12VDC UPS Project

The comparator and switch are not really doing anything useful.

Two diodes will do the business quite easily.

Both cathodes to the output, one anode to the 12V DC in and one to the battery.

It is a good idea to disconnect the battery from the load if its voltage drops too low but your circuit doesn't do that.

MK

michaelkellett said:

It is a good idea to disconnect the battery from the load if its voltage drops too low but your circuit doesn't do that.

The dropout voltage of the battery is managed by the BMS. It's not in the circuit.

Hi look here, an idea:

This is a very simple "UPV".

The switching time is not the big problem if you make a buffer (Elko-accu) during toggling the relay.

It's just an idea.

I did near-identical circuit with my home router a few years ago. I just used 8xAA cells (non-rechargeable) since it is very rare here to have outages. I would possibly use Ni-MH if I were to do it again today.

The goal is to do it with a solid state switch. No relay. I'll try to modifier this circuit to a solid state.

The goal is to do it with a solid state switch. No relay. I'll try to modifier this circuit to a solid state.

Solid-state relays are okay—a perfectly viable option. Solid-state changeover switches are a bit komplexer.

But please don't forget about the leakage current (return current in µA) inherent to semiconductor switches; the battery will feel that.
True isolation (via a mechanical contact) would be better; that way, you avoid equalization currents, which are also highly temperature-dependent. Leakage currents drain the battery without providing any benefit.

Don't be afraid to use a relay.

The "only" potential issue is:
the contact sticking shut even after the relay has de-energized; the contact joint can melting, current switch off lightning.
With a DC relay, the iron core can become permanently magnetized if the relay remains energized by direct current for extended periods.
One possible solution for monitoring this is to use a second contact with an LED to indicate whether the circuit is closed or open.

A more complex approach involves applying a short negative pulse to the relay coil upon switch-off to collapse the magnetic field.
This is achieved using a second winding around the iron core (utilizing back-EMF).

->> One more point regarding your circuit:
Make sure to include deep-discharge protection and a charge cutoff for the battery; otherwise, the battery will eventually be ruined.

OK, now I understand why all the USV I came along have a relay. Maybe I shouldn't reinvent it. 

I understand you perfectly. I’m a tinkerer too—someone who always wants to improve everything, and I mean *everything*.
My hammer is always within reach, "pangs!", and I already have new ideas.Bit by bit, a drop of knowledge...