I’m not sure if this is a good idea or a bad idea… I know that one can make fairly crude PCB resistors (like 3D printer heated beds, and the PCB power resistor used in the PowerCore EDM). I am working on a circuit that has some crude requirements for low value resistors (0.1 to 1 ohm) to go in series with decoupling capacitors (0.1 to 1 uF). I don’t need an exact resistance, and a simple PCB trace resistor might work nicely.
I realize that many factors of PCB fabrication will alter trace resistance, but I’m wondering if anyone else finds the idea of PCB trace resistors useful as a library component, or perhaps they’re already implemented somewhere in KiCad? It would always be a crude approximation of the value, and one would need to know the copper thickness to calculate the approximate resistance. Whether a serpentine or linear path, I don’t really care… though it would be cool to have some auto-generated options.
I realize that I can create a custom footprint to accomplish this. Today I manually calculate trace width & length, then put that trace line between two pads, and save as a component footprint. I think this might be a great feature for some minor automation (somehow have the footprint editor input the copper layer weight, list the total trace length between pads, and output the expected resistance). I have no idea how that might work nicely with Kicad’s current footprint library architecture / interface.
This feature might also be useful for adding a footprint feature to verify the copper layer thickness on a manufactured board (drop in a PCB trace resistor with an estimated resistance, and see what actually gets fabricated on your board). I’m guessing some manufacturers already do something like this, I’m clearly not the first to think of this general approach.
Please let me know if this is already implemented somewhere, and I’m just completely missing it!
I have used some mm of track when I have read that for stability I need at least 15 mΩ and that typical ceramic capacitors can have ESR as low as 10 mΩ. So, for sure I needed to add 5 mΩ.
5mΩ resistor has a price about 30 times higher than 1k.
Another use case: current measurement for power electronics. Current shunt resistors are also expensive (and often voluminous). A pcb-resistor is not presice (in fact its very imprecise), but it worked good enough for overcurrent-detection.
To be fair: if possible I try to avoid the pcb-resistor-solution, I have used it only few times in the past.
Cheap power resistors, low cost temperature measurement (resistance change versus pcb temp), heaters, making inexpensive but accurate resistance ratios, not including another tiny component on a pcb when unnecessary (less to pnp for manual work), not needing to buy a variety of small resistors or power resistors (which are more costly than you might expect). I realize that resistors are generally cheap.
I didn’t know the name for this. Thanks for the reference!
I guess it will come down to making custom components by hand for this, but at least NetTieResistor or something similar is a better name than PCBResistor.
On reddit someone posted a couple of relevant script links for this:
I was hoping there might be a more elegant solution than manually creating the components right as I go to use them. It’s niche enough that I wouldn’t expect dev’s to spend time on this. Thanks for the suggestions!
I think I have seen at least 5 of such projects, with various complexity and intentions. For example, one of them is for generating coils for a BLDC motor, another handles inductors on multiple layers.
The power resistors are not so cheap, because PCBs cost money, and the board would have to get HUGE in order to make a resistor with a reasonable resistance and able to dissipate significant power.
Temperature measurement is a possibility, although it would require an accurate current source and high-resolution ADC. A 10¢ diode or thermistor seems like a better idea.
An accurate resistance ratio is interesting, but the extremely low resistance means either high currents or huge areas. Seems like a bad solution to the problem,
I’ve worked in the business for 40 years, and I’m always looking for ways to save a few cents on a board, but I’m not convinced this design technique is very useful.
I agree that it’s usefulness is limited, but there are some niche applications. I have seen it used on those EUR2 SMPS boards from China. Due to the temperature coefficient of copper (and also production tolerances) accuracy will be very moderate (maybe 5% or 10%. 20%?) But it may be good enough for current measurements and protecting against gross overloads. On those boards a track of about 1mm wide and 3cm long was used.
The main advantage of this method is not the (only) the cost of the resistors themselves, but it also costs time on the machines (bot setup and placement of each). Also, you can scrap an item from your BOM, which is one less part to order and keep track of inventory.
For the same reasons sometimes PCB tracks are used as fuses by deliberately making a narrower section. Below a video from mikeselectricstuff about the results of such a cheap pcb fuse in a relatively expensive inverter microwave oven. Panasonic microwave destroyed by blown bulb. - YouTube
If you follow the design rules, then the power resistor cannot be made from a copper track… The shunt must be of a certain alloy and, when heated, minimally change the resistance… Low-power circuits use conventional resistors with dual power margins…
While the cost of the component is minuscule, the purchase, inventorying, placing it on the board and testing it can add up to much more. A buddy of mine worked for Black & Decker. He told me the things they would go through to save a fraction of a cent.
