Thoughts on my design?

[Tangent_Tracker]

Ok folks, been pondering this for a bit....

So I have boxes that are say 20m apart. Connected with cat5 cable, data(RS485), ground and "power" which leads to this...

each box has 3 18650s in series to give 11v1 nominal. Drain will be fairly light at no more than about 100mA per box when active.

There is a master box, a slave, and some supplimentry boxes that do other functions.

Everything to be controlled by master box... And I want all batteries in the boxes to "normalise" or balance to the same charge ultimately. Every so often I will bring a car battery with a Li-ion charging cirtuit to trickle charge things over a period of a week.

So, here is my circuit. each unit controlled by a PIC, 1829 or similar with power from a small MP2338 switcher, nice and efficient...

There is an issue I have tried to overcome with the "ideal diodes", that being that you cannot exceed Vs of the multiplexer switches, so the voltage from the highest module is used as Vs.

PIC samples voltage from it's own battery and that of the remote's battery to determine if firstly the voltage difference is reasonable, and secondly, if normallising falls within reasonable limits, whether a current limit resistor should be switched or if the supplies can be directly connected. Bear in mind, the current can flow either way...

So, program aside, are there any serious issues with my idea? I am still concerned about the multiplexer, of course even an "ideal diode" will exibit a voltage drop, and the data sheet does suggest max switchable voltage is Vdd.. I just cannot find another solution that can allow me to get round this and remain compact/simple/low overheads (quiescent current etc..)

Not too worried about the connection of the batteries directly... As long as I am careful and don't allow too much current, hence the low fuse connecting being the link between both sets of batteries, lower than the maximum rating of the multiplex.. One thing is for sure, they won't be overcharged amongst themselves!

Thoughts appreciated!

**Ignore the divider values, not decided if PIC will run from 3v3 or 5v yet!

[Tangent_Tracker]

Just in case it's been missed folks..?

[Gyro]

First thought, do you actually need the slave box to be battery powered? If the master box goes down, is the slave box capable of doing anything? At 100mA you will get a bit of some voltage drop across 20m of cat5, but you could probably compensate this.

[NiHaoMike]

Keep it simple and just use PTCs to limit the current.

[Tangent_Tracker]

Both master and slave box's are the same, just the mode they run in. It's for a signal system so I want everything to be self powered....

[Tangent_Tracker]

Keep it simple and just use PTCs to limit the current.

Hmm I wanted to disconnect power for various reasons, mostly power conservation, as there are units wihout batteries...... You have got me thinking now though... I thought about using a small relay, giving enough current to activate the armature then providing just enough current to keep engaged. Alternative is to use two power switches in parallel... But the latter use mosfets with diode so that would not work..

[Doctorandus_P]

On my Home network, I use 4 wires of cat-5 cabling. One wire pair for RS485 data, and another wire pair for 24Vdc and GND. Each node has a local regulator, so a few volts drop over the cat-5 cable is not a problem. It's one of the nice things of RS485. It is officially capable of handling negative inputs down to -7V on it's input. So it still works if the local GND of a remote node is lifted a few volts because of voltage drop over the cable.

[mariush]

Do you actually USE that 11v produced by the 3 cells in series for something?

 A cheap suggestion : use a charger IC like MP2672 - 4v to 5.5v input, can charge 1 or 2 cells in series (charge voltage configurable between 8.4v and  8.8v)

MP2672GD : https://www.digikey.com/en/products/detail/monolithic-power-systems-inc/MP2672GD-0000-Z/13159530

MP2672AGD (newer recommended revision, pin compatible with previous) : https://www.digikey.com/en/products/detail/monolithic-power-systems-inc/MP2672AGD-0000-Z/13572801
 
The IC also has a system output pin, during charging it outputs the voltage required to charge the 2 cells (for example minimum something around 6v max 8.4v for 2 4.2v cells in series, when power is gone it connects sys pin to battery ... so you can use a buck converter on the sys pin to produce 5v or 3.3v or whatever.

Put 4 18650    two 18650 in parallel, in series with another two in parallel, add your buck regulator job done.

You can send 5v through the ethernet cable ... you won't have 1v drop on 20m of cables ... for example let's say AWG24 wires so 84mOhm per meter, 20 meters so 40 meters in total .... let's say you want to charge batteries at 100mA, so let's say 200mA at 5v  .. V = I x R = 0.2 x 40m x 0.085 =  0.68v drop, you'll still have 4.4v at the end of the cable.

[Tangent_Tracker]

Do you actually USE that 11v produced by the 3 cells in series for something?

 A cheap suggestion : use a charger IC like MP2672 - 4v to 5.5v input, can charge 1 or 2 cells in series (charge voltage configurable between 8.4v and  8.8v)

MP2672GD : https://www.digikey.com/en/products/detail/monolithic-power-systems-inc/MP2672GD-0000-Z/13159530

MP2672AGD (newer recommended revision, pin compatible with previous) : https://www.digikey.com/en/products/detail/monolithic-power-systems-inc/MP2672AGD-0000-Z/13572801
 
The IC also has a system output pin, during charging it outputs the voltage required to charge the 2 cells (for example minimum something around 6v max 8.4v for 2 4.2v cells in series, when power is gone it connects sys pin to battery ... so you can use a buck converter on the sys pin to produce 5v or 3.3v or whatever.

Put 4 18650    two 18650 in parallel, in series with another two in parallel, add your buck regulator job done.

You can send 5v through the ethernet cable ... you won't have 1v drop on 20m of cables ... for example let's say AWG24 wires so 84mOhm per meter, 20 meters so 40 meters in total .... let's say you want to charge batteries at 100mA, so let's say 200mA at 5v  .. V = I x R = 0.2 x 40m x 0.085 =  0.68v drop, you'll still have 4.4v at the end of the cable.

Thanks Mariush, I do sadly, as there are LED's that require about 12v, some Panel switches and Signal LED's. They are happy down to about 9-10v, and I did consider using boost circuits but it seems things are getting a bit more complicated in that case... I will look at your suggesting and think on...