Ripple current in input capacitor too high?

[eem2am]

Hello,

When on Maximum Load (600W) and 90VAC input, the ripple current in the input electrolytic capacitors of the SMPS400A180 power supply is over 9 amps RMS.....



the input caps are two Samwha 820uF....there is a voltage doubler......these caps are not rated to anywhere near 9 Amps.

Samwha , 820uF, 200V electrolytic capacitors:
http://www.samwha.com/electric/product/list_pdf2/HY.pdf

SMPS400A180:
http://www.hypex.nl/docs/SMPS400_datasheet.pdf


So how can this be so.?

Am i missing something here?

Is an audio load never on maximum load for very long?....

How long is the longest it would be on maximum load?

Page 4 of the SMPS400A180 datasheet states its got a maximum load capability of 600W....
......but at 600W load, there is 9 Amps of ripple current in the input electrolytics, which is well above their rating, so how can they be stating 600W?

[janneman]

Indeed one part of the answer is that your amp will probably average 10W or less even when listening at elevated levels. I've seen many demo's with reasonable equipment loafing about at 5W average with an occasional peak to 100W+. So when benchtesting at max power your amp is stressed more than ever in music reproduction use.
So if the SMPS is rated for use in audio amps, that's a much easier environment.

The rating of those caps, is that at a specific temperature? When the caps are adequatedly heatsinked or otherwise never get above say 50 Celcius, the rating may be relaxed.

jan didden

[sawreyrw]

I don't understand your question. If you really have a 600 W load on a 90 VAC source with a rectifier and 1640 uF cap, the cap ripple current will be huge. What did you expect? Post a schematic for more help.

[Ron AKA]

The ripple current value sounds about right. However, the capacitance value seems extremely low. One rule of thumb is about 3,500 uF per amp of power supply current (50 Hz), and 3,000 uF per amp (60 Hz), on each side of the power rails.

[janneman]

Originally Posted by Ron AKA

The ripple current value sounds about right. However, the capacitance value seems extremely low. One rule of thumb is about 3,500 uF per amp of power supply current (50 Hz), and 3,000 uF per amp (60 Hz), on each side of the power rails.

This is an SMPS, so they can generally work with very high input voltage variations. They get away with relatively low input capacitance.

jan didden

[Eva]

Audio SMPS are not required to provide full output power all the time. Worst case average power is typically 1/6 to 1/8 of the peak power capability of the amplifier modules at the desired load impedances.

Ripple current in these capacitors is rated at 105 deg C and for 7000h life. These are high quality long life capacitors, much better than the caps used in other SMPS (typ. 2000h life at 105 deg C). At 55 deg C allowable ripple current for 7000h life should be considerably higher, and even higher if shorter life is allowed.

[eem2am]

janneman:, thanks for this....

I've seen many demo's with reasonable equipment loafing about at 5W average with an occasional peak to 100W+

The million dollar question is, however, what is the maximum duration of a peak power interval?

....are we talking 1 second, 10 seconds, 1 minute, or longer?

[AndrewT]

If you are listening to normal music, then near full power duration is likely to be less than one tenth of a second but repeating fairly often, or one second but repeating much less often.

If you are listening to compressed music then the near full power could be almost continuous with some music types.

If you are powering a dedicated bass speaker with disco style music then the near full power is continuous.

[eem2am]

i would imagine that a heavy metal guitar is on peak power continuously?

Also, if its a guitar amplifier, then theres nothing to stop the customer buying a 10 way multiplex box and playing 10 guitars through one amplifier.....this woul dsurely mean being on peak power all the time. (say if all 10 are being "speed-strummed" by the 10 musicians)

Sorry i dont know what is meant by compressed music,? is it something like when music signal is first put through DSP?... so that peaks are lowered and troughs are raised?

This is all significant, because if there is any , even remote chance, that the music will be on peak power continuously, then a PFC stage must be used in the supplying SMPS.

[Eva]

If a signal is on peak power too often or for too long, it will definitely not sound like music.

The waveforms from most instruments, synthesizers and voice have a low peak to average ratio due to their complex frequency content. Otherwise they would sound very close to sine or square waves, the two ideal "maximum power" signals.