If you are just interested in a battery backup (which is not a UPS) there are much simpler ways of doing this. Am I understanding this correctly ? Most modern digital alarm clocks have battery backups.

 

Wow, I didn't know there was a CMOS version of 555. That's good news !
Yes, please... You can post that circuit. I never worked with 555 before, and I can't find a Spice model to study it's functionality in simulation... It is equivalent to the old NE555 ? 'Cause this is available in the simulator...

 

I never worked with 555 before, and I can't find a Spice model to study it's functionality in simulation... It is equivalent to the old NE555 ? 'Cause this is available in the simulator...

I don't have a model for the CMOS 555 that works, so I use the NE555 idealized model in LTspice, which is close enough for most simulation purposes (its output goes to the V+ rail for example).

Below is the sim of an example one-shot circuit with the 555:
The 555 triggers when the TRIG input level goes below 1/3 of the V+ supply voltage.
The input is capacitor coupled, since, if the TRIG voltage stays low after the one-shot times out, it will immediately retrigger (output stays high).
The one-shot time is approximately 1.1 times the R2C1 time-constant.

I attached the .asc LTspice file for the circuit.
Here's an explanation of the 555 operation I wrote, if interested.

 

Since the purpose of this project is to have a 12-Volt DC UPS,
various topologies that may simplify it's construction and operation should be considered.

The following block-diagram is how I would suggest building it ..........
.
.
.

.

 

Which kind of illustrates my point of backup batteries.

 

I designed, built, and delivered a 24 volt UPS for a friend. The external functionality was similar to your system, in that with the mains active power was available to the load as required, and a float charge was supplied to the battery. When the mains failed the power was provided from the internal battery. The changeover was accomplished by means of a diode that became forward biased when the DC from the mains powered supply dropped a small amount. I do not have a scope fast enough to monitor whatever disruption was possibly due to the diode turn-on delay. The entire circuit consisted of a resistor and 4 diodes.
My point being that the similar functionality could be delivered with a great deal less complexity.
A 12 volt version would be no more complex, as I see it.

 

I checked again the initial post and, indeed, I didn't ask anything about circuit complexity nor how I can simplify it. I just ask for help to make it work as it is. So, I don't understand why so many of you spam me with this kind of messages about complexity.

Below is the sim of an example one-shot circuit with the 555:

Thanks ! Now I know how a 555 works.
Let me study it so I can understend how can I integrate this in my circuit...

 

I mention complexity because it usually is not cost effective in producing an improvement in results, and because it always increases the probability of hardware failures, and it always increases cost.
This insight comes from 40 years of engineering experience designing equipment that must never fail.
Aside from that, every component added also increases the number of items that must be, and stay, within specifications to provide satisfactory performance. That reality is well understood and documented to exhaustion by folks dedicated to assuring reliability.

 

There is no component under stres in my circuit, it works at the uA level. So it doesn't have any reason to fail.
And the cost is no problem to me. I do not product this in series to sell them, I just built one for my needs, and I use only quality components. Why some of you want to help me by force (in a way that I don't want) ?

 

The offer of a simpler way to get similar results comes, in my case, from "being an engineer 24/7/365, and not just somebody who does engineering as a job." That often includes seeing other ways to accomplish an end.

 

Part of designing any Circuit,
even before the first device is chosen or purchased,
should always involve the following things .........
1)
Trying to figure-out what could possibly make it fail.
2)
Trying to figure-out how to simplify the design without compromising reliable functionality.
3)
Always assume that You may have missed something.
4)
Start-over at number-one.
.
.
.

 

Part of designing any Circuit,
even before the first device is chosen or purchased,
should always involve the following things .........
1)
Trying to figure-out what could possibly make it fail.
2)
Trying to figure-out how to simplify the design without compromising reliable functionality.
3)
Always assume that You may have missed something.
4)
Start-over at number-one.Adth
.
.
And the time for making mistakes that demand corrections or a start over is while the design is still on paper, not as it is built. Errors at the sketch stage are much cheaper to fix. When I was at Sun Electric, ISD, the mechanical design group had a mantra oof "Never time to do it right, ALWAYS TIME to do it over." That was rather disturbing.