Fanless PSU build-up, first phase

[gmat]

My goal: power an uber-PC with a silent (i.e., not quiet) PSU. That means, no fan, no coil noise.
Options:
1 - Buy a 'standard' PSU, and mod it, with huge passive heatsinks, and get rid of the case.
2 - The same standard PSU, but this time watercooled.
3 - Buy a real fanless PSU

Let's consider each option pros and cons
1: standard, passive
(+) easy, riskless
(+) no impact on the watercooling loop
(-) bulky, to say the least. The needed passive heatsinks would be huge.
(-) difficult placement, due to convection -> nothing on top of it...
(-) heavy (must loan a truck to go to LANs)
(-) heatsinks not easy to find (ok i already have them i admit )
(-) only the MOSFETs will be cooled. If anything else heats up, problems can arise.
2: standard, watercooled
(+) good results with only 1 waterblock
(+) no heat produced
(+) easy placement
(-) adds heat, and restriction to the watercooling loop
(-) difficult mod - extra care needed in WB mounting, insulation etc.
(-) expensive (need 1 waterblock)
(-) only the MOSFETs will be cooled. If anything else heats up, problems can arise.
3: real fanless
(+) the easy way, nothing to do
(+) PSU designed to work this way - no extra hot componentry (coils, etc)
(+) tight box, easy placement
(+) high quality componentry (to produce less heat)
(-) expensive

Weighting all factors, i decided to save money and opt for the solution (3).
Now who are the contenders:
* TK power 300W -> in USA, $300 + shipping
* Deltatronics 360W -> Germany, cannot order it anymore (??)
* RSG Electronics 300W -> Germany, 210 euro + shipping

Obvously i'll head to the RSG Elec one.
I am in email contact with them, they told me it's available now !

I then looked at the PSU datasheet:
http://www.rsg-electronic.de/en/date...CP300Wvorl.pdf

As this PSU wouldnt handle my complete setup, I'll power the CDRW+CDROM+2 HDDs block with a *separate* PSU.
That means providing them 12V and 5V.
Here are the power requirements:

Device			12v		5v		comments
Cheetah 10K6	0.98A	0.97A	power idle=9W
IBM 18ES		2A		0.75A	power idle=6.9W
UltraPlex		1.7A	0.8A
Yamaha			(1A?)	(0.8A?)	writing=11W

i can also power the rad fan:
PAPST 4412 N/2GL -> 1.25W / 12V (0.11A)

counting 11W for the yamaha on both 12V and 5V, and taking a security margin, i have:
12V: 67.4W -> ~70W
5V: 23.75W -> ~25W

So, again at RSG Elec we have:
* for the 12V, the LEA100F - it can endure the current rush at HDD startup and has active PFC.
* for the 5V, either the LDA30F or the S40

Here's a list of prices (per 1 unit) they sent me:

RCP300W		1pc.	EUR 209,50	approx. 4-5 weeks
LEA100F-12-RS	1pc.	EUR   94,98	approx. 3-4 weeks
LDA30F-5-S		1pc.	EUR   49,99	approx. 3-4 weeks
S40-5			1pc.	EUR   29,98	approx. 3-4 weeks

the '-R' means 'remote start', and the 'S' means 'with enclosure'
The S40 is more bulky than the LDA30F. The specs are there:
http://www.rsg-electronic.de/en/datenblatt/s40.pdf
http://www.rsg-electronic.de/en/datenblatt/lda30f.pdf
The LDA is more protected, and has greater precision. The S40 has 10% tolerance which is too much to my taste.
I'll then take the RPC + LDA30F + LEA100F total=355 euro for 430W of fanless PSU ! Not so bad, compared to the TK Power 300W ($300 + shipping from the US..)

(i'll check the price for the LEA75F-12-RS to see if i can save a few euro)

I'll order the stuff in December if i have some money left. Either way i'll get nothing before 2003... I'll keep you updated.

[bigben2k]

For that kind of money, I'd watercool a PSU, but hey, that's just me

[gmat]

BB2K, watercooling a PSU is a difficult challenge, involves a custom waterblock or at least a HDD waterblock (=expensive), and runs the PSU with zero airflow. Right, the MOSFETs will be watercooled but other components may get hot as they're not designed to work this way.
Net result = same money, more risky, less reliable. And it will add another WB requiring maybe a 2nd heatercore...

[gogo]

I was going to (read I did, then freaked out) do something similar.
The computer I did it to is as follows:
Aopen MB (can't recall model, & I'm not at home)
Maxtor 40 gig (I think its a 640x or something)
Creative 52x carom
Celeron 1.0a
Zalman flower cooler
compusa no name 250Wpsu

anyhow, I noticed that the metal casing on the psu was never all that warm, most times it was cool to the touch. (ambient of house between 23 - 27C depending on where and when) I decided to see if it really needed the fan. So I took the fan out. let it sit for about 10 min and thought, hmm, maybe I should be measuring the temps instead of just watching it. I had my wife drive out to the store and pick up a thermometer. (I being in no condition to drive. Which is probably why I attempted said project without one in the first place. Hind sight is 20/20 or in my case 15/35 ) In any event, played with the probe in numerous positions letting the computer do what it normally does and serve me files and occ. browse the web. It never got above 43C in any spot. I let it run over night having set the alarm to go off it hit 48C. (When I woke up I think it was down around 30C, it's been awhile) Being sober at this time (that morning) I decided I should do it all again stressing the system.

I turned it off as soon as it hit 50. It did take a while though. And none of the other components really seamed too much warmer.

I hooked a fan (vantec stealth) to a 5v line and then ran the stress tests again. never got higher then 40C. and usually is nowhere near it. It's usually about 2 - 3 degrees over room temperature. but then again it usually just stores files and occasionally surfs the web. I have the fan set to go on when the temp on the Mosfet hits 40 and it seems to shut off when it hits 35.

The other components even without the fan didn't seem all that much warmer, but I think with any more wattage, or prolonged exposure they would definitely be a concern.

Personally a leak in the PSU doesn't sound good to me. No leaks sound good, but that one sounds really bad, and it would be a PIA to check for leaks.

[gmat]

I'll answer first to PSU watercooling:
It's not an easy mod, but leaks are not the problem. Use a HDD waterblock and have the barbs stick out of the PSU frame. It's the way BladeRunner did it:
http://www.zerofanzone.co.uk
That way, no leak problem, and it's easy to check for them.

Now for the fanless operation:
First i must admit you've got ballz. Being drunk may help.. Anyway your setup is typically low power, i know laptops that draw more current ! No wonder your PSU does not produce any heat...
My storage subsytem draws about 100W by itself, my OC'ed tbird draws around 100W as well.. I've got a GF2 GTS which draws lots of current, an adaptec 29160 (heavy hitter)... My 300W PSU is running at its limits (and since it's getting old, the 'limit' is lowering dangerously). With a powerful fan (Sunon), and thermally controlled, it runs at max speed nearly all the time and spits a very hot air. I wouldnt even dare placing a low-rpm fan in this PSU.

[bigben2k]

Way to go, Gogo!

PSUs with today's system really need a fan, or some kind of cooling solution, because the power draw is ever increasing.

What you observed sounds right. As the power draw increases (with the CPU load), temps will rise. A bigger PSU should help alleviate the damage that this heat causes (i.e. get a fat PSU, it'll run cooler).

[gogo]

hmm, what about just a higher quality PSU to start with?
i.e. pcpower and cooling silencer 400 here
It's allready designed to run cooler, ahh but no fans to suck any heat it would still get a little toasty. I see the problem again. but I wonder how hot the components would get?

Maybe ramsinks attached to the Caps though you would still have the problem of getting the heat out of the PSU in the first place.

Shuttle's heat pipe? mmmm, probably get to warm just for it to even work properly.

[gmat]

Originally posted by bigben2k
(i.e. get a fat PSU, it'll run cooler).

Whoops. That's a common misconception, BB2K.
Heat is produced by inefficiencies, aka losses internal to the PSU. Usual PC switching PSUs have about 70% efficiency. The 30% left are mostly heat.
Higher rater PSUs have simply fatter transistor and lines to sustain heavier current. But as power increases, losses do as well.
So in brief:
- fatter PSU + same setup = same heat
- fatter PSU + heavier setup = more heat
Also, higher wattage PSUs have often lower efficiency (to maintain decent costs)... = even more heat !

The only real way to get low heat from the PSU is to use a low power system such as gogo's one, C3 or Celeron, integrated GFX and sound, 1 IDE drive and 1 CDROM...

(edit) gogo replied at the same time.. well you're right, a "better quality" PSU at equal wattage will produce less heat due to higher efficiency.

[gogo]

oh yeah, I'm sure you read it, but this is what 'inspired' me in my state to take out the fan. silent pc review

I figured I wasn't putting out that much power so it might work.
Had I been sober I may never had attempted it, as it stands I got lucky and learned a little and got a temp sensor.

[gmat]

I'll email RSG to know prices of more products:
PAA75F
SPS-025
and LEB150F - would produce both 5V and 12V.

(edit) yes i'm a SPCR reader

[gogo]

Tk power $195 here

They also do custom.

here

Not sure about over seas though.

well off to work

[bigben2k]

That would be correct, about the heat.

My point (which I didn't state clearly) is that the heat density will be lower in a fatter PSU. (i.e. same heat, but dissipated into more components).

I think...

[gmat]

Then there's the ATX form factor constraint. So any PSU has roughly the same amount of components... The most expensive one being the MOSFETs, doubling em would send the price through heavens. All they do is put bigger MOSFETs (and higher power rated components). So don't take for granted that this 550W PSU will run cold with your system if it was previously powered by a 300W... You're better off with a more efficient 300W to begin with.

[jtroutma]

gmat:

after reading the whole thread and thinking a bit......why dont you just add "beefier" heatsinks to a standard ATX PSU and then replace the fans with Vantec Stealths?

I mean, your 10K HDD is already your loudest component in your entire setup minus your PSU fans. It would just make sence to me to simply increase the surface area of the PSU heat producing sources and simply use a lower rated fan.

Am I completely off my rocker here or is there something I dont know?

Good luck

[gmat]

Originally posted by jtroutma
gmat:

after reading the whole thread and thinking a bit......why dont you just add "beefier" heatsinks to a standard ATX PSU and then replace the fans with Vantec Stealths?

Because those are fans, and make more noise than my PAPST 4412... And this is the only fan left apart from the PSU one.
Besides adding 'beefier' heatsinks would lead me to option (1).

I mean, your 10K HDD is already your loudest component in your entire setup minus your PSU fans.

Wrong, there's the 7.2K HDD before (going to replace that one with a 15K3)
And the HDDs are going to be silenced anyway ,with quite an extreme solution.

It would just make sence to me to simply increase the surface area of the PSU heat producing sources and simply use a lower rated fan.

Read the initial goal (first post, first line)... Fanless is *no fan*, no noise. Even a PAPST or a Vantec makes noise. And yes i can hear it.

Am I completely off my rocker here or is there something I dont know?

Yep. My goal is to supress fans and go 100% watercooled and passive. So adding fans is not an option to suppress fans...

Good luck

Thanx

[gogo]

Yeah I can hear the fans over the Maxtor dive in mine (it's suspended with elastic with a few HS attached) even at 5v. the only time I can't hear the fans is when the furnace kicks on or when the moniter goes to sleep and buzzez. (old KDS14in)

What about completly modding a PSU. Something similar to what pH did.
Maybe something like the koolance Exos, but a PSU.

[jtroutma]

Well then Gmat It looks like you have two options then..

1) spend MAJOR bucks and get the fanless PSU

2) just watercool the SOB <--my choice

(Apparently I must have ear damage becuase I thougth that the Vantec Stealths did a very good job at being "quiet." DAMN them diesel engines!!!! )

[Arcturius]

Have you thought about building your own linear power supply? I know their efficiency is down around 40%, but at least it isn't variable like in a switching supply, and they are simple to build. Also, you would have control over all the components going in, so you could plan how to cool it ahead of time.

Also, even though it would be fairly bulky, it would still probably be less bulky than three other supplies taking up space in your chassis.

Plus, you can make exact-length power runs.

Oh, and did I mention the power is _extremely_ clean?

[gmat]

Yep. Actually i already built linear power supplies when i was younger. But the bulk of a 450W linear PSU would be nowhere near 2 small fanless switched PSU's... Not to mention the weight !
Moreover i don't trust my own soldering / electronics skills enough to build a PSU for a $2500+ PC... And the sum of component prices would be quite high as well.

I'll put my trust (& my money) into those neat RSG Elec boxes. They're small and designed for industrial applications - it means rugged. I think i'll take the 150W 12V+5V one, waiting for the price.

[MadDogMe]

About the argument/opinions that H2o'd PSU's 'other' components will get hot. what about the fanless PSU's that only have a whacking great HS hanging off the back?, they work OK with zero airflow don't they?, have they got 'higher rated' caps', restistors, diodes and whatever?...

Does anyone understand the 'ratings' for components?, I'm wondering if you could replace the existing mosfets with ones that are 'rated' for a much larger 'ampage/current'?, that'll run cooler as they are nowhere near their limit?. do 'other' components have simular ratings but the same 'settings' so to speak?. if not then the 'other components' getting hot is just a myth...

Most components are rated to run very hot anyway are'nt they?, 60DegC plus?...

Has anyone spoken to BladeRunner to see if all's been OK with his PSU since?...

[gmat]

Originally posted by MadDogMe
About the argument/opinions that H2o'd PSU's 'other' components will get hot. what about the fanless PSU's that only have a whacking great HS hanging off the back?, they work OK with zero airflow don't they?, have they got 'higher rated' caps', restistors, diodes and whatever?...

Take a look at those RSG power supplies. No heatsink. Just a different architecture (more stages...)

Does anyone understand the 'ratings' for components?, I'm wondering if you could replace the existing mosfets with ones that are 'rated' for a much larger 'ampage/current'

Well i'm an EE so i think i understand them
Mosfets rated for a higher current simply have a 'broader path'. If you swap a mosfet with a higher wattage one - and drawing the same current of course -, you'll get *exactly* the same heat produced. Why ?
Because it's produced by the difference between input (switched + filtered AC) and the output (DC). The more important the difference, the more heat it will produce. It's not related to the MOSFET itself.
If you want to reduce its heat, you must add stages before. Linear conversion, or multiple filters all help getting a more precise pre-regulated voltage.

if not then the 'other components' getting hot is just a myth...

Most components are rated to run very hot anyway are'nt they?, 60DegC plus?...

Has anyone spoken to BladeRunner to see if all's been OK with his PSU since?...

It's not a myth. Coils,and more frequently power resistors are getting hot. If PSU designers size em thinking they'll be aircooled actively, they don't pick the same ones as if they had to be cooled passively...

Ah try a search on SPCR forums (PSU related one i think) BladeRunner is going fine with his watercooled PSU, although he had to add a HS to a component...

[MadDogMe]

Thanx for the clarification, I know nowt' about electric...

My next question was going to be which components actualy get hot. the resistors and mosfets for sure, but transformer?( that's what BR put the block on as well is'nt it?) coils?, ect. do they get HOT or do they get hot ?...

A quick question on my Enermax 350watt PSU, my 5&12v rails are slowly detiorating, I do have a problem with the ATX connector being a poor~connection, but even if I 'jiggle' it it does'nt improve (it used to), I'm down to 4.76 & 11.98 idle, used to be 4.90 & 12.13~18 when new. the mobo is relativley new but readings were same/simular on my old mobo (as was dodgy ATX connection, but not as bad).

What would cause them 'both' to deteriorate?, I could understand one or the other from a poor ATX connection , but both?. I understand the 12&5v rails are somehow tied together. would the common earth having a poor conection cause this?, or is it the PSU giving up the ghost?, OR is it just a bad ATX block?(gold plated Ennermax jobbie).

I'm thinking of soldering a block connector (3AMP Mains rated) to the ATX pins on the back of the mobo and bypassing the ATX connector. if I bundle the +12v, +5v & +3.5v's together respectivly I can use one ten~joint sized connector instead of a twenty. what do you think?. I've been told the 5v mosfet mobo mod is'nt such a good idea as it bypasses some ferous~ring, volt cleaning?, I've been told at another place that it bypasses the mosfet alltogether!!(by soldering +5v ATX/MOLEX wire to middle pin/cut off stump of mosfets)...

Gee!, was'nt such a quick question after all!! ...

[gmat]

Originally posted by MadDogMe
coils?, ect. do they get HOT or do they get hot ?...

Coils are made of great lengths of wire, huuhhhh.... COILED around a core. Just any wire, even pure copper, has resistance... Basically this makes a transformer an equivalent of a coil + a resistor in series. If the material used for the coil was a supraconductor it wouldnt heat (ask your neighbour particle accelerator). But the smallest supraconductor coil i've seen was helium cooled and tall as a house.
So as a resistor it gets hot, with the good ol' formula P=RI^2. Since all the PSU current passes through one coil (in our small ATX units) thats a lot of heat, so yes they can get *uncomfortably* hot.

A quick question on my Enermax 350watt PSU, my 5&12v rails are slowly detiorating

Life is a b*tch, aint it.. The same here. Even had to reduce the OC.. :/

What would cause them 'both' to deteriorate?, I could understand one or the other from a poor ATX connection , but both?. I understand the 12&5v rails are somehow tied together. would the common earth having a poor conection cause this?, or is it the PSU giving up the ghost?, OR is it just a bad ATX block?(gold plated Ennermax jobbie).

Well firstly Enermax arent known for 'voltage fidelity'. For their price they're even quite imprecise. They're very good at marketing products and make shiny golden plated grilles, but their electronics are just about average. The PSU i got is another 'average' product, and it's been several months since i've seen that.
I think it's because of consumer-level rated components, slowly 'burning in' as we OCers would call that. But 'burning in' for a power component is bad...
The original reason ? Heat, and lower quality components. Guess why you pay only $100 for a 400W PSU, and to have a real linear lab PSU (which wont die or degrade in 2 years) of the same power you have to shell out $800 to $1000...

(mobo mod)
what do you think?

Bypassing the ATX connector by soldering directly another connector to its back pins *could* work, provided that:
- you've got good soldering skills
- the other connector is good enough
- you use shorter, thicker wires
Don't forget shrink wrap and soldering flux.. make a decent work

Now bypassing mobo MOSFETs is a dangerous proposition. Make sure you know where you're going. What you are bypassing. And what you're powering, really. Make sure the components you'll power that way can endure the variations or power surges of your direct-to-consumer mod.
Ferrous cores (with coils around) are there to balance the effect of capacitors. Even common PSU wires going to your mobo have a small capacitance (along with connectors), and it may be a good thing (for CMOS components) to cancel it. Usually the power part of the mobo has an impedance adapter to nullifiy the impedance of PSU wires and connectors.
In the end you may lose hardware for a sum greater than that of a shiny new PSU... You've been warned.

[Axle]

*head spinning* //

WELL. I was thinking real hard just yesterday about what to do with my PSU, and finally decided on WCing it (mainly because I don't have $300 ).

But these MOSFETS, they are LIVE, am I correct? I mean why else would most sinks be live then? So screwing them onto a copper block (another reason I wanna do it- I think i could make quite a decent WB), would make the block live, and therefor the water? Can't be right, in fact I know it isn't, so what's going on here?

I basically want to "pull a bladerunner", with some minor differences. Even if anything else is getting hot, I'm going to cut a hole in the PSU box, and then the top of my case, seeing as heat rises. But I realize this may pose some electromegnetic issues, I'm just not sure what.

All these fanless PSUs seem like a great idea, I just dont have the money.

Oh (feel free to laugh-I'm really not an EE), what about replacing the MOSFETS with Active PFCs? Probably around the supidest thing you've ever heard, but I had to ask.

Thankyou sir!

[bigben2k]

What's a PFC?

I'm looking into building a PSU array, and I have to somehow fit a chip in the PSU that adjusts the output (it also has an output for the array controller). I think it works with the regulator, but I'm not sure...