Electrical – Current limit of switch mode power supply

switch-mode-power-supply

I have a premade SMPS, it is 12V and maxes out at 8.5A. I have a TEC1-12706 on one pair of terminals and there will be a fan on the other terminals (I only have a total of 4 terminals for out). I would like to limit the current for the fan. How would this be done?

The PWM controller inside of the SMPS is the OB2269AP along with a coupler ic.

Best Answer

As outlined in some of the comments, there are several ways to limit the current. One of them is to force a constant-current operation (CC) directly at the secondary side. Your SMPS is naturally designed to operate in constant-voltage (CV) mode and goes into short circuit protection if the feedback information is lost in the primary side (the opto collector is open loop). A simple way to limit the output current and force a CC operation is to break the ground path in the secondary side and insert a shunt resistance (\$R_3\$) in the below schematic:

enter image description here

Then the voltage developed across the shunt drives a bipolar transistor \$Q_1\$. When this voltage reaches the transistor \$V_{be}\$, it will start pulling the optocoupler LED to ground, forcing current to maintain a "constant" drop across the shunt. I say "constant" because the gain is very poor and it is a cheap system. If we consider a 0.65-V \$V_{be}\$ voltage at 25 °C, then for a 1-A constant current, \$R_3=\frac{0.65}{1}=0.65\;\Omega\$. \$R_2\$ can be dropped (unless you want to fine-tune the CC point) and \$R_1\$ typical value is \$100\;\Omega\$. In normal operation, when the output current is below the CC value, \$Q_1\$ is blocked and the added block is silent: the SMPS operates normally in CV. As the output current increases, the two loops being ORed, the CC takes the lead while the CV gives up and stop pulling LED current as \$V_{out}\$ goes down. As I said, it was a cheap way of doing CC in cell-phones chargers at the time Nokia was still number one, so quite some time ago : ) but it did the job well.

If I misunderstood the question and you actually want individual dc branches with their own current limit, then you have to go for a linear active series shunt solution but with the penalty of extra power dissipation when the current limit is reached.

Good luck with the mod and carefully test your changes with an isolated current-limited dc power source set at the lowest input voltage to limit damages in case something goes wrong.