Yes, certainly. Big flat surfaces moving generate plenty sound pressure… So you’re right, these signals may be entirely irrelevant to the cut.
Jösses! So the whole machine is essentially mounted on springs! Not saying that this makes it worse, but it sure as hell doesn’t simplify an analysis…
Ah,
There was quite a thing for “isolating” 3D printers a while ago with lots of people trying wibbly feet for their Prusa and other printers. One of the better YT channels, Thomas Sanladerer I think, did a bunch of tests and discovered they mostly made the print quality awful by allowing vibrations and flex in the machine. Putting the printer on a heavy slab of something solid was determined to be the starting point for low noise and high print quality.
Not saying a closed cell rubber or, better, a visco-elastic damping material like Sorbothane doesn’t have it’s place, but you need enough mass and rigidity in the sprung component on top of it.
Shouldn’t that be the case for a CNC mill?
Well, any moving machine really, but the Shapeoko seems to have been designed expecting to be placed on a stable fixed surface rather than having sufficient structural rigidity and mass to free float. Haven’t seen a Nomad in the flesh but the same is likely true there.
I’ve finished talking to Carbide 3D support, which was unhelpful:
Q:
At this point, I’m not looking for a recipe, I’m trying to determine whether the Nomad 883 Pro’s MRR is limited by its spindle power or by another factor like frame rigidity. My testing so far indicates the latter.
Even if you can’t tell me how to achieve maximum MRR, can you tell me whether Carbide 3D has been able to reach top spindle power while cutting Aluminium?
I don’t care about the endmill, whether you use coolant or airblast, even the feeds and speeds, I just want to know whether, in Carbide 3D’s experience, the Nomad 883 Pro is capable of reaching its rated spindle power in Aluminium without being limited by another factor inherent in the machine itself.
A:
As I said this subject is beyond the scope of our support, theoretical maximums. I did ask the development group and they responded no data for this exists at this time.
And earlier in the thread:
Our advice is here: https://docs.carbide3d.com/support/supportfiles/Nomad883_feeds_125.pdf
This response is extremely disappointing to me, because it means that Carbide 3D’s specs are close to worthless. They specify a 70W spindle, which is bad enough on its own, but to make matters worse, they only support a 2.4W cut (that’s the calculated cutting power for 6061 AL with those feeds and speeds).
If they were operating in a vacuum, I might say this is okay, as there are no standards but there is competition here and it looks to be handling this much better. Bantam Tools has recommended feeds and speeds for their machines and though they’re similarly conservative with the base recommendations, they also specify much more aggressive feeds and speeds, which they mark as “advanced”.
Carbide 3D’s unhelpfulness pushed me to give up on the Nomad 883 Pro as-sold, I’m not going to bother trying to push it any further, I’m going to upgrade it or replace it.
Anyhow, ranting about C3D aside, I also contacted a linear motion store to ask about their opinions on upgrades:
This led me to thinking linear rails and ballscrews are really the right answer for me. I’ll add screws and rails on all axes and just keep the Nomad’s frame. If I do that, I should be able to use a real spindle no problems.
I also thought about replacing the machine with a different one but despite all its flaws, the Nomad’s footprint is really what I need. There are drastically more capable machines at the same price point but they’re big and bulky and not appropriate for my apartment.
Anyhow, I’ve started working on the design for the upgrade and the main challenges are look to be:
I think producing the backing plates is the most troublesome problem. I can get a 3rd-party service to do it but the cost will be ~200€ per plate. This would be a lot cheaper if I could reuse the same plate for the X and Y axes but it doesn’t look like that’s possible, as the Y-axis is a lot longer.
The positioning edge needs to be as straight as you want to ever be able to align your axis to, might be expensive to get that machined. The other option is to bolt it down to a flat face and use a reference edge to align it.
The rails are not structural, they expect their mounting surface to do the support work. You might be better off with an extrusion of some sort, or you may need some fairly thick plate to avoid deflection. Time to calculate deflections of different cross sections and compare them with the steel rods already installed.
Those are pretty chunky items, even on a Shapeoko they’re robbing me of machining space by adding the extra 28mm thickness of rail and block (for HGH15 not 20). I’m also losing space to the mounting plate for the linear blocks so that’s another 8mm.
Might be worth comparing with the MGW15 class rails, these are a bit lower in their rated loads but the Nomad isn’t a very big machine.
Big HGH class rails;
Smaller MGN class rails;
Worth doing a proper CAD model to find out how much space you’re going to use for the supports, rails, blocks, ballscrews, ballnuts etc. I suspect your choice will end up being a compromise between ultimate performance and having and X, Y, Z travel left.
Will the exiting controller and motors cope with all the extra rotating and moving mass of ballscrews, plates and linear rails?
Hasn’t Vince already done a heavy duty Z that would be a nice place to start?
There’s no shortage of manufacturer data on what loads ballscrews are rated for;
Not a certain vendor starting with X then…
There are pretty good, flat faced, extrusions available from people like Misumi which many people use as CNC parts. If you want to get really OCD about the flatness and straightness you could do an epoxy pour or shim under the rails to flatten them fully?
Despite all of these inconsistences and lack of matching the specs, many of us are cutting aluminium just fine on a Nomad. You know, making things with the machine rather than looking at numbers.
It cuts as well as it cuts. If you’ve stuff to make, make it how the Nomad makes it. If it doesn’t make it well enough, move on to something else.
Yeah, I was thinking machining would be difficult so I was going to go with a reference edge. Those seem to be much cheaper.
I was going to go for thick plate, as it’s cheap and easy to obtain pre-milled Aluminium plates with decent tolerances. I figured 15-20mm (thicker than most of the Nomad’s parts) should be more than enough. Any thoughts?
Yeah, I started looking at the design in CAD, which is when I noticed just how big these rails are. I’m thinking now that I should probably start from the Z-axis and work my way back. I was looking at those ones in particular because if I go for 1605 ballscrews, I need ~40mm of clearance anyway, so why bother with spacers when I can just use waaay oversized rails? 
If I go for the oversized rails though, I think I’d only have space for one carriage block on each rail, which feels a bit weird to me, so the spacers (or smaller ballscrews) might be the better choice.
Controller yes, motors maybe. I’m fine with replacing the motors though, steppers are cheap. I’ll that into account when looking at the dimensions I have to deal with.
Vince’s Z is bolted onto the Nomad’s stock X-axis carriage, which I’d like to avoid. I do like his mounting plate though.
Ah, interesting, I didn’t know you could buy pre-milled extrusions. I don’t think I can buy from Misumi though, I think they only accept business customers? Do you know of any other vendors?
I’m not “looking at numbers” but thanks for the condescension.
I bought the Nomad because I wanted to mill Aluminium and it was advertised as being able to. I later found that it doesn’t mill Aluminium nearly as well as I would like it to. Aside from the more basic issues (it’s loud and doesn’t have any chip management), it doesn’t have the precision I’d like it to have (backlash on the axes, plus open-loop steppers) and it doesn’t have the MRR I’d like it to have (even basic projects require a lot of material to be removed, so take many hours to complete). Precision I could deal with but the usable power of the machine is so far below what was advertised that not only am I disappointed, I feel misled and deceived.
It’s like I ordered a car and it broke down going up a hill. When I took it to the dealer they told me “Sorry, that’s for advanced users, we won’t support it. You should go talk to the folks at the motor club instead, our recommendation is to only drive on flat roads”.
You’re welcome to be happy with the machine. I’m under no illusions that my experience is universal and if you enjoy the machine I’m honestly happy for you. I’m not happy with the machine as-sold and no amount of condescending comments is going to change that.
Anyhow, this thread is, as the title suggests, about achieving more challenging cuts on the Nomad. If you have advice on how to do that or experiences related to it, please feel free to share. If you have anything else to say, please start a new thread.
You probably own a car with a speedometer that shows a velocity way beyond what you’ll attempt to drive the car at. I advise you to return the car or modify it until it can reach the speed on the speedometer.
Condescension is rife in this thread, along with a sense of righteousness, indignation, and deservedness. And not all of that is from me.
If you don’t like the thread, feel free not to participate.
This is my final response to you. Engaging with you is clearly not going to be productive.
Acknowledged.
…
While the round rails are definitely not as ridgid as square, after seeing what type of cuts can be made, it was clear they weren’t really a big issue. My No-Mod beats a Shapeoko in aluminum MRR.
Oh a using the @spargeltarzan Millaylzer this past week really opened my eyes on depth of cut and the forces and ramps change dramatically.
Thanks, that’s a helpful comparison.
Honestly, I think the linear rails are overkill but I just want to cross off everything in the list of “things I worry about”, even if I end up overpaying for an overengineered machine.
Though you have given me an extra push to try upgrading the machine more piecemeal. I could do it in two stages:
That would also give me some experience using linear rails for Z before I go and put them everywhere on the machine.
Very much agreed. Make sure to read the docs in the help menu too!
The thing I found most interesting is that increasing depth of cut can lead to basically perfectly consistent cutting forces:
One thing to think about is the extra time needed to square and re-tram an entire machine instead of just the Z. Carbide3d already did the work, i guess depends on if you want more performance than just what the low hanging fruit will give ya.
Yeah I’ve gone over those help docs quite a few times. Its interesting trying to predict a real world cut quality based on computer perfect results. I ran hundreds of cuts on the S3 with seven or so different tools in the past week. Those two graphs are single flutes so until you get past a certain depth of cut, a shallow/wide cut will always outperform.
Yeah, I’ve been thinking about that for sure, especially where I can put what sized holes to make sure there’s a bit of jiggle room for tramming.
I might be optimistic but I’ve been thinking that tramming and squaring the machine might actually be a good learning experience.
Have you done it before? Is it something that’s time-consuming but pretty straightforward, or is it something that’s going to make me bang my head against a wall?
If you do decide to do this…
There’s a bunch of online calculators for motion of intertia of various cross sections;
Which gets you the characteristic strength in both axes of the cross-section you’re using.
Then use a beam deflection calculator for a point load on a simply supported beam, load in the middle is worst case, Aluminium is about 70GPa for the Young’s Modulus
That way you can compare different options and see what you’re trading. With the very short beams on the Nomad reasonably small stuff will do the job.
That’s a lot of torque to apply to a single block, when you have multiple blocks the torque is mostly converted into radial load on the blocks, with much smaller values.
I googled and ended up reading posts from folks on forums like mycncuk.com to see who’d ordered from which suppliers in my region. Haven’t bought any yet so I can’t recommend a vendor.
The AliExpress vendor I used for the ballscrews and rails OTOH I can recommend. BST Linear Motion is his store, PM me for his email if you want to talk to him.
You inspired me to check and it turns out the version of SolidWorks I’m using has a basic simulation module that I can use to simulate loads on my parts. Here’s what the deformation of my current X-axis mock looks like if I make it out of EN AW-7075 Aluminimum and put 1000N of force of downward force into one of my screw holes and 1000N of force pulling against the screw hole (so in total a ~140kg diagonal force):
I was surprised the deformation was so small but it is a pretty hefty chunk of metal right now.
Does that seem realistic?
Yeah, I was thinking that too. The rating for moment on the 25mm rails I was looking at is 43407 kgf-mm. If the spindle nose is ~200mm from the rails, that leaves me with 217kgf of force before the bearing is overloaded, which is less than it could be.
I had a look at some AliExpress pre-assembled axes and it looks like they mostly go for small bearings and carve out some space under the carriage, so I think that might be the way to go.
I had a look around and I think it’s going to be difficult. If we weren’t in Corona times I could buy some extrusion and then mill it flat myself at one of the local fablabs but right now that’s not possible. I think plate is the best option for me right now. I think extrusion would have better mechanical properties for sure but in this case I think the plate is going to be easier to work with.
100kg of Force, that’s quite a bit.
What are the dimensions of that aluminium plate?
How are the Nomad Y axes built? I’ve not seen how a Nomad is put together but the Y plates frequently seem to end up being the weak point on machines where flex creeps in. As the Nomad does Y travel on the table you at least have a fixed gantry to stabilise.
376mm x 130mm x 15mm
The linear motion configuration is very similar to the X-axis: 20mm steel rods with linear bushings embedded in a hefty chunk of Aluminium that has a lead screw going through the middle. The steel rods are embedded in the Aluminium plates at the front and back of the machine. The front and back are also connected to each other with what seem to be the beefiest structural components of the Nomad: 50mm x 25mm x 430mm Aluminium bars. These bars run down the side of the machine.
