philscomputerlab wrote:

kithylin wrote:

Please show me your dual power supply adapter thing... I'm interested a lot in that.

When I looked for it, I expected something really complicated, mini nuclear reactor style. But turns out it's just a little gadget 🤣

Here is a eBay article link: http://www.ebay.com/itm/24-Pin-Main-Second-Du … =item43c8165319

OH! That... yeah. That thing it's self doesn't actually distribute load. It just makes the second one switch on when the main one does / at the same time. Runs the ATX-POWER-ON signal off to a second atx plug that plugs in to the second one. Useful still. Thanks! I might actually need this as I'm considering some plans for my i7 in the near future to require 1000-watt + 750 watt double units, and found a double-psu case for it.

Well, isn't that what you want?

Drive the motherboard with one PSU, and the graphics card, hard drive and optical with the other one. And there you have your load distributed.

Why would you need 1.75 kW of power for any i7-based system? If you've got that much power to burn, I'd be going with one of the Xeons with 14 to 18 cores.

philscomputerlab wrote:

Well, isn't that what you want?

When I've read about "Dual PSU adapter. Allows using two Power supplies to spread the load" at first place, I've thinking about module, that can drive modern mobo from two cheap psu. Or able to hotswap psu in cheap home server. :) Unfortunately this is not so easy, as just connector to simultaneously power two psu from one mobo.

Lukeno94 wrote:

Why would you need 1.75 kW of power for any i7-based system? If you've got that much power to burn, I'd be going with one of the Xeons with 14 to 18 cores.

Who said it's for an i7? 🤣

philscomputerlab wrote:

Lukeno94 wrote:

Why would you need 1.75 kW of power for any i7-based system? If you've got that much power to burn, I'd be going with one of the Xeons with 14 to 18 cores.

Who said it's for an i7? 🤣

kithylin did.

Also, easy_john, I hope by "cheap" you're not referring to those junk $20 or so units, with no certification, no regulation and that can't produced anywhere near the kind of power they claim to be able to do.

Ah, yes, I see it now.

Lukeno94 wrote:

Why would you need 1.75 kW of power for any i7-based system? If you've got that much power to burn, I'd be going with one of the Xeons with 14 to 18 cores.

Those new ones with those cores are multiplier locked, can't overclock, and thus no good for gaming.

And I'm not intending to actually use all 1.75 KW, power supplies are most optimal on efficiency when they're in the center of their load. I was thinking, some day, something like 3 x gtx-770's in a system and try to bring back some of the 15,000 rpm sas drives I used to have back in to here. Even with all of that I was estimating total actual at-outlet draw only around 1KW - 1.2KW, and that presents a choice. Buy a $400+ 1.4KW - 1.5KW power supply to handle it all in one, or use two cheaper ones and use that for less than half the price. (which I already own the 1KW one, and just yesterday got the 750 working for free).

Besides that, with big platter drives it's more for providing amperage than it is actual power-draw-at-outlet, it's weird to explain. Like my NAS server requires a 750 watt power supply for it's hard drives, but only actually pulls 95-115 watts @ outlet on the input side.

Just because we have big wattage power supplies, it's not always that we're actually going to pull that much power, I'm learning some systems need bigger ones even if the power pull @ outlet is low.

Roland SC-55, under the Focusrite. Never had one before.

kithylin wrote:

power supplies are most optimal on efficiency when they're in the center of their load

This seems like one of those statements that are way too general that no one should believe at face value. Just looking on google for atx power supply efficiency graphs shows that most of them only have 3 data points at 20%, either 50% or some other peak, and 100% load. Only 3 data points are bound to make some kind of artificial peak so that could be where this statement comes from but that's just speculation on my part. Other graphs with more data points usually show a much larger flatter range of load with good efficiency but every powersupply is going to be different and if you need to know how yours performs then you need to test it yourself.

Formulator wrote:

Roland SC-55, under the Focusrite. Never had one before.

IMG_0197.JPG

an excellent choice my good Sir, enjoy! 😎

TeddyTheBear wrote:

kithylin wrote:

power supplies are most optimal on efficiency when they're in the center of their load

This seems like one of those statements that are way too general that no one should believe at face value. Just looking on google for atx power supply efficiency graphs shows that most of them only have 3 data points at 20%, either 50% or some other peak, and 100% load. Only 3 data points are bound to make some kind of artificial peak so that could be where this statement comes from but that's just speculation on my part. Other graphs with more data points usually show a much larger flatter range of load with good efficiency but every powersupply is going to be different and if you need to know how yours performs then you need to test it yourself.

The reason for those points is simple; that's what the 80+ standards are based on. 😀

I bought these random CPUs for ~6 euro.

Im hoping for a PIII 866(133) or 650(100) and a 533(133) as I lack Coppermine CPUs with 6.5 and 4 multiplier which creates holes in my P6 scaling project.

I wish CPUs were so cheap here... 😒

Anyway, some new stuff:

Radeon X1900 XTX

Radeon HD 2900 XT

Intel S5000PSL - dual socket 771 board supporting all CPUs starting with netburst all the way up to 45nm quad-core Xeon (FSB 1333). At the moment I'm testing it using single Xeon L5410 (4x 2,33 GHz, 12MB L2) and 8GB DDR2 FB-DIMM (quad channel) 😈 Working perfectly and very power efficient compared to Socket F K8 Opterons.

havli wrote:

I wish CPUs were so cheap here... :blah: […]
Show full quote

I wish CPUs were so cheap here... 😒

Anyway, some new stuff:

Radeon X1900 XTX

Radeon HD 2900 XT

Intel S5000PSL - dual socket 771 board supporting all CPUs starting with netburst all the way up to 45nm quad-core Xeon (FSB 1333). At the moment I'm testing it using single Xeon L5410 (4x 2,33 GHz, 12MB L2) and 8GB DDR2 FB-DIMM (quad channel) 😈 Working perfectly and very power efficient compared to Socket F K8 Opterons.

Awesome stuff.

If you get two X5470 that board will last you a long time performance wise. They are still somewhat expensive though because of the S775/S771 mod.

I'll start with two Xeon 5080 - the ultimate NetBurst king. And later some quads as well, but I prefer more cost effective models. Maybe even older 65nm Xeons - X5355 is for sale at local auction, very good price at the moment.

Work asked me to recycle some old computers 🤣

This will be ongoing / slow process, so I Can my time, test it all, chuck the non-working gear and hoard the good stuff 😀

Intel G41 chipset board with E4500 CPU and two 1 GB sticks of DDR2 (removed). Also got a SATA optical drives. HDD I still need to sort out a process (wiping) that keeps the brass happy.

35A on the 5V and also -5V old school PSU:

philscomputerlab wrote:

35A on the 5V and also -5V old school PSU: […]
Show full quote

35A on the 5V and also -5V old school PSU:

I assume the MPT-4012 is a 12v design and the one under it is a 5v design?

Yes, very likely. Tried with my Athlon XP 3200+ machine, and the voltage readings are much improved. Around 4.7 - 4.8 under full load.

What I don't understand is, why color coding is different while both are ATX power supplies (although apparently different ATX revisions).

+12V : Yellow vs. Yellow/Black
P.G. : Gray vs. Orange
P.S. on : Green vs. Gray
5V sb : Purple vs. Orange

Messed up sticker?