Is it because the charging current would be too high (like the inrush to a big toroid in a typical linear supply) and might exceed overload protection? Or some other reason?
It could be because too much capacitance may cause the voltage regulation loop to go unstable. It's going to be power supply architecture dependent if a large amount of load capacitance can be added or not.
SMPS designs of years ago could depend on the Effective Series Resistance of the output capacitor for feedback loop stability.
SMPS designs of years ago could depend on the Effective Series Resistance of the output capacitor for feedback loop stability.
If the SMPS sees too big a load on power up it assumes an over current event and turns off.
You dont need massive extra smoothing with an SMPS due to its high frequency.
You dont need massive extra smoothing with an SMPS due to its high frequency.
I tried the hypex smps 400 with 0.6F and a DC soft start on the rails consisting of 2x10 ohm resistors and a relay closing at around 95% of rated rail voltage. The Hypex is not regulated and seemed to work well in terms of startup and stability but the SMPS was still a bit noisy under load. Not sure if this was due to the large caps or perhaps the 2.2mH I had inbetween. Tried a 1 ohm resistor as well but it was sort of the same.
Yes, crazy I know but wanted to test supercaps. I am not done testing either but will try with a traditional supply next.
There is the risk of a marginally stable psu, once soft started.
Check with a additional dummy load, on and off, whether it goes into oscillation and does stabilizes without ringing.
Check with a additional dummy load, on and off, whether it goes into oscillation and does stabilizes without ringing.
Aside from too much startup current triggering protect circuits, too much output capacitance may cause discontinuity burps, oscillations as the control loop tries to maintain the output voltage stability, overshoot/undershoot particularly under light loading. SMPS typically operate within a tight margin compared to mains frequency transformers and there is an optimum value of output capacitance for a given situation. In this case, bigger is not necessarily better. Consider the mains at 50Hz, a bridge with 1,000uf reservoir cap. Ripple is at 100Hz, where a SMPS with Fs of 50KHz would have a ripple at 100KHz. The ripple rejection of a 1uf cap would be the same as a 1,000uf cap at 100Hz.(discounting lead and parasitic inductances) Same with inductors. 2.2mH (2,200uH) is a HUGE inductor for SMPS frequency range. An output inductor for a typical off-line half bridge might be in the 10's of uH range depending on the desired output peak ripple current, with regard to the output capacitance and load range.
Taking the Hypex smps 400 being unregulated as an example I assume it's like a traditional supply but with a lot higher frequency so are these issues still relevant?
Meaning there is no voltage control loop just a passive mains follower so to speak.
The impedance of the reactive load is important to the output filter. Xc=1/(2piFC) and XL=2piFL. So at high F, the large capacitor shows very low Z, and the large inductor shows very high Z. The output filter uses series inductor and parallel capacitor. I suppose without regulation, if the large capacitor is decoupled it might charge the cap and operate. But, it may still affect the operation. I'm sure that the cap values were chosen to be optimal for the intended output. At such a high ripple frequency, the small capacitor is still a very large reservoir. 🙂