Mad Dog Multimedia (Super Flower) MD-500SCPS

[Pentium4]

This is a project that I'd like to get to at some point soon. I bought it 4 years ago and it's been sitting in the original box and even has a sticker on it dated April 6, 2006 from CompUSA. Unfortunately, it did not come with all of the modular cables; but does have enough for functionality. One 2 SATA power cable, and one 3 molex + berg cable. 20+4 pin and 12V 4 pin are hard wired.

It's sad that almost every associated company with this unit and the box advertisements are out of business: CompUSA, Mad Dog Multimedia, ATi, BFG, and Maxtor... lol

The box does have some interesting info on it:





I really do like the case of this unit. It's unique looking and the 3 different sized fans definitely adds to that. The fan switch is also pretty neat. In my opinion, I think only Auto is tolerable, noise wise. Standard is loud and Turbo is REALLY loud.

Label:


Shots to show the fans and modular connectors:




As you could have probably guessed, all of the fans are not ball bearing, but sleeve bearing Globe Fans.

Overhead shot:


On the input side, you can see the large Jenpo caps. There are 2 MOVs behind them. I like that the switchers are on the other side of the heatsink as to not get the main caps hot. They are two 2SC3320s. I think I will leave the main caps when I work on it.


There is a 30A schottky on the 3.3V and 40A on the 5V. Unfortunately, I can't read the 12V output rectifier because the numbers on it are faded. As of now, I also can't tell which rail that 10Ω resistor is on.

What do you all think of this unit?

[eccerr0r]

I have one of those "Mad Dog" branded PSUs as well, but it's 400W. Yes I hate it when the "modular" cables are gone...sigh. Though I probably can attach some adapters to the molex 4 pin, it's useless for PCIe...

However I did find that it seems the modular sockets on the case, at least the 4-pin, seem to fit the 12V for the ATX CPU connector. But of course I don't have any surplus cables, not to mention the actual 12V connector on this power supply is toast. So I need to get a whole bunch of these connectors, plus get one for SATA/PCIe...

I didn't look inside mine yet however, but this 500W one looks a bit overrated, seems to be fine for 300W or maybe 400W tops, it at least has 1000µF primaries (though would be half since they're in series). Transformer looks tiny however, so not 500W. But it has one heck of a lot of fans, maybe it will work...

[Pentium4]

Originally Posted by eccerr0r

I have one of those "Mad Dog" branded PSUs as well, but it's 400W.

I believe I had that one too! Was it a Macron unit with a 120mm Nspire fan?

Originally Posted by eccerr0r

Yes I hate it when the "modular" cables are gone...sigh. Though I probably can attach some adapters to the molex 4 pin, it's useless for PCIe...

I do have some extras, maybe they will fit. There's no way each manufacturer is purely proprietary, right??

Originally Posted by eccerr0r

However I did find that it seems the modular sockets on the case, at least the 4-pin, seem to fit the 12V for the ATX CPU connector. But of course I don't have any surplus cables, not to mention the actual 12V connector on this power supply is toast. So I need to get a whole bunch of these connectors, plus get one for SATA/PCIe...

Man, I wish I knew. Back in June I recycled a massive bag of salvaged molex/Sata/PCIe cables

Originally Posted by eccerr0r

I didn't look inside mine yet however, but this 500W one looks a bit overrated, seems to be fine for 300W or maybe 400W tops, it at least has 1000µF primaries (though would be half since they're in series). Transformer looks tiny however, so not 500W. But it has one heck of a lot of fans, maybe it will work...

I do think the 1000µF caps wwould help the switchers reach their max potential. The 39 transformer isn't that small and I think Super Flower designs PSUs pretty well. I do know that they've got 500W out of a half bridge unit with an ERL39 transformer before in a Kingwin branded unit. As for this unit, I think you're correct. The heatsinks couldn't handle it, especially with that 10W resistor adding heat and taking away some of the surface/air ratio.

I'm really curious about the 12V rectifier. Either way, nothing I will put this in will pull more than 150-175W so I think it'd be a great unit for that and even help get rid of some of the chassis air

[eccerr0r]

Not sure what the specifics are of the 400W I have, would have to go dig it out of the computer it's in...

I saw a place sell a whole bunch of "modular" cables by themselves, but they did not match the Mad Dog PSU unfortunately. Seems most "modern" "modular" use 6-pin so they can have two grounds, one 12V, one 5V, and whatever else is necessary like an extra 12V and 3V3 for SATA which seems to have turned out to be disused...

[goodpsusearch]

Wow this power supply has many fans! (pun intended)

[momaka]

Wow, really cool save, Pentium4!
I'm glad you were able to get ahold of such PSU, as I know you'd restore it perfectly.

The box is such a trip from the past with those old companies and pictures.
IDK why, but it makes the inner geek in me smile when I see stuff related to old PC hardware.

Originally Posted by Pentium4

I really do like the case of this unit. It's unique looking and the 3 different sized fans definitely adds to that. The fan switch is also pretty neat. In my opinion, I think only Auto is tolerable, noise wise. Standard is loud and Turbo is REALLY loud.

Yeah, the case is indeed quite cool with that anodized mirror finish.
As for the loud fans - that was pretty normal for back in the early to mid 2000's. If anyone disagrees, go fire up an old Athlon XP box with a stock cooler, then come back and let me know what you heard.

Originally Posted by Pentium4

As you could have probably guessed, all of the fans are not ball bearing, but sleeve bearing Globe Fans.

Nothing wrong with that.
They'd probably be even louder if they were BB's.
Globe Fans are OK, IMO... most of the time.

Originally Posted by Pentium4

Overhead shot:
https://www.badcaps.net/forum/attach...1&d=1639451065

Looks decently packed with components.
I can't see the output toroid to tell you more about its capabilities, though.
If it has ~1" (25 mm) diameter output toroid... output power will probably be capped to around 300-350 Watts. With the next size up (1.3"... or ~33 mm diameter), that can stretch things out to 500-550 Watts.

Originally Posted by Pentium4

On the input side, you can see the large Jenpo caps.

2x 1000 uF - nice!
Shame they are Jenpo brand, though. Probably would still last a while, due to lack of APFC. I just feel a little uneasy about Jenpo, based on what I have seen with that brand used on the output of PSUs - almost always fails. But again, these being on the primary side should be OK.

Originally Posted by Pentium4

There is a 30A schottky on the 3.3V and 40A on the 5V. Unfortunately, I can't read the 12V output rectifier because the numbers on it are faded.

Is it in a TO-247 or TO-3P case? If so, it's very likely a 30 Amp part... which for half-bridge topology means you can actually pull the full 30 Amps out of it. So if that's the case, then at least the label for the 12V rail is not a lie.
Only a matter if the output inductor can handle that kind of current, particularly when cross-loaded with a light 5V-rail load.

Originally Posted by Pentium4

As of now, I also can't tell which rail that 10Ω resistor is on.

More than likely, it's *not* the 12V rail (that would equate to 14.4 Watts of power dissipation. ) And probably not the 5V rail either... but no guarantees. Usually such low values (20 Ohms and under) are on the 3.3V rail.

Originally Posted by Pentium4

What do you all think of this unit?

Classic half-bridge design with a cool/funky case. Should do nicely for systems under 200 Watts total. It's got a strong 5V rail, so can also power retro Athlon XP rigs or dual PII/P3 machines as well.

Originally Posted by eccerr0r

Yes I hate it when the "modular" cables are gone...sigh.

My biggest pet-peeve about used modular PSUs!

Originally Posted by eccerr0r

Transformer looks tiny however, so not 500W.

I think that's only how it looks in the picture, due to the case of the PSU being larger and because the heatsinks and primary caps are also big.
If the main transformer is a "true" 39-size (meaning, it measures 39 mm across the top, length-wise), then it could / should be able to technically handle 500 Watts... though probably peak power. EE/EI/ERL35 can usually do up to 350 Watts and ~400 Peak. So EE/EI/ERL39 is just a step-up from that, extending to about 400-450 Watts generally.

The primary heatsink looks big enough to handle that power. The secondary one... maybe! Regular Schottky rectifiers waste a lot of power / heat up a lot, due to forward voltage drop, compared to almost no voltage drop on modern-day synchronous rectification designs (almost anything high-power nowadays.)

But I think the biggest limit for the output power is the output toroid/inductor. See remark I made above. If it's a small T106-26 core, forget it to do more than 300 Watts. Even a T106-52 will run rather toasty at that power level. For power above 300 Watts (but not more than 500 Watts), usually a T130 core size is needed (preferably a -52 core for lower losses.)

Originally Posted by Pentium4

I do think the 1000µF caps wwould help the switchers reach their max potential.

Yeah, the 1000 uF caps (500 uF total capacitance due to series connection) should be plenty for 500-600 Watts.

I mean, many modern cheap 500 Watt APFC PSUs have only 270-330 uF on the primary side. Sure the APFC boosts the voltage, so that means higher energy storage despite the lower capacity. So in the case of older PSUs like this with no APFC, the voltage is lower, but the capacity is higher... so energy is somewhat similar.
A 500 uF cap charged to 340V DC has about 28.9 Joules.
A 330 uF cap charged to 385V DC has about 24.5 Joules.

Originally Posted by Pentium4

The 39 transformer isn't that small and I think Super Flower designs PSUs pretty well.

They do indeed.

Speaking of Kingwin units... I got one labeled 650 Watts... but it's not really a 650 Watt PSU at all. :\
Will post about it some day soon I hope.

[Pentium4]

Originally Posted by momaka

Wow, really cool save, Pentium4!
I'm glad you were able to get ahold of such PSU, as I know you'd restore it perfectly.

The box is such a trip from the past with those old companies and pictures.
IDK why, but it makes the inner geek in me smile when I see stuff related to old PC hardware.

Thank you, sir. I do 100% plan on doing a full refurbish but it's getting harder as I get older. Especially when there are such annoying tasks as removing conductive glue. Recapping (especially modular PSUs) is a task enough and the glue just makes it feel like much more of a task.

The box is bittersweet for me. It's cool and interesting to see but it also makes me sad because I miss when computers were more simple.

Originally Posted by momaka

Yeah, the case is indeed quite cool with that anodized mirror finish.
As for the loud fans - that was pretty normal for back in the early to mid 2000's. If anyone disagrees, go fire up an old Athlon XP box with a stock cooler, then come back and let me know what you heard.

I actually didn't even wanna touch it cause it's been flawlessly clean for over 15 years and all it takes is one fingerprint to "ruin" it

The old Athlon XP machines were the first thing I thought of too lol. Sure, it kept them cool but they were maddening to listen to and usually got very dusty and killed the fans more than it normally would. I have some old Pentium 3 machines where I use an oversized cooler and then hardwire a molex>3 pin connector to 5V or 7V.

Originally Posted by momaka

Nothing wrong with that.
They'd probably be even louder if they were BB's.
Globe Fans are OK, IMO... most of the time.

Agreed! I'm glad that they're sleeve bearing. It's the false advertising that I don't like. I haven't checked yet, but I will get some fresh oil in there right before I assemble it.

Originally Posted by momaka

Looks decently packed with components.
I can't see the output toroid to tell you more about its capabilities, though.
If it has ~1" (25 mm) diameter output toroid... output power will probably be capped to around 300-350 Watts. With the next size up (1.3"... or ~33 mm diameter), that can stretch things out to 500-550 Watts.

I took some better pictures so you can take a look. It's worth noting that the fan control thermistor was stuck in the middle of the main output toroid.

Originally Posted by momaka

2x 1000 uF - nice!
Shame they are Jenpo brand, though. Probably would still last a while, due to lack of APFC. I just feel a little uneasy about Jenpo, based on what I have seen with that brand used on the output of PSUs - almost always fails. But again, these being on the primary side should be OK.

Yeah, their caps are very low quality. However, it's amazing how well any can do in a voltage doubler. I think it's from what Wester547 mentioned how the higher voltage keeps the electrolyte much healthier. Not having any stress and not getting hot also helps.

Is it in a TO-247 or TO-3P case? If so, it's very likely a 30 Amp part... which for half-bridge topology means you can actually pull the full 30 Amps out of it. So if that's the case, then at least the label for the 12V rail is not a lie.
Only a matter if the output inductor can handle that kind of current, particularly when cross-loaded with a light 5V-rail load.

5V and 3.3V are TO-247. The 12V is TO-3P. I was finally able to check it... it's a 16A ultra fast This actually severely irritated me, damn it, Super Flower!
Seriously. In an era of Prescott Pentium 4s and the 89-110W TDP Ahlon 64/FX CPUs, you seriously thought 16A was enough??? Just the CPU alone would use up most of that part under extreme loads. Not even counting fans, motherboard, VRM efficiency, GPU, hard drive(s), DVD drive(s). The other annoying part is I don't know how it will react to a shottky. I only have up to 20A ultra fast parts. I have plenty of brand new 30A schottkys but sometimes half bridge units don't like them. So I have to take apart the stupid thing, put one in and test it. Worst scenario, I will put in a 20A ultra fast and just use it with a lower TDP CPU.

More than likely, it's *not* the 12V rail (that would equate to 14.4 Watts of power dissipation.) And probably not the 5V rail either... but no guarantees. Usually such low values (20 Ohms and under) are on the 3.3V rail.

You're probably right, it's most likely on the 3.3V. The leads are oddly separated. I won't be able to tell which rail its on until I take it out and trace it. I would love to just remove it. You think it would go haywire if I fully removed it? I could also just reuse those heatshrunk leads to put in a 50Ω or 100Ω resistor there.

[eccerr0r]

Interesting they put a warning that the modular connectors are not hot plug

[Topcat]

Originally Posted by Pentium4

...but it also makes me sad because I miss when computers were more simple.

Nail, meet head! I really miss the 'golden era' of PC; from ~1995 through maybe 2010. Of course that may be subjective & debatable by some....but it was IMHO a period where the PC really got into the mainstream markets and all the glory & neat things that arose during those times....much like the 'golden eras' of radio & television.... Today, they're just another PC; quite unexciting & stale...the 'wild west' days of the PC & internet are a thing of the past....

Didn't mean to hijack the thread; and good save on the PSU!

[momaka]

Originally Posted by Pentium4

The box is bittersweet for me. It's cool and interesting to see but it also makes me sad because I miss when computers were more simple.

Yeah, I miss those times too! (late 90's / early 2000's). Now everything is too dumbed-down on the surface, but extremely complex and impossible to get to underneath ("sandboxed" might be the word here?) Also, everything is too uniform/streamlined - i.e. if manufacturer A makes a board with features X and Y, manufacturer B will make almost the same board with the same exact X and Y features without much deviation. In contrast, old hardware was a lot more diverse... (maybe even "wild" at times.)

Originally Posted by Pentium4

I actually didn't even wanna touch it cause it's been flawlessly clean for over 15 years and all it takes is one fingerprint to "ruin" it

That's funny!
Understandable though.

Perhaps wear gloves when handling it?
I know it sounds weird, but at some point when I used to work at Microcenter and was cleaning returned PCs and TVs, I had to wear gloves - especially for the TVs. Once I got a good exercise from moving them TVs, it would be almost impossible to stop my hands from sweating and leaving fingerprints on everything that I just cleaned. So I started wearing gloves. The downside is, I really disliked the rubber smell that my fingers got from the gloves. But on the upside, no more dead/decaying skin on some of my fingers from using the cleaners all the time.

So anyways... maybe pretend this PSU is your patient and wear gloves?

Originally Posted by Pentium4

The old Athlon XP machines were the first thing I thought of too lol. Sure, it kept them cool but they were maddening to listen to and usually got very dusty and killed the fans more than it normally would. I have some old Pentium 3 machines where I use an oversized cooler and then hardwire a molex>3 pin connector to 5V or 7V.

Yeah, I do the same for the P3 CPUs. Heck, even with the stock heatsink and fan, most P3 CPUs will still run cool enough with fans on 5-7V.
Sadly, this doesn't work too well on the Athlon XP's. Well, perhaps on my lowly-low Duron Applebred (Thoroughbred B) it might work, since it's only about 50-ish W TDP. But on my XP 2500+, no way!

Originally Posted by Pentium4

Agreed! I'm glad that they're sleeve bearing. It's the false advertising that I don't like.

True.

Originally Posted by Pentium4

I took some better pictures so you can take a look. It's worth noting that the fan control thermistor was stuck in the middle of the main output toroid.
https://www.badcaps.net/forum/attach...1&d=1643594884
https://www.badcaps.net/forum/attach...1&d=1643594884

Thanks!
Main output toroid doesn't quite seem to reach the height of the secondary heatsink. And the secondary heatsink isn't too tall due to the overhead fan... so I think the output toroid probably uses a T106-26 core. (The "106" specifies the outer diameter - i.e. 1.06", or 25 mm. And -26 is the core material type: powder core up to 50 KHz with average losses. This, you can tell by the color: yellow core with white paint on one side is -26 type. In contrast, light green core with blue paint on one side is -52 type. -52 type cores have very similar specs as -26 cores, but with lower losses at higher frequencies, so they tend to be more efficient.) T106-26 is probably only good for up to about 250 Watt PSUs (maybe 300W peak.) T106-52 core would have been better here and could allow probably 300-350W designs. For anything above that, I usually see manufacturers (the good ones like Delta, HiPro, and LiteOn) jumping to a T130 size core.

Originally Posted by Pentium4

Yeah, their caps are very low quality. However, it's amazing how well any can do in a voltage doubler.

JEE, you're telling me?
Actually, I have even something worse: VIVA primary caps in an old 250W Deer PSU... and YES, they are still holding up just fine! The ones on the output, of course, have failed.

Originally Posted by Pentium4

5V and 3.3V are TO-247. The 12V is TO-3P. I was finally able to check it... it's a 16A ultra fast This actually severely irritated me, damn it, Super Flower!

Oof!
Yeah, that's very underwhelming, considering the label claims 30 Amps on the 12V rail.
And it's even more of a shame, because the primary side is perfectly adequate for the task: 1000 uF input caps, TO-3P switching BJTs, and ERL-39 transformer.

Originally Posted by Pentium4

Seriously. In an era of Prescott Pentium 4s and the 89-110W TDP Ahlon 64/FX CPUs, you seriously thought 16A was enough??? Just the CPU alone would use up most of that part under extreme loads. Not even counting fans, motherboard, VRM efficiency, GPU, hard drive(s), DVD drive(s).

Hey, I've ran a system with an Athlon FX-57 (103W TDP) and a GeForce 6800 XT (~50-60W TDP) on a HiPro PSU rated for 16 Amps on the 12V rail just fine. It did run quite warm, but it has the heatsinks for it (though it probably didn't help that it was a "transfused patient" PSU, using the guts of a HiPro in a case with fan from another PSU that had much worse fan.) That HiPro has only a 20 Amp rectifier on the 12V rail and it's a single-transistor forward config... so the 16 Amp rating really is 100% accurate. I also pushed the same PSU (prior to this setup) with an Athlon 64 3500+ and a ATI Radeon HD3870 video card.

So 16 Amps was actually acceptable for a PC back in that era... though kind of mid-range-y. 22-25 Amps would have been what the "high end" PCs back then would need (unless we are talking about GeForce 8800 GTS/GTX/Ultra - those were power monsters! )

But then there's the output toroid too - if it really is only a T106-26 part (try to measure its outer diameter), there won't be a point to go with a very highly-rated 12V rectifier, because the toroid will probably overheat way before you reach the limits of the rectifier. I actually ran into that issue with my Inno Power (Macron) MPT-301 when I used it on a very heavy 12V system (Pentium 4 Northwood + GeForce 6800 XT AGP) - the PSU got very very hot. It didn't shut down or die or do anything weird. But I can tell you it wasn't happy running such a 12V-heavy cross-load. I did a similar load test later with a NiCr heating element wire, and both extreme 5V-heavy and 12V-heavy cross-loads got the output inductor very very hot - over 65C!

I think that's why they stuck the thermistor in the output toroid of this Mad Dog PSU as well - SF probably knew that output toroid will be one of the hottest-running parts in the PSU at high loads.
Of course, if you have a type-K thermocouple and don't mind playing with the PSU a little while under load, you can see if that really is the hottest part or not.

Originally Posted by Pentium4

The other annoying part is I don't know how it will react to a shottky. I only have up to 20A ultra fast parts. I have plenty of brand new 30A schottkys but sometimes half bridge units don't like them. So I have to take apart the stupid thing, put one in and test it. Worst scenario, I will put in a 20A ultra fast and just use it with a lower TDP CPU.

Should be OK if you use a 100V (minimum) rated rectifier. 60V reverse voltage may not be enough, though. There are also 30 Amp ultra-fast rectifiers if the Schottkies don't work.
But pulling 30 Amps through an ultra-fast... oh boy, things will certainly get toasty on that secondary heatsink.

Originally Posted by Pentium4

You're probably right, it's most likely on the 3.3V. The leads are oddly separated. I won't be able to tell which rail its on until I take it out and trace it. I would love to just remove it. You think it would go haywire if I fully removed it? I could also just reuse those heatshrunk leads to put in a 50Ω or 100Ω resistor there.

Hmmm... not sure how I missed it the first time, but I do see it now - it's that big resistor attached on the secondary heatsink, isn't it? If so, it might actually really be on the 5V rail then. TDP-wise, it makes sense: 3.3V on a 10 Ohm resistor will make it dissipate about 1 Watt... so they could have gotten away with using a small 2-Watt resistor on the PCB. Since this isn't the case here and that mounted resistor looks like a 5 Watt part, then it might be on the 5V rail... in which case, it would be dissipating (5^2)/10=2.5 Watts. Certainly not on the 12V rail, though, because 10 Ohms on 12V is 14.4 Watts P_d!

Oh, and you *can* tell on which rail this resistor is attached to - simply use your multimeter and probe the resistance to ground on the 3.3V, 5V, and 12V rails. That should tell you if there are any other dummy load resistors and approximately what values.

Since that big resistor is mounted on the secondary heatsink and looks overall decently far away from the output caps, you probably can/should leave it alone. There's a chance the PSU may not like the 5V rail lightly-loaded if you use it on a 12V-heavy system (though from what I have read in the past, SuperFlower has always had one of the best cross-load performing PSUs.)

Originally Posted by Topcat

Nail, meet head! I really miss the 'golden era' of PC; from ~1995 through maybe 2010. Of course that may be subjective & debatable by some....but it was IMHO a period where the PC really got into the mainstream markets and all the glory & neat things that arose during those times....much like the 'golden eras' of radio & television.... Today, they're just another PC; quite unexciting & stale...the 'wild west' days of the PC & internet are a thing of the past....

LOL, are our minds synchronized or something? I started typing my post just about 40-ish minutes ago with the same idea, as you can see with one of the quotes I replied to above.
This is both creepy and funny, I suppose.

[goodpsusearch]

Originally Posted by momaka

Oof!
Yeah, that's very underwhelming, considering the label claims 30 Amps on the 12V rail.
And it's even more of a shame, because the primary side is perfectly adequate for the task: 1000 uF input caps, TO-3P switching BJTs, and ERL-39 transformer.

Might be 2x 16A rectifiers in parallel, seen that many times on Jou Jye psus and others and the label rated for 30A. Maybe pentium4 need to check the other side of the heatsink for a second 16A rectifier

[Per Hansson]

Originally Posted by goodpsusearch

Might be 2x 16A rectifiers in parallel, seen that many times on Jou Jye psus and others and the label rated for 30A. Maybe pentium4 need to check the other side of the heatsink for a second 16A rectifier

Seems likely, the PSU did manage a 26A load in this review:
https://web.archive.org/web/20050416...scps/print.php

[Topcat]

Originally Posted by momaka

LOL, are our minds synchronized or something? I started typing my post just about 40-ish minutes ago with the same idea, as you can see with one of the quotes I replied to above.
This is both creepy and funny, I suppose.

I believe they are when it comes to this stuff apparently!! Strange, nah....more along the lines of great minds think alike!

Just as an example of the difference in era from then till now, look at the packaging this PSU came in versus today.....

[PeteS in CA]

Just as active PFC has trickled down from late 1980s mainframes and minis into higher end desktop computers (ditto ASIC PFC-PWM combos of the 1990s), I think there is a good chance that microcontrollers and phase modulation could trickle down (are trickling down?) into higher end desktops, with all the fun that will come with PSU manufacturer programmed microcontrollers.

The days of "simple" TL494-clone and 384x based PSUs are numbered (especially the low switching frequency voltage-mode TL494-clones).