turning on switch mode PSU

[duncan webster 1Participant @duncanwebster1]

In the latest issue of ME Mr Editor describes a mains to 48v switch mode power supply and how he has arranged for it to have a load before it is switched on. Is this general for all such devices? On my Topsy organ I have one giving 24v (5A max) which is permanently connected to a 24-12v converter, which in turn drives a permanently connected microprocessor, so the 48v one has a load on start up, albeit very small. It only takes serious load when the 24v fan starts up and the 12v valves start operating.

I could fairly easily arrange a bigger load which switches off after 5 seconds, or attach a somewhat bigger load permanently, but I'd rather leave it alone!

Edited By duncan webster on 19/08/2018 17:34:45

[duncan webster 1Participant @duncanwebster1]

[SillyOldDufferModerator @sillyoldduffer]

Not a general feature but it's certainly true of many. I was led a merry dance when the halogen bulb on my mill went intermittent due to a dodgy spot weld. I was convinced that the switch mode PSU was faulty because it made no volts when switched on. With a replacement bulb it works perfectly. Many computer PSU's need a load too.

On the other hand the various LED PSUs I've owned all fire up without a load.

My working hypothesis is that low power types don't need a load (or have one built-in) and high power models do. I reckon switch mode might generate excessive voltage spikes unless a load pulls them down, and that the PSU protects itself.

Could be talking complete rubbish, I've not looked at a circuit.

Dave

[Anonymous]

It really depends on the topology and how the switching circuit works. If we take a simple buck converter it will have a minimum on time. During that time an amount of energy is transfered to the output. If that energy is not used before the next switch cycle then more energy will be transfered and the output voltage will rise without being controlled. So you need a load to absorb a minimum amount of energy or the control loop loses the plot. That's one reason a lot of modern buck converters can operate in a hiccup mode, albeit at the expense of more output voltage ripple. Flyback converters are a whole different ball game, especially if you have one controlled output and several unregulated outputs following along, or so you hope.

Datasheets and apps notes should specify a minimum load, if needed. It can vary from a low value that can be consumed by the feedback resistive divider to around 10% of the maximum current rating.

True LED switch mode supplies are controlling current, not voltage, so they work in a different way and aren't so bothered by zero output current.

The reverse is also true and a switch mode supply may never start properly if it sees a maximum load at startup. To some extent it depends upon the capability of the incoming supply. For a simple buck converter the input can be modelled as a negative resistor. For constant output power as the input voltage drops the input current increases. For very low input voltages if the supply cannot provide the current the switcher will never start properly. That's one reason a lot of ICs these days have undervoltage lockouts. So nowt happens in switching terms until the input voltage is at a sensible level.

So, like everything else – it all depends……………..

Andrew

[duncan webster 1Participant @duncanwebster1]

Thanks, I get that. I'll attach my oscilioscope to it and see what wave form I get out with the fan and valves off. It won't be soon as it's not easy to get to at the moment, but I'll let you know what happens.

[Author]

Posts