Questions regarding battery type and what voltage moving forward

I am not finished with it. I am still trying to improve and add some filtering and noise suppression components (capacitors), but this is what I have thus far. I took some time to try and keep the ground plane clear. I do have a few vias, but the traces on the ground plane are very short and are not chopping up the ground plane too much. The components I have on there are screw terminals(12v), dc barrel jack(12v), usb ports (5v), power relays controlled by the wemos over wifi (still haven't made the plunge to the esp32 device). A passive POE rj45 ethernet port powered by 12v for my ubiquiti device. I have two seperate 12-5v converters on the board. One for the relay power and the other for usb power. Although, I am now looking at using 7805 dc regulator instead of the large more powerful ones I found here: https://www.amazon.com/dp/B077TC3812...T8Z4A880F&th=1

I have added a 1.5M resistor to the ADC pin of the wemos D1 mini A0. This is in addition to the internal voltage divider on that particular pin. This will allow me to monitor the voltage of the particular source. If the primary source falls below a threshold, I will have the wemos turn the SSR off and turn on the secondary source. I discussed the anti parallel issue on an electronics forum. Without using expensive mosfets, I was advised to use Schottky diodes to prevent the two sources from becoming parallel if a fault occurs and both SSR's are allowing current to flow. I will obviously have voltage drop through the diodes, but it will hopefully be minimal. I still need a short circuit protection device. I am looking at the simple crowbar configuration for that.

Here are a couple of images (top and bottom) of the PCB. I have circled how I handled the surface mount grounding issue. I simply made a small trace on the top of the board to get away from the pad and then a via to the ground plane. I am not sure if this is the best way to handle this, but it is what I did.

Thanks!

Looks fine. Don't just use one Via when you could use 2 or 3, just don't go crazy.

In the PCB I posted way back (it was this thread?) I was handling high frequency digital signals, so I put ground plane both top and bottom then stitched everything everywhere so there were no unstitched islands. Took a few hours, but it should pay dividends. Not in your case, no need.

On that note, take a back up of what you have and experiment with filling the top layer with a ground or "Power" plain and see if you like that.

The only thing I noticed other than that was the closeness to the edge of some components and vias. Some PCB houses will reject those or move them for you if they are too close for their tooling.

On voltage drops, SSRs and Mosfets have you done the thermal calculations and are you fine with bare (PCB plain cooled) or do you need heatsinks (or bolt to a chassis style) to handle any switching? If I understand your use case there shouldn't be much back and forth switching, so the fets should stay cool. The diodes... it's voltage drop * current = heat. You can buy larger SMD diodes for 20-30W disipation.

By the way. The great thing about making your own circuit PCB is you can fore see it not working properly.

So you can add things which will be very useful later when it doesn't work. Consider adding small power injection points for headers and even the odd "unmasked" Via or bare pad as test points.

Keeping tracks spaced well apart makes them easy to scratch off some solder mask and solder in some botch wires.

On "mechanical foot prints", stuff like switches, pots, barrel jacks.... if your first PCB has a blopper and they don't line up.. Don't throw them in the bin... solder wires on and mount those components with sticky glue

If you have the choice, pick the "hand soldering" foot print for resistors and capacitors. They provide a lot of room for error

I know this is more of an electronics question, but what size capacitor would you use for filtering/decoupling the incoming and outgoing 12v supply? I placed a 100nf capacitor near all the I/0 pins of the mcu, usb outlets, adc pin, and the input/output of the 12-5vdc regulator. I’m not sure if that is the best, but I figured they couldn’t hurt. I would assume the capacitors need to be larger for the main 12v power supply incoming and out going at the barrel jacks, but I’m not sure what value they should be? I’m not sure they are that critical in a completely dc system, but I’m not sure. Any advice?

Thanks

It is more of an electronics question and I think if you ask it on an electronics forum you will get some pretty scary mathematics based answer.

The purpose of "bulk" caps is to provide a supply of current locally to the load for sudden demand.

Consider a scenario without any bulk capacitor. Somewhere in your circuit a digital line goes from low to high in under 1 nano second. Even though that is only a signal line which wouldn't be considered to carry current, it still has capacitance and it's the fight between the current available and that capacitance which will govern the rise time.

So that little sub nano second current spike, could be 10 Amps in the case of no bulk cap, will have to be sourced from the 12V source... at the end of the the battery cables, lugs, connectors etc. All of those have capacitance and inductance and they literally take time to supply the current. Usually this results in noise on the power rails, slow rise times and dodgy signal integrity. All of those current spikes create little voltage drop spikes as well.

Putting a bulk capacitor in the circuit provides a VERY, VERY high source of current right next to the circuitry. So any current spike that would cause a voltage drop on the incoming DC lines will be supplied from the capacitor. In practice this basically "low pass filters" the current on the "source" side of the cap. On the digital circuit side the current is bouncing all over the place (and why those 100nF caps are necessary), but on the power supply side it's pulling a much more steady and milder current.

It's calculating the size that involves all the maths. Having a true understanding of what the loads are and how spikey their current demands are. I personally wouldn't know where to start.

What I do, is, for most general small DC loads like "the local PCB" I just put a 100uF, 100nF and 1nF in parallel on the DC jack. Same again post regulator. 100nF and 1nF on IC power pins. Some components, like mosfets and drivers have specific requirements, so do op-amps and DC converters. Usually I work with what they recommend in the dataset.

The trouble with capacitance is "in rush". While we are familiar with this in terms of blue or orange sparks on battery terminals when input capacitors are involved... your little 5V PCB is not immune. If, like I did, you put a whole bunch of 220uF capacitors all over your board, because you have them and want rid of them.... you end up with something stupid like 2mF of capacitance in parallel supplied by an 800mA regulator. That regulator stands absolutely no chance of charging all of those caps up and not taking damage. It died.

Bulk caps are often located close to the "sink" of high but short pulses of current, to present a much smoother current flow to the outside power source. A mosfet driver in a mobile phone can source 20A of gate current, but the phone does not need a 20A power supply. As long as the mosfet is not charged more often than a sensibly sized resistor can recharge the cap at maybe 1 Amp, it only needs a 1 Amp power supply. So if you DO want large bulk caps for spikey loads, you can current limit the charge side of them.

venquessa I finally go around to trying out the setup. I ordered my pcb, but it will still be a few weeks before it comes in. However, I wanted to see if I could charge the 24v and 12v systems at the same time and only pull from the 12v system. I couldn't get it to work. I tried a few different methods. I have 4 100 watt panels. If I wire them all in parallel and connect them to a charge controller this will charge the 12v system. Coming out of the 12v system, I tried a 12 to 24v converter to supply 24v to the second charge controller to charge the 24v system. The problem is the 12 to 24v (or 24 to 12v) converters supply closer to the nominal voltages than the actual voltages the batteries need to fully charge. So, when I supply the 12.5ish voltages to the charge controller this is obviously not enough to charge a 12v lead acid battery. Should I look for a dc 12 to ?? converter that would produce a higher voltage? 18v? 24v?

The next thing I tried which I am probably going to get scolded about, was I connected two pairs of panels in series and then the two series wired panels together in parallel. 2S2P maybe? So now I have 24v panels. I ran that to the charge controllers. One charge controller will auto determine whether the battery is 12v or 24v. Although the panels were producing voltages to charge a 24v battery pack, the charge controller seemed to be charging the 12v battery fine with the higher voltages coming in. Is this a NO NO?

I then tried to connect a second charge controller to the same solar array and charge the 24v pack, but it would not work if I had the other charge controller connected.


I am looking for some advice as to the best way to move forward?

Thanks!