LowRider v4 Rebuild – Turns out I did mess up the first time

LowRider CNC Your Builds
1

LowRider v4 Rebuild – From Full Sheet to Quarter Sheet

After living with my full-sheet LowRider v4 for quite a while, I’ve started seriously considering rebuilding it around a quarter-sheet (24" × 48" usable) work area instead of a full sheet.

This isn’t because the LR4 has been disappointing—quite the opposite. It’s because actually using the machine has taught me where my priorities really are.

Why I’m Considering Rebuilding

When I originally built the machine, I assumed I would regularly put full sheets of plywood on the CNC.

That hasn’t happened.

The only time I’ve actually machined a full sheet was when I cut the strut plates used to build the machine itself.

Everything else starts the same way:

  • Buy a full sheet.

  • Break it down with a track saw.

  • Put manageable pieces on the CNC.

The track saw has essentially replaced the need for a full-sheet CNC in my workflow. It’s faster, safer, and easier than wrestling a 4×8 sheet around the table.

Meanwhile, the machine occupies roughly a 5’ × 10’ footprint in an already crowded shop.

The LowRider was designed so the machine can be removed from the table and stored when not in use, but my build doesn’t really work that way. Early on, I designed and published a full Unistrut table system for the LR4, and that’s the direction I chose for my machine. It has proven to be extremely rigid, but it’s also a permanent installation. Between the Unistrut skeleton and the 3/4" MDF spoilboard, the table is a two-person lift, so the original “store it against the wall” concept no longer applies. Here is the link to where I started designing the Unistrut adapters:

Credit Where It’s Due

I also want to give credit where it’s deserved.

The more I’ve thought through this rebuild, the more I’ve come to appreciate just how well Ryan optimized the LR4 for its intended audience.

There are countless little engineering decisions throughout the design that don’t immediately stand out until you’ve built and lived with the machine for a while.

The printed parts are a great example. The internal load paths, hidden geometry, wire routing, and thoughtful nut and bolt placement make the machine remarkably easy to assemble and maintain while remaining surprisingly strong.

The moving gantry design is another one. By raising and lowering the entire gantry instead of building a tall Z-axis, the cutter stays close to its support structure, keeping the Z assembly simple, compact, and rigid.

Even the removable machine concept deserves recognition. I completely ignored that design feature by building an oversized Unistrut table that weighs a ton, but that’s on me—not the design. If I had built a lighter table like Ryan intended, I could simply lift the entire machine off and store it on a shelf when it wasn’t needed.

None of the ideas in this thread are meant as criticism of the LR4.

They’re simply an attempt to adapt an already excellent design to my own shop, my own workflow, and the tools I now have available after several years of using the machine.

I think that’s actually a testament to the design. The LR4 has given me a platform that has been reliable enough to make me think carefully about where my compromises should be, rather than where the machine’s compromises are.

Why Quarter Sheet Makes Sense

Reducing the machine to quarter-sheet size solves several problems simultaneously.

First, it dramatically reduces the amount of floor space dedicated to the CNC.

Second, it shortens the gantry considerably.

That matters because the gantry is the largest structural member in the machine and therefore the largest source of deflection and torsional twist.

Shortening the span improves stiffness before changing a single material. That single change is what made me step back and rethink the entire machine. Once the gantry became shorter, several ideas that weren’t practical on a full-sheet machine suddenly became realistic.

My goal is also shifting toward aluminum machining. Every increase in rigidity directly improves surface finish, depth consistency, and chatter resistance.

The only capability I would lose is slab flattening or occasionally machining long stock.

Fortunately, my existing table construction already suggests a solution.

Unlike the standard LR4 table, my machine references everything from the Unistrut frame. The Y rails are located from the frame itself rather than the spoilboard. That means if I keep those mounting locations fixed, the frame becomes the machine’s permanent datum. The spoilboard can be replaced, and a bolt-on extension can be added, without disturbing the primary alignment of the machine.

My current thinking is to build the quarter-sheet machine first, then create a bolt-on extension that allows the machine to return to roughly a 24" × 98" work area when needed. Alignment features between the Unistrut sections would ensure repeatable assembly while allowing the everyday machine to remain compact, rigid, and optimized for the work I actually do.

Even if I ultimately decide not to build the extension, I don’t consider that a failure. The goal of this rebuild is to make the machine better for the work I do every week, not preserve a capability I rarely use.

Hidden Belts

The table redesign also gives me the opportunity to incorporate the hidden belt modification.

Since I’m rebuilding the frame anyway, I can position the Unistrut specifically around the belt path instead of designing around the original exposed layout.

Beyond simply protecting the belts from dust, chips, and accidental damage, routing them inside the structure cleans up the machine considerably and removes one more exposed component from the work area. It feels like one of those improvements that makes sense to include while everything is already apart instead of treating it as a future retrofit.

Rethinking the Gantry

The standard LR4 relies on printed triangular strut braces connected by front and bottom strut plates.

Those plates are more important than they first appear. Together with the braces they create a deep structural section that resists gantry twist.

The EMT tubes provide both the bearing surface for the bearings and part of the structural beam.

For most builders, I think Ryan found an excellent balance between cost, simplicity, and performance.

My situation is a little different.

I already own a plasma cutter, drill press, welder, and enough 1/8" × 3" steel flat stock to fabricate replacement strut plates.

Rather than asking the CNC to machine replacement parts, I can use my existing MDF strut plates as precision templates.

My fabrication plan is:

  • Cut the steel into two matching blanks.

  • Tack weld them together.

  • Transfer punch the first pair of locating holes.

  • Drill those holes through both plates.

  • Bolt the MDF template to the steel using those holes.

  • Continue transferring and drilling the remaining holes while the template is positively located.

  • Plasma cut the outside profile.

  • Finish to the line with grinders.

  • Separate the finished plates.

The hole locations are the critical dimensions. Everything else can be fit and finished from there.

The outside profile simply needs to provide proper clearance and enough material around the mounting holes.

Steel also changes how the assembly carries load.

Instead of relying primarily on plate geometry, the plates themselves become significant structural members capable of carrying much more of the load.

Tube Selection

Forum testing has shown that thicker-wall tubing is noticeably stiffer than standard EMT, particularly on longer machines.

Because this rebuild shortens the gantry so dramatically, the economics change.

Instead of needing roughly ten feet of tubing, I only need enough for two rails around 30.6" long.

That suddenly makes higher quality tubing much more practical.

At the moment I’m leaning toward staying with the standard two-tube design while sourcing the straightest, roundest, thickest-wall tubing I can reasonably find, whether that’s DOM or another suitable mechanical tube.

I’ve also looked into the three-tube modification discussed on the forum. Based on what I’ve read, and Ryan’s comments over the years, I don’t think it’s worth adding the extra complexity for this build. On a much shorter gantry, I think better tubes combined with stiffer plates is probably the cleaner solution.

Why I’m Chasing Rigidity

One thing that often comes up in CNC discussions is diminishing returns.

I think CNC machines are a little different.

The controller only knows how many steps it commanded.

It has no knowledge of structural deflection.

It doesn’t know if the gantry twisted.

It doesn’t know if the belts stretched.

It doesn’t know if the frame moved.

Every source of mechanical error increases the difference between commanded position and actual cutter position.

None of those errors exist in isolation.

A couple thousandths of frame movement combined with a few thousandths of gantry twist, belt stretch, spindle runout, and cutter deflection become one larger positioning error.

Reducing any one contributor reduces the total accumulated error.

I’m not trying to build a machining center out of a 3D printed CNC.

I’m simply trying to remove as many unnecessary sources of error as practical while staying within the spirit of the machine.

Current Direction

At this point my plan is to:

  • Rebuild around a quarter-sheet working area.

  • Retain the Unistrut frame concept while making it modular.

  • Incorporate the hidden belt modification as part of the table redesign.

  • Keep the machine optimized for daily use rather than occasional large jobs.

  • Fabricate new 1/8" steel strut plates.

  • Source higher quality tubing if reasonably priced.

  • Improve rigidity wherever practical without adding unnecessary complexity.

Looking for Real-World Experience

Before I start cutting steel, I’d appreciate feedback from anyone who’s gone down a similar path.

I’m particularly interested in:

  • Quarter-sheet LR4 builds

  • Steel or aluminum strut plates

  • Better tubing on short gantries

  • Modular tables

  • Hidden belt implementations

If you’ve tried something similar—or decided not to—I’d like to hear why.

I’m not trying to redesign the LowRider from scratch. I’m trying to build the version that best fits the way I actually use the machine.

2

One thing this planning process has taught me is that there are a lot of possible upgrades for the LR4. The trick isn’t figuring out what can be changed—it’s deciding what actually provides enough benefit to justify the added complexity.

Here are a few ideas I’ve considered and, at least for this rebuild, decided against.

Three-Tube Gantry

I spent quite a bit of time looking into the 3-tube gantry mod as well as Ryan’s comments from the LR3 development and forum discussions.

From everything I’ve read, the strut plates and triangular braces already do most of the work in resisting twist. On a quarter-sheet machine with a much shorter gantry, I don’t think the added complexity and weight of a third tube is justified.

I’d rather put that effort into better tubing and stiffer strut plates.

Completely Redesigning the Strut Plates

The temptation is there to start from a blank sheet of paper.

I’ve decided not to.

The existing geometry has thousands of machine hours behind it. I’m keeping the important features exactly where they belong, especially the hole locations and printed part interfaces.

The steel plates will simply be optimized for my fabrication methods rather than trying to reinvent the design.

Linear Rails

I’ve thought about replacing the EMT with linear rails.

I still think that’s a great long-term direction, but not for this build.

The current machine architecture works well, and I want to see what a shorter, stiffer version can do before abandoning one of the core design features of the LR4.

Ballscrews

The shorter machine suddenly makes ballscrews much more practical from both a cost and sourcing standpoint.

Even so, they solve a different problem than I’m trying to solve today.

Belts have proven to work very well on the LR4. Before replacing them, I want to eliminate the larger structural sources of error first.

Rack and Pinion

This is another idea that keeps coming back.

Long-term, I think rack and pinion has a lot of potential on the Y-axis.

For now, though, it falls into the same category as ballscrews. It’s a much larger architectural change than I think is necessary to accomplish my current goals.

Moving the Controller Off the Gantry

I’ve also considered relocating the controller and all associated wiring to a fixed location on the table instead of carrying it on the gantry.

There are definitely some advantages:

  • Less moving mass.

  • Easier access for maintenance.

  • Simpler cable management.

  • Better protection from dust and accidental impacts.

The downside is significantly more wiring to accommodate the moving axes.

For now, I don’t think the performance gain justifies redesigning the electrical system, so the controller will probably remain on the gantry. It’s something I may revisit in the future if I decide to substantially redesign the electronics.

Why I’m Stopping Here

One thing I’ve learned from this project is that it’s very easy to keep chasing “just one more upgrade.”

At some point, improvements stop making the current machine better and start turning it into a completely different machine.

My goal with this rebuild isn’t to redesign the LowRider.

It’s to build the stiffest, most capable version of the existing design that fits MY shop, MY workflow, and the tools I already have.

3

Another area I’m taking the opportunity to rethink is everything that supports the machine. One advantage of rebuilding from the ground up is that I can incorporate features from the beginning instead of continuously bolting things onto the outside.

Air System

Since aluminum is one of my primary goals, compressed air becomes much more important than it has been for woodworking.

The machine will have a dedicated air supply integrated into the table itself:

  • Shop air from my existing hard-plumbed system

  • Through my existing dryer before reaching the machine

  • A dedicated filter/regulator mounted directly to the table

  • Quick disconnect for easy removal if necessary

That gives me one clean, dry, regulated source specifically for the CNC. From there, a directed nozzle — most likely Loc-Line — keeps air on the cutter where it actually matters for aluminum work.

Lubrication

Since the air infrastructure is already there, I’m planning to add an inline fogger on the air blast line for mist lubrication at the cutter during aluminum cuts. It won’t replace proper maintenance, but it makes consistent lubrication a feature of the machine rather than something that depends on me remembering to do it manually.

Dust and Chip Collection

This rebuild is also a good opportunity to separate woodworking and metalworking chip collection.

The CNC will move to a dedicated shop-vac style extraction system. For wood it still handles chip collection well. For aluminum it keeps metal chips isolated from the main dust collector — and away from an impeller that also sees fine wood dust.

The large dust collector stays dedicated to stationary woodworking machines where it belongs.

4

I think your analysis is pretty thorough, but I don’t agree with those points being advantages, but rather disadvantages. One power cord to the gantry makes way simpler cable management then 5 motor cables and endstops along the Y axis. The back of the gantry can be fitted with the tilting hood covers and that is super easy access for maintenance. Drive the gantry by hand or by its own power to an easy spot to access and it is up on top of the table not mounted below the table off in the distance that is difficult to access. With the speeds and acceleration we run on these, the mass of the controller and power supply isn’t a huge factor in comparison to the router or spindle. My system has skipped more on the X axis than the other axes and the controller doesn’t have any bearing on that.

I didn’t mean for this to be nitpicky, but perhaps it came off that way. I think your priorities are well stated and I’m interested to see what you find as you implement it. My system is roughly 20x30 and while it has cut aluminum and worked ok, I think your upgrades would help it be better, so I watch with interest.

5

Most of the “hidden belt modifications” I’ve seen on the forum modify the geometry in a way that will reduce machine rigidity. If you’re aiming for increased rigidity as an objective, you should give that serious consideration.

What evidence do you have that the strut plates themselves are the limitation (vs say plastic deformation in the braces)? You’re about to embark on a lot of fabrication where any error stacks up in a way that further degrades the machine.

Over in the lounge / LR beta areas we’re actively discussing what things might significantly improve the already excellent design of the LR4. Switching from MDF strut plates to steel strut plates isn’t one of the items we’re thinking through.

That’s backed by at least some measurements Ryan has been making with instrumentation to identify where the rigidity can be improved reasonably.

Your thesis that shortening the gantry is a major win is spot on. I’m not at all convinced that going to steel strut plates is going to make any meaningful difference to your end machine.

That’s in the same category. One of the LR4 major improvements comes from going one size up on EMT. OD of the tubing does most of the work. You’re not going to get the major improvement you think you are by trying to substitue the tubing.

Engineering without numbers is just an opinion.
I’m hoping @vicious1 will jump in here and tell you about some of the numbers he’s been seeing…

Chasing zeros is why that comes up. CNC machines aren’t different, they’re the primary example.

Edit:

I can tell you that some future machine may well be using R&P drives. Belts do represent one area for improvement, and after playing with a LOT of belts, it looks like what’s on LR4 is already about as optomized as it can be. Ryan did a great job.

Orob is spot on here. For cutting aluminum or hardwoods or other materials where we need all of the torque on the X axis, this is an area for optomization. Y and Z are dual motor, and Z has a screw so torque there isn’t an issue. X is single motor, so looking at drive current and more powerful NEMA 17s is an area based on your description of machine needs might be worth looking into.

6

Firstly, thank you @spsully2582 for creating this topic, and openly sharing detailed considerations, ideas and progress on various aspects your journey towards a setup that’s right for you, and maybe others too.

Looks like you’re planning to mill metal with air/lubricant. So, consider conformal coating controller board if you decide to keep on the gantry, since metal chips flying onto the controller boards is known to kill them.

Rob’s suggestion resonated with me. After one too many jobs failing because power cord was yanked (was previously running router and PSU cords to stationary power outlet extension, apparently a few wraps of duct tape wasn’t enough to keep the cord plugged into the outlet).

Ended up less compact, but very functional setup…

17824387303251983810718511641849
178243873032519838107185116418491920×3413 1.2 MB

17824388262345960690480568752290
178243882623459606904805687522901920×3413 1.04 MB

Slightly oversized box mounted (with 2020 attached to bottom strut) for future gcode activated relay switched router. But honestly, turning router on/off manually hasn’t frustrated me enough to bother with that.

I didn’t see the data/summary of the belt investigations. Feels like investigations are early stage, nothing is decided, changing or being released anytime soon. My hope was, similar to EMT, data would show that moving up to larger belt size and/or doubling belts would be good bang for buck upgrade (easier to pack, ship, assemble, calibrate than R&P?). Vaguely recall seeing a kick ass 3D printer at 2024 RMRRF (near rackrobo crew back when RMRRF was just one hall) that had impressive perf using wider belts (15mm or 20mm?).

7

This is the most important aspect of this type of machine. Any machine in any price category. Size matters. So many people build a giant machine, “just in case”. Well that sacrifices speed 99% of the time for that one time you might need to cut something giant. If that comes up I suggest using a tiling method, pins or something, or just making it a two piece design. I have owned both 1/4 sheet and full sheet and 1/4 is significantly faster. This topic used to be in the instructions for the MPCNC until the LR was created. Everyone just kept adding size “just in case” I have to sit back worried that they make it so big it is not as fun for the normal projects.

The Beta testers get all the credit here. They never let me push a decision through until we have all voiced our opinions…on every tiny little aspect. It ends up with a great machine but it can be a grueling process for us all.

Torque more than twist but we are working on that.

It is a balance. The screws transfer a lot of the load to the braces, followed by a tight fit, but ultimately the printed braces carry most of the load. You can use 2" steel plates and you will just be adding mass at some point. Upgrading one piece will help some, but it will take a redesign to really get it down further.

If we were fighting a 30mm flex it is one thing but we are fighting tenths of a MM, I found my table arms actually flexed 0.1mm under normal test loads…I was absolutely shocked.

Steel plates will help a little, but will also cost you acceleration.

That is an odd statement to dissect. Yes thicker wall tubing is more rigid as a statement that is correct. Adding ID to the beam rails has an extremely fast diminishing return on the beams overall structure. The beam is a lower steel rail, then printed braces, then screws and nuts, then strut plates, with a tiny bit of added support from the top rail…from the rail the other way we have bearings, bolts, core, screws tool mount, tool, motor shaft, collet, endmill. Each one only adds a tiny bit to the equation. Adding a twice as rigid tube will not double the stiffness of the beam to tool assembly.

You can do tube load calcs and watch the rigidity with a 9lb load central on the rail.

I am 100% for this, but know there are sooooo many factors stacking up there is 100% no one smoking gun. As I mentioned before, I found my Y min table wind to be a 1/6 contributor to my overall error at max tested load. If we had a smoking gun, low hanging fruit we would have fixed it.

With that said please please please keep going with this. I am very actively knee-deep in LR5 R&D. I know I can not see it all and can’t test it all so the more you test the less I need to. Just please be extremely cautious “announcing your findings”. People used to release parts as a new revision EG “better LR4” or “the LR5”. I suggest sticking to very carefully logged numbers and descriptions of how you tested them. As an example of how this can commonly go wrong, a fresh build is more rigid than a used and abused build. That can be a large false positive.

This is simple.

Smaller builds are more rigid, shorter belts, smaller table, smaller beam.

Metal strut plates do help, how much is undocumented, but will obviously help.

Tubing any thicker than the 1.5 to 2.5 wall thickness we already suggest will probably not result in a very measurable improvement on a quarter sheet build. a rail that is 1/4 as long is 64X as rigid.

The closer the belt is to the load the less of a lever arm it creates on all the other parts. Moving the belt away from the endmill is a bad idea.

8

You either need to mount them on something to make them more stiff, or get very large linear rails so they are at least as stiff at the Conduit.

I am pretty sure we are going this way for X and Y. Ballscrews are not magic, they are a great option that cost a lot of money and need a lot of maintenance.

That is only okay on a small machine, a full machine has too long of wire runs to make that much of an option.

I’m excited.

9

Haha, yes. Not looking at anyone… :face_savoring_food::face_blowing_a_kiss:

10

Whoa!!! Thats a lot to process….. I guess I need to go in chunks.

That sir, is the reason I refuse to get on a pedestal and preach anything as fact. I have never run a fancy industrial machine or even a high end hobby machine. My experience started with an abandoned LR3 build, then this machine. All of my evidence even if I had it down to microns, would still be based on what I “feel” and theories I come up with. “I think my low rider should cut steel” is a great theory, but its not a fact nor do I think its even possible yet. So please, take my ramblings as just that. The ramblings of a guy who wants to try and use premium materials to try and get more out of the only machine he has ever used.

That being said, the strut plates are not why I’m changing things, just one of the elements getting changed while I’m at it. More on this in a later chunk

Now that we have established I’m not some whiz kid, my question on these two responses are about clarification. I don’t plan on changing the X belt, just the 10ft long Y belts. I’m not sure if I made that seem otherwise, but I have no intensions of even trying to figure out the X belt.

It could be that you are talking about the Y belts as well, at which point, I dont understans how those impact the gantry. I guess having them lower might create an inward pull at the bottom and in turn almost try to pull the gantry apart length wise. But I would love to hear more about this.

My theory on this was simply to get electronics away from dirt, oil, debris, etc. I haven’t really looked into it too much. But I seem to remember there being a handful of guys running table mounted boxes on the LR3 and that seemed to make a lot of sense to me. I have seen numerous algorithm based ads showing machines costing several thousands, and they all seem to have the control box separate and send these aircraft looking cables through the drag chains. Again, this was a thought and I quickly moved it to the “not right now” category. But the shorter 1/4 sheet machine would make it more feasible than the giant in my garage right now

I mis-spoke when I said replace the emt with linear rails. I was more thinking about adding the linear rails to the strut plates and using that instead of the EMT as a bearing surface. I wouldn’t remove the emt all together. That would require a completely different core and motion system Not to mention it would put them closer together than the emt and make the core more prone to deflection. Even if the linear rails are no question better than emt, the overall system would likely get worse.

The floor space in my shop vs the size of my machine is the real catalyst to this change. I am taking the opportunity to change what I think I cheeped out on or should have done differently. I dont think that the MDF strut plates are the smoking gun. I dont think even the length is the smoking gun. I do think that by going the easier route before and making the machine so large, that I hurt the performance in ways that can be easily be fixed now.

As for the Steel plates. I was thinking aluminum was the way to go at first. But the availability locally and the lack of confidence with it pushes me to steel. I know steel is something in the ball park of 3 times as stiff as aluminum, but after stiff enough, the left overs are just extra weight. At this point if I wanted to do aluminum, I’d have to buy an old truck bed toolbox of marketplace and cut out my blanks from that. On the flip side I have the steel sitting in my shop right this second. The strength of the steel does give me room to remove material strategically if I wanted to lighten the load. But I’m not sure that heavier, more vibration dampening aren’t a good idea.

The biggest that dawned on me during thinking through the redo is the cost. First build was “Who cares about a little more EMT?” And a full sheet of MDF as a spoil board was just part of the game. But looking back, I needed 60” x 2 and no way would a higher end DOM stainless or something like that make sense financially if I had to buy so much. But now, I need half as much. That makes it worth looking into at least from a financial aspect. The same factor comes into play with the wiring and moving the control board off the gantry. Before it would have been miles of wiring to get all the Y and all the X motion without yanking things apart. The smaller machine makes it so much less that it has crossed my mind.

11

Same thing, the further away the belts are from the endmill the more things will flex and move.

12

In turn , know that I have the utmost respect for all the community that experiments and shares back here.

I’m only offering advice, no harsh criticism or condescension meant.

I’m also looking forward to seeing what you do next.

13

Something that just came back to me from when I built this.

I was printing with a 0.6 nozzle and laying down.0.8 lines. But when i set my profiles, I still used the same wall count and infill. Without actually weighing each part, I would guess i have roughly twice as much plastic as I’m supposed to. I was talking to @Jonathjon back then about the insane amount of filament I was going through and he pointed out I didn’t change the rest of the profile for the fatter lines. I’m going to chalk that up to a “happy little accident”

14

Rails

You will need to self source the rails.

You are looking for either 29.5mm, 30mm, or 32mm Outside Dimension (OD) steel tubes. In the U.S. and parts of Canada this will be 1” EMT Conduit (29.5mm), off the shelf hardware store conduit (not “rigid” conduit). In other countries you will either be 30mm or 32mm OD.

If you want to find a different material for rails, look for:

  • OD ±0.2mm of 29.5mm, 30mm, or 32mm.

  • At least 1.3mm thick wall.

  • Steel, Stainless Steel, DOM Steel (aluminum and carbon fiber will not work).

  • Super thick walls, solid rods, and filled tubes will gain little to no rigidity and just add mass and cost.

Putting thoughts down:

standard 1-inch trade size Electrical Metallic Tubing (EMT) conduit being 1.45 mm is the standard. The docs say above 1.3mm and the 1.5-2.5 is above both of those.

According to Ryans 1/4 length = 64x rigidity… thats not likely wrong, but strange numbers to wrap my head around

emt ridigity (alone) should increase exponentially as the section of tube gets smaller.

(L)^3 where L= Original Length / new length

  1. 62" vs 31": 62 ÷ 31 = 2× longer → stiffness ratio = 2³ = 8×
  2. 63" vs 21": 63 ÷ 21 = 3× longer → stiffness ratio = 3³ = 27×

That is unsupported spans only.

Current length 55.625

24” machine 30.625 length ratio 1.8163 or 5.9918x stiffer

20” machine 26.625 length ratio 2.0892 or 9.1188 stiffer

Tube OD Wall ID I (mm⁴) Ratio vs EMT
1" EMT 29.5 1.45 26.6 12,618 1.00x
DOM 0.083" wall 29.5 2.11 25.3 15,731 1.25x

Formula: Total stiffness ratio = (I₂/I₁) × (L₁/L₂)³

Tube Length I (mm⁴) Stiffness vs current EMT
1" EMT 55.625" 12,618 1.00x (baseline)
1" EMT 24" 12,618 5.99x
DOM 0.083" wall 24" 15,731 7.48x
1" EMT 20" 12,618 9.13x
DOM 0.083" wall 20" 15,731 11.39x

Ryan was simplifying it a bunch, but only 25% from the thicker wall is chump change compared to go to a shorter rail over all. I just did the math one single run of tube, no gantry math, and while over 11X stiffer is a bunch….. its just numbers not practical application like Ryan and the Team have likely test over and over.

1" EMT: I_total = 2 × (12,618 + 127.6 × 50.8²) = 2 × (12,618 + 329,454) = 684,144 mm⁴

DOM 0.083": I_total = 2 × (15,731 + 180.6 × 50.8²) = 2 × (15,731 + 466,127) = 963,716 mm⁴

Tube Length I_total (mm⁴) Stiffness vs current setup
1" EMT 55.625" 684,144 1.00x (baseline)
1" EMT 24" 684,144 5.99x
DOM 0.083" wall 24" 963,716 8.44x
1" EMT 20" 684,144 9.13x
DOM 0.083" wall 20" 963,716 12.86x

so even figuring in for the beam formation that is created with the separated tubes, the length is dramatically more important than thickness.

15

Conclusions from previous thoughts:

I sat down and actually ran some beam calculations after reading Ryan’s reply because I wanted to understand why shortening the machine seemed to matter so much more than upgrading tubing.

Looking back at the documentation, the recommendations already make a lot more sense.

Standard 1" EMT is roughly a 1.45 mm wall.

The documentation recommends anything above roughly 1.3 mm wall thickness and notes that once you get into the 1.5–2.5 mm range, thicker walls mostly add mass and cost with very little additional rigidity. How little they add is so about the math.

For a simply supported beam, stiffness is proportional to the second moment of area divided by the cube of the unsupported span:

Stiffness ∝ I / L³

where:

- I = second moment of area (tube geometry)

- L = unsupported span

Using my current machine compared to the quarter-sheet dimensions:

- Current gantry tube: 55.625"

- Quarter-sheet gantry tube: 30.625"

Length ratio:

55.625 ÷ 30.625 = 1.816

Cube that ratio:

1.816³ ≈ 6x stiffer

before changing a single thing about the tubing.

Using a common 0.083" wall DOM tube instead of EMT only increases the tube’s section stiffness by roughly 25% so looking only at a single tube:

- Current EMT = 1.0x

- Shorter EMT = ~6x

- Shorter DOM = ~7.5x

I also modeled the two-tube beam by including the tube spacing in the calculations. The absolute numbers changed, but the conclusion didn’t.

The shorter machine absolutely dominates the improvement.

The thicker tube still helps, but it’s almost a rounding error compared to simply removing two feet of unsupported span. That lines up with what Ryan and others have been saying…… for years actually

The thing this exercise ignores is that simple tube bending isn’t the real enemy anyway.

The beam math is useful because it explains why shortening the gantry matters so much, but the cutter isn’t just hanging a static weight from the middle of the rail. The cutting forces are applied at the bit, below and in front of the gantry structure. That creates a torsional load through the entire assembly.

So the real question isn’t just:

“How much does one tube bend?"

It’s more like:

“How much does the entire gantry and core rotate under cutting load?”

The LR4 isn’t just two tubes. It’s a structural system made up of:

- Tubes

- Strut plates

- Printed braces

- Fasteners

- Bearings

- Belts

- Core

- Router

All of those components contribute to the final stiffness of the machine.

So my takeaway has changed quite a bit. I’m no longer looking at premium tubing as something that will transform the machine by itself. Though honestly I didn’t think it would be magic. I did think it would be more of a difference though. Instead, I see the shorter gantry as doing almost all of the heavy lifting. Better tubing, steel strut plates, and other upgrades simply become incremental improvements layered on top of a machine that is already dramatically stiffer because of its size.

At least for me, this exercise wasn’t about proving anyone right or wrong. It was about understanding why the documentation recommends what it does. I still think it’s worth improving multiple areas of the machine, but I no longer expect any single component to be as advantageous as i had previously thought.

16

I had an idea earlier…

I’m starting to take the machine apart, and thinking about what I can only describe as torture testing.

If i bolt one of the xz plates to my bench, I could put my digital torque wrench on the opposite end of the gantry and set up a dial indicator. Then measure total twist at a set value. Then repeat it again with a few different set ups. That should give measurements that more closely resemble the real stress on the machine. It’s not quite the same as the core trying twist the gantry in the middle, but it’s a more telling number than sitting here and doing theoretical math.

Not sure if that data would help others or not. And maybe there is a better way to set up the test to get more accurate numbers. But I’d be game. Maybe start with black pipe if I can find the right diameter. It’s about the thickest heaviest stuff I can think of at Lowes or HD. Then cut the emt and black pipe in half, adjust the plates and redo it.

I would be tempted to try without the plates, but that’s not fair to the gantry. The plates locking those faces in plane is where the torque gets eaten up.

17

Trying to isolate parts is just going to lead to incomplete data with you guessing or assuming the load actually is. It is far easier to test with a full built machine and check at each point under consistent loads. Remember the load is at the point of the endmill that is cutting material.

Part isolation can be done with FEA, but that tells you almost nothing about the behavior of the assembly.

From my tests, shorter belts, and shorter beams result in the most rigid machine. The rest of the parts stack up to the rest of the error but not a significant number.

18

The machine is designed with a 9Lb load at the endmill. So testing at 50lbs also results in bad data.

19

Did this make me an OG? Lol

20260627_231809
20260627_2318091848×4000 1.79 MB

20

Old Guy? Old Grump? Those are badges to wear with pride.