When to load a switch mode power supply unit?

Thread Starter

PlasmaT

Joined Feb 19, 2015
61
To my understanding, switch mode power supply (SMPS) units tend to draw in a high current at start up when the capacitors charge up. This inrush current can be significantly more than normal operating current.

Therefore is it safe (or correct) to connect the load to the SMPS unit before its plugged into the mains? or should I wait for a while for things to stabilize and then connect the load? Appreciate your valuable advice.

Thanks in advance.
 

MaxHeadRoom

Joined Jul 18, 2013
28,617
It has been my experience that there is far more inrush in Txfmr/linear supplies than SMPS.
Although they both exhibit inrush of some kind.
It is common to find both types connected to load before switch on.
Max.
 

Reloadron

Joined Jan 15, 2015
7,501
I have seen some SMPS driving some pretty heavy loads at turn on, even things like a large computer system or network. They come up just fine under a load.

Ron
 

ErnieM

Joined Apr 24, 2011
8,377
The load is probably irrelevant for the inrush current as the inrush is only banging into a cheap & dirty cap bank for a very short time to give these caps their initial charge. It is only after the cap bank is charged that the house keeping comes alive and the output comes alive and into regulation after the soft start up time expires.

Note there is nothing inherent in the design of a switcher making it draw this huge current. The last one I worked on has a series resistor to limit the inrush current; this resistor is shorted out then the house keeping supply comes up so the efficient stays up too.

It also has power factor control so it runs nearly at unity. Switchers have come a long way since I first saw them.
 

Thread Starter

PlasmaT

Joined Feb 19, 2015
61
...It is only after the cap bank is charged that the house keeping comes alive and the output comes alive and into regulation after the soft start up time expires.

....The last one I worked on has a series resistor to limit the inrush current; this resistor is shorted out then the house keeping supply comes up so the efficient stays up too...
So the power supply output is ready only after the capacitors are charged up. Seems logical.
It is also understood that the load can be connected to the power supply unit before its plugged in to the mains.

Using a resistor in series seems a wonderful approach. I have always wondered why are the capacitors directly connected to the mains via the diodes only. Thanks for sharing ErnieM.
 

AnalogKid

Joined Aug 1, 2013
10,986
There are a lot of different switcher topologies, and many variations for small vs. large supplies. But generally speaking for industrial type supplies of a one hundred watts or more, they start up better with the load attached because the outputs do not regulate well with zero load. In fact, many have a minimum load requirement of 5% or 10%. Also, as mentioned above, the input current is shaped by the control IC with a technique called "soft start." The inrush current still can be 10 times the load current, but the two are almost completely independent in those first few milliseconds.

ak
 

Thread Starter

PlasmaT

Joined Feb 19, 2015
61
Very interesting...So connecting a capacitive load or an inductive load will have different effects on the power supply unit at start up.
 

AnalogKid

Joined Aug 1, 2013
10,986
No. They might cause the supply output to break into oscillation, but as stated they will not affect inrush current. Most of inrush current happens before the switching supply starts switching, so there is no regulation translating output load characteristics into input current.

ak
 

ian field

Joined Oct 27, 2012
6,536
To my understanding, switch mode power supply (SMPS) units tend to draw in a high current at start up when the capacitors charge up. This inrush current can be significantly more than normal operating current.

Therefore is it safe (or correct) to connect the load to the SMPS unit before its plugged into the mains? or should I wait for a while for things to stabilize and then connect the load? Appreciate your valuable advice.

Thanks in advance.
Many commercial SMPSUs have a NTC inrush limiting thermistor in series with the mains fuse, generally speaking - if it ain't broke don't fix it, but on ones I have to repair, I usually check whether they have one, if not I add one.

A very tiny minority of SMPSUs can go bang if you fire them up with no load - but that is the exception to the rule.
 

Thread Starter

PlasmaT

Joined Feb 19, 2015
61
...
A very tiny minority of SMPSUs can go bang if you fire them up with no load - but that is the exception to the rule.
Yes. I had similar experience with a 12V 2A SMPS. I used to power up a arduino nano based device which drew less than 25 mA during normal operation. The power supply adapter was used since i happened to have it at the time of test run, perhaps it was not the optimum power supply to use. Anyway after a few days of continuous operation, the power supply adapter gave up. When I opened it up, a power transistor had "roasted" and noted it had no heat sink as well.
 

Evil Lurker

Joined Aug 25, 2011
116
With SMPS the idea is to keep the inductor or transformer operating in continuous current mode, i.e. the current flowing through the inductor never reaches zero. If the SMPS switching and feedback loop cannot handle this the PSU goes into discontinuous or "burst" mode where it will switch literally in small bursts. When a PSU which is normally designed for running in continuous current mode is forced to run in discontinuous mode ripple on the output shoots up and the voltage on the output is much "dirtier". I'm guessing that this is also quite hard on the output filter caps as well. Some SMPS PSUs, especially low power wall warts are perfectly ok with this, and even designed with that taken into consideration so the problem is usually limited to larger supplies in oh I'm guessing the 100W+ range.

As far as surge on the startup, most SMPS have current control mode meaning each switching cycle is pulse by pulse current limited with a set maximum peak inductor current. In theory the output should not experience any voltage spikes upon start up. What you do have to worry about is surge going into the SMPS from the mains. You take a PSU with oh, say 1500uF+ @250V capacitance and fire that bad puppy up without some sort of choke or NTC and well let's just say it will make the lights flicker and possibly even trip a breaker.
 
Top