garage upgrades

41 views
Skip to first unread message

A J

unread,
Nov 23, 2025, 2:29:30 PMNov 23
to HomeBrew Robotics Club
Hey All,

I have never really gotten into CNC or even 3D printing but I have feeling that people are going to want to upgrade their workshops by summer.

s...@lig.net

unread,
Nov 23, 2025, 3:31:31 PMNov 23
to hbrob...@googlegroups.com
Several interesting machines are now coming out that solve problems we've wanted solved for a long time, finally at more or less reasonable prices.

Order here:

https://www.kickstarter.com/projects/959480926/nestworks-c500-next-gen-smartest-cnc-with-industrial-power

This is the other one recently:
https://www.kickstarter.com/projects/1763117873/fibreseeker-3-the-first-personal-continuous-fibre-3d-printer

Chris Albertson

unread,
Nov 25, 2025, 1:20:32 PMNov 25
to hbrob...@googlegroups.com
I was excited to read about the continuous fiber printer.  Then I looked into it.   (1) I don’t think you need a special printer to use it, although I doubt it would work in a basic “Ender” 3D printer. and (2) the raw material is about $150 for 50 cubic centimeters. 

To put cost in comparison.  I pay about $0.02 per gram for basic plastics. ($20 for a 2000 gram spool)   If we assume the continuous stuff is as dense as water, then 50cc = 50ml = 50 grams.   So it is $3.00 per gram.   About 100 times more expensive.    I’m sure that I am wrong about the density, so the cost is higher. 


As for milling machines,  Harbor Freight is not bad for under $1K.   I think anyone who would want this would have the mechanical ability to attach stepper motors to the hand cranks. Then Open Source software will control those motors.   The harder part in both cases is learning CAD and then even harder, mechanical engineering.





--
You received this message because you are subscribed to the Google Groups "HomeBrew Robotics Club" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hbrobotics+...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/hbrobotics/9a5ffc40-0dd5-4909-89a3-4ec87bad5ecd%40lig.net.

Stephen Williams

unread,
Nov 26, 2025, 2:39:04 PMNov 26
to hbrob...@googlegroups.com

The point of the continuous fiber printer is to feed carbon fiber (which is expensive, but goes a long way) immersed in plastic fiber (which may or may not be expensive here) plus normal plastic fiber (whatever you want to use that is compatible with the first plastic).  You only need a little carbon fiber to make a print very strong.  Shouldn't be a significant cost most of the time.  Also, I think I saw a mention of using glass fiber, soon.  That can be very inexpensive while also being strong, weighing a bit more.

A full milling machine is useful, but takes a lot of care & feeding.  In particular, getting precise positioning I expect to be difficult.  For larger jobs, probably fine.  The all-in-one CNC units that are extremely precise are useful for small parts that need to be precise, like gears, bushings, frames.

It would be good to work out a combination that converts the cheap milling machines into a nice CNC.  I don't have time to solve that.


sdw

A J

unread,
Nov 26, 2025, 3:39:33 PMNov 26
to HomeBrew Robotics Club
That sounds exciting, having a 3D printer that creates a composite material.

Chris Albertson

unread,
Nov 26, 2025, 6:34:09 PMNov 26
to hbrob...@googlegroups.com
OK, so they are not using 100% fiber.  Just selective reinforcement.    This must be why I see they use a dual head printer.  I’m slowly working on this.  I have the electronic “guts” of a multi-hed printer on a literal breadboard right now.  OK, I use DIN rails, so I can move parts from breadboards to real devices and not have to worry about mounting hole ring right.     The 3D printer and the milling machine are all DIN rail and “never finished”.  The breadboard in the photo is for software dev and testing.

SOme new printer designs can offload the head from the X carage and park it on a rack, this mean you can have as many different materials as you can build out rack space.    Can also be combined the automated filament loading.   This is all open source design.
Voron printers are very well designed.    They work abut as well as the Bamboo printers but are modifiable.   Here is one build where the guy install 6 heads.



About the HF milling machine.    There are several “levels” of conversion.     First, I will say this is a cheap mill, but it can cut carbon steel and, of course, aluminum.   If you use it by hand with cranks, there is backlash in the lead screws, but machinists know how to deal with this (they never reverse a crank direction while cutting). But CNC does reverse the direction.      So the levels are:

1) just put a motor on the crank and live with the backlash.   Try and deal with it in software.  This works “good enough” for some cases.

2) Next, I tried buying a linear scale.   This is just like an electronic caliper; the sensor slides on a track and works about as accurately as a cheap digital caliper; the sensor also costs about the same.   Use the sensor for closed-loop frigid tappingal with backmovements) Buy a more expensicannottal scale made from glass.   These are verspindle,ate but $100.     But feedback is stlead scservo-drivenin the lead screws is not 100% cCurrently,ut.    Glass does not better than thlearningtainltop-of-the-lines above.

4) Finally, the “real fix”prettyreplace the lead screws with ball scLCNC-basedall screws have zero measurable backltools;nd Chinese screws are good enough unless you are machine optopen-source machineical parts.  

What makes  the CNC work is LCNC, Linux-based open source machinbe tool software.   Look at Tormach mchine tools, they are LCNC based.   It is bpretty much top of the line pro-grade but has the associated llearning curve.  Currentluy I have leadscrews for X.Y and ball screws on Z.  I have not yet installed a servo'd spindle so I can not do rigidd tpping.    A CNC machine is basically just a robot designed for very accurate movments and with structure stiff enough not to deform with the reaction force, so "real machines" are made from cast iron.


As for milling gears.     With the right tooling you can do that on a manual mill.   Backlash is not an issue.  You do need a good rotary table and involute cutters for the gear pitch.


 

On Nov 26, 2025, at 11:38 AM, 'Stephen Williams' via HomeBrew Robotics Club <hbrob...@googlegroups.com> wrote:

The point of the continuous fiber printer is to feed carbon fiber (which is expensive, but goes a long way) immersed in plastic fiber (which may or may not be expensive here) plus normal plastic fiber (whatever you want to use that is compatible with the first plastic).  You only need a little carbon fiber to make a print very strong.  Shouldn't be a significant cost most of the time.  Also, I think I saw a mention of using glass fiber, soon.  That can be very inexpensive while also being strong, weighing a bit more.

A full milling machine is useful, but takes a lot of care & feeding.  In particular, getting precise positioning I expect to be difficult.  For larger jobs, probably fine.  The all-in-one CNC units that are extremely precise are useful for small parts that need to be precise, like gears, bushings, frames.

It would be good to work out a combination that converts the cheap milling machines into a nice CNC.  I don't have time to solve that.


sdw

On 11/25/25 10:20 AM, Chris Albertson wrote:
I was excited to read about the continuous fiber printer.  Then I looked into it.   (1) I don’t think you need a special printer to use it, although I doubt it would work in a basic “Ender” 3D printer. and (2) the raw material is about $150 for 50 cubic centimeters. 

To put cost in comparison.  I pay about $0.02 per gram for basic plastics. ($20 for a 2000 gram spool)   If we assume the continuous stuff is as dense as water, then 50cc = 50ml = 50 grams.   So it is $3.00 per gram.   About 100 times more expensive.    I’m sure that I am wrong about the density, so the cost is higher. 


As for milling machines,  Harbor Freight is not bad for under $1K.   I think anyone who would want this would have the mechanical ability to attach stepper motors to the hand cranks. Then Open Source software will control those motors.   The harder part in both cases is learning CAD and then even harder, mechanical engineering.

Chris Albertson

unread,
Nov 26, 2025, 6:51:21 PMNov 26
to hbrob...@googlegroups.com
I forgot to show my DIN rail-based prototype system. You can mount the rails to anything you like. Professionals will mount them to the inside of a waterproof steel box, or you can just use a board. Then you can move and swap parts with no tools. Here is a setup for driving up to 8 stepper motors and some I2C-based sensors. The silver mount upper left is for a 1KW AC relay for later. DID makes this like building Lego once you mount the rail clips to all your parts.

DID is mostly used in Europe to build electrical panels for a building with circuit breaks and such. but it can be used for anything. Already my prototype is filling up, so it looks like the printer will need three rails.

IMG_1944.jpeg

Chris Albertson

unread,
Nov 26, 2025, 8:04:24 PMNov 26
to hbrob...@googlegroups.com

On Nov 26, 2025, at 12:39 PM, A J <aj48...@gmail.com> wrote:

That sounds exciting, having a 3D printer that creates a composite material.

Simply ceating composite material is not new or hard.    I can buy ABS with glass fiber mixed in for not much.   It makes a very stiff part with good mechanical properties.      But what is new here is the abilty to orient the fibers and to control where they are placed.

For most of us there are cheaper way to reenforce plastic, one idea I use is to print the part with a long threaded internal hole down the entire length of the part.  then I thread a long screw down the hole.  It acts like a steel reenforcing rod.    For good measure I can put some ca glue down the hole before the screw goes in.

The other thing is that I see a lot of very poor printed designs.   Simply doing better would drastically improve the strength.   So many people make 3D prints in the same shape they would use metal.   3D prints typically have solid shells about 1mm thick and then very low density interiors so unlike solid metal the streght is a funtion of the surface area, not the cross section area.     Yes you can print solid but this is structuelyy ineffescent and goive sub optimal stregth to weight ratio.   The better printed designs will look “puffed up” and have compound curves where possible.   3D prints work well for a strees skin desin robot and very poorly for one that has thin bones and soft skin.   I like to use the example of a Milwaukee cordless power tool as the BEST way to design with plastic,  the structure is in a curved shell with no flat spots and there is no interior chassis or frame.   using this style of design we can use cheap plastic and not have to resort to expensive fiber or even CNC’d metal parts.       But for a larger dog-bot or a human, you will need the metal parts as platic parts need to be larger then the allowed space for them.

If you have wheels, non of this matters as they are not senitive to power/weight ratios.  But walking or flying is very different.

I am slowly working to improve my CNC mill.  But then I fond I can simply crank it by hand and it’s done in 15 minutes.   Back in the 1940s they build mass produced turbo-charge airplane engines and car/truck transmisions with hand-cranked machine tools.  You really can get by without CNC if you have enough labor hours to burn.



Chris Albertson

unread,
Nov 26, 2025, 8:48:09 PMNov 26
to hbrob...@googlegroups.com
On Wednesday, November 26, 2025 at 11:39:04 AM UTC-8 Stephen Williams wrote:

It would be good to work out a combination that converts the cheap milling machines into a nice CNC.  I don't have time to solve that.


I have started to place such a solution on GitHub.   I’ve been lazy and not kept it up.    Here you can see my first attempt that work surprizingly well.  Most of thre conversion parts were 3D printed.    The plan was to come up with a boot-strap where first you do a plastic CNC, then use that to re-make the parts in metal. Then I found that the plastic was good enough because for example a plastic motor mount only has to handle the reaction torque from a motor.    All the high-stress parts come with the HF mill.   
Attach is a render from Fusion360.     It works well enough but there are a few more things to improve.   None of the added parts are expensive.  Even the big vertical ball scew is only about $50 or so.      I could make this available but some of it could be better designed

The video demo looks like it is moving slow, but only if you are used to looking at 3D printers.   The mill is dating with a Sharpie faster then you woud ever cut metal.  Maybe 10X after.   
CNC Converted Sieg X2 Mill v57.png
Here is a Youtube if when I first powered it all up some years ago.    It is old, but shows it working



Stephen Williams

unread,
Nov 27, 2025, 3:56:51 AMNov 27
to hbrob...@googlegroups.com

MarkForged has had continuous carbon fiber FDM printers for years.  But they started at $20k.  I have their demo parts they used to send out: A motorcycle brake lever with clear nylon + carbon fiber.  It is like cast aluminum but much lighter: Zero give no matter how much you try to bend it.

https://markforged.com/3d-printers/fx10


Chopped carbon fiber (CF) filaments are soimewhat stronger than plain plastic.  But they are a small fraction of the strength of continuous carbon fiber, either lay up + epoxy or with one of these printers.  It is the long, strong, light fibers that provide the full strength.  FDM continuous carbon fiber still doesn't get you full woven or wound composite strength because we are still stuck with flat layers, but it seems to come close.

These new printers might be a pain to work with.  But it seems promising.  The 80-90% reduction in prices is very nice.


sdw

Reply all
Reply to author
Forward
0 new messages