Building your own Flight Controller (FC) is a major milestone for any drone engineer. While gyroscopes and accelerometers (the IMU) provide stability, a magnetometer often called a digital compass to provides the "sense of direction" necessary for autonomous flight.
Here is a guide on why and how to integrate the LIS3MDL magnetometer into your custom flight controller project.

Why Do You Need a Compass in a Flight Controller?

A gyroscope can calculate orientation, but it suffers from drift over time. Without a reference point, the drone's heading (Yaw) will slowly rotate in the software even if the drone is pointing straight.

• Heading Accuracy: The magnetometer uses the Earth's magnetic field to provide an absolute North reference.
• Autonomous Navigation: For features like "Return to Launch" (RTL), Position Hold, or Waypoint missions, the FC must know exactly which way it is facing to move toward a coordinate.
• GPS Limitation: A GPS can tell you where you are and your ground speed, but it cannot tell you which way the drone is "nose-in" while hovering.

This image illustrates that a drone can move forward while facing sideways; the compass tells the FC the "facing" direction.

The Challenge: Magnetic Interference

The biggest mistake in FC design is placing the magnetometer "on-board" near high-current components.

• High-Current Wiring Effects: According to the LIS3MDL datasheet, high current in wiring and printed circuit traces can cause significant errors in magnetic field measurements.

• The 10mA Rule: Conductor-generated magnetic fields add to the Earth’s field. You must keep currents higher than 10 mA at least a few millimeters away from the sensor.

• EMI Sources: Motors, ESCs, and battery leads create massive electromagnetic interference (EMI).

Design Pro-Tip: Follow the "Cube Orange" or "Pixhawk" architecture. They often use a multi-PCB stack where the sensitive sensors (like the LIS3MDL) are separated from the power distribution and noisy IoT/Microcontroller logic.

LIS3MDL Magnetometer

The LIS3MDL is an ultra-low-power, high-performance 3-axis magnetometer. It is ideal for custom FCs due to its flexibility.
Key Specifications:

• Selectable Full Scales: ±4/±8/±12/±16 gauss.

• Digital Interfaces: Supports both I2C (standard and fast mode) and SPI interfaces.

• Operating Voltage: 1.9 V to 3.6 V.

• Resolution: 16-bit data output.

Hardware Installation & Schematic Guide

To integrate the LIS3MDL into your flight controller, follow the standard I2C implementation as shown in my provided schematic.
Wiring the LIS3MDL (I2C Mode)

To enable I2C mode, the CS (Chip Select) line must be tied high to Vdd_IO.

• Decoupling Capacitors: You must connect a 100nF capacitor (C1) between pin 4 and GND. Additionally, the Vdd line requires a 100nF high-frequency ceramic capacitor and a 1uF low-frequency electrolytic capacitor.
• Pull-up Resistors: Both SCL and SDA lines must be connected to Vdd_IO through external pull-up resistors (typically 2.2kOhm to 10kOhm. My schematic uses 2.7kOhm (R8, R9), which is excellent for Fast Mode I2C.

This diagram in the datasheet provides the "golden standard" for your PCB layout.

Using an External Compass (GPS Combo)

If your on-board PCB layout is too cramped or noisy, the standard industry practice is to use an External Compass.
Most modern drone GPS modules (like those based on the u-blox M8N or M10) include an internal magnetometer.

• The Technique: You bypass the on-board LIS3MDL and connect the I2C wires (SCL/SDA) from the GPS module to your FC's I2C port.
• Placement: The GPS/Compass is then mounted on a "GPS Mast" high above the motors and battery.
• Software Configuration: In firmware like ArduPilot or PX4, you simply select the "External" compass as the primary (Compass 1) and disable the "Internal" one to avoid interference.

Resources

• Official Datasheet: STMicroelectronics LIS3MDL Product Page.
• Journal/Documentation: Refer to the ArduPilot Compass Calibration Guide for details on how to handle "Soft Iron" and "Hard Iron" offsets in your custom code.
• PCB Design Jurnal: Search for "Minimizing magnetic interference in small UAVs" on IEEE Xplore for academic papers on trace routing for magnetometers.

While you can design for controlled impedance, the only place you can verify that design is by measuring your PCB traces with a Vector Network Analyzer (VNA) for how they actually perform at high frequencies as opposed to how you think they should.

You should start by getting calibrated and performing a standard Short, Open, Load, Through (SOLT) calibration to eliminate or correct for any systematic error introduced by the ends of your test cables. Calibration is very important; if you do not perform this process before measuring your test cables, your measurement will be invalid.

VNA SOLT calibration

After calibration, connect your PCB for measurement using appropriate fixtures such as SMA connectors or high quality probes (ensure you have a solid ground and avoid moving the cables, even minor changes can affect measurements). In general, for single-ended signals, you will measure S11 (return loss); for differential pairs you will use S-parameters such as Sdd11 and Sdd21.

VNA measurements

You can derive the impedance of a trace from the reflected data, i.e., the amount of reflected power will provide evidence on whether your 50 Ω trace is well-matched (low reflections equal high return loss). If you are seeing notable reflections, then it's likely that your trace's impedance is not the expected value. In addition to this, you can also use Time Domain Reflectometry (TDR mode) on the VNA to visualize changes in trace impedance over its length; this will help you find points of discontinuity such as vias or changes in width.

TDR mode on VNA

Also, Polygons, Launch Profiles, and Test Fixtures have an impact on results. Transition issues can appear to be an impedance mismatch when the Trace is actually correct.

Finally, you want to compare the measured data to your respective design targets. Small variations should be acceptable, but large differences usually mean that there is an issue with your stack-up, Trace Geometry, or manufacturing tolerance.

To summarize, the VNA will show you what an impedance looks like as an actual measured parameter versus as a theoretical parameter. It is the divergence from "should work" to "does work".

High-speed signals do not stop working in a trace. They keep reflecting and ringing and doing nothing but cause problems if there is no proper termination. Termination techniques are where series and parallel type of termination come into play.

With series termination, you place a resistor (usually in the range of 22 to 100 Ohms) very near the signal source. The purpose of the resistor is to match the source impedance to the trace impedance and reduce the reflection back into the source at the very beginning. This is a simple and low power solution that works very well for point-to-point digital interfaces like SPI or GPIO. The downside is you will have slightly slower edge rates due to the resistor creating an RC delay with the capacitance of the transmission line. However, with many designs this is an advantage rather than a disadvantage.

Series termination

For parallel termination, you place the resistor at the receiving end and typically match the trace impedance (50 Ohms to ground or reference voltage). The parallel termination absorbs incoming signals and keeps them from reflecting back down the line. This type of termination works best for very high speeds and/or long trace implementations where reflected signals would corrupt your data. The downside with parallel termination is when the line is driven you will have constant power through the resistor.

Parallel termination

To determine which option to choose, if a trace is short and the speed is moderate, series termination will usually work well and will save you power. If a trace is long, the speed is high, or the need for signal integrity is tightly controlled, then parallel termination will provide you with better control of reflections.

You should also think about topology. Series termination will perform best for point-to-point connections; however, parallel termination can accommodate the most demanding environments. In many instances, combining the two types of termination creates the best performance possible.

High speed signal termination comparison

Conclusion: at high speeds, you must terminate your signals; therefore, you need to choose wisely or you will have an oscilloscope display that is a train wreck.

To preserve signal quality over HDMI using high-speed differential pairs, a 100Ω differential impedance is required. HDMI's differential pairs require tighter coupling, constant spacing, and a continuous reference plane to maintain signal integrity. Signal traces should be as short as possible with no stubs, and all trace lengths should be matched within their tolerances. The HDMI connector should be placed at the board's edge, only using direct routing of differential pairs to minimize signal loss. Low-capacitance ESD protection devices should be added to HDMI connectors as close as possible to protect the circuit from voltage surges generated by ESD before they reach sensitive integrated circuits. A poor connection here will decrease video performance and could lead to ESD-induced signal failure.

To properly work for USB 3.0, high speed signal pairs need to be designed with a 90 ohm differential impedance. This requires an accurate control of trace width, spacing and dielectric layer stack-up in order to accomplish this level of precision. Differential pairs need to be routed together with good coupling and equal spacing, and no stubs or sharp bends. In order to achieve an accurate differential impedance and stable return path, a solid reference plane needs to be present below the differential pair trace. Length matching is important with respect to the signals, however, avoid over-exciting the length by using too many serpentine routings. The number of vias should be minimized, and if used, should be as symmetrical as possible. Inadequate routing techniques here will not only degrade the overall performance of the USB communications, but also will cause the loss of communication altogether between devices.

The return current path for any signal always has a return path even when disruptions occur in the return current path like when there is a split ground plane or there is a gap between different planes. If there is a disruption in the return path, the current will take a longer different route than it normally would and this will cause an increase in EMI, noise and loop area. Therefore, now you have turned your clean signal into a small FM transmitter. Keeping return paths continuous and below the traces is extremely important with high-speed signals. Do not run traces across plane splits and do not use stitching vias to change layers to maintain continuity of a path back to its source. Good integrity of a signal source is based not only on the physical characteristics of the trace itself but also on the physical characteristics of the physical return path to the source.

Charging the 12V LEAD ACID Beast with Beauty

If you own a motorcycle, you already know the pain. I was little out of station for a couple of weeks and when I came home and gave a push to my bike, there was nothing. The battery is dead. Lead-acid batteries self-discharge over time, and if you don't ride often enough, the battery just slowly dies on you. Buying a new battery every few months is not an option. So, I used to charge manually on a workstation which costs me time and money. Even though most of them are using just a transformer with a rectifier, no regulation, no protection, just brute force current being dumped into the battery. But I am an electronics engineer and that’s how I decided to give a try to lead acid battery charging.

The one that actually understands how to charge a battery properly. With trickle charge for deeply discharged batteries, constant current for bulk charging, over-charge for topping it off, and float charge to keep it maintained without overcharging. The whole charging profile, done right. That's where the CN3767 comes in. It's a dedicated 12V lead-acid battery charger controller IC from Consonance Electronics that does everything I just described. I have designed a PCB in EasyEDA and fabricated it from JLCPCB and tested out the final prototype.

PCB Design

You can download the Gerber files along with BOM and CPL from here. I have used JLCPCB for manufacturing because their services are available in a wide domain with reasonable prices. And because I have designed the PCB in easyEDA online which has an integration with JLCPCB. At least this can give me a peace of mind over the files.

Testing & Results: I set up a DC bench power supply set to 18V and took a 12V/7Ah lead-acid motorcycle battery. For monitoring put the multimeter on battery terminals, current clamp on charge line. There may be drop in wires and across device so the voltage at power supply and battery end may vary.

Constant Current (CC) Mode Test:

With the battery at around 12V, I connected the charger. The red CHRG LED immediately lit up, confirming the charger entered the charging state. The current stayed rock-steady at approximately 1A throughout the CC phase. That's the CN3767 doing its job, regulating the current via the sense resistor feedback loop.

Constant Voltage (CV) Mode Test:

As the battery voltage approached 14.8V, I observed the transition to over-charge mode. The voltage locked at 14.8V and current started tapering down from 1A. This is the critical phase where the charger is topping off the battery. The voltage holds steady while the current gradually decreases as the battery reaches full capacity.

Outro: Building a proper battery charger is not that hard when you have the right IC. The CN3767 takes care of the entire charging algorithm. Even though I am going to sell some pieces to my motorcycle repair shop, so everyone who needs a proper solution can get this design. I will increase the rating to maybe 3A in CC mode for faster charging. Motorcycle batteries can easily handle up to 5A. We have seen that the transition from CC to CV mode is smooth, the regulation voltages are accurate, and the LED indicators give clear feedback on the charging state. The best part? That MPPT input means I can slap a small solar panel on the bike someday and have a self-maintaining battery system. Charging the beast with beauty, indeed.

#PCB#

Tired of weak plastic threads? JLC3DP now offers Threaded Insert Service for professional-grade assembly strength.

✅ Specs: 8 types available (M3, M4, M5).
✅ Durability: High torque resistance & wear-proof for repeated use.
✅ Materials: Supports most Resins, Nylon & Plastics (excl. WJP, X-Resin, Metal).
Stop tapping, start building. Upgrade your parts today! 🚀

👉 more details: https://jlc3dp.com/help/article/threaded-insert-service
👉Try it now: https://jlc3dp.com/3d-printing-quote?from=fiona

#JLC3DP# #3DPrinting# #SurfaceTreatment# #PostProcessing# #RapidPrototyping# #ThreadedInsertsService# #hardware# #3DPrintingCommunity# #3dprints#

Using unconnected copper wires that have been left behind after routing changes (dangling traces) can lead to confusion, DRC errors, and unintended antennas. Run a Design Rule Check (DRC) to identify all unconnected and floating nets, and the copper that remains unconnected. Use your EDA tool’s net inspector or highlight feature to trace these unconnected and floating connections. Any floating or unused traces should be removed and/or terminated correctly, and each pad should be assured to be connected to its intended net. Having a neat layout helps to manufacture the board, reduces risk of EMI, and aids in debugging the design. A clean PCB design can look good, but also shows good engineering discipline.

When a single PCB does not meet your design needs, multi-board systems become necessary. Board-to-board connections depend on various connector types (such as mezzanine, pin header, edge connector), along with how they are aligned for signal integrity, to determine if your end product will be seamless or fall apart completely.

The connector type will largely determine the board to board connection. When it comes to mechanical stress during assembly, the board alignment must be precise. Even the slightest misalignment can result in poor electrical contact and/or early failure of a connection due to an inadequate connection time and/or temperature. Consider that shackles or mounting hardware (standoffs, screws) are required for an accurate spacing and that no flex should be introduced during assembly.

PCB with good alignment and connector type. Shackles or screws for accurate spacing.

PCB with good alignment

Signal integrity becomes increasingly difficult when transitioning through any method of connectors. To maximize signal integrity from the circuit to the connector, signal traces should always be kept as short and direct as possible. At any point in time, the geometry of the differential pairs should be observed and maintained when going into the connector. Additional vias and other components that can cause signal deterioration should be avoided near the connector. Finally, every high-speed connector has been designed to interleave ground pins to minimize the effects of crosstalk- therefore, utilize these pins with care.

PCB with interleaved ground pins

The way power is distributed among your components should also be considered. Do not use a single connection for the high current; try to use more connections to both lower resistance and lower heating in your system. To handle fast switching a good practice is to add local decoupling capacitors at the connectors on both boards to stabilize the voltage.

You should also consider how straightforward it is to join and separate two or more plates. Do you have pins or does each plate have a socket to make joining together easy? If you used connectors, will those connectors hold up to all of the assembled/disassembled action that will occur?

PCB with reliable board to board connections.

Summary: Multi-board systems do not only have an electric component, but also mechanical and common usage components. If you make good connections, you will end up with a fully functional system. If you make a bad connection, you will have a very expensive jigsaw puzzle.

Decoupling capacitors are invaluable components of reliable PCB design. They provide noise suppression, voltage stability, and keep your integrated circuits from operating chaotically.

Well placed decoupling capacitors

All digital integrated circuits will draw current in bursts when switching. Without local energy storage, this will create noise and voltage drops on the power rails. This is where decoupling capacitors come in. They are physically small capacitors connected in parallel with the power (Vcc) and ground (Gnd) that store charge at a local level. When a digital IC switches and requires current, the decoupling capacitor can provide the current immediately through a short trace instead of having to travel along a length of noisy power trace.

Decoupling capacitors should be placed very close to the IC's power + and ground - pins. Connecting the decoupling capacitor to the Vcc and Gnd using short traces is essential to limit inductance. If there are long traces connecting the decoupling capacitor to the Vcc and Gnd leads, inductance will increase; thus causing the decoupling capacitor to function poorly. Ideally, a via to a solid ground plane adjacent to the decoupling capacitor should be used.

Capacitor value is another important consideration. Many designers will use multiple values of decoupling capacitors in their designs:

Values of decoupling capacitors:

0.01 μF - 0.1 μF to suppress high-frequency noise

1.0 μF - 10.0 μF for low-frequency noise stability

Using small and large value capacitors assures that all frequencies between high and low are suppressed. Designers typically use decoupling capacitors of at least one 0.1 µF per power pin, and generally use bulk devices for low-frequency noise dampening.

Decoupling capacitors to suppress noise at high and low frequencies

The type of capacitor should be considered as well. Multi-layer ceramic (MLCC) capacitors are highly recommended; MLCC's have low ESR and ESL and are preferred for decoupling applications. Mount MLCC in accordance with their orientation, and keep them as close as possible to the IC.

PCB with multi-layer ceramic (MLCC) capacitors

Ground return is very important. Without a solid, continuous ground plane, decoupling won't work.

Decoupling capacitors are critical, not optional! Properly placed decoupling capacitors and proper capacitor values will yield a clean and stable operating circuit.

Backplane printed circuit boards handle many different types of high-speed connections between daughter cards, requiring strict layout discipline. Differential pairs should have controlled-impedance routing, and the stack-up should be consistent between all layers - every effort should be taken to minimize stubs using back-drilling as required. Via transitions should be well managed to help minimize reflections. Connectors should be aligned properly and spaced evenly, with higher pin counts having increased potential for crosstalk. Ground planes must be continuous to provide good return paths for signals. Mechanical and airflow concerns should also be considered during the design of a backplane, since they are typically used in cramped systems. A good backplane design will provide the most reliable signal quality/ integrity (SI), power distribution (PD) & mechanical reliability.

Two-dimensional layouts can be deceiving; however, the use of a three-dimensional (3D) PCB viewer in an electronic design automation (EDA) tool helps verify the heights of components as well as their alignment with connectors and fit into the housing (enclosure) prior to production. Using a 3D print model can also help in identifying other physical concerns such as possible collisions between tall components; inadequate edge clearance next to the edges; and misalignment of mechanical parts (housings/cases or heat sinks) and connections to the printed circuit board layout. Confirm all clearances for connectors, buttons, and mounting hardware. Properly routed PCBs can also fail mechanically if the 3D model has been ignored. Performing a simple 3D inspection prior to production will help eliminate expensive redesigns and assure proper fit and function of a PCB.