Polymodding: Does it actually work?

[Wester547]

Was the fan much quieter in the "before" test than in the polymod test?

[c_hegge]

I can't remember that far back, but I doubt if it would have been any different

[LongRunner]

What I'd be interested in is - wait for it - a second poly-mod test. But this time covering specific scenarios, such as fitting only the standby output of a PSU with polys of much lower capacitance (but still adequate ESR and ripple ratings) and seeing if that works. Try it with both two-transistor and IC-based +5VSB circuits.

The theme here being circuits that probably don't actually require the capacitance of the original electrolytics, but instead made them as large as they were to get sufficiently low ESR.

After all, a 220µF 6.3V 6.3mm Chemi-con PSA has the same ESR as a 2200µF 6.3V 10mm Panasonic FR - which is itself high-grade for a PSU.

[Wester547]

I can't remember that far back, but I doubt if it would have been any different

Do you remember if the fan was as loud as it was in the polymod test, in the 320W test in the HP-D3057F3H review?

[c_hegge]

No, I don't. I doubt if it would have been any different, but I can't remember that far back.

[Wester547]

I find it very fascinating how the ripple voltage and voltage regulation results are so similar with polymers that have much lower ESR ratings and much higher ripple ratings than the Teapo SCs on the output before it. I understand that the microfarad ratings are much lower but even still. It does explain how flexible that particular output filter is, though. Maybe it has to do with the feedback loop and PWM control, or more to do with the toroids and PI filter coils. It makes me wonder how much a contributing factor ESR rating is to ripple voltage (or even capacitance to voltage regulation), though maybe that design is just an exception.

[c_hegge]

I'd say so. Remember what happened to the Antec Neo 500 (Seasonic Super Silencer based). That thing went crazy with the polies

[Wester547]

I see before and after results for the SP-400 test, yet it was stated you couldn't get it to boot with a PC, unless those results were tested when connected to the load tester... ?

[c_hegge]

The results were indeed what it did on the load tester. My guess is that the polies affected it's transient response (ie. the way the unit handles changes to its load). The load tester provides a more stable load, so even though it looked OK there, it didn't work properly with the constantly changing load on the PC.

[Wester547]

I wonder how Hipro managed to get linear regulation grade ripple voltage (<4-5mV) and voltage regulation with just the magamp method on the +3.3V rail. I understand that the +3.3V rail uses saturable reactor control which makes both voltage regulation, cross loading, ripple suppression, and efficiency easier, and the fact that the +3.3V toroid only has to regulate one rail and not up to five rails, and that the power draw is much lower, but even still.... I suppose it would have lots to do with how the feedback is designed on all accounts.