README.md
View this project on CADLAB.io.

TODO


use 2 current sense resistors in parallel
fugu reverse battery protection
The metal cover ESP32-S3-WROOM can be a condensation trap in humid environments. Especially when power saving at night and the chip cools down.


Design Principles

Keeping it simple, minimum number of components
Only 2 layer PCB so this can be made at home
Not using too small SMD components, mostly 0805
Using TO220 switches for the buck for easy maintenance
Keeping it dense. This reduces parasitic L and R.

Compared to the original design:


Multiple input caps to increase life-time, efficiency and (hopefully) reduce EMI
2 HS switches in parallel to reduce heat and increase efficiency
Using ESP32 ADC to sense input voltage, INA226 for bat voltage
Using current sense resistor with INA226 (immunity against magnetic fields, less drift)
Current sense on the battery side (less noise, can detect bat reverse current)
Low-side current sensing because it supports up to 60V battery voltage (for 16s lifepo4 48V batteries)
Backflow switch on battery side (bat eFuse if sth goes wrong, can safely handle a short of the high-side switches)
Backflow switch powered by HS charge pump from the IR2184 (no extra DCDC)
Much higher 1.4 A gate drive current (previously 130mA) to reduce switching losses (less heat but more ringing)
Introducing snubber circuit to reduce EMI and MOSFET stress
TVS protection circuit
off-board 3.3V and 12V power-supply
off-board programmer
Improved voltage and current sense PCB Design (voltage divider & filter caps close to ADC)


PCB Design Notes


Current sense: INA226 datasheet ]]>https://www.ti.com/lit/ds/symlink/ina226.pdf#page=30]]>
Current sense: INA226 datasheet ]]>https://www.ti.com/lit/ds/symlink/ina226.pdf#page=30]]>


Schematics Design Notes


Current sense filtering ]]>https://www.ti.com/lit/ds/symlink/ina226.pdf#page=14]]>


TI: Fundamentals of MOSFET and IGBT Gate Driver
Circuits ]]>https://www.ti.com/lit/ml/slua618a/slua618a.pdf?ts=1691532999585]]>

High Side Mosfet


suffers from switching stress
short switching times to lower switching losses but increase potential ringing (EMI)
design switch node RC snubber circuit
Consider switching loss to be higher than i2r loss. when choosing the component, consider switching times, and reverse recovery characteristics of the body diode
Driven in quadrant ?
Choose a High-Side switch with low Q_rr (TODO src) ]]>how to compute qrr]]>
Qrr rises with increasing temperature
don't use CDS19505, better TK3R3A06PL


Low Side Switch

Current through the LS Switch always flows from source to drain (4th quadrant of their V-I plane),
which makes the gate drive signal rather irrelevant. It is much easier to switch than the HS, ringing is generally
not an issue. ]]>gate drive fudamentals]]>
Choose a MOSFET that is designed for synchronous rectification. Q_rr can be high. 
Prefer low body diode forward voltage
CDS19505 is a suitable choice.

Coils

Inductivity


Choose inductivity L (link TODO)


lower inductivity, higher ripple current
higher voltage -> higher inductivity
max flux (core saturation)



Core Geometry


toroids have low stray inductance but a hard to wind.
PQ-Core


Core Size


Depends on the power needs and switching frequency
The Bigger the better


reduces flux density and core loss (but higher copper loss)
less thermal issues
but: more expensive, need more space

toroid sizes that make sense: T130, T184, T225/T226


Geometry


Choose core geometry ( and size (depends on power needs). Choose core materials.
Sendust aka KoolMu is a good choice. It is an alloy powder, composite of metal and plastic, distributed air gap.


sendust has high saturation current, so T130 works. however, wire diameter is limited because it just doesn't fit
through
smaller cores have smaller A_l value, so need more turns => more copper loss

Choose wire gauge and strands (consider DC loss and skin effect)
Compute num windings with A_L value and target L
Designers: ]]>micrometals]]>
The bigger the better (usually)
Wire easier to cool than core. Power loss ratio core/wire:
20/80 ]]>micrometals core design considerations]]>
Higher initial permeability increases A_L, reduces num windings, moves loss from wire to core
Sendust Toroid 60u - 125u
T184 (OD=1.84in/46.7mm) 125u A_l=281 ]]>https://www.semic-shop.de/ljf-t184-s-125a-bk-de/]]>
17-20 turns of 5xAWG15 (1.45mm)
T225 / T226 (OD=2.25in/57.15mm)


T130


a single T130 with A_l=61 needs a lot of windings
stack 2 cores to reduce windings by 2^.5. for same inductivity
using 2 strands of 1.8mm wire (140cm length each) works
TODO test cores:


Ljf T130-S-125A BK
Ljf T130-S-075A BK
Ljf T130-NF-125A BR



Materials

three opt spots: saturation vs loss vs costs


Sendust (black, power supply, 10,5kGauss)
Super Sendust (PV inverter, 12kGauss
Sendust Plus
Neu FLUX


Advantages of bigger cores


bigger volume => less magnetic flux density
more surface => better cooling
usually higher A_l => need less windings


less copper loss due to reduced length and thicker wires

can fit thicker wires and/or more strands


disadvantages


more loss, scales proportional to core size
cost, size, weight


]]>https://www.micrometals.com/design-and-applications/material-selection-a...]]>

Coil designs


T184 sendust (u_i=125, A_L=281)
N=17
5xAWG15 (d=1.45mm) or 3x (d=1.9mm)
130cm wire length for 17 turns on T184 (17 * 65mm)


Strands Formular: d_b = (d_a2 * n_a/n_b).5 # d_b = diameter, n_b = num strands

Gate Drivers

Higher gate drive current -> faster switching -> less switching losses/heat, but more EMI/ringing


Careful with mosfet body diodes that have a high Q_rr
Measure ringing from switch-node to GND and place snubber if needed
IR2104 has a rated output current of 290mA, which is suitable for 33ohm or 47ohm gate resistors. For power >500W this
is a bit slow.
IR2184 has 1.9A. still can still be a bit weak, causing voltage bounce at the driver output
2ED2748S01G


Eff Opt


losses in sync buck ]]>https://www.ti.com/lit/an/slvaeq9/slvaeq9.pdf]]>
main losses are switch loss and inductor loss
use thermal imager and start with the components that produce the most heat
cheap caps with higher ESR can get hot (especially C_in). Find better caps, place caps in parallel to reduce ESR
Check coil core material loss in datasheet. Use bigger core.
Use thicker copper wires. To reduce AC losses (skin effect) use multiple strands in parallel
Reduce switching times: smaller gate resistors, stronger driver. Make sure there is no severe ringing at switching
node and at the gates. Consider using a faster Misfit (low Qgd, low t_rise, low Qrr)
Place Schottky diode in parallel to LS switch
Use a second HS switch in parallel (with separate gate drive resistor)
Use short & wide PCB traces, maybe 4-layer PCB.


#

Current Sensor

Temperature drift of components (resistor and amplifier) might be greater than accuracy of the sensor, so choose a small
burden resistor. Use large copper areas and thermal vias to increase heat dissipation.
For 30A it should be smaller than 1mOhm, e.g. 0.5mOhm . The INA226 is quite precise (16bit) but can suffer from temperature
instability if placed close the burden resistor (and potentially any other components that heat up).
With a 1mOhm I experienced a temp drift of 80mA at about 70°C (30Amps) which is about 2W at 24V.

Reject Noise


50hz inverter
39 kHz pwm
380Hz Power supply


IR2184 and TK6R8A08QM

I want the charger to be efficient, small and without a fan. No fan and high efficiency is both reducing loss. Si
Mosfets waste most of the energy during switching. The faster the switching, the less energy is wasted. Need high gate
drive current, because the gate is like a capacitor. For simplicity and safety, want a driver with dead-time, Infineon
strongest driver from the IR series is IR21x4, almost 2 A current. Considering power level of 800W. Choose a fast MOSFET
with low Qg and Qrr.

Snubber


C0G caps are more temperature stable than X7R.
Resistor should handle 2W power
Load output


Scaling up input voltage


input caps
remove PV supply diode
bigger inductivity
change mosfets HS & LS V_GS(breakdown) to 150V
voltage sense resistors


Water Robustness Considerations


Add a clear varnish coat to the PCB and components (especially the current sensor circuit, Supply DCDC converters)
Choose low R_filt for the anti-aliasing filter at the current sensor (INA226) inputs to avoid destruction of the chip's amplifiers.
Take extra care with installation of the 3.3V and 12V supply.
If using a (almost) sealed enclosure (aluminium box), add a bag silica to compensate internal humidity. When the converter is cooling down at night it'll suck water into the enclosure. The ESP32 WROOM metallic case is not sealed. Consider placing a membrane? Sticker on its breathing hole.


Board Connectors


Solar +/- (screw terminal)
Battery +/-
USB 4 pin connector (2.54mm) flashing, jtag, mass storage, usb host


USB type-c connector (DNP)

UEXT (SPI and I2C) for displays, additional sensors, can etc