PSU company is giving poor advice.

[cupoftea]

Hi,
The Vicorpower "DCM design guide" is giving very poor input filter advice.

DCM design guide:

https://www.vicorpower.com/documents/design_guides/DG-DCM-Design-Guide-VICOR.pdf

On eg page 12 it is saying that the output impedance of the input filter should
stay MORE than 10 times less than the input Z of the SMPS, up to the crossover frequency
of the SMPS.
Then it shows a "nonsense" diagram of this in Fig 2.2 (a) again on page 12.

This diagram only achieves "more then ten times less than Zin" due to the ideal voltage source
that is shown there. But no real world source is going to be an ideal voltage source.....it is more likely to be the
output of an upstream SMPS, which is more like a current source followed by a CLC filter....and when you put
such a "real world " source into this diagram, you certainly do not achieve "more then ten times less than Zin".

So why are they showing this very poor advice?

Contrary to what they are saying, the input filter can in fact have an "Output impedance" which is actually even equal to the SMPS zin, -as long as this only occurs at frequencies which are less than 10x the smps crossover frequency.

 

[aliarifat794]

The input source to an SMPS is not an ideal voltage source but rather the output of an upstream SMPS, which behaves more like a current source followed by a filter. In such cases, the input filter's output impedance may not achieve the desired "more than ten times less than Zin" requirement as suggested by the guide.

 

[cupoftea]

Thanks..
Also, the attached shows that the "law" that Zout of the SMPS input filter
should be more than 10x lower than the SMPS Zin is not even accurate.
The attached shows an SMPS with a crossover of 6700Hz. At up to
6700Hz, its Zin is -3.42 Ohms. The Zout of the shown input
filter is 0.56 Ohms at the crossover, and yet there is no instability.
The attached LTspice sim confirms this.

 

[cupoftea]

Also, suppose the Power supply which supplies the SMPS is a mains transformer with diode rectifier and smoothing cap.....most of the time, the diodes arent even conducting...so its just going to look like its big output capacitor.....Ie , no ideal voltage source at all......it will be virtually impossible in this case, for the SMPS's input filter to have an output impedance which is more than ten times less than the SMPS Zin at all frequencies from 0 to crossover.
...as the attached shows

 

[Easy peasy]

Certainly - if there are low Z electro's feeding the converter in question - then there will be few issues

220uF - 470uF per amp drawn by the converter is a good guide.

these days you do not get 100's of mH in an upstream filter causing low frequency effects - which require more knowledge to deal with.