Is there a better way to protect my power supply other than the efuse?

Question

I use fairly cheap standard power supplies with no protection in my circuit experiments, usually with output of 5V, 12V, 18V, 30V and a current up to 1A. I use standard glass fuses to protect these power supplies, however short circuits and overcurrent occurs too often and I am looking for a permanent way to protect these power supplies. I found an efuse such as tPS259621 to be a good solution, but I was wondering if there's a simpler way or other suggestions/advices.

Answer 1

1. E-Fuses

e-fuse is the best choice but it may be expansive in some applications. You just have to short their output and then measure the chip temperature. There won't be any thermal runaway. e-fuses are relatively new products on the market but they are getting momentum.

2. DC/DC and AC/DC converters

All DC/DC and AC/DC converters are power limited and they won't blow when you short their output. The chip temperature will stay low and there won't be any thermal runaway.

3. BJT based power supplies

If your power supply is based on an npn BJT than it's surely power limited thanks to the SOA (Safe Operating Area) of the BJT that limits the output current once the junction temperature reaches a certain value. The problem here is that the temperature may rise up to 90 degrees and there's a concrete possibility of a thermal runaway.

4. Current limiter circuits There are tons of circuits out there. Just Google "current limiter circuits" or "current limiter circuit that don't gets hot". Simulate them with LTSpice or build a prototype and measure the temperature the reach.

5. Resattable fuses or PPTC

I quit using them back in 2000 in my projects after discovering that once tripped you have to physically disconnect the wire from the terminal block in order for them to get back to the untripped state.

They are resettable provided you physically disconnect the wire from the terminal block after tripping.

Ciao

Answer 2

I suggest getting a "bench power supply" with constant current and constant voltage modes (i.e., you get to set the maximum voltage and current independently).

Once you have your circuit working, use one of those single-voltage supplies for it, if you want.

A search for "30v 3a bench power supply" should lead you to several at less than USD100.

Answer 3

The preferred alternative is to purchase a CV CC (Constant Voltage Constant Current) power supply online. They sell adjustable ones on Amazon for around $90. The ones I've bought are 0-30V 10A, 0-60V 5A, and 0-120V 3A. You can directly short the output of these power supplies all day long and it does not hurt the supply.

Another solution would be to check to see if the power supplies that you're using are current limited or have short circuit protection built into the output. If the power supplies you are using are the wall wart style. They most likely use a flyback topology. A flyback (in general) will inherently have short circuit protecting due to the way the power supply works. With that being said, the only way to know for sure is to open up the power supply and know how to identify the topology that's used.

This next part is only an option if you are willing to sacrifice a power supply in the event that it doesn't end up having over current protection. You could run load tests to see if the output is current limited / protected. You could start out with resistors to test different loads. Make sure the power rating of the resistor is sized to the amount of power you're drawing. You'll need to measure the voltage across the load resistor during the test. A clamp DC current meter/probe is optional, but nice to have if you have one available. Do not use a regular multimeter with the leads in series to measure current. This could blow the fuse in the meter and possibly damage it if the power supply doesn't have over current protection.

Try a load at 90%, 100%, 110%, 120%, 130% and so on. Slow and steady is the key.
Measure the output voltage across the load resistor and measure the current for each test with either a probe or by using ohms law I = V/R. V is the voltage across the resistor and R is the resistance of the load resistor. Look for any drops in the output voltage as you increase the size of the load each time. For example if you notice it went from 5V to 4V.

If the power supply has overcurrent protection it will either shutoff after a certain current threshold, or it will run in a constant current mode. When the power supply is maxed out in constant current mode the current will remain the same for example at 1A. This means it's no longer a voltage supply and is now a current supply. The output voltage will be V = IR. The voltage will be determined by what resistor value is connected since the current is now constant.

If you observe it running in a constant current mode, you can continue with loading it down more and more and verifying that it's still running in a constant current mode. Eventually once you've got the load down to a few 100mOhm, you can try a direct short circuit test. Again don't do this unless you're willing to sacrifice the supply.

Answer 4

One great, non-destructive current limiting approach is incandescent light bulbs. The resistance goes up as they get hot, so initially, it will look like a short circuit, but the impedance will quickly rise with time in the event of a short, and the light bulb will light up. You can pick the bulb you need (smaller bulb = lower current limit).

Answer 5

You can also check the resettable fuses or PPTC. They basically increase their resistance with the temperature to protect the circuit from overcurrent. Like a normal fuse but without having to replace it.

You can choose the PPTC with different trip currents, so at which current it will interrupt the current flow. They are cheaper than e-fuses and easier to use, but not so precise and fast. But faster than normal fuses.

You can watch the explanation in the Great Scott channel: https://www.youtube.com/watch?v=sF0KOVWj9p8

Answer 6

I'd recommend getting a bench power supply like this one.

Advantages of the linked one (KD3005P):

• Pretty good voltage and current range (30V 5A)

• User settable voltage and output current limit

• Linear (no switching noise)

• Controllable via USB using a simple protocol, so you can script some experiments with python.

• Easy to use

If you want overcurrent protection, you must consider what type you want:

• Something that trips and shuts down like a fuse

• Foldback, hiccup, like switching power supplies

• CC/CV (Constant current/constant voltage), like bench supplies and LED drivers

These have very different behaviors in the following cases:

• If you connect it to something that has large decoupling capacitors, like the board you are experimenting with, there will be a large current spike.

The CC/CV supply will limit current while charging the caps, then return to CV mode. Everything works fine.

The others are not intended for this. They're meant to be used as part of a complete system where the load is already connected when powering up, so the capacitor charging current spike is smoothed by the power-on slow start. If you connect a board with caps to an already powered-up SMPS, then nothing will limit the charging spike, which occurs between the charged output caps of the supply and the discharged caps of the load.

If you use an eFuse like you mention in your question, it will trip and you will have to power cycle the supply to get it to work. Effective, but not very convenient.

If you use something slower like a polyswitch then it should work, but that's only because it's slow, so it'll ignore the current spike.

• If you connect it to something fragile like a LED, or a short

If the supply has output caps after its current limiting circuit, then the energy stored in the output caps will be dumped into the load before the current limit circuit can do anything about it. This is why bench power supplies have the minimum possible amount of output capacitance that still guarantees stability of the internal feedback loop. This is not the case with power bricks and other SMPS, you usually have a substantial amount of capacitance at the output.

Another factor if you don't have CC/CV is ringing and overvoltage. When you connect the output to a load that has caps, the large current spikes will excite the LC circuit formed by the supply's output caps, wire inductance, and load capacitance. This will cause it to ring and can create enough overvoltage to fry your board.