0
\$\begingroup\$

I'm new to PCB design (First one) and I want some advice and tips from more knowledgeable people in regard to basic design elements and especially the power consumption part. I want to make the hardware consume as little as possible during sleep. I spent a lot of time researching, but would like some expert proofing before I pull the trigger.

I'm trying to design a battery powered esp32 that will turn on the motor when the NFC breakout board connected through SPI is reading the correct code. It only wakes up by a built in touch sensor and is in deep sleep most of the time. I choose the ESP32 to add some IoT capabilities. I'm using low quiescent current LDO voltage regulator (4uA) for 3.3V. I have connected ground of the motor, voltage divider for battery measurement , and the NFC module to N-Channel logic level (>1.8V) MOSFET to operate and save power by turning off NFC module and the divider constant drain. This specific MOSFET has a low 1uA leakage IDSS.

Capacitors are the same as the data sheets suggest. Motor current is not more than 1.4A for 5 seconds. Batteries will be 2x2 lithium-ion (~6V-8.4V). The NFC module will be connected by male 2.54 1x6 header on the back side of the board (GND to SCK). Also note that the pcb and breakout board's top,left, and right sides are meant to match. Only the bottom side on my pcb is shorter. There are external UART connectors. Also I'll be soldering all components by hand.

Old Schematic: Schematic

Old PCB: enter image description here

Project link on EasyEda for PCB and Schematic for close up inspection

Feel free to request any info that I could've missed.

Update: I've reworked the voltage divider/motor circuit. It Now it uses a pmos connected to battery in both drain and gate (Motor is a 6 Ohm resistance so the gate is connected to VB) If my theory works, the traces have so little resistance in relation to the motor (about 0.1 Ohm to 6 Ohm of the motor) the voltage at the gate would be still much lower than the drain (1V vs 8-6V if my calculations are correct, threshold is 1.2V difference) so pmos will be turned on. There's no direct connection between any pin to VB when the circuit is off.

I divided the schematic into smaller parts. Added a diode for reverse voltage protection, reworked the SPI lines to make them adjutant since they were just software SPI. Made battery traces slightly thicker. I've decided to not put any terminations resistors. UART have one internally, while the breakout board have resistance for each line. Of course, ideally you would put the resistors at the sender, but given that they seem to work fine with 20cm of ribbon cables, I think I could skip it? Also used smaller caps for 100nf values. I've bought separate BMD for my Li-ion so I don't have to fiddle with that.

Ver0.2 Schematic: enter image description here

PCB: enter image description here

Link to new

If you have any tips, questions, or corrections, feel free to leave them in an answer. I'll mark Umar response as the correct answer for now.

\$\endgroup\$
4
  • \$\begingroup\$ What is the battery capacity? \$\endgroup\$
    – User323693
    Aug 21 '19 at 6:20
  • \$\begingroup\$ Typically you would use NFC as the wakeup source. Your power monitoring divider cutoff is unworkable. \$\endgroup\$ Aug 21 '19 at 8:08
  • \$\begingroup\$ @ChrisStratton with 4.7K, OP can read value of about 3.4 or 3.3V. there is no margin I would say when the batteries are new \$\endgroup\$
    – User323693
    Aug 21 '19 at 11:02
  • \$\begingroup\$ The point is that the cutoff is unworkable, not that a divider cannot work. \$\endgroup\$ Aug 21 '19 at 13:20
2
\$\begingroup\$

I will try to answer

1. Aesthetics:

Please never draw a schematics like this. I wish your second schematics will be nicer one.

The ground should always be in the bottom and the supply should always be on the top. Poor drawing simply sets off readers because it is so difficult to comprehend and needless to say it doesn't look good. You should really follow the tips in the link below. Rules and guidelines for drawing good schematics

The pulldown is very difficult to identify.

enter image description here

The free wheel diode can be connected directly across the 2 pin connector, it will then speaks it purpose to the reader by itself.

enter image description here

2. Battery reverse voltage protection:
One wrong connection would fry the board. There are options to protect the circuit from reverse voltage without the loss of voltage drop or battery usage. Example is a PMOSFETconnected in such a way they the body diode comes in series with the battery.

3. Battery voltage monitoring Pin safety

When the Q2 Transistor is off the battery voltage (~7V) will be directly applied to the IO4 pin of ESP. though the current flow is limited it is still the violation as per of the datasheet. Also, plan a clamp diode so that in such a case the voltage will remain clamped to a safe level across the pin. Other option is to use a PMOS switch.

5. Misc

  1. Pull-up and Pulldown resistor values can be 100K.
  2. How is touch realised?
  3. What is the expected interval of motor re-activation?
  4. The switches for individual modules is a good idea. You are trying to bring down the quiescent current to a lesser value which is good.
  5. Have you calculated the trace width of the PCB trace for power pins especially the motor current path? There are free online calculators
  6. Please provide some de coupling capacitors near the connector for NFC section
  7. LED forward voltage is unknown. When is the LED blinked? Will it be ON always?
  8. The capacitor C4 should be placed closer to the VCC pin (in the PCB)
  9. How are you communicating with NFC module. SPI? I2C? Have you considered pullups or line terminations for those?
  10. Series resistors for UART, NFC communication lines if possible will help limit the IO current in case of accidental errors.
  11. What is the purpose of IO21 GPIO. It seems like you have missed putting a switch for the NFC module
  12. Vbattery trace is dangerously thin. The same track goes to drive the motor.
  13. How long should the product be alive in your application?
  14. The battery voltage divider needs your attention. The resistor combination should be in such a way that the voltage developed across the botttom resistor (which is connected to Q2) should be measurable. Always less than ADC Reference voltage for example. Consider use case if fully charged battery which is about 8 V according to you, 5% regulator tolerance, tolerance of resistors too
\$\endgroup\$
4
  • \$\begingroup\$ On the reverse battery protection: if connectors are used (as opposed to straight pin headers), the OP can pick one which enforces the correct polarity. JST-type connectors are pretty common for such tasks. \$\endgroup\$
    – anrieff
    Aug 21 '19 at 6:45
  • \$\begingroup\$ Agreed. Totally. I would still keep a safety section on as second level protection. During development it is common to supply with a bench power supply for example \$\endgroup\$
    – User323693
    Aug 21 '19 at 6:47
  • \$\begingroup\$ Yes, it's best if both protections are in place, mechanical and electrical. \$\endgroup\$
    – anrieff
    Aug 21 '19 at 6:54
  • \$\begingroup\$ Thanks for your time! I'm currently short on time to redesign it now but I will try to answer some of your questions and cast some more of mine: pastebin.com/PmKPgcZJ \$\endgroup\$ Aug 21 '19 at 21:33

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.