro on sf ) but the current code just displays the values direct.

As for the base mill, there are a number of ways of buying the necessary parts. Personally I supply the kress spindle rather than the ER16 one, so don't need that anyway, but the price is such that a complete package is often easier to pick up. We don't see second hand ones often, since that would be the cheapest option, even if you then fit a new set of lead screws and nuts ... the parts are not expensive anyway.

[efinkg]

What is the difference between the Kress and ER 16 spindles? Mr. Werby of Computer Sculpture said that the CNC Ready mill cannot be used by hand. Is this true? Is this due to the adjustable leadscrew nuts or the lack of handles supplied?

[cameraguy]

What is the difference between the Kress and ER 16 spindles? Mr. Werby of Computer Sculpture said that the CNC Ready mill cannot be used by hand. Is this true? Is this due to the adjustable leadscrew nuts or the lack of handles supplied?

just the lack of handles. my dad uses one of my extras with hand cranks purchased separately and fitted back on. Just remove the motor mounts first.

[Jeff-Birt]

I've had several folks purchase the CR mill PLUS hand cranks. Then they have used it as a manual mill for a while before upgrading to CNC. The CR version of the mill comes with the much better spindle motor as well as the stepper couplers. Taig is now only supplying the split lead nuts so you will likely get them on either a manual version or CR. If you add up the cost of purchasing the upgraded motor and stepper couplers separate (and shipping them separately as the motor is heavy) you a little better off buying the CR version IF you might want to upgrade later.

[efinkg]

So I have spoken with Andrew Werby and he tells me that the motor mounts + couplers would be $106 and the larger motor $80. This is still less than the CR model, unless my math is completely off.