Alright, v2.0 is in the making. I finished 95% of the schematic and rough component placement. Now I have to route the traces.
Features are: NFC antenna connector, connector for solar charger, dual 18650 madness, SD card holder, possibility to connect the usual OLED screen as well as two different transflective sunlight-readable ones, USB type C connector, GPS connector (uart pins plus high side switched VCC), battery temperature, humidity and altitude monitoring and piezo buzzer. Furthermore, the user can choose from a variety of different ebyte modules (sx126x) as both a footprint of the small as well as of the big version are on board. All in all the board is 50x110mm in size but itās worth it. Battery life is gonna be insane.
5 Likes
There are some batteries produced by Sony a bit bigger than 18650 with specs like 5000mAh 10C discharge and 3.6V nominal power. Iād give the choice of the battery type to the makes. While for the GPS, some users for privacy reasons may choose not to have a GPS. Iād just leave a socket.
Itās up to the people anyway. I mean, just donāt populate the 18650 holder and use the giant pads to connect any battery you want. And of course GPS is not really onboard but just a socket.
Sometimes you have to make decisions. Remember, life is a giant ocean of possibilities until reality manifests itself through decisions 
1 Like
Though for users nervous about a GPS: A better solution than a socket IMO is the software is open source and they (or someone they trust) can validate that if the GPS is disabled in software it is not used and exposes no ārisk.ā
Also the global clock provided by gps gives great options in the future for protocol improvements (scheduled frequency hopping or turning off all receivers except during brief windows to see if someone wants to talk)
2 Likes
The socket gives you flexibility however. If the module is only tied to the board via a digital interface cable you can connect an integrated tiny module like the BN220 (popular in drone scene) and place it further away from other antennas and metal to improve reception. Better than having the GPS antenna connected via a lossy thin coax to the receiverā¦
or simply get rid of the GPS if you donāt need, it is an extra part with a relative high price compared to the other components
Great to know that. My batch will ready by the end of this week, I am looking forward to build the minimal SX126x+nfr52 from chips, I hope I could deliver that in lower cost and works well with current firmware.
2 Likes
Wow thatās quick! Do you assemble them by hand?
Yep, but I will need by BGA soldering station for the module.
I used a mini PTC heater hot plate for 4 bucks from Aliexpress
Thatās a great hack. Iāll look into that.
Dual battery (via some sort of ideal diode OR, right?) will be great way to increase capacity. LFP chemistry can be an alternative for harsher environments (higher temps, etc)ā¦ One other advantage of LFP is the operating voltage, one can āskipā regulation and use the batteries from 2.8-3.5V-ish volts, I know it works with the ESP32.
1 Like
The concept will be like in most powerbanks: 2 or 1 unprotected cells in parallel and the protection circuitry will be on board. LFP is not bad, but I like having a 3.3V rail that you can rely on. Regular LiIon 18650 are ubiquitous, safe and good capacity per buck
1 Like
Nice work. What solar solution did you settle on? Iāve been going around in circles researching the many different options!
I started developing a cost-efficient true-ish MPPT charge controller based on an ATtiny85 and INA219 (on the input side to maximize input power from the panel). I configured it to output 240kHz PWM to drive a P-ch-mosfet-based buck topology. The tiny itself is just connected to the lithium cell measuring the battery voltage and terminating the charge if necessary. Unfortunately in the insane 64MHz PLL mode necessary to provide the high freq PLL the MCH draws 10mA, so this solution is only good for bigger solar-panel-battery-combos. The tiny85 can sleep at 5ĀµA so it wonāt drain the battery too much as soon as it is full, but while itās PWMing it draws those 10mA. Iām going to complete it anyway and make it open source as soon as there is time left.
Recently I tested this:
and found it work reasonably well, while being dirt-cheap and compact. It is one of those pseudo-MPPT chargers which adjusts the charge current to keep the input side around the pre-set maximum power point, which is actually about 90% of the nominal voltage of the panel. I ordered both the 12 and 6V version and found the 6V version to be more efficient with a 12 element (ā6Vā) solar cell. Maybe it gets even better if I tweak the MPPT voltage divider. This will be my charger of choice in future solar node experiments. It will connect externally as an option to the new 2.0 boardā¦
2 Likes
Have you experienced any issues using a switching charger around RF circuitry? It makes me wary but I havenāt much experience in that field.
I also wondered the importance of a power pass feature of a charger to allow the load to continue operating without upsetting the charge termination of the battery.
Iām planning to connect it as an external module and itās not carrying particularly high currents so EMI should be negligible. No nubers here though. But Iām running the LORA module from a boost converter, which worked without any problems.
The small solar panels in solar nodes wonāt deliver very high currents (let it be 750mA or so if sun is optimal, but it wonāt be the case most of the day). Continous drain from the battery is 20mA which wonāt change the voltage of one or two 18650 much as itās a tiny load in comparison to their current handling capability. The energy spikes for lora transmissions should come from the generously distributed bypass capacitors for the most part and therefore should not confuse the charger either.
1 Like
Please note that these charges are set for 4.2V(?). I would set the charge point to something lower like 4.1V or even 4V so that the battery doesnāt get stressed out daily to 4.2V. Greatly increases the battery life if charged to 90% (instead of 100%). When pulling high current (like 1.3A) you also want to be at 90%, not 100%. 6V solar solution ā please check charging efficiency in shaded situation (under a tree), it is important to have some usable output with the 12 element panel, you need more than 3.7V with this buck converter.
2 Likes
switching chargers should be no problem if decoupled correctly, watch the PWM switching fronts for any UHF noise, I doubt it, but yes, MMPT chargers can be 'all over the place". I donāt like charging to ācharge terminationā voltages in single Li-ion battery setups. There is no need to torture the battery with 20mA in and out currents around 4.2V.
2 Likes