Problem with loadsharing and TP0456?
Opened this issue · 22 comments
Hi,
i stumbled upon this post: https://www.instructables.com/Load-Sharing-Use-Solar-Panel-Safely-With-TP4056/ do we have a prolem here?
Regards,
notecp
This is good stuff (in a bad way). I need to think this through and understand What I would want to do with the design. We don't want to shut off the load while the battery is charging, per se. Maybe we need to have a switch for the panel itself possibly, monitor charge cycles in software, etc...
Fortunately, Our load is not a big current consumer (duty cycle for WAKE vs. sleep), but I'll need to investigate.
okay.
they even provided a fix for this on the site. Maybe it is worth a try?
I don't like the idea of the load (weather station) being turned off every sunny day when the battery is charging. Any daylight hours when battery is needing charge, the station will be down. I'm OK with a FET switch, just not placed there (for our application needs.
I still want to do some research on the full module we are using to understand more before coming up with a proposal.
i think it will siwtch automatically around the tp chip when the sun is shining.
the load will stay powered all the time i think
I stand corrected. I missed the diode feeding the load directly in the schematic. Yes, we should investigate this.
I'm going to look into this some more. It intrigues me...1. Does the TP4056 work normally with a USB power supply plugged in (not a solar panel)? I'm thinking it's more a function of the wimpy solar panel, we end up current starving and VBAT+ never reaches 4.2V. I have a 20V/25W solar panel and a high current buck converter between the panel and the TP4056. 2. Don't we determine the battery is charged with monitoring the voltage at VBAT (again, healthy supply handles the battery and the load without current staring)
I keep trying to rationalize that the circuit will work well when a USB supply is connected. Your thoughts?
i tested it with a powersupply (capable of up to 2.5 A at 6V) attached to the input.
The board finishes the loading and turns blue (while the ESP32 is off via the switch). As soon as i turn on the ESP it immediately starts charging the lithium pack. The power drawn went up from approx 0 mA to about 100 mA (sadly no digital gauge).
my conlusion is, that my linked articel semms to be true and we have to investigate further.
I'm not saying that the observations are bad. I actually like the circuit modification. I don't like the idea of immediately trying to top off the battery every time the system turns on and I like the idea of letting the battery actually get the opportunity to discharge.
Are we actually over-charging though? Do we have a fire hazard on our hands? My analogy: An over zealous waiter, waiting to refill your water glass as you pick it up to drink from it. He does not over fill it, but he is there to fill the glass as you drink.
If the battery voltage is > 4.2V and the system stops charging at 4.2V then are we truly over-charging? Are there any documented cases of this happening where batteries are getting damaged or systems dying, or fires? I don't think the problem is just related to solar charging.
Thinking ahead: I'd also like to include the ability to kill the solar charging unit entirely at low temperatures (inside the case). I'll probably use the same type of FET switch for that also, but drive that FET from the ESP so I can make decisions in software. We could also initiate "full discharge cycle" on the battery periodically, if we desire. This assumes you are using the newer design that does not aggressively discharge the battery nightly.
Ian, transmitter battery voltage does not ever change for you. No sign of discharge, even at night. Am I missing something?
I've updated the v4 PCB design goals to comprehend these inputs.
Ian, great article! I can now get my head around the "why". I did like the conclusion at the end of part 2, LOL. Nobody's ever had a problem with it, but not the right thing to do. I'm convinced that it is the right path to take.
My take on the solution to the problem. I'm envisioning a "pass through" PCB that sits below the 4056 module. This could be a useful solution for many projects. We can keep it off the weather station PCB altogether. It would be available for any project though. The only "hack" would be connecting the TEMP pin to the lower PCB (next paragraph).
I also found it interesting that the 4056 has a pin for TEMP control, but it states that the NTC resistor is part of the battery. Users will not want to purchase a more expensive battery, so we would take care of the low temp case and high temp case by tacking a thermistor onto the battery? No, we are not measuring the temperature in the battery itself, but can still control high/low temperature settings on the charger circuit itself.
On to your specific application: How do you manage the charging of your 3S LiPo pack? You are not loading a battery into the weather station PCB at all, right?
Very raw and unchecked at the moment, but I'm thinking of something like this for the load switching...
an interposer PCB with basic socket compatibility to the existing footprint.
Ok, doing final checkout of PCB before fabrication, but here's the 3D view.
Gerbers have been created and PCB is in fab now.
FYI, this retro-fit PCB may be a tight fit to the weather station PCB. You may consider mounting it from the underside of the PCB. I'll be making the gerbers available as soon as I receive and verify the design.
Hi James,
Interesting thoughts, earlier today I stumbled across a video handling the same issue. The schematics are quite similar, and well explained. https://www.youtube.com/watch?v=37kGva3NW8w&t=32s
Thanks for sharing the post! I respect Andreas work and publishing. I think this confirms we are on a good course. PCBs have cleared US customs and headed to Denver, CO today.
- Finally got time to build up the insanely simple load switch PCB. I'll try to test it this weekend.