Stability of power module with non-spec'd ESR in output capacitors?

[Flyback]

Hello,
We are using a PRM48BH480x250A00 power module (48Vin to 48Vout) to give a 1.5V, 95A output by using a VTM48EF015T115A00 “electronic transformer” (NP/NS = 1/32) downstream of it.

The actual load on the 1V5 rail is about 600 ASICs.

Page 31 of the PRM48BH480x250A00 power module datasheet states that there are definite limits on the amount and type of capacitance that is fitted to its output, otherwise instability will result.

It states that its output capacitance should be no more than 47uF, and that no more than 25uF of this can be ceramic capacitance. Aluminium Electrolytic capacitance can be used to make it up to 47uF total if wished.

It is also stated that any ceramic capacitance used must have ESR in the range of 2milliohms to 200 milliohms, and that “Aluminium Electrolytic” capacitance should have ESR in the range of 0.1 to 1 ohms.

The problem is, that at the output of the VTM48EF015T115A00 “electronic transformer” we will have some 10mF of distributed ceramic capacitance, around all the ASIC loads. The thing is that there will obviously be PCB trace resistance going to all these ASIC loads, and the PRM48BH480x250A00 power module will not know the difference between “PCB trace resistance” and “ESR” and so will “think” that this is ESR……..the thing is that due to the turns ratio of the “electronic transformer”, just 2 milliohms of trace resistance will appear to be 2 ohms of “ESR” to the PRM48BH480x250A00 power module. ….that is outside of the required ESR “band” of 0.1 to 1 ohms.

-So how can we evaluate stability with this capacitance with its too_high ESR value?

I cannot see any way of using a gain phase analyser such as AP300 with the PRM48BH480x250A00 power module. For a start there is no access to the output of the error amplifier internal to the PRM48BH480x250A00 power module, and one needs to look at that during the frequency sweep of the gain/phase analysis measurement.

The feedback compensation components internal to the PRM48BH480x250A00 power module are not given so we cant calculate the feedback loop for it.
Do you know if Vicor give a guide on how to assure stability in such cases?

PRM48BH480x250A00 power module datasheet:
http://cdn.vicorpower.com/documents/datasheets/PRM48BH480T250A00_ds.pdf

VTM48EF015T115A00 “electronic transformer” datasheet:
http://www.vicorpower.com/products?productType=cfg&productKey=VTM48EF015T115A00

 

[Tony Stewart]

You can monitor V,I on the output using a sweep active load current to get the closed loop response or use a variable step load and examine time domain ringing. Then repeat with a low Q static load with an LC filter in between such that the breakpoint is approx 1/10 f or <100kHz since it switches around 1MHz.

The LCR LPF will raise input impedance and lower output impedance anywhere near the lag/lead compensation filter of the the regulator to effectivly decouple the higher c capacitive ultralow ESR that you want. This could be as low as the trace inductance.

 

[AnalogKid]

While all Vicor components are weird, Vicor has excellent tech support IMO. There have been times when I have not liked their answer, but their answers are almost always correct. They will review schematic, board layouts, whatever you send them.

ak