+12V on Macron MPT-301

[momaka]

Re: +12V on Macron MPT-301

Even using schottkys with 100 reverse voltage I have horrible luck, they usually don't start up. Is there something not suitable with the VF30100S?

The only thing I can think/guess might be the problem is relatively slow recovery time. If the rectifier is still "ON" when the voltage from the transformer reverses in the second half-cycle, then the rectifier might appear shorted to the PSU's SC protection.

Since the VF is lower, I assume it'll still increase the efficiency.

Yes.
Vf times the current through the rectifier gives you the heat which the rectifier has to loose. So for example, if you have a rectifier with Vf = 1.3V and the current going through it is 10A... then it has to dissipate 1.3 x 10 = 13 Watts. That's quite a bit. Compare that with a 0.5V schottky passing the same current: 5 W - almost 1/3 the heat output just from that rectifier.

I believe Momaka has the same power supply (with a 16A ultrafast or fast recovery on the +12V output), and with an AMD 95W TDP CPU he couldn't get it to kick out any noticeable heat even in hot weather, even at full load.

Correct. It was actually with a 105 W TDP AMD FX-57 CPU. I also tested a friend's i7 quad-core with a 95W TDP. Although, I should note that my tests weren't longer than an hour or so and not under 100% CPU load all of the time. I just tested a few games for a little bit. I should also note that I have drilled a few extra holes in the case of my Macron PSU.

When idling, the 12V sits at 12.20V, at 100% load it drops to 12.02V. I thought that was bad but I guess that's just typical group regulation. When the CPU usage is jumping up and down a lot, the 12V does bounce around quite a bit...not sure if I should worry about that or not.

Nothing to worry about. Just about every brand-name computer from the Pentium 4 era does this. On my Dell OptiPlex GX170L, for example, the bounce on the 12V rail is so bad that you can hear the system fan revving up and down when the CPU load changes (this is with a 250W HiPro PSU). IIRC, the voltage on the 12V rail is bouncing between 11.88V and around 12V. But this is not my only computer that does this. In fact, I think my HP Pavilion 8756c is even worse - and that thing is an ancient Pentium 3 -based system that I still use today when at home and not in college. Works perfectly fine to this day with its original IBM Deskstar HDD.

You'd think HDDs hate the bouncing on the 12V rail, but what they actually hate the most (in my experience, anyways) is high ripple. I had an HDD nearly cook itself from a crappy L&C PSU. Same HDD ran just normally warm with a decent PSU.

^ That's if you parallel them without a heatsink. If they're in parallel on a good heatsink (if they're on opposite ends of the same heatsink and have adequate thermal pads), they probably won't be more than a couple of degrees apart in temperature (celsius), so it shouldn't be a problem.

Exactly. It's crucial to have the rectifiers near each other AND on the same heat sink.

^ Two 15A bridge rectifiers inside a 800W PSU, both on the same side on the heatsink . Wouldn't the external one get much hotter?

Probably.
But I don't think it matters here this much anyways because even one of those bridge rectifiers will be sufficient to handle the full load from the PSU. I've seen microwave ovens rated for over 1 kW with just one of these bridge rectifiers. I think it's more of a gimmic/hype that they have bridge rectifier heat sinks in those PSUs.

[budm]

Re: +12V on Macron MPT-301

800W, that is about 6.67A of current, Vf of the power rectifier can be as high (or more) as 1V, two diode Vf drop in each 1/2 of AC cycle, so power will be about 13.33W/2 = 6.67W (for 2 Diode Vf which will conduct only 50% of the cycle). The other 2 Diodes in the bridge will dissipating heat in another half of the AC cycle. So it can get hot.

[Wester547]

Re: +12V on Macron MPT-301

Synchronous rectification helps even more as far as secondary side efficiency goes.

Nothing to worry about. Just about every brand-name computer from the Pentium 4 era does this. On my Dell OptiPlex GX170L, for example, the bounce on the 12V rail is so bad that you can hear the system fan revving up and down when the CPU load changes (this is with a 250W HiPro PSU). IIRC, the voltage on the 12V rail is bouncing between 11.88V and around 12V.

Really? I thought Hipro's voltage regulation (in those 250W Dell units with a 3300uF 16V capacitor before and after the ferrite coil and either a 20A or 30A schottky on the +12V rail.... but some of the later models had no ferrite coil at all and two 3300uF 16V capacitors and a 1000uF 16V capacitor in parallel on the +12V output) was better than that. I remember you saying you couldn't get your HP-P2507F3P to crossload... unless you're talking about another unit?

Exactly. It's crucial to have the rectifiers near each other AND on the same heat sink.

Is that because the further they are apart, the greater the difference in temperature (IE, paralleling them on a heatsink where they're right next to each other is better than paralleling them where one is on one side and the other on the other side of the heatsink)?

[c_hegge]

Re: +12V on Macron MPT-301

But I don't think it matters here this much anyways because even one of those bridge rectifiers will be sufficient to handle the full load from the PSU. I've seen microwave ovens rated for over 1 kW with just one of these bridge rectifiers. I think it's more of a gimmic/hype that they have bridge rectifier heat sinks in those PSUs.

The reason is that bridge rectifiers are generally rated for what they can deliver with a heat sink installed, unlike discrete diodes.

[TELVM]

Re: +12V on Macron MPT-301

... I don't think it matters here this much anyways because even one of those bridge rectifiers will be sufficient to handle the full load from the PSU. I've seen microwave ovens rated for over 1 kW with just one of these bridge rectifiers. I think it's more of a gimmic/hype that they have bridge rectifier heat sinks in those PSUs.

I suspect this pairing of way oversized bridge rectifiers may have something to do with overall efficiency. The pic comes from a 80+ Platinum PSU, and the same can be seen in many gold/platinum PSUs.

It maybe not critically neccessary on some PSUs, but I think heatsinking the bridge rectifier is a healthy custom, it wouldn't harm .

[momaka]

Re: +12V on Macron MPT-301

but some of the later models had no ferrite coil at all and two 3300uF 16V capacitors and a 1000uF 16V capacitor in parallel on the +12V output) was better than that. I remember you saying you couldn't get your HP-P2507F3P to crossload... unless you're talking about another unit?

It's another 250W unit... one of the newer 250W ones.
But keep in mind it's running a Dell OptiPlex GX170L with a 2.8 GHz P4 Prescott CPU. Those things have a TDP of 89W and can pull a max of 100 W. With nearly no load on the 5V rail, you can see why that is. If it was the other way around with a large load on the 5V rail and almost none on the 12V, I think the PSU would cross-load a lot less.

Is that because the further they are apart, the greater the difference in temperature

Yes.
Behind each other with a heat sink in between is the best way to do it.

800W, that is about 6.67A of current, Vf of the power rectifier can be as high (or more) as 1V, two diode Vf drop in each 1/2 of AC cycle, so power will be about 13.33W/2 = 6.67W (for 2 Diode Vf which will conduct only 50% of the cycle). The other 2 Diodes in the bridge will dissipating heat in another half of the AC cycle. So it can get hot.

Woops, big brainfart on my part there. Yes, without a heat sink, they will.
I think what i meant to say was that having two bridge rectifiers is a gimmick, not that the heat sink is (don't even know why I said that.... the microwave ovens I opened all had their bridge rectifiers heat sinked... what I wanted to point out is that they all had just one bridge rectifier).

[Wester547]

Re: +12V on Macron MPT-301

It's another 250W unit... one of the newer 250W ones.
But keep in mind it's running a Dell OptiPlex GX170L with a 2.8 GHz P4 Prescott CPU. Those things have a TDP of 89W and can pull a max of 100 W. With nearly no load on the 5V rail, you can see why that is. If it was the other way around with a large load on the 5V rail and almost none on the 12V, I think the PSU would cross-load a lot less.

Good point. IIRC you had tested your HP-P2507FP3 with a R300/Radeon 9500 Pro and because of the fact that the card puts a good load on the +3.3V, +5V, and +12V rails (despite not really being a power hungry card at all, even for its time), it keeps the rails balanced very well in terms of voltage regulation, so that probably helped lots.

[TELVM]

Re: +12V on Macron MPT-301

And how about three 10A bridge rectifiers inside a 1275W platinum PSU?:

http://www.techpowerup.com/reviews/T...X-1275M/4.html

That's theoretically enough rectifying for 7KW @230VAC . But I don't think max total power is what that mega-overkill trio is about, but efficiency.


I bet in the not too distant future we'll begin to see mosfet bridge rectifiers inside top-notch "80+ Unobtainium" PSUs.