LowRider v4 Community Hog-Out Test

LowRider CNC Troubleshooting
1

I’ve been kicking around an idea for a while, and I’d like to see if the community is interested.

One of the questions that comes up over and over is:

  • Are metal strut plates actually better than MDF?

  • Does adding more printed braces make a measurable difference?

  • How much does gantry length really affect performance?

  • Is a heavier spindle worth the tradeoff?

  • Does PET-CF, ASA, PLA, or another printed material matter and on which parts?

Everyone has opinions, but we don’t have much actual data.

The goal of this project is to change that.

Rather than asking “How fast can your machine cut?”, I want to build a community dataset that shows which design choices actually increase machine stiffness and usable cutting performance.

The idea is simple:

Everyone runs the same Hog-Out test generated with Vector76’s excellent test pattern generator, using the same basic cutting parameters, bit type, and material. We then record both the machine configuration and the results.

Each machine starts with an easy feedrate and increases in fixed increments until the machine reaches its practical limit. Think of it like a volumetric flow test for a 3D printer—the goal isn’t to win, it’s to identify the point where performance starts to fall off.

For each submission we’d record things like:

  • Gantry length

  • Tube material

  • Strut plate material

  • Number of printed braces

  • Printed part material

  • Router/spindle model

  • Bit stickout

  • Feedrate ladder results

  • Highest clean feedrate before failure

  • Failure mode (deflection, chatter, skipped steps, etc.)

Over time, we should be able to answer questions like:

  • Does ⅛" steel outperform ¼" MDF, and by how much?

  • Is the difference larger on longer gantries?

  • How much do additional braces actually help?

  • Is there a point of diminishing returns?

  • Which modifications provide the biggest improvement per dollar?

I’m not trying to create a scientific laboratory test. I’m trying to create a repeatable community benchmark that anyone with a LowRider v4 can run in an afternoon.

I’ve put together a draft testing protocol and data sheet, but before I ask people to start cutting I’d really like feedback from the community.

What would you change?

Are there variables I haven’t considered?

Is there a better way to measure performance while still keeping the test simple enough that lots of people will actually participate?

If we can get enough submissions, I think we’ll end up with something that’s genuinely useful—not just for people modifying existing machines, but for anyone building a new LowRider and trying to decide where to spend their time and money.

I can’t load the document directly here so I made a drive folder and made it shareable.

2

Before anyone starts filling out spreadsheets, I’d like to get a couple people’s blessing.

First, I’d really like Ryan’s thoughts on the idea. This project exists because of the LowRider, and if this turns into a useful dataset I’d rather it belong to V1 Engineering than to me. My hope is that it becomes something the community can continue to build on, and if the information ends up being useful for future LowRider development—whether that’s incremental improvements to the LR4 or ideas for an eventual LR5—even better.

One of the strengths of the V1 community is simply its size. Even with a dedicated beta team, there’s no practical way to test the huge variety of machine sizes, materials, modifications, routers, spindles, and build styles that are already out in the wild. The community has collectively built hundreds, if not thousands, of machines. If we can gather data from even a small percentage of them using the same benchmark, we’ll end up with a dataset that’s impossible for any small beta group to produce.

I’d also like Jamie’s permission to use his Hog-Out test generator as the foundation for this benchmark. It already provides a consistent, repeatable way to progressively increase cutting load, and it seems like a perfect fit for collecting comparable data across many different machines. There’s no point reinventing something that’s already well thought out.

If both of you are on board, I’ll clean up the protocol based on everyone’s feedback, organize the data collection, and let the community do what it does best—build a dataset that answers questions with measurements instead of opinions.

3

*mods

I didn’t know where to put this, so please feel free to move it wherever best suits.

4

One thing I wanted to explain is why the protocol standardizes so many variables.

The goal of this project isn’t to find the fastest possible cutting recipe. It’s to compare LowRider machines. Every variable we can standardize is one less thing that can influence the results.

If one person cuts plywood, another cuts MDF, one uses a compression bit, another uses a 3-flute end mill, everyone chooses different depths of cut, and everyone generates their own toolpath, then we’re no longer comparing machines—we’re comparing machining strategies.

Instead, I’ve tried to standardize as many variables as are practical:

- Same test pattern

- Same cutting material

- Same style of end mill

- Same depth of cut

- Same feedrate increments

- Same test location on the machine

That leaves the machine itself as the primary variable.

The things we’re actually interested in become the differences between the machines:

- Gantry length

- Tube material

- Strut plate material

- Printed part material

- Number of braces

- Router/spindle choice

- Bit stickout

- Other machine configuration differences

Obviously, no community benchmark can eliminate every variable. Two sheets of MDF won’t be perfectly identical, and no two routers or end mills are exactly alike. But by removing as many variables as reasonably possible, patterns should begin to emerge as more machines are added to the dataset.

This is also why the protocol isn’t necessarily optimized for maximum material removal. Every machine probably has a different “best” recipe. That’s not what we’re trying to measure.

The philosophy is simple:

If two machines differ by only one design choice, I’d like that design choice to show up in the data.

If we can keep everything else reasonably constant, then differences in the results are much more likely to reflect differences in the machine itself—not differences in the cutting recipe.

5

Yes, a year or two back a user made extensive studies and posted them. No need to discuss this one. :grinning_face_with_smiling_eyes:

6

My gut feeling is that with so many variables, you would need a ton of data to converge on a useful conclusion. Testing one variable at a time is the gold standard.

2 Likes
7

That is actually something I missed on my edit. Lol. Ai helped me write it and I didn’t get the right filaments changed out

Scientifically, I completely agree. If you want to isolate a single variable, changing one thing at a time is the gold standard.

The challenge is that no one person has the time, budget, or shop space to build and test every meaningful combination. There are simply too many variables: gantry lengths, tube materials, strut plates, brace counts, printed materials, routers, spindles, and countless other small modifications.

That’s where a community dataset can become valuable.

Think about how the automotive industry works. Engineers spend years testing prototypes under controlled conditions before a vehicle is released. Once it’s in customers’ hands, though, manufacturers continue collecting data from warranty claims, service records, and real-world use. That data isn’t more controlled than engineering tests—it simply represents a much larger and more diverse sample of real-world conditions.

I see this project in a similar light.

Ryan and the beta team did the engineering that produced an excellent machine. What I’m proposing isn’t intended to replace that process. It’s an opportunity to learn from the number of LowRiders that have been built since.
The strength of the community is the sheer number of different configurations that already exist.
With enough submissions, we can begin filtering the data instead of trying to test every combination ourselves.
For example:
Show only 30" gantries. Does strut plate material make a measurable difference?
Now expand that same comparison across all gantry lengths.
Do additional braces help equally on every machine, or only on longer gantries?
Does one modification become more valuable as the machine gets larger?

I don’t know what the answers are, and that’s exactly why I think it’s worth collecting the data.

The goal isn’t to prove one design is “best.” The goal is to identify trends that only become visible when you have enough real-world machines contributing data.

8

I’m just not sure it’s practical. There are already a lot of details and it’s missing some. I don’t think there would be enough participation. Some of these things are subjective. Some of it depends on how well the machine is calibrated. I rarely use 1/4 inch endmills. There can be a lot of variation in a single flute upcut endmill. Some things like steel strut plates just aren’t practical for most users. It’s a system so these things are all interconnected like strut plate material and number of braces. Some decisions have other consequences like if you add too many braces, you can no longer fit the controller on the beam. Sometimes the consequence is cost. It may be slightly better but a lot more expensive. Maybe they configured it to increase acceleration and amps but that requires adding a fan. The only “mod” I have planned for my eventual LR4 build is to use a larger NEMA 17 on the X axis since its the only axis with a single motor. That decision is based on that logic and that it’s not that expensive.

I think the test would need to be a lot simpler to get the participation required. But then, does a simpler test provide enough value?

Also, if we get all this data, this might lead to more users deviating from the yellow brick road which comes with its own set of challenges.

I’m curious about others opinions. Maybe I’m just grouchy today. :laughing:

2 Likes
9

I started to write something similar. I’m not sure I’m grouchy :slight_smile: , just not sure how to evaluate one machine/builder against another.

I’ve seen someone run one of my machines who had a much better understanding of the feeds and speeds, was able to do an operation that I wouldn’t have even considered. They made it look easy. I’m pretty sure I’d have broken my machine if I’d attempted the same. That’s a long-winded way of saying the biggest limitations for most of my machines is… Me.

I’m also always leery of a comparison that sort of presupposes that there’s something “Wrong” with the LR4. It’s a great machine, and given its price point and ease of build it’s actually an exceptionally great machine.

10

That’s actually the main reason I was thinking in terms of a fixed benchmark instead of asking people what their machine can do.

If we ask, “How fast can your LowRider cut?”, then user skill becomes a huge part of the answer. Feeds and speeds, toolpath strategy, CAM choices, router RPM, depth of cut, workholding, and experience all get mixed together.

That’s not what I’m trying to measure.

The idea is to remove as much of that as possible:

- Same generated gcode

- Same material class

- Same bit type

- Same depth of cut

- Same feedrate ladder

- Same pass/fail criteria

- Same basic measurement method

Then the user is not really being asked to optimize anything. They’re just running the benchmark and recording where the machine starts to fall outside the acceptable range.

That should make it less about “who is better at feeds and speeds” and more about “what does this machine configuration do under the same test load?”

It still won’t be perfect. Nothing community-run will be. But I think a fixed test with objective pass/fail measurements gets us much closer than collecting subjective impressions or asking people to report their best cutting settings.

11

Just to be clear, I don’t have any emotional investment in this becoming a community project.

If people think it would be useful, I’m happy to put in the time to organize it. If the consensus is that the value isn’t there, that’s perfectly fine too.

At the end of the day, I already have my answer for my own shop.

I really like my LowRider v4, and it has been a great machine for what it was designed to do. But through using it, I’ve also learned that most of my work involves materials and cutting forces that fall outside of its design goals. That’s not a criticism of the machine—it’s simply a mismatch between the tool and the work I enjoy doing.

My long-term plan is still to design and build a machine that’s better suited to the kind of fabrication I do most often. When that happens, I fully expect my LowRider to stay in the shop and probably become a dedicated plasma table, because I think it’s an excellent platform for that application.

The reason I proposed this benchmark wasn’t to justify my own redesign. It was because I thought the community might benefit from having real data showing which modifications make a measurable difference and which ones don’t.

If that turns out not to be something the community is interested in, that’s completely okay. I’ll spend the time designing and building the next machine instead.

12

I agree completely.

If someone has to spend an hour filling out forms after making chips, participation drops dramatically.

That can be simplifying the process.

13

I agree, and even worse, things like print quality, etc are even tougher to control across the wider community.

What if someone doesn’t realize they have a slightly cracked core, or some layer separation?

Or someone saying they used PLA, but it was really PLA+?

There’s just so many variables in the way each person builds a machine, and every one of them affects the machine differently.

The bad part would be publishing data that says “By doing X, you will get Y benefit” when you can’t guarantee, because of all of those variables, that it will be the same for everyone.

And now you have other things like X length and table type to consider as well. Were they V1 motors or self-sourced, etc., etc.

I think, in the end, you might end up with an n-dimensional dataset with exactly one value for each variation. It’s probably not likely to be able to gather much value from it.

I think Ryans goal is ultimately to design a machine that doesn’t require mods as a general rule, and should only be considered on a case-by-case basis depending on how the user plans to use the machine.

I think it would be difficult to generalize mod performance gains in a way that doesn’t risk sending some users down the wrong path for their use-case

14

On the flip side, it could definitively show that mods hurt more help. Or maybe it shows that 1 extra brace is by far stronger and the build guide gets updated.

It is also entirely possible that that we find the sweet spot for each use type instead of a general best. Laser/ plasma guys don’t need the same stiffness as someone trying to mil aluminum, but could benefit from extra speed and acceleration. “Hey guys I’m building this to cut vinyl” could have a much more use oriented response.

Again…

The idea here was not to rock the boat. I just thought the data might be useful to the community as a whole. It was never about crowning the best machine builder or most skilled operator. Very simply to try and quantify all the x factors

1 Like
15

Don’t misunderstand the responses you’ve gotten.

I don’t think anyone is meaning to beat you down, and it’s not about rocking the boat.

It’s just that sometimes, more data can be more harmful than helpful, because the analysis of that data can be way off without taking every variable into account.

I don’t think anyone read it that way.

I don’t think anyone is saying the machine is perfect, either, or that it couldn’t use some analysis. It’s just that any reasonable outcome from these types of tests, IMO, probably require a more controlled test environment in order to get some type of insight from the results.

It could increase the support burden on the forums, rather than reducing it.

That’s just my $0.02… but I’ll defer to @vicious1 on it.

16

I completely get where you are coming from, the variables seem untested, some are, some have been, some have been verified with data. We have already discussed the tests I have been running this last year.

Here is my suggestion, start it. Collect all the data you can, and share it. Let’s see if anyone wants to join in, I welcome it. But, please be very careful before you make any claims, statements, recommendations as to how it should be built. Just compile data.

Your Goals

1- You have been pushing for a better machine. What exactly is your goal, as in what exactly would your ideal machine be? We might already be able to help. I looked through your posts but I am not sure what you are trying to actually accomplish in terms of machine performance other than just “better”.

My Goals

My goals have been met since the day I launched the MPCNC, I milled aluminum before I released any files.

At some point early on I realized we could easily max out a 1/8" endmill, We then moved on to 1/4", 5 years ago, https://studio.youtube.com/video/qNdod5hKfhs/edit At that point I was no longer concerned with speed (material removal rate) in wood or plastic.

Most of my projects work best with a smaller radius 1/8" endmill so we have already maxed out my project use case. Years ago.

My new machine motivation

What I spent the last 11 years doing, the only reason I spend time making new machines.

Can I make it;
Easier to build
Easier to use
Easier to source
Cost less
Higher material removal rate in aluminum.

The harsh truth

We are now chasing zeros. Chasing zeros is not really fun. Making a machine twice as rigid will no longer make it twice as fast, we might gain 5%. Severe diminishing returns. Making it twice as rigid might make it twice as expensive and twice as hard to source (steel plates), for a 5% gain.

If there is a fancy filament that makes it 5% more rigid, does twice the material cost and the need for a fancy printer with a heated chamber and ruby nozzle worth it?

If you look at the machines that cost more than ours, look at the price versus material removal rate. For some people an extra $2,000 is worth 20% faster, for me, it is not, so that makes it hard for me to want to spend any time developing or supporting a machine like that. I enjoy running a machine I am not worried about making mistakes on and can easily and cheaply fix.

So, any improvements kind of have a tough set of criteria to hit to even be considered.

2 Likes
17

But just for the records, you used to publish “Drag race” videos :wink:

Not trying to “corner” you btw, I’m totally fine with the current policy you’re trying to achieve
Hobby CNCs do not need to have extremely fast removal rates, if you want/need industrial grade results (for speed or precision), jlcpcb might be a better alternative than any DIY CNC

1 Like
18

That is the video I linked in my post.

That is the point I realized speed was no longer an issue. From then on we focused on other things. I felt like the MPCNC had hit its goal, I was happy with it’s capabilities and cost.

We moved on to
Bigger machine - Lowrider
Better firmware for our use case, easier to make edits - Fluidnc
Board quality and cost - jackpots
Easier instructions that I can get help with - mkdocs / github
Kit costs - Partner with LDO

The question for me is what is next, besides the revamped Zen that I have been working on.

1 Like
19

That’s actually a really good question, and I think it’s where my thought process diverges from the design goals of the LowRider.

Where my goals differ is in the work I enjoy doing.

Most of my projects revolve around fabrication more than woodworking. Steel, aluminum, welding, plasma cutting, and machining are what I naturally gravitate toward. Plywood and MDF are usually just fixtures or prototypes that help me get to the metal.

Instead of trying to optimize cost per performance, i am trying to optimize absolute capability within my shop constraints. Because of that, I’m willing to accept tradeoffs that wouldn’t make sense for the average LowRider builder.

If spending more money, adding weight, making parts from steel instead of printed plastic, or making the machine more difficult to build gives me a meaningful increase in rigidity, that’s a trade I’m personally willing to make.

I don’t think that makes it a better LowRider. I think it makes it a LowRider optimized for a different use case and a target audience of 1.

The LowRider has taught me a lot about what I actually want from a CNC, and it’s also shown me where my own priorities differ from its design philosophy.

I don’t see that as a criticism of the LR4. If anything, I think it highlights how well the machine accomplishes the goals it was designed for. Nor do i think the LR4 is a bad machine. Quite the opposite—I think it’s an exceptional machine for the audience and goals it was designed around.

For me, the logical next step probably isn’t trying to turn the LR4 into something it was never intended to be. It’s designing a machine around my own priorities while letting the LowRider continue doing what it already does well. The new machine when it happens will have done sweet aluminum parts from the LowRider.

Ironically, I wouldn’t know any of what I want and need if it weren’t for the LowRider and this community.

Whether this benchmark ever goes anywhere or not, I’ve already gotten far more out of this community than I’ve put into it.

For now, I’m going to let the idea sit and see if there’s enough interest from the community to make it worthwhile. If there is, I’ll gladly put in the work to organize it. If not, that’s okay too. My original goal was simply to find a way to give something back.

1 Like
20

Being the CNC gateway drug for people is EXACTLY what I want. For some they never use it, for some it is everything they wanted, for others they go to a $20k machine with confidence knowing it is not a waste of money. No matter what the outcome the financial outlay was minimal and I know it is a serious machine, not a tiny cnc with a 12v brushed dc hobby motor that looks like a CNC but can’t actually do much.

2 Likes