[time-nuts] Re: Z3801A questions

[ed breya]

Cool. There is a fairly simple option where you can run it from the 
Vicor SMPS, but startup has to be in two stages. You can't readily do a 
soft-start mode with a SMPS - that works with more conventional loads 
that are mostly resistive and capacitive, to ease the current surge. 
When the load is other SMPSs, you have to get through their negative 
input resistance region without the driving SMPS tripping or stalling in 
some foldback mode.

A way to get around this is to store enough startup energy in a fairly 
large capacitor - you'd need to experiment and figure out how big. You 
could run the original SMPS say, into the large capacitor, with maybe 
some small resistance to soften the charging surge, and only dropping a 
little bit in normal operation. There is a possibility of stability 
problems in certain SMPSs with large C, but most should have no problem 
at all if designed right (with AC feedback from ahead of the output 
filter, and DC from the terminals or remotely).

It would have to be a two-stage affair. First, bring up the 
SMPS/capacitor section. Then the power switch would connect the Z3801A 
to the bus and start up. This could be automated too, with appropriate 
sequencing and time delays. The main thing is to first store enough 
energy, then dump it in. The very low effective source resistance 
(mostly ESR of the capacitor) swamps out the negative resistance of the 
load until things get up high enough for normal operation. The negative 
resistance will still be there, but will have less effect at normal 
voltage, and swamped by the SMPS's low output R, after it's in regulation.

If you do a conventional transformer/rectifier/capacitor supply, you're 
mostly transferring the surge into the low impedance of the power line 
at power-up - same effect, more or less. The parts have to take a bit of 
abuse during the surge, but usually it's no problem if built to handle 
the normal load.

Ed