We've reached the next color, purple, which also coincides with the largest PCB I've made at JLCPCB, about 200mm*70mm. Initially I experimented with a homemade PCB, obviously aesthetically it didn't turn out too well, but I wanted to make sure the project didn't have any bugs (I still make mistakes 😅). It is a panel with five 5x8 matrices and MAX7219 drivers; I am going to use it with an ATmega328 microcontroller.

Who else has chosen this color?

#PCB# #Color#

I wanted to try out all the PCB colors, and I did just that. Continuing on from my previous post, I come with a PCB that I have made over time at JLCPCB, the color is green, probably the most common option and yours, I could say basic option. I won't talk much about the PCB, it's a simple adapter for the ESP-01/S module. 🤓

*I know my phone camera isn't great, so the color might appear a bit too dark to you.

#PCB# #Color#

In case anyone is still wondering what color to choose for a PCB, below are photos of my PCB (made by a hobbyist).

Just curious—does the chosen color have anything to do with your project's intended use? 🤓

Since the PCB I designed is related to temperature, I chose the color red.

#PCB# #color#

Overview

This circuit is a linear, temperature-controlled fan driver. Unlike simple on/off thermostats or PWM (Pulse Width Modulation) controllers, this circuit varies the DC voltage supplied to the fan smoothly. This "linear" approach eliminates the high-frequency switching noise often associated with PWM and prevents the "coil whine" that occurs when fans are run at very low voltages where they struggle to spin.

Key Functional Blocks

1. Temperature Sensing (The Input)

• Components: TH1 (10k NTC Thermistor) and RV1 (10k Potentiometer).

• Function: These two components form a voltage divider connected to the base of transistor Q1.

  • TH1 (NTC): An NTC (Negative Temperature Coefficient) thermistor decreases its resistance as temperature rises.
  • RV1 (Pot): This allows the user to manually set the temperature threshold (calibration).
  • Logic: As the device gets hotter, the resistance of TH1 drops, changing the voltage at the base of Q1. This signals the circuit that cooling is needed.

2. Control & Error Amplification (The Brain)

• Component: Q1 (2N5401 PNP Transistor).

• Function: Q1 acts as the control element.

  • As the temperature rises and the base voltage drops (due to the NTC), the PNP transistor turns ON more heavily.
  • It amplifies this small change in temperature into a control signal that drives the power MOSFET.

3. Power Driving (The Muscle)

• Component: Q2 (IRF540N N-Channel MOSFET).

• Function: This is a high-power transistor that handles the actual current for the fan.

  • Instead of acting as a simple switch (fully ON or fully OFF), it operates in its linear region (acting like a variable resistor).
  • As Q1 conducts more current, it increases the voltage at the Gate of Q2. This lowers the resistance between the Drain and Source, allowing more current to flow through the fan.
  • This results in the "smooth ramping" mentioned in your text. The fan speed increases gradually rather than in steps.

4. Hysteresis (Stability)

• Components: R1 (100k) and R5 (220k) (and potentially the interaction with R3).

• Function: The text mentions "shutdown hysteresis prevents frequent cycling."

  • Without hysteresis, if the temperature hovered exactly at the threshold (e.g., 40°C), the fan would rapidly turn on and off (chatter), which is annoying and wears out the motor.
  • The feedback resistors (likely R1 and R5) create a "lag." Once the fan turns on, the threshold temperature for turning it off is lowered. For example, it might turn on at 45°C but won't turn off until the temperature drops to 35°C. This creates a stable "dead band."

5. Coil Whine Prevention (The Zener Diode)

• Component: D1 (5.1V Zener Diode).

• Function: The text states the circuit "avoids applying voltage just at the threshold... preventing... coil whine."

  • Fans often make a buzzing or whining noise at low voltages (e.g., 1V–4V) because the magnetic field isn't strong enough to spin the rotor smoothly; it just vibrates.
  • The 5.1V Zener diode likely sets a minimum voltage threshold. The circuit is designed so that the fan stays completely OFF until the temperature is high enough to drive the voltage past this 5.1V point. Once it turns on, it starts at a healthy speed (silent operation) rather than a slow, whining crawl.

6. Filtering

• Component: C1 (100uF Capacitor).
• Function: This capacitor is placed across the fan terminals (parallel to the fan). It smooths out any voltage ripple and provides a burst of current when the fan motor demands it (motors are inductive loads), ensuring stable operation and reducing electrical noise.

Summary of Operation

1. Cold State: The NTC resistance is high. Q1 is OFF. Q2 is OFF. The fan receives 0V (or is disconnected) and is silent.
2. Warming Up: Temperature rises. NTC resistance drops. Q1 begins to conduct.
3. Activation: Once the threshold is passed (and the Zener voltage is overcome to avoid whine), Q2 begins to conduct. The fan voltage ramps up smoothly from ~5V to 12V.
4. Hot State: Q2 is fully ON (low resistance). The fan receives the full 12V for maximum cooling.
5. Cooling Down: As temp drops, the fan slows down smoothly. Due to hysteresis, it stays on slightly longer than it would in a simple on/off circuit, preventing rapid cycling.

The circuit description was created using A.I.

Source: *https://www.youtube.com/watch?v=YGpXFbxD_Wk&ab_channel=REDShade*

I did not create this circuit, I just designed it for my own use.

After working on my heated cup holder, I spent some time checking the 3D-printed housing before sending it to print.

And I nearly missed one small thing.

There are quite a few parts packed into a relatively small space — the heater, temperature sensor, wiring, and the cup itself. When everything is visible at once, it’s surprisingly easy to overlook whether a specific part is actually sitting where I intended.

So I started checking the assembly in a 3D Viewer from different angles.

Being able to highlight or isolate individual parts was pretty useful here. Instead of rotating the whole model around and trying to figure out which part I was looking at, I could focus on one component and check its position.

It sounds like a small thing, but catching a layout issue before printing the housing is definitely better than finding it afterwards.

Anyone else had this happen — you only notice a small layout issue after the part is already printed? 😅

My commute has gotten longer recently, and I’ve started noticing one small problem:

My coffee gets cold before I finish it. 😅

I drink something hot almost every morning, so I decided to try making a simple heated cup holder instead of just accepting it.

The basic setup is pretty simple:

• A flexible heater for the heat source
• A temperature sensor to keep an eye on the temperature
• A simple controller
• A 3D-printed housing to hold everything together

The part I found more interesting was the heat distribution. I didn’t want one small area to get really hot while the rest of the cup stayed lukewarm, so getting the heater placement right took some experimenting.

It’s definitely a small DIY project, but I’m curious to see how well it works once I use it through the colder months.

Anyone else working on small winter DIY projects lately?

I’ve been looking into flexible heaters recently, and something I find interesting is that reaching the target temperature doesn’t necessarily mean the heating is uniform.

From what I understand, a few things can affect where hot spots show up:

• Trace spacing
• Current density
• Sharp corners in the trace
• Uneven contact with the surface
• Insulation and thermal conductivity
• Power density

I’m still learning about the design side of this, so I’m curious how others look at it.

When you’re trying to get more uniform heating, what do you usually pay attention to first — the trace layout, power density, or the surface the heater is attached to?

I’ve often heard the rule: “Never use 90° traces.”

But I’m curious how much this actually matters in real-world PCB designs.

My current understanding is that, for many low-speed signals, a single 90° corner is unlikely to cause a noticeable problem. The corner does slightly change the local trace geometry and may introduce a small impedance discontinuity, but whether that matters seems to depend on things like:

• Signal rise/fall time
• Trace width and stackup
• Controlled impedance requirements
• The rest of the signal path

For faster signals, RF layouts, or tightly controlled impedance traces, I can understand being more careful. But for ordinary routing, is avoiding 90° corners mostly just good convention rather than a strict technical requirement?

I still tend to use 45° bends, but I’m interested in how experienced designers approach this in practice.

Do you avoid 90° corners in every design, or only when the signal requirements justify it? Am I missing any important considerations?

We’ve seen situations where two sheet metal panels were made to the same drawing, with the same material and finish, yet they didn't look quite the same when placed side by side.

The geometry wasn't the problem.

Material grain and the direction of forming or finishing can influence how light reflects from the surface.

This becomes much more noticeable on large visible panels, especially when several parts sit next to each other.

It’s one of those details that may never appear as a dimensional failure but can still affect how a finished product is perceived.

For visible sheet metal, “same specification” doesn't always guarantee “same visual result.”

💬 Discussion:

When appearance matters, how do you control grain direction and visual consistency across multiple sheet metal parts?

#SheetMetal# #ManufacturingEngineering# #SurfaceFinish# #ProductDesign# #Fabrication#

We tend to think of laser cutting as a simple profile-making operation.

But heat is part of the process too.

On thin or heat-sensitive sheet, a dense pattern of closely spaced features can change how the material behaves during cutting. The issue may not show up as an obvious dimensional failure immediately; flatness and edge condition can tell a different story.

This becomes especially interesting when a large panel contains many small cutouts.

The geometry may be perfectly reasonable from a CAD perspective.

The cutting sequence and heat distribution are where manufacturing reality starts to matter.

💬 Discussion:

Have you ever changed feature layout or cutting sequence because of heat-related issues?

#SheetMetal# #LaserCutting# #ManufacturingEngineering# #Fabrication# #MechanicalDesign#

Integrating PEM fasteners into sheet metal looks straightforward until the part reaches assembly.

Hole size, material thickness, proximity to bends, and installation force can all affect the result.

We’ve seen designs where the sheet itself was strong enough, but the fastening point became the limiting feature.

The interesting part is that the problem wasn't really with the fastener.

The surrounding sheet geometry hadn't been designed around the way the fastener would actually be installed and loaded.

With sheet metal, the connection deserves as much attention as the component being connected.

💬 Discussion:

When specifying PEM or other pressed-in fasteners, what tends to cause the most trouble in your assemblies?

#SheetMetal# #PEMFasteners# #MechanicalDesign# #Manufacturing# #DFM#

90°? The Metal Disagrees.

You bend sheet metal to 90°… and it springs back a little. 📐

That’s springback—the material releases some of the stress from bending and tries to return toward its original shape.

So the press brake may bend beyond 90° to get the final angle you actually want.

Metal has opinions. Engineering has to negotiate. 🛠️

#SheetMetal# #SheetMetalFabrication# #DFM# #Engineering# #Manufacturing#

We once saw a CNC part where the surface finish gradually changed during production, even though the cutting tool was still within its expected life.

At first, it looked like a tooling problem.

It wasn't.

The real difference came from tool engagement. As the machining path changed, the cutter was removing material under different cutting conditions. The same tool could behave quite differently depending on how much material it was carrying.

This is easy to miss when looking only at tool life.

Tool condition matters, but so does what the tool is actually doing at each stage of the cut.

That experience changed how we evaluate surface consistency in production. Tool replacement alone doesn't always solve variation.

💬 Discussion:

When surface finish starts drifting during a run, what do you check first — tooling, cutting parameters, or the machining path?

Hashtags:

#CNCmachining# #ManufacturingEngineering# #MachiningProcess# #CNCProduction# #Manufacturing#

🛠️ CNC pros!

Let's settle a classic shop debate before the weekend. You can only keep ONE tool type in your carousel for the next project. The rest are banned.

What’s staying in the machine?

1️⃣ End Mill (The multi-tasker) 2️⃣ Face Mill (The cycle-time saver) 3️⃣ Chamfer Mill (The part-finisher) 4️⃣ Drill Bit (The hole-maker)

Drop your choice and defend your livelihood in the comments! 👇

#JLCCNC# #CNCMachining# #MachinistLife# #Manufacturing# #Engineering# #MechanicalEngineering# #Workshop# #Tooling # #Metalworking #CNC#