+12V on Macron MPT-301

[Pentium4]

Re: +12V on Macron MPT-301

I'd work on improving case ventilation, all the heat from the Preshott flamethrower is cooking the PSU.


We need to let Preshotts know, in no uncertain terms , that we're serious about cooling for them to behave.

I think I may just need to adjust the fan settings in the BIOS to be more aggressive. As it is, the motherboard is pretty stupid. Until the P4 gets to that scorching temperature, the case fans are only running at 400RPM. It has a 80mm intake, and 92mm exhaust. When it gets to 69C or hotter, they rev up from 400RPM to 3,000RPM I'll also check the thermal paste.

Oh, and......NICE HEATSINK!

Edit: With the same amount of load, the thing kicks out noticeably less heat. I measured before and after and the exhaust temp is about 4C cooler. I'll take that as a success. 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. There is another slot on the other side of the heatsink to put another rectifier in parallel with the main one. It's being partially blocked by one of the 3.3V rails magamp coils so I'd have to extend the leads on that so the rectifier could fit.

[TELVM]

Re: +12V on Macron MPT-301

... The P4 idles around 57C, and under load gets to 70C ... ... The 92mm exhaust fan spins very slow as well unless the P4 hits 69C ...

I'd work on improving case ventilation, all the heat from the Preshott flamethrower is cooking the PSU.


We need to let Preshotts know, in no uncertain terms , that we're serious about cooling for them to behave.

[Pentium4]

Re: +12V on Macron MPT-301

Success! It liked the MBR20100CT and the voltage is reading a little stronger at 12.02V

[Pentium4]

Re: +12V on Macron MPT-301

Possibly. The top of the heatsink of the Prescott is about 1 inch below the bottom of the Macron PSU. The P4 idles around 57C, and under load gets to 70C. Too hot, but it has a copper heatsink and I don't want to invest in a higher end cooler, I even reapplied thermal paste and dusted the heatsink. I know the P4 can handle the heat (it doesn't even throttle down when hitting 73C on rare occasion), and there's Chemi-Con polymers on the VRM. The 92mm exhaust fan spins very slow as well unless the P4 hits 69C, so you may be right.

*EDIT* Worth noting: the 12V rectifier was fairly loose on the heatsink. I'm going to try an MBR20100CT

[Wester547]

Re: +12V on Macron MPT-301

It should make a tangible difference. It is baffling that the heat output was that great, though - 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. Perhaps the heat is partially coming from that toasty Prescott? (or the fan in his revision of the power supply spins faster by way of a more aggressive fan controller, maybe, or his CPU fan spins faster)

[Pentium4]

Re: +12V on Macron MPT-301

In half-bridge topology, a great deal of the time, Schottkys with a 45V reverse voltage or less are not enough - goodpsusearch shorted an ESAD83-004 in a gutless PSU (a half-bridge unit) he was rebuilding sometime ago by putting it on the +12V rail. You can get away with it in forward topology (as far as the the reverse voltage goes, the voltage it has to block when it isn't conducting is much higher) because the pulses from the transformer are usually lower in voltage in single and double forward than half bridge. Usually speaking, the rectifier with the highest reverse voltage rating is on the +12V (or -12V) rails and the ones with lower reverse voltage ratings on the other outputs (+5V, +3.3V, -5V if present, +5VSB), but yes, the underside always tells the truth. You are correct to observe that an ultrafast would work better than a schottky in a number of gutless power supplies but it wouldn't hurt to try with this one. If it doesn't work, you would probably need to replace the main transformer with another compatible one for it to work with schottky on the +12V rail. Either way, I'm sure the heat output will be improved greatly when you replace the 10A ultrafast with a 30A part.

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? And okay I'm kind of stupid/blind From the angle I could see the rectifier while installed in the PSU, it looked like a 10 but after I desoldered the heatsink and wiped off the dust, it's actually a STPR1620CT. But still, to power a Prescott P4 system, I still don't like the idea of a 16A ultra fast there. I want to get it back in commission so I may just add the 30A ultra fast. Since the VF is lower, I assume it'll still increase the efficiency.

[Wester547]

Re: +12V on Macron MPT-301

In half-bridge topology, a great deal of the time, Schottkys with a 45V reverse voltage or less are not enough (60V should be minimum as a rule of thumb) - goodpsusearch shorted an ESAD83-004 in a gutless PSU (a half-bridge unit) he was rebuilding sometime ago by putting it on the +12V rail. You can get away with it in forward topology (as far as the the reverse voltage goes, the voltage it has to block when it isn't conducting is much higher) because the pulses from the transformer are usually lower in voltage in single and double forward than half bridge. Usually speaking, the rectifier with the highest reverse voltage rating is on the +12V (or -12V) rails and the ones with lower reverse voltage ratings on the other outputs (+5V, +3.3V, -5V if present, +5VSB), but yes, the underside always tells the truth. You are correct to observe that an ultrafast would work better than a schottky in a number of gutless power supplies but it wouldn't hurt to try with this one. If it doesn't work, you would probably need to replace the main transformer with another compatible one for it to work with schottky on the +12V rail. Either way, I'm sure the heat output will be improved greatly when you replace the 10A ultrafast with a 30A part.

[Pentium4]

Here I have a classic Macron MPT-301 (Rebranded as EVER)

I recapped it a while ago, and back then I glanced ( ) at the alignment of the rectifiers and *thought* it used the 30A rectifier for the 12V but I looked at the traces and it actually uses just a 10A ultra fast recovery diode... What I should have done was just look at the location of the filtering capacitors. The 3.3V trace crosses over the 12V trace via the PI coil for the 3.3V rail. And I *think* the 12V rail uses a PI coil that is strangely over by the -12V wire.

Thing is, I'm surprised it ran as long as it did. It was powering a Pentium 4 650 Prescott, 4 RAM modules, single hard drive, 915 motherboard, and a Radeon X1050 for months. When that P4 was installing updates, the PSU was kicking out an insane amount of heat, especially since it has a conservative fan controller, it doesn't spin very fast. Although I did cut out the original punched grille and replaced it with a nice wired one That rectifier might have been getting close to its max, but even at full load the voltage never dropped below 11.89V which to me means it's doing okay.

Okay enough rambling and get to the point...Since this thing has no discoloration whatsoever, should I worry about that rectifier? Since I've been having horrible luck replacing 12V rectifiers on half bridge PSU's lately (Usually don't start up) I really don't want to keep soldering and desoldering the secondary heatsink because it pisses me off doing it more than once

What I have are some usual MBR schottkys, some Vishay VF30100S rectifiers (which I have bad luck with) and I have a 30A ultra fast, which might be the best option, It's a BYV42E-200. What do you guys think? And sorry for the long post, here are some pics







Overhead shot before recap: