Homann Designs TM20LV (BF20,G0704) conversion

[phomann]

The ballscrew free end mount FF12 was assembled onto the endplate.

[phomann]

The fixed end mount end is a bit more complicated as it need to also accommodate the stepper motor mounting plate. As can be seen the FK12 is mounted to the endplate, a spacer plate is then attached, then the stepper plate. And finally the stepper motor is assembled onto the endplate.

[SCzEngrgGroup]

Peter,

I'm curious - how expensive is Corian, compared to aluminum? How does it machine? Did you pick up some cheap remnants?

Regards,
Ray L.

[phomann]

Hi Ray,

Corian is expensive. $800 to $1000 for a 1800m x 900mmx 12mm sheet. I get small offcuts such as the sink cutouts from a kitchen bench top.

It machines beautifully. Quick and forgiving and easy on the machine. On my Taig mill I can do 1mm cuts at 400-600mm/minute. It taps OK but it is easiy to strip a M5 thread by over tightening.

I'm using it a bit like a 3D printer to make trial parts.

For instance, on the Y-axis bearing bracket, all the measurements need to be taken off the existing parts, so it is not a simple job (for me anyway). When I initially made the bracket, the mounting hole for the FK12 bearing was 2mm too high, so I just adjusted the drawing and made another bracket, all in under 20 minutes.

[phomann]

These images show the trial fit of the endplates and ballscrew to the table.

[phomann]

The next thing to do was to mount the table onto the saddle.

The ballnut mount is positioned by just taking up the 4 bolts without tightening them. Then the saddle is wound fully to one end and the endplate is aligned, then the saddle is wound to the other end for the remaining endplate. Once free movement is obtained across all the table range of movement the nut holder is finally fastened assuring alignment.

The endplates are then removed and the ballscrew nut unscrewed from the holder (but remaining fastened to the saddle) as the table needs to be taken off the saddle to gain access to the Y-nut holder.

The middle image shows the countersinking of the bolt holes as there was insufficient clearance to clear the tip of the saddle when the table is fully forward and over hanging the saddle. It only needed less than 1mm but I countersunk down 2.5mm to be sure. As can be seen I bolted the Saddle to the Taig mill table by a couple of the existing holes. The cast iron of the saddle did not seem to stress the mill at all.

[phomann]

The next step was to test the stepper motor mounting onto the table while mounted onto the mill saddle and base to check the clearances.

All looks good so far .

[phomann]

Peter,

I'm curious - how expensive is Corian, compared to aluminum? How does it machine? Did you pick up some cheap remnants?

Regards,
Ray L.

Hi Ray,

here is a crappy video of the Taig cutting an endplate with a 4mm high helix slot mill.

[phomann]

The following images show two of the three pulley set-ups I intend to try.

The 1st is a 20 to 28 teeth set-up giving a ratio of 1.4:1 using a 570 oz-in motor.
The 2nd is a 14 to 28 teeth set-up giving a ratio of 2:1 using a 387 oz-in motor.

I also have pulleys and belts for a 20 to 20 teeth set-up giving a ratio of 1:1 using a 570 oz-in motor.

All the pulleys are steel XL flanged pulleys. Not sure why I chose steel over aluminium. Maybe they will be more durable.

I'm hoping that the 2:1 setup with the 387 oz-in motor will be adequate. That way I can use a G540 as the stepper driver. I think it will be Ok.

With this setup;

5mm travel per rotation of the ballscrew.
200 steps per revolution of the motor
2000 steps per revolution of the motor using a Gecko 10 microstep drive.
4000 steps per revolution of the motor using a Gecko 10 microstep drive and a 2:1 pulley ratio.

This ideally equates to;
A resolution of 5mm/4000 or 0.00125mm per step (0.0492 thousands of an inch)
An accuracy of 5mm/400 or 0.0125mm/step (0.492 thousands of an inch)

A 60ipm feedrate requires the stepper motor to rotate at;
60 ipm = 60*25.4 = 1524mm per minute

that is: (1524)/5 = 304.8 revs/minute

and with a 2:1 ratio 609.6 rpm

which is 609.6 *2000 /60 = 20,320 steps/sec

600rpm should be ok for the stepper but we will see. My taig mill rapids are set at 60IPM, so with the 20tpi leadscrew, the steppers are running at 1200 rpm. If it is in the torque delay area, the gearing will become less effective.

As to rapids of 60 ipm, I know there are lots of conversions that run a lot faster. For me 60ipm should be fast enough.

I really want to see if these mill conversions can be driven by the G540.

[109jb]

I would not recommend running a reduction on the steppers. I had a 2:1 reduction (4 steps for each 0.001" movement) on my z-axis and found that the since the reduction required the motors to run faster , the motors would lose steps because the torque had dropped off too much at the higher rpm. I switched the pulleys around to a 1:2 ratio ( 1 step for each 0.001") and I have more than 4 times the speed and no lost steps.

Remember that a stepper is going to have its highest torque at 0 rpm and it drops off from there. Making it turn faster will reduce the torque available at any given feedrate.

[phomann]

I would not recommend running a reduction on the steppers. I had a 2:1 reduction (4 steps for each 0.001" movement) on my z-axis and found that the since the reduction required the motors to run faster , the motors would lose steps because the torque had dropped off too much at the higher rpm. I switched the pulleys around to a 1:2 ratio ( 1 step for each 0.001") and I have more than 4 times the speed and no lost steps.

Remember that a stepper is going to have its highest torque at 0 rpm and it drops off from there. Making it turn faster will reduce the torque available at any given feedrate.

Hi,

Your statements no strictly true as the rated torque is held until the corner speed. But once that is reached, the torque does drop off.
http://www.geckodrive.com/support/st...or-basics.html

The other thing to note is that microsteps cannot be relied on for accuracy. Keeping in mind that the detent force is trying to putt the rotor to the fullstep position, the microsteps will not position the stepper motor rotor evenly between the full step position.

When quoting accuracy is is the fullstep position movement that mut be used not the distance a microstep theoretically moves. That said gearing will increase the accuracy.

I've allowed for gearing to do some testing to see what the difference is. I can use 1:1 and 1.4:1 with the 570 oz.in motor and 1:1, 1.4:1 and 2:1 with the 387 oz.in motor.

Cheers,

Peter

[109jb]

The textbook answer and the real world are not necessarily the same. The example I gave is a real life example to illustrate an example of where gearing is not desired.

Also, I did not say anything about microstepping or accuracy, the numbers that I posted were to give a sense of overal gearing of my z-axis, nothing more. On my machine as I originally had it, one full step would have theoretically moved the z-axis 0.00025", and after my gearing change one full step would move a theoretical 0.001". Those were not statements of accuracy, but of how the final drive gearing worked out.

As long as you bring it up though, the motor positioning accuracy at full and 1/2 steps will be dead-on, at least to the overall accuracy of the motor. It is the steps in-between the full and 1/2 step that may not be linear. Take for example a setup where there is one full step for each 0.001" machine movement. Theoretically the machine should be able to be positioned dead on every 0.0005" using microstepping. Now lets say I want to position to 0.00025". Even though you can't rely on perfect motor positioning, the microstep motor position is still going to be closer to the commanded position that the full or 1/2 step position. When considering benchtop machines like we have here, my opinion is that due to the machine quality itself, ballscrew accuracy, etc. expecting more than 0.001" accuracy is a pipe dream anyway. My machine is set up for 1/8 microstepping, but this is only because it runs smoother in 1/8 mode than in ful or 1/2 stepping modes.

Finally, the machine you are building is just like many many others that have been built previously and it has been shown that direct drive of 0.2" or 5mm lead screws works and works very well with good accuracy, speed, etc.

In the end, it is your machine and your choice as to how to build it. Good luck in whatever you decide.

[phomann]

I think your missing the point of what I'm doing here.

The purpose of the gearing is to be able to test the adequacy and suitability of using 387 oz.in motors. This will allow the use of a Geckodrive G540.

You only need to look at the number of problems people have trying to wire up controllers out of discrete components for these type of machine conversions. The G540 is a complete 4 axis stepper controller that requires minimal wiring and setup and may be a much more straight forward way of dealing with the controller electronics.

If I don't get the performance I'm after with this setup I can change to the larger 570 oz.in motors I have, along with a controller built of of discrete components based around the G210x Geckodrives.

Oh, I almost forgot. Thanks for reminding me it's my machine and my choice as to how I build it.

[109jb]

I'm not missing your point but you may be missing mine. Regardless of what stepper you use, I still believe direct drive would be the preferred method to link the stepper to the drive screw. That is all I am saying.

Hoss already demonstrated that a 570 running 3A and 24V will drive the G0704 at 50 IPM. That combo is probably running at less than 1/2 of the motor's 570 oz-in rating, so there is no reason a 387 oz-in wouldn't run better at its optimum input voltage and amperage.

[phomann]

Yes, I understand all of that. By using pulleys and this setup, it will allow me to test various ratios, motor sizes and even motor types.

Pulleys also allow the conversion to be more compact rather than having the steppers project so far out from each of the axes.

After testing, I should be able to provide information on the different configurations so that others can use it as part of decision making when doing their conversion.

Cheers,

Peter

[Fastest1]

Peter, I hadnt been back over here in a few days. I see you using the Corian. My families business has been fabricating Corian countertops since its inception. I have used it for lithophanes of my children and it was great to work with. Dusty and messy. Be warned that it is very abrasive on tools. I used to try to save the Porter Cable and other branded routers the fabricators would throw away almost weekly. I finally gave up trying to resurrect them, there was a reason the fabricators were throwing them away. Interesting use of the product anyway. In all fairness it will work as you have it for quite a while before needing replacement (if it ever needs to?) Also it was mentioned about Hoss attaining a 50ipm @ 24v. I got 331ipm at 48v. Hoss was probably in the 500's.

[phomann]

Thanks for the information on Corian. I mainly use 4mm high helix solide carbide slot mills designed for acrylic. Running around 1500rpm, or possibly slower. I do see the mills start to get blunt after quite a while (months) and I just change the endmill when I notice.

The main reason for using Corian on the conversion is so I can quickly make the parts on the Taig mill and without stressing the little Taig. Once I get the conversion running the first job of the new mill is to make the parts out of aluminium.

I use Corian a lot for making jigs for PCB assembly. It has proved to be very successful for this.

Selective Soldering Machine Jig Making Process

I also made a nice tool tack for my Taig mill as well.
Taig Mill Tooling Rack

As to the speeds, I want to find out what performance I can get out of the G540 and 387 oz.in motors. I also have some 570 oz.in motors as well which I will try on the G540.

[phomann]

The Y axis is coming along.

Once again it is measuring the sizes and mounting locations by measurement. What I doing is taking measurements with vernier callipers, then with the understanding that it is a metric machine and assuming that the dimensions are in metric, round the measurement to a reasonable values based on what I feel an engineer would have chosen.

So far it has worked out pretty well. The X_Axis end plates were spot on, The mounting holes for the plate were 160mm apart. The ballscrew centreline is 50mm below the mounting hole centreline. The ballscrew centreline is 15mm above the saddle, etc.

The Y-nut holder has a 15mm boss that locates into the saddle and is held there by a M8 capscrew. The original parts can be seen below.

[phomann]

For the new Y-nut holder, I used 25x40mm aluminium bar. It is just wide enough to slip down through the slot in the base.

Since the screw is 16mm compared to the 25mm one that came with the machine, the hole is the front of the base where the screw attaches to the bearing block has quite a lot of clearance for the new screw.

As the saddle extends a long way (40%-50%) over the front of the base I decided to take advantage of the smaller screw diameter and lower the new screw centreline by 4mm. This allows me to have more clearance and fit a larger pulley along with the pulley and belt cover without worrying about any binding resulting in reduced travel.

To make the Y-nut holder, I chucked the 40x25mm stock into the Taig 4 jaw chuck mounted on my 7x14 CNC lathe. Once the stock was centred, I turned the boss, drilled and tapped the M8 mounting hole.

While still in the chuck, I used the tool-bit to scribe the centreline down the side of the stock. This was to ensure that when the work piece was transferred to the mill I could easily locate the mounting boss centreline.

Once in the mill The centre hole and mounting bolt hole pattern was centre drilled. The bolt holes were then drilled countersunk and tapped for M5 bolts.

I then used a drill press to drill the centre clearance hole, stepping up in drill sizes until the 20mm hole was drilled.

Unlike the X-Axis ballnut which went smoothly, I had all sorts of chatter with the 20mm drill bit (previous drill was 1/2"), with the MT3 taper chuck holder dropping out of the quill, and/or jamming in the work- piece.

As you can see, I has to flip the ballnut over on the screw to mount it the way I need it.

[Fastest1]

Nice work. I think you will find comparable results with all controllers if the voltages are the same. I have 3 different ones and I have run a few of my machines on all of them. I cant tell a difference. They have different options and packaging. Hobby Cnc Pro, G540 (2 actually) and a PMDX-126/5056D combo.