Linuxslate.com Forums

Attempt #4 to Create a Hardware Solution for the Heltec V4 Solar Power Low Voltage Issue (Best So Far)

Note: Posts are out of order. Read the 2nd post in this thread, and then return here.

In the previous post we outlined a rather complex, but effective way to allow the HelTec V4 LoRa board to be opperated from battery/solar power without Boot Looping, or Resetting to Factory Defaults during a low light, or depleted battery situation.

This post will detail the modifications to the second of the two TP4056/DW01A boards used to realize this solution.



This Details the Modifications to the Second TP4056/DW01A board have been made: UPDATED:

1. Carefully heat and lift pin 5 of the TP4056 chip. This will disable and isolate the TP4056 charging circuit. Only the DW01A (with associated circuitry) and MOSFET are used on this second TP4056/DW01A board.

1a. (Optional) The USB-C connector will not be used on this second TP4056/DW01A board. It should be taped over or removed.

2. UPDATED Carefully cut the trace between pin 6 of the DWO1A and B-. This trace should be cut just below C2, so as to retain C2 in the circuit. When looking at the physical board with the markings right side up, cut this trace is to the left of the FS8205A dual MOSFET and just below C2. Leave a few mm of the trace on the lower side (markings right side up) of C2, and scrape some of the solder mask off so that a wire can be soldered here.

2a. The cut trace done above will be reconnected via anm SMT diode (see below for diode selection). It is probably best to solder the Cathode of the diode vertically to the edge of the B (-) terminal, and then solder a small wire from the trace we exposed on the lower side of C2 (which is actually the DW01A pin 6 side) to the Anode of the diode. It may be possible to solder a through hole (wire leads) diode point to point. Do not allow the Anode of the diode to touch the B (+) Terminal or anything else. (See description and warning below).

3. Short (Solder Bridge) Pins 7 and 8 of the MOSFET Chip. (This will actually mean that Pins 6,7, and 8 are all connected, and can be bridged.) This bypasses the current protection mode of this board. While in actuality, this probably does very little, I did it for 2 reasons: This half of the MOSFET chip provided battery charging over current protection. We do not need or want over current protection for the 2nd board. Again, we are only using the voltage comparator portion of this second board. Additionally, while I do know how efficient MOSFETs are, we don't want any avoidable loss in a solar powered system.


OPTIONAL Further Modification:

--- Carefully remove the TP4056 Chip. (Note that we had previously lifted pin 5 of the TP4056 to disable and isolate it). It is very difficult to remove the TP4056 due to it being attached to the board to provide heat sinking. If you do not have SMT rework equipment, it is suggested that you destroy the TP4056 in place, and then carefully remove the pins so as not to damage pads/traces.

UPDATED Solder a small wire to the pad that used to be pin 6 of the TP4056. (This is one end of resistor R1). Solder the other end of the wire to OUT (-), or to Pins 6,7, and 8 of the MOSFET that we previously bridged. Jumper the original (+) Input of the board to the OUT (+). This restores (+) to the Anode of the Blue LED.

Doing this enables the Blue LED to tell us when the output of this 2nd board is ON. The convoluted way we do this allows the Blue LED to work the way we want it to without adding additional components and with soldering that can be done with a minimum of SMT skill.

Yes, this will cause a very slight additional drain on the battery, but allows quick assessment of power status. And besides that, every project is better with more blue LEDs! Do this optional mod at your own discretion.



At this point, we have pretty much butchered this second TP4056 Board

All that is left working is the Under-voltage comparator inside the DW01A, and the respective half of the dual MOSFET chip. This is exactly what we want.

Furthermore, we have biased the comparator inside the DW01A (actually the entire DW01A) up approximately 0.7 volts from normal, which ensures we have sufficient voltage to run the Heltec V4 before applying power to it.

WARNING: Do Not Connect a Battery directly to this modified TP4056/DW01A board.

This is why we need 2 separate TP4056/DW01A boards -- The one that handles the battery must be left unmodified.


Diode Selection:

Pretty much any small diode can be used. I used a SMD diode that is likely comparable to a 1N4007. This is probably a little overkill. A 1N4001 equivalent is probably sufficient. Even a signal diode such as a 1N4148 is likely sufficient. Different diodes can be tried to arrive at a turn-on voltage that always provides sufficient voltage to the Heltec V4, while minimizing premature shut down in low battery or low sunlight conditions.


Why are we doing it this way?

The goal here is to allow reliable operation of the Heltec V4 in a low budget/low barrier to entry manner. TP4056/DW01A boards are available for about $2.00 (lower in quantity,) and the diode is a few cents. Two of the same TP4056/DW01A modules are used to minimize the types of circuits need, and to allow taking advantage of quantity discounts. A minimum of SMT soldering skill is required, and the modifications can be done in a short time.

Butchering a $2 board is a far cheaper and easier solution than a custom PC board, and a BOM with dozens of required parts.

The Boost Board is similarly readily available, and costs a few dollars.


Is the Boost Converter board discussed in the previous posts still needed?

UPDATED Early testing indicates the Boost Converter Board is still needed.


Can the USB connector on the *first* TP4056/DW01A board still be used for a +5VDC output solar panel, or other +5VDC USB power source?

Yes. This first TP4056/DW01A board is unmodified, and the USB connector works as normal. If the unit is operated from a grid-derived source of +5VDC USB power, and the battery is used only for backup, the benefits of the TP4056/DW01A board to provide LVC, and protection against boot looping when power is re-applied is still realized.

Temporary Solution to the HelTec V4 Solar Charging Issue

This thread has been split off from the Meshtastic?! More like Meshtragic thread. The Previous posts got way out of hand, and are significantly trimmed here.

IMPORTANT NOTE: What is described here seems to be only a partial fix. Please see subsequent post for another possible hardware solution.

Ref Meshtastic firmware - Heltec power management. Sleeping Beauty mode. (Issue #9911)

Problem Summary:

As covered in the referenced Meshtastic Github Issue and elsewhere, The Heltec V4, with any current Meshtastic firmware, cannot be used as a solar node due to the fact that it is intolerant of a low voltage situation (as well as being a bit of a power hog.)

The symptoms include constantly restarting (boot looping) during or after a low-voltage condition, and a factory reset once stable voltage is reestablished. Recovery from this condition is impossible without physical access, which may be difficult if the node is installed in an outdoor elevated location such as a tree, roof top, or tower.

A second symptom is an over-discharged battery, as the HelTec V4 contains no under voltage (LVC - Low Voltage Cut-off) circuitry.

While the Meshtastic developers work on a firmware fix (or at least improvement), it is desired to come up with a simple hardware fix. This fix should utilize inexpensive, and available modules that require only simple wiring. The solution should prevent the Factory Reset issue, and prevent serious under discharge of the battery.

Solution:

Two (2) TP4056/DW01A boards, and an adjustable boost module are being used. One TP4056/DW01A board is left in it's standard configuration and handles battery charging, including Overcharge protection, and battery Low voltage Cutoff (LVC), which the Heltec V4 on-board charger lacks.

The output terminals (OUT (+) and OUT (-)) of the TP4056/DW01A charger board are connected to the B (+) and (B-) terminals of a second TP4056/DW01A board. This second TP4056/DW01A board is heavily modified, and acts as a simple voltage comparator and switch.

The OUT (+) and OUT (-) terminals of the second TP4056/DW01A board are connected to a Boost Converter. The Boost converter is used to insure that the HelTec V4 always receives adequate voltage. A 100µF capacitor in parallel with the output of the second TP4056/DW01A board (Input of the Boost Converter) helps reduce transient load conditions when the Boost converter, and or the HelTec V4 starts up. In this configuration, we minimize the conditions under which the HelTec V4 experiences a low voltage condition during battery recovery, or low sunlight conditions.


Side Effects:

This solution is imperfect, and should be considered temporary. It suffers from the following:

-- Battery voltage indication may lost. See below for possible workarounds.
-- While battery LVC is provided, the Cut-off voltage of the TP4056 module is not adjustable, and is lower than ideal to preserve battery health.
-- We are trading battery charge life for battery longevity, and reliable restart of the HelTec V4. For a given battery and Solar Panel capacity, run time will be Reduced.
-- Since any added circuitry must cause additional losses, overall efficiency is lost. More solar charging, and/or battery capacity may be necessary
-- Since the comparator in the second DW01A and the MOSFET do not act as a perfect switch, and the Boost Converter still passes some voltage when OFF, there are still conditions where the HelTec V4 may expirience low or fluctuating

Please see the end of the Meshtastic?! More like Meshtragic thread for the development of a solution. That information will be trimmed and moved here shortly.

Temporary Solution to the HelTec V4 Solar Charging Issue -- Status Update


I have completed modifications to a replacement second TP4056/DW01A board, and replaced the one on the prototype board shown in the previous (now deleted) posts. I have also switched to using a normal 18650 battery in a proper battery holder.

Since making these changes, we have had mostly overcast skies here. My $1 test solar panel is not in an optimum location, and does not get full sunlight until almost noon. As a result, I have not gotten more than approximately 1 hour of uptime from my test node.

I have procured a box of loose solar cells, and I intend to wire up my own, much larger solar panel. It will be rated at about 10 watts at 12 volts. A buck converter will be included to keep the output of the panels at ~ 5 to 6VDC.

The above described Solar Panel has been constructed. I also discovered that one of the 18650 cells I was using was bad. I'm looking forward to tomorrow's sun.

My original node will be retrieved from the tree, and I will use that waterproof case to make a test node that will be in a better spot for both signal and sunlight, but will still be accessible for now. This test node will also have 2 high capacity 18650 cells.

Check back for more results.