Making PCBs for a hobbyist can be frustrating because of lead times. While it is incredible that we can get high quality PCBs from manufacturers within a week, there are still some significant gripes to contend with.
-
Shipping can be expensive
Sometimes you just want a small, simple PCB which will cost 5 dollars, but then you get a shipping fee of 20 dollars which can really make you think twice about whether you actually need that PCB. -
Rapid prototyping is not possible
A one-week lead time is impressive, but can still be too slow. If you need the PCB to continue the project, you really can't do anything until they arrive. If you made a mistake, you then have to repeat the process which can become incredibly time consuming and frustrating. And paying for faster shipping just makes the whole process more expensive.
Our solution to this would be to make a highly accessible and inexpensive PCB mill that almost any mid to high level electronics enthusiast could use. The purpose of this mill would not be to replace purchasing of PCBs from actual manufacturers. This mill would serve as a way to make "draft" PCBs that the user could test and verify before sending the design to an actual manufacturer. (If users are satisfied with the quality of the draft PCBs they could just use those in their projects too if they wish.)
There are those that enjoy tinkering and fixing their machines, but there are also those that just want something that works. Our machine would mostly target people who want a PCB mill mostly to use as a tool, not as a hobby. This would mean creating a mill with a solid workflow that doesn't require much intervention from the user.
In order of importance, aspects of out machine that could distinguish it from others
- Ease-of-use
- Lower the tedium of using the mill as much as possible
- Reliability
- Keep the failure rate as low as possible.
- Accuracy
- The ability to mill better PCBs
- Convert a 3D printer gantry into a pcb mill.
- Salvage the parts off of already existing 3D printers for parts to make our own gantry.
- Modify a desktop CNC mill to tailor it for pcb milling (increase its accuracy, etc).
Each category is on a scale from 0 (worst) to 2 (best). Each column has a weight. Higher multipliers means the category is more important.
- Viability
How practically possible the idea is. - Risk
How many engineering and development issues have been forseen or could be encountered. - Complexity
How complicated would development be. - Time Commitment
How much time it would take to develop the design. - Upgradability
How upgradable to the end-user the final design would be. - Flexibility
How easy the final design would be to customize or modify. - Cost
How much the final design would cost to the end-user. - Accessibility
How easy it would be for an end-user to use construct or otherwise acquire the product for themselves. - Usability
How easy it would be for the end-user to use the design.
| viability | risk | complexity | time commitment | upgradability | flexibility | cost | accessibility | usability | total | |
|---|---|---|---|---|---|---|---|---|---|---|
| Weight | 3 | 2 | 2 | 3 | 1 | 1 | 2 | 2 | 1 | - |
| Convert a 3D printer | 1 | 0 | 1 | 2 | 1 | 1 | 2 | 1 | 1 | 20 |
| Custom Gantry | 2 | 2 | 0 | 1 | 2 | 2 | 1 | 1 | 2 | 23 |
| Mod an existing CNC | 1 | 1 | 1 | 2 | 0 | 0 | 1 | 0 | 1 | 16 |
From this decision matrix and other feedback, we've decided that it's best for us to make a custom gantry. In addition, since 3D printers are so ubiquitous nowadays, it would be beneficial if we could try designing a custom gantry that uses salvaged parts off of an old 3D printer (an Ender 3, for example).
Our mill should include these hardware features
- Lead screws for all axes
- Increases accuracy
- Steppers could be geared down for even higher accuracy at the cost of movement speed.
- Linear rails for X/Y axes
- Reduces vibrations
- Better than cylindrical rods
- Linear Scales for feedback control
- Increases accuracy
Our mill should be able to produce PCBs with these minimum specs and tolerances. This should satisfy milling requirements for a QFP which is the smallest package that most hobbyists would reasonably encounter.
- 2-layer milling
- Minimum track width: 0.25mm
- Minimum track spacing: 0.25mm
- Minimum via size: 0.8mm
- Track width tolerance ±20% (±0.05mm at smallest width. Should be possible with lead screws)
- Minimum build area: 100mmx150mm
- Maximum CNC base dimensions: 500mmx500mm
- Maximum CNC height: 400mm (not strict)
- 80 W per hour power consumption
- At most 2 hours per PCB, typically.
- Minimum achievable spindle speed: 10000 RPM
Some of these specs are modeled after a typical desktop 3D printer.
- Make a working gantry
- Single layer milling
- Through-hole drilling
- Solder mask scraping
- Make a custom spring-bit to remove rolled on solder mask to expose pads.
- Two layer milling
- Finding a reliable way to align the two sides well is super important.
- Via insertion?
- Pick-and-Place (PnP) attachment
- Solder paste printer
- Ricky Cui (EE 2025)
- Azam Khan (ME 2025)
-
Nasty chemicals
A low-tech way to make PCBs is to mask the traces and etch them using an acid to dissolve away the copper, leaving the traces behind. The masking can be done with a light sensitive photoresist layer on top of the copper that gets removed by a laser or UV light, or some sort of ink that is transferred to the copper. Either way, this is frankly a terrible way to make PCBs. It involves large amounts of chemicals that can be harmful and makes a huge mess. Often the quality of the traces isn't very good and thin traces are very easily destroyed in the process. -
Other commercial PCB mills
Decent commercial solutions to at-home PCB manufacturing (i.e. mills that are actually marketed as being able to make PCBs) are pretty much aways in the range of about $3000 to over $5000. This is out of reach for almost all hobbyists. -
Additive PCB Manufacturing
This involves printing the PCB traces with some sort of conductive paste. Some commercial machines on the market use this technique but it comes with some downsides:Additive PCBs are exotic and seem to be meant only for niche experimental use. The manufacturing process differs so much from a normal PCB you'd order from a regular manufacturer that it is difficult to believe that you could use one to verify your normal designs meant for subtractive manufacturing.
Also, we don't know what the conductive paste is. It's not solder paste but something else that acts as PCB traces. It could be a proprietary thing and, you might be forever chained to buying the materials from the original manufacturer. Their nominal resistance might also be higher than that of regular copper traces.
- Create a PCB design and generate a gcode file for both the front and back of the PCB.
- There already exists software to do this.
- Calibrate and zero the CNC to a known point.
- Probably will use end stops.
- Clamp down a sheet of copper clad FR4 in a pre-determined location.
- Mill the first layer. Do not drill holes yet.
- Flip the board along the diagonal. This should keep the same corner at the origin point.
- Mill the second layer.
- Drill holes.
- Use thin wire through the via holes to create vias.
- Solder and clamp them flat.
- Use small copper rivets to plate through holes.
- Clamp them down and solder.
- Done
- Create a PCB design and generate a gcode file for both the front and back of the PCB.
- There already exists software to do this.
- Calibrate and zero the CNC to a known point.
- Probably will use end stops.
- Clamp down a sheet of copper clad FR4 in a pre-determined location.
- Mill the first layer. Do not drill holes yet.
- Flip the board along the diagonal. This should keep the same corner at the origin point.
- Mill the second layer.
- Drill only via holes.
- Use thin wire through the via holes to create vias.
- Solder and clamp them flat.
- Apply soldermask on first side.
- This will cover everything including vias and undrilled through holes.
- Use CNC to expose surface mount pads and undrilled through holes.
- Repeat 9 and 10 for the other side.
- Use small copper rivets to plate through holes.
- Clamp them down and solder.
- Done
- Mandated use of safety glasses while the machine is running.
- CNC will be enclosed in some way to prevent particles and shield from debris.
- A vacuum will be used to remove particles created from cutting.
- An alternative is to do the milling in a puddle of oil to trap all of the particles.
20W-Fibre Laser
This uses a powerful laser to literally burn away the copper and even the FR4 to make the PCB. It's very accurate and very fast. However, this is still quite expensive at $1500-$2000 or greater. Also it releases a lot of fumes from the burning that definitely require a good ventilation system.
- https://hackaday.com/2021/01/11/laser-blasts-out-high-quality-pcbs/
- https://www.kurokesu.com/main/2021/01/07/making-fine-pitch-pcb-prototypes-with-fiber-laser/
Other no-name CNCs
Some of these are advertised to be able to make PCBs but the cheap ones usually have dubious accuracy. You don't see many people using these since they're not really meant for this purpose.
(Example video) (No links to specific products since you can find them basically everywhere.)
Makera Carvera
At $6000, most people couldn't afford this. It is designed to be able to make good PCBs.
Voltera V-One
An additive PCB manufacturing machine. Also not very affordable. And as mentioned earlier, additive PCBs are too different from normal PCBs, so it's use is constrained to niche experimental designs.
X. C. Wang and H. Y. Zheng, “High quality laser cutting of electronic printed circuit board substrates,” Circuit World, vol. 35, no. 4, pp. 46-55, Nov. 2009, doi: https://doi.org/10.1108/03056120911002415.
This research paper from Wang and Zheng, from the Singapore Institute of Manufacturing Technology, explores fabrication of PCBs using different laser cutting settings on a diode laser. The paper aimed to explore fabrication that minimized charring, delamination, and the heat affected zone that are often found on laser cut PCBs.
This paper was useful to compare PCB milling to laser cutting methods. Laser cutting appears to be very viable as a process, however it can be difficult to get the settings correct and the required equipment is expensive.
D. Wise, "PCB Fabrication," Technology Interface International Journal, Fall 2007. [Online]. Available: https://tiij.org/issues/issues/fall2007/01_Wise/Wise-PCB%20Fabrication.pdf. [Accessed: 30-Sep-2024].
This is a paper from a Departmental Engineer at the University of New Mexico (UNM). It discusses the use of a mill to cut printed circuit boards(PCBs). The paper is written somewhat informally, and the references listed are not the most reliable. However, most of the information within the paper describes the hands-on process at UNM.
The paper walks the reader through the process of manufacturing a PCB on a desktop mill, all the way from design to settings up the machine and cutting. It discusses common issues and how to effectively set up a FR4 board.
A. Nae and B. Toma, "Study on Equipment for Manufacturing PCB," Journal of Petroleum-Gas University of Ploiesti, vol. 1, pp. 9-15, 2010. [Online]. Available: http://jpgt.upg-ploiesti.ro/wp-content/uploads/2024/02/9_T_1_2010_Nae-Andrei-BT.pdf. [Accessed: 30-Sep-2024].
This is a paper from two researchers at Petroleum and Gas University of Ploiești. It discusses the use of stepper motors for creating CNC routers. It highlights design principles and details, and different ways in which routers are frequently designed. There is also some discussion on electrical design and software controls.
This is useful as a guide for designing a rigid motion system that can carry a spindle. There is good discussion on different design choices such as moving gantry vs bed, as well as different components and how they affect the motion system.
Y. Crama, O. E. Flippo, J. van de Klundert, and F. C. R. Spieksma, "The assembly of printed circuit boards: A case with multiple machines and multiple board types," European Journal of Operational Research, vol. 98, no. 2, pp. 457-472, 1997. [Online]. Available: https://www.sciencedirect.com/science/article/pii/S0377221796002287. [Accessed: 30-Sep-2024].
This paper by Crama et al. Optimizes the process of placing components onto a PCB in a production line.
This paper was not very useful to this project; while it provided some insight into PCB assembly, it was highly focused on a factory setting rather than a custom production/hobbyist environment.
"Principles of Mechanical Design," MIT Center for Bits and Atoms, 2021. [Online]. Available: https://fab.cba.mit.edu/classes/865.21/topics/mechanical_design/index.html. [Accessed: 01-Oct-2024].
A guide from MIT on mechanical design principles for building rigid and precise systems. It discusses important concepts such as stiffness, backlash, pretension, and more.
This is a useful reference for any potential designs in the future. It outlines many important fundamental concepts, both theoretically as well as physical design choices.
K. K. Tan, T. H. Lee, and S. Huang, Precision Motion Control: Design and Implementation, 2nd ed. London, UK: Springer-Verlag, 2008. Available: DOI: 10.1007/978-1-84800-021-6
This is a textbook on motion control systems. It covers everything from actuators to gantry designs, to control systems, and more.
It is very useful as an in depth guide to designing a motion system. While it is a bit old, many of the fundamental concepts are still applicable.