Power Supply Issues with RFID Access Control Backup System

[charcowless]

Hello everyone,

I'm new to electronics and need help troubleshooting an issue with a backup power system for an RFID access control reader in my residential complex.

System Components:
- RFID reader (12V, 6W consumption)
- 12V/8A power supply unit (PSU) with 13.8V output
- 12V/7AH VRLA battery as backup
- MIC 6A8 diode installed on PSU +12V line

Setup:
The PSU, battery, and access control system are connected in parallel. The PSU provides 13.8V to both the battery and the access system.

The Issue:
When there's a power outage, the system works as intended - the battery takes over and powers the RFID reader. However, when power is restored, the PSU fails to take control and charge the battery. Instead:

1. The PSU output voltage fluctuates between 1V-10V
2. The battery continues powering the system and discharging
3. The system only works properly if I:
   - Disconnect the battery (which turns off the RFID reader)
   - Wait for the PSU to stabilize (returns to 13.8V)
   - Reconnect the battery

After this manual intervention, everything works normally - the PSU maintains 13.8V, powers the system, and charges the battery.

I've already tried installing a diode to prevent reverse current, but the issue persists.

Any suggestions on how to fix this so the system recovers automatically after power is restored?

Thanks in advance for your help!

[nctnico]

Replace the PSU. It either is broken or the system never has worked as intended. The first person to call is the one who installed it.
If you want to try the diode fix, install a load resistor across the PSU. The problem is likely that the PSU doesn't start with voltage across the outputs; a load resistor fixes that but you'd have to test what the maximum resistance value is and then use half that value (=twice the current) to have some chance of the system working reliably.

[tom66]

Yeah, I concur.  Most likely the startup capacitor ('Vcc capacitor') on the PSU primary side is bad.  So, when it tries to start up after a power outage it can't get the output up quickly enough before that capacitor fails.

It is a common issue with cheaper power supplies because they are typically small electrolytic capacitors that dry out due to the high temperature of nearby components.

When there is a load on the output such as a flat battery, the startup capacitor doesn't get as much current from the tertiary transformer winding, so it runs out quicker. With no load on the power supply the output hits regulation quickly so the cap can be charged up before it fully discharges causing the controller chip to shut off.

[floobydust]

It helps to know the make/model of power supply. Is there a fuse on the DC side?
When power comes back on, the battery will draw high current charging but the PSU should not go into hiccup mode I think.
The PSU is too big at 8A for a 7Ah battery, that can be dangerous. There are also PSU's designed for 12V battery like Meanwell LAD-120A

What I do is use two isolating diodes plus a PTC fuse to limit charging current, and a fuse. I'll add a schematic later.

edit: added schematic of what I am presently using.

I find with old SLA batteries, one cell shorts so you have to limit charge current or the battery melts. This is a danger.
I was going to use a 4Ω 5W resistor but that is big and can be a lot of heat. Goal was C/10 max.
The RXEF030 polyfuse resistance limits charging current, at the expense of recovery time charging.
Of course it usually trips during higher charge currents but as the battery charges it holds around 0.3A, it eventually cools down and resets.
You could use a dual Schottky diode in a TO-220 package. Silicon Vf is over twice the heat.

[charcowless]

Thank you all for helping! In the end, I installed a small charge control module to limit the current drawn by the VRLA battery when the power grid reconnects. This prevents the power supply from being overwhelmed by the sudden current demand as soon as it begins operating.

[nctnico]

I was going to use a 4Ω 5W resistor but that is big and can be a lot of heat. Goal was C/10 max.
The RXEF030 polyfuse resistance limits charging current, at the expense of recovery time charging.
Of course it usually trips during higher charge currents but as the battery charges it holds around 0.3A, it eventually cools down and resets.
You could use a dual Schottky diode in a TO-220 package. Silicon Vf is over twice the heat.

You shouldn't use PTC fuses as current limiters. They aren't made for that purpose. The datasheet says:  These devices are intended for protection against damage caused by occasional overcurrent or overtemperature fault conditions and should not be used when repeated fault conditions or prolonged trip events are anticipated.

IOW: a better solution is to put a current limiter in place using a transistor or MOSFET.

[tom66]

I was going to use a 4Ω 5W resistor but that is big and can be a lot of heat. Goal was C/10 max.
The RXEF030 polyfuse resistance limits charging current, at the expense of recovery time charging.
Of course it usually trips during higher charge currents but as the battery charges it holds around 0.3A, it eventually cools down and resets.
You could use a dual Schottky diode in a TO-220 package. Silicon Vf is over twice the heat.

You shouldn't use PTC fuses as current limiters. They aren't made for that purpose. The datasheet says:  These devices are intended for protection against damage caused by occasional overcurrent or overtemperature fault conditions and should not be used when repeated fault conditions or prolonged trip events are anticipated.

IOW: a better solution is to put a current limiter in place using a transistor or MOSFET.

Indeed.  I believe it was older Dell motherboards that used PTC current limiters for the USB ports; this was fine for momentary shorts, but it wasn't uncommon for some USB ports (especially on school PCs) to get vandalised or damaged, shorting VBUS to GND permanently.  Such a short tends to lead to the PTC device burning a hole in the PCB eventually, likely destroying the motherboard in the process.

eFuse devices are pretty cheap these days.  I use MP5016H in a few designs, though it's a small SMD device so maybe out of the reach of hobbyists.  It's sub $1 in volume.