Project Frankenmill: Sieg X3+Tormach 770 Spindle Head

[Spk64]

I purchased this from McMaster.

http://www.mcmaster.com/#8082k85/=pvrodj
16awg 4 conductor, shield and drain. 3.86 per foot.
Has been working well on my router with lots of movement in the cable.

Sent from my iPad using Tapatalk HD

[TangentAudio]

That stuff looks good, thanks for the pointer. I might consider their continuous flex 18AWG 4 conductor if I get around to rewiring my X/Y steppers. When I first built the machine I was cheap and used 4 conductor 18AWG stranded in-wall speaker wire from Monoprice. It's held up OK for four years, but it's really not the right cable for the job.

[TangentAudio]

Tried a little shimming of the head this morning, and I think I managed to correct the fore-aft error. Before fixing it was out 0.004 to 0.005" over 7.5". Now, with 0.003" brass shim stock under the lower bolts, it is better than 0.0005" over 7.5".

[uawgmsmco]

I love living vicariously through other peoples projects. This has been an awesome build

[ChrisAttebery]

Nice job. I'd love to have a 10k spindle on my machine, but that head is out of my price range.

[TangentAudio]

Thanks for the kind words, guys. I'm not sure my workmanship is worthy of such praise, but I'm trying new things and learning and that's what matters to me.

ChrisAttebery: Definitely... I don't recommend this as a cost-effective way to go. It's a silly project in a lot of ways, but it's fun at least.

[TangentAudio]

Apparently I was good this year, because Santa showed up a little early. This actually was delivered on a Sunday afternoon by UPS. Guess they are working some overtime to get holiday packages out!


As you can see it's a customized enclosure from Hoffman (brand new old stock). It has a bunch of custom cutouts: (12) 7/8" holes, and four cutouts for what look like oversized electrical outlets. I am thinking that I will make some blank cover plates for some of the openings, and possibly connect/enlarge some of the other ones to fit a nice big aluminum panel for all of my connectors. I also tracked down a good ebay deal on an internal plate, which mounts down in the bottom of the enclosure and is intended to be where you mount all of the guts that go into your enclosure.

The two lengths of Thomson ball screw, and a new ball nut have also arrived. I have to start making a plan for how I'm going to do the end machining. I bought an extra length of ball screw so I could practice and make mistakes on one if necessary. The ball nut is to potentially replace the one on my Z axis, which has never seemed quite right.



I hope everyone that is celebrating has a happy and safe holiday. Cheers!

[nateman_doo]

Dumb Question... what did you do with your X3 head?

[nateman_doo]

I ask because I am going to rebuild my X2. Its not really a "rebuild", more like replacing the X2, with another very similar build to my tormach build. If you want to sell it, I hope you keep me in mind. I live in Jersey, so I am sure the shipping would be brutal.

[TangentAudio]

Wow. It's been almost a decade since my last update on this thread.

Not long after I grafted the Tormach 770 head onto my X3, I ended up getting seriously involved in my local maker space. I helped the space purchase a Tormach 1100 in 2016, and I taught CAM and CNC machining to members for a few years. Having access to the 1100 meant I had less need for my own machine at home, but I still used it occasionally, mostly when I didn't feel like making the drive to use the 1100. But the flaws, particularly in the Z axis dovetail ways and poor gib fit, continued to plague the machine.

I've been dusting off old projects lately, and I got the urge to continue where I left off all those years ago and see if I could pull off a linear rail conversion on the column to help this issue. On the plus side, I've improved my machining and design skills over the last decade. On the minus side - is an old converted X3 even worth the time or money anymore in 2023? Probably not. But the machine has been a great platform for learning, so pushing my design/machining/metrology skills with a machine that already mostly works, that I already have in my shop seemed sensible. The alternative path I was considering was a new conversion of a different machine, or even scratch-building a machine. Both more ambitious projects for sure.

Another plus is that I now have access to a lot of machines that I did not back in 2013 - the aforementioned Tormach 1100, a ProtoTrak BridgePort mill, and an ProtoMax waterjet... This would be a good project to put those all to use.

So a bit of a review... The X3 has a poorly designed Z axis drive setup for CNC use, with the leadscrew/ballscrew located in the back of the column and plates that wrap around the column to drive the slide in the ways on the front of the column. This, combined with the generally poor fit of the dovetails and gib, create a situation where the Z gib adjusters must be left fairly loose to avoid binding. The ball screw tends to lead the head due to the cantilevered nature of the setup, and with the loose gib, it tends to make the head "nod," or rock in the slideways when Z moves. It's a major contributor to poor surface finish.

Here are a couple photos from my initial conversion in 2009.





And here we are with the head removed again in 2023.




I briefly considered making a whole new column from scratch, but abandoned that idea in favor of modifying the existing one. And instead of milling off the existing dovetails, as is common, I got to thinking about spinning the column around 180 degrees and mounting rails to the column sides somehow. This is the basic starting point with the flipped column.



Initial considerations as of March 2023...

• how/where to mount linear rails
• re-use old Thompson 16mm dia ballscrew or upgrade?
• new location for Z axis motor?
• possible to improve torsional rigidity of the "open box" design of the original column by partially closing it in?
• possible to improve overall rigidity by improving/expanding mounting surfaces between column and base? adding stiffening members to the column itself?

The project has been moving forward since then - I will make a series of posts to catch up with where I am now.

[TangentAudio]

I mainly considered two approaches to mounting rails to the column.

Approach 1 - add some kind of additional side supports to the existing column and mount the rails to the supports. A simple plate would be all that's needed to mount the head to the blocks.



Approach 2 - mount the rails directly to the column and build a more complex saddle assembly to attach the head to the blocks.




Ultimately I decided to go with approach 1 even though it uses more material overall. It seemed like it would be quite challenging to get the rails aligned in approach #2.

I spent some time debating using steel vs aluminum for the added supports. After searching for drops from metal vendors, I ultimately settled on using 2024 aluminum rather than 1018 steel. Some compromises obviously - different coefficient of expansion than the cast iron as the parts change temperature, for one. And not as easy to do traditional scraping for the rail mounting surfaces in aluminum vs steel. But availability, weight, cost, and ease of cutting and machining all won out for me. I figure that it's a low precision hobby machine - and all of the other conversion parts are already made out of aluminum. Any thermal problems I create are likely to be overshadowed by all of the other issues the machine will continue to have even after fixing the Z axis.

Based on what materials were readily available as drops from the metal yard, I ended up going with significantly larger dimension material for the side supports than the initial sketch above - settling on 1.5" x 2.5" bar stock. I evolved the design to this point:



This approach might add some rigidity to the upper part of the column (especially if I can partially close in the front side) - but it's limited by the weak bottom section where I needed to allow room for the original bolts to attach to the base. I have some things I'd like to explore that might help with rigidity in that area - basically some stiffening plates that would mount on the back of the new column, and potentially attach to the base at some newly created mounting points.

[TangentAudio]

Materials in hand. 1.5" x 2.5" 2024 bar stock, and a 1" thick chunk of 2024 plate for the Z saddle.



Machining flat (ish) one side of the bar stock on the Bridgeport with a 3" shell mill.



My milling setup could have been better for sure. I did get some 'lift' at the two ends that were outside of the vice and held up by 1-2-3 blocks and machinist jacks. Also potentially an effect of relieving stress in the bar stock. The side rails will be held on with ten M10 flathead cap screws so there will be significant force to hold it to the column and flatten it back out, hopefully not distorting the column too much in the process. Shown here against my 24" straight edge.



Layout for the bolts.



Drilling and countersinking.



Transfer punching the bolt pattern to the cast iron column.



Drilling the column bolt pattern.

[TangentAudio]

Drill, drill, drill...



Tapping the M10 holes in the cast iron column, using the new side bars as a tap guide.



With both new side rails attached.



And a detail showing the same old weak interface between column and base. It does look a bit silly with the huge side bars on it now.

The base is hollow underneath the column, as this was originally the place where the bevel gears for the Z screw meshed. Adding a plate and/or possibly filling with E/G might help a bit with rigidity and damping.

[TangentAudio]

Before attaching the side bars, I had done some measuring of the column on a granite surface plate and was able to confirm that the original precision ground surfaces on the front and back of the column were in fact parallel and true within reason. One corner had a slight dip, but clearly these surfaces had been machined and ground in the original setup to cut the dovetails. Years ago I had confirmed that the column was relatively square to the Z travel within reason, so I was able to use these ground surfaces as references for setup and machining.




For extra security, I decided to epoxy the side bars to the column so they would not move. I figured the epoxy would also help fill in the troughs of the roughly fly cut cast iron surface and provide a more solid connection. I just used JB Weld here, nothing exotic.



Once cured, I set up a 2-4-6 block with an indicator to slide along the front (formerly rear) reference surface of the column, so I could measure the position of the side bars relative to one another and find the high and low spots. Since I did the bolt hole work in a low-precision way with traditional layout, transfer punches, and a drill press, I knew it would not be that precise. But I also knew the final surfaces would be machined for flatness. These measurements were used so I knew how much I needed to machine off to make both rail mounting surfaces parallel to the reference and with each other.



I set up on the Bridgeport and machined clearances on the back side so I would be able to flip the column and support it on 1-2-3 blocks on the old precision dovetail surfaces, without interference from the newly added side bars sitting slightly askew.



The extra holes on the back side of the side bars are for potential future stiffening plates or fixturing bolts for E/G.

[TangentAudio]

Milling the clearance on the back side of the side bars. Using a 1 inch indexable end mill with two APKT aluminum inserts, and the old ProtoTrak in basic power feed mode for the long cuts.

One "gotcha" is that the X travel on our Bridgeport is nominally only 24" (610mm), limited in part by the motor for the ProtoTrak setup. I chose 640mm linear rails, and the aluminum side bars are even longer still. So it's not possible to mill the whole back side for clearance in a single setup.

Thankfully after doing this operation, I discovered that it's possible to squeeze about 26" of travel out of the X axis with care. And the 1" diameter end mill gets some additional clearance at the two ends. That was a relief, as it would make the critical machining for the linear rails much more simple.

[TangentAudio]

Now on to the critical milling to create the mounting surfaces and reference edges for the linear rails.

First thing in the morning, I spent a ton of time tramming and re-tramming the head of the Bridgeport to get it as close as I could within the limits of my measuring tools, skills and patience. Once that was done, it took quite a while to get the column positioned on the bed of the Bridgeport in the right spot so I could both securely clamp it down and also access everything that needed to be milled in one setup. Our Bridgeport only has a 42" table so this was a bit trickier than it sounds.

Of course, it also had to be set up with the column reference edge square to the X axis travel. I used a combination of 1-2-3 blocks and my 24 inch straight edge to "project" the ground dovetail surface out so I could run a DTI along.



I must have spent 3-4 hours working on the setup, but it's certainly the most critical part. The actual milling and drilling is easy once the setup is done right. I was so focused on the task that I didn't snap many photos, but here's one during tapping of the rail mounting holes.



And the end result rail mounting surface and reference edges on the two side bars. I made extra clearance at each end to squeeze out every drop of X travel + end mill diameter to fit the rails.



Test fitting the rails.





Rough checks for flat and coplanarity.

[TangentAudio]

I did an initial mounting of the rails to check for parallel when using the two reference edges that I milled. With minor tweaking of the secondary rail, things look good at first glance. This is only a 0.0005" indicator, and there is a tiny amount of movement probably on the order of 0.00015 (~3.8 um).



I didn't spend a lot of time with this because I had to pull the rails back off to do more machining for the ball screw.

[TangentAudio]

I considered re-using my original Thomson 16mm 0.200" pitch ball screw from the original CNC conversion kit that I used, but the ball nut wasn't sounding particularly good after the years of use. The screw also seemed to have developed a slight bend, perhaps as a result of the buckling load of the heavier Tormach head. The final straw was stumbling across the original Acme lead screw from the machine's manual days, and it was much larger - about 25mm.

A 20mm ball screw probably would have been a sufficient upgrade to prevent buckling, but given the small price difference to step up to a 25mm, I opted for that based more on gut feel than calculations. The screw radius has a 4th power effect on the inertia of the system, so it can have a large impact on required motor to achieve the desired acceleration. I already planned on an upgraded Z motor for this conversion, so hopefully I'd be able to account for that. I chose a 5mm pitch, which is close to the original 0.200" pitch of the previous screw.



I opted for an affordable rolled screw, figuring that this machine is really not worthy of expensive ground screws. I did buy from a vendor that matched ball nut to the screw, claiming 0.001" backlash as shipped, presumably by hand-sizing the balls loaded into the nut. I ordered the screw with the ends machined to match a pair of genuine Taiwanese SYK BF20/BK20 end supports.

My intent was to use the original support mounting positions, but I'd have to account for the different bolt hole spacing of the larger supports. I decided to try filling the original holes with JB weld epoxy and then drilling/tapping new holes. There would be some overlap of the new and old holes, which caused some minor headaches.



Due to the tight quarters inside the walls of the column, I couldn't come up with a workable combination of collet or chuck to hold a center drill to start the new holes in the proper bolt locations. I forged ahead and tried it without center drilling first and had a host of nightmares with the drill wandering. Surprisingly the wandering was not when making the initial hole, but as I drilled deeper into the cast iron. I think some of it was related to the epoxy-filled holes as mentioned above - it tended to wander to the softer material. I also think some was my inexperience with feeling out the required quill pressure and peck depths deep-drilling in cast iron. Suffice it to say the holes wandered more than I was happy with, but I did manage to get them tapped and I think they are still true enough to use. I had to abandon one hole and 'erase' by plunging a 0.5" end mill to remove the attempt. I will have to plug this with a turned piece of steel and loctite and try again.



Next up, I needed to create clearance for the end support and ball screw to project through the top of the column. I was reluctant to do this, since I knew removing that much material would reduce the rigidity of the column, but I was confident I'd be able to gain that back by adding a plate on the new face of the column afterwards.



After that, this is what I was left with.



And here it is with the ball screw, nut, end supports loosely test fit.



The original CNC conversion had a NEMA23 stepper located on the bottom of the column, driven by a belt. I'll be moving to a directly-driven NEMA34 at the top of the column.

[TangentAudio]

I chose a 4.5Nm (637oz-in) NEMA34 closed-loop stepper with an electromagnetic brake from StepperOnline. I'm hopeful that my back-of-the-napkin motor calculations will work out, and I'll be able to hit my desired acceleration/velocity and meet the torque needs.

The brake will be a nice addition - in my original conversion I did not have a brake, so the head had a tendency to drop when the power was removed. I had a precisely calibrated stick that I would jam under the head to stop this, but it was a nuisance.

I tested the motor and driver out on the bench by writing a quick bit of Arduino code.



I took a chance and ordered a $17US coupler from Amazon, to couple from the 14mm shaft of the motor to the 17mm turned end of the ball screw. Unfortunately the runout was truly awful and I returned it. I ended up ordering a much more expensive one through Misumi to get something that ran true. Obviously the OD of the coupler doesn't strictly need to be concentric with the shaft bore, but the whole thing wobbled under testing, so I suspect this runout might be real. I have a feeling these might be made with a standard bore size and then tossed into a 3 jaw chuck and bored to different ID's, with little attention to concentricity.

[TangentAudio]

Now, making the stepper mounting plate for the new Z stepper.

We recently got a fun new toy at our makerspace, a ProtoMax waterjet. Limited to cutting in a 12" x 12" area, up to 1" thick. A wee baby waterjet, for sure, but perfect for a project like this.

I figured I'd test out the machine to rough the blank for the stepper mount, and do finish machining such as final bores and counterbores on our Tormach 1100.

I bought a piece of 1" thick 2024 plate to make the plate that mounts the head to the linear rail blocks. The piece was large enough that I could also make this stepper mount from it. I would have opted for thinner stock - 0.500" would have been fine, but I didn't have anything convenient on hand. Besides, this would be a good test cut, since I intend to use the waterjet to rough the head mount plate soon.

This took 81 minutes to cut, but that's mostly just babysitting the machine to make sure it doesn't run out of garnet and there are no other problems.



Cut quality was much better than I expected.



Better taper on the cut than I expected from a simple 2 axis machine too.



And here it is sitting roughly in place, before final machining.




This plate is designed to allow two-axis adjustment via slots, to make the stepper and ball screw shafts concentric. Coaxial error can be corrected by shimming if necessary, if it's outside of what the coupler can handle.




And there we have it, we're caught up to present day. A couple months of work, after ten years of silence. Will anyone even notice this thread that came back from the dead? Stay tuned...