Why does full bridge smps have a resistor in series with cout?

[treez]

The following (attached) is a 250W full bridge smps as used by a huge Telco in a PSU.
Spec is fsw=300khz, vin=48v, vout=26v,current mode, CCM , Np/Ns = 1

The main cout cap is C54 (68uF) and you can see it has resistor R37 (220mR) in series with it.
Why is this resistor there.?
Why would you want to decrease the ESR-Cout zero frequency?....surely one would always want the esr zero to be as high as possible for stability's sake?

 

[mtwieg]

Increasing the ESR of Cout can help attain a higher crossover frequency, since it provides up to 90 degrees of extra phase boost. I've done this in some designs where bandwidth it more important than ripple voltage.

 

[treez]

So supposing that ripple isn't an issue, are there no disadvantages to adding such resistance.....I sometimes do smps led drivers with only ceramic output caps....there is very low esr and its got big tolerance, is a best to add small series resistance in order to improve stability?

 

[schmitt trigger]

Some hysteretic controllers also require some small amount of ripple voltage to operate. Although this ripple can be synthesized from the inductor, some people prefer to a actually generate it directly from the output voltage.

The resistor in series with the main output capacitor is a simple way to achieve that.

 

[mtwieg]

So supposing that ripple isn't an issue, are there no disadvantages to adding such resistance.....I sometimes do smps led drivers with only ceramic output caps....there is very low esr and its got big tolerance, is a best to add small series resistance in order to improve stability?

In general I wouldn't add ESR unless you really need the higher bandwidth and can't find any other way to do it. In my case I was trying to achieve fc>200kHz, so I would set the zero around 100-150kHz. But for most designs this isn't necessary to achieve the desired fc.