Hypex SMPS DC Error detection and DC blocking cap bypass

[MaartenJ]

Hi Hypex enthousiasts,

I'm currently building my miniDSP based crossover integrated in a slim-line case with single Hypex SMPS400A180 switching mode power supply, feeding four Ucd180HG modules. This will be directly feeding the drivers in my Linkwitz Pluto speakers.

I have two questions:
The SMPS400A180 contains two DC Error monitoring inputs (J6 pin 1 and 2) each to connect to LS+ terminal of two attached UcD modules. Following the datasheet, when DC is detected on either of this inputs, SMPS will switch to Standby mode quickly to protect speakers. This is fine, however I would like to monitor (and protect) four UcD modules. Is there a recommended way to achieve this? I was thinking of connecting two Ucd's LS+ output's per "DC Error Input" by e.g. 10k resistors in series? Would this work? If in this situation, one UcD would fail and the full rail voltage would be present on the LS+ (worst case), this config would act as a voltage divider and only half that voltage would be sensed. Would this still work? Is there a treshold to the DC detection limit? Any other ideas to really select "max out of two"?

My second question is about (a future tweak) bypassing one of both DC blocking caps in each of my four signal paths. Since the DAC's in the miniDSP have DC on their outputs, these are blocked on the miniDSP print as well. So I have a choice to bypass on the UcD input circuit or at the miniDSP print. Since I am digitally attenuating at the miniDSP, only signal cable is between the outputs of the miniDSP and UcD inputs.
Bypassing which cap would yield the biggest sonic improvement?
miniDSP is using "Nichicon Audiophile capacitors" according to their datasheet, while the Ucd180HG modules are also using Audiograde types, I don't know the brand/type, value is 684nF, OM3 is listed on the caps, nothing more. They are designated C1 and C2, (colored dark-red) left site of the input pins if I'm correct?

Would appreciate any suggestions!

gtrz, Maarten

[meanman1964]

Maarten did you asked this question at zelfbouwaudio.nl?

[MaartenJ]

No, why?

[Yves Smolders]

I asked the same question to hypex directly.

I worked around the problem in another way, but if you look on the downside of the DC protection pins, you'll that it's already 2 big resistors (I think 10K) per input leading to the same trace on the pcb. it's SMC, but according to hypex you can just add another 10K resistor to add a third DC protection.

[MaartenJ]

Yves, thanks!

So when I need to monitor DC of 4 UcD's, 2 adittional R10k resistors could be soldered to the joining location of the 2 existing SMD resistors attached to J6_1:2? That would at least give a symmetrical situation, only sensed DC would be divided by 4 (instead by 2 as in the original design already foreseen). The thing I'm worries about is whether the protection is designed to become active only if a rail voltage is sensed or that it would be triggered ealier. If the latter is the case I might be fine I think!

Any ideas, comments?

regards, Maarten