My Strange experiences with Fans and my first post :D

~El~Jefe~ · Post by **~El~Jefe~** » Mon Feb 28, 2005 4:52 pm

I never posted before, so this is kinda cool. However, I have read at least a thousand entries on here and some on other sites as well. Been a lurker i guess. Now I have something to say!!

My Rig:
754 3000+ chip with cool and quiet on
Zalman al/cu for the CPU (not the hugish new one, last years model)
1 Gig chip
Spinpoint 80, 160 Gig sata (samsung)
400W Seasonic Tornado (have to pop it open tonight to find version if its even on it)
AIWonder 9800 Pro with the blue, latest sandwich plate Zalman and a 5v 92mm zalman fan on it. (note that AIWonder and 256 pro have faster clocks on the ram than 9800 pro non-AIW, but I have sinks on them anyways)
Case: Coolermaster Cavalier, with the big blue dial on it that never worked.
Pabst 120 on the exhaust, at 5v.
the cpu fan is always on max low except when i play half-life2 and have the speakers on ten as i shoot people.

I told all of this in annoying detail to explain my system. Its a good system by most quiet standards, being very powerful for gaming yet not producing a lot of heat out the back.

Here are some things that are freaky. Initially I had the pabst at 12volts full, for like a few months, wasnt that loud actually, a really smooth fan!! (recommend it over nexus 120 fan, nexus is really a bad cfm mover for a higher end system)

Then, I hooked it up finally, with some nervousness to the zalman big blue controller, which is the ultimate fan controller in my opinion. I noticed that system temps at like 7~ volts were the same, so I dropped it down to 5volts for a few hours. My house is warm, about 76 degrees F. I noticed that my Tornado spun slower and slower the more I turned down the single Pabst in the back. I keep it on 5volts now all the time, my PSU spins up less now. BTW, as I am posting in the PSU forum, Seasonic tornado isnt really quiet unless your landlord shuts your heat off or your parents refuse to put the temperature in the house to something comfortable. it likes to whine up and down with the slightest push, more anoying than a slightly louder yet more even sound.

So, I basically think I have proven that more air sometimes means warmer temps

I only think that it works if you push more than pulling, something mentioned once as being more efficient, I know, but not mentioned as to what I found, that more air sucking out from next to the psu causes the psu to not be able to remove its own heat as well.

JanW · Post by **JanW** » Tue Mar 01, 2005 3:02 am

Welcome to SPCR!!

Sounds like to have a RevA2 Seasonic (I have them, too. Really annoying), and a case with restricted air intake. I noticed that my SuperTornados don't like working against even moderate backpressure at all.

~El~Jefe~ · Post by **~El~Jefe~** » Tue Mar 01, 2005 7:55 am

JanW wrote:Welcome to SPCR!!

Sounds like to have a RevA2 Seasonic (I have them, too. Really annoying), and a case with restricted air intake. I noticed that my SuperTornados don't like working against even moderate backpressure at all.

wow well you kinda figured out all my problems thanx. i guess i need to ditch this power supply i stil have the box. also, i probably need some extra holes in my case, somwhere.

JanW · Post by **JanW** » Tue Mar 01, 2005 10:27 am

Either that, or you can mod it (the PSU), if you're so inclined. That's what I did. And opening up the air intake if it is restricted will help with any PSU, and it will allow you to run fans slower.

StarfishChris · Post by **StarfishChris** » Tue Mar 01, 2005 11:12 am

I remember reading about the SPCR PSU test box having an exhaust fan added - for the Fortron reviewed, at least, the results showed that fan speed & volume rose with a back exhaust - clearly the extra pressure isn't good for bottom-fan PSUs. Turning the fan round to be an intake was similar to having none at all, but the best option depends on the case (side ducts, filters etc). No harm trying it out - a bit of unheated air for the PSU may even prove helpful.

~El~Jefe~ · Post by **~El~Jefe~** » Tue Mar 01, 2005 7:24 pm

StarfishChris wrote:I remember reading about the SPCR PSU test box having an exhaust fan added - for the Fortron reviewed, at least, the results showed that fan speed & volume rose with a back exhaust - clearly the extra pressure isn't good for bottom-fan PSUs. Turning the fan round to be an intake was similar to having none at all, but the best option depends on the case (side ducts, filters etc). No harm trying it out - a bit of unheated air for the PSU may even prove helpful.

Hm, I am wondering how i can get rid of negative pressure in my case? i dont want to add another fan, that defeats all purposes.

the tornado has a bottom 120 mm fan, nice cooling effect just hums up and down though with a small change, which is annoying. The psu is really really cool for a psu on a higher end setup.

JanW · Post by **JanW** » Wed Mar 02, 2005 1:40 am

Ok, I might have to take back what I said. I wasn't familiar with your case and just had a look at coolermaster's website. It looks like there are lots of air intakes on the back and sidepanel. Front air intakes can't really be seen well on the website. But unless you taped those intakes up, this does not really seem to be a restricted airflow case.

It looks to me like you have three options: Since turning the case fan voltage down so far had only benefits (lower noise without higher temps), the logical thing to do would be to continue in that direction until benefits in noise level off and/or temperatures start to rise too much. Of course depending on how far you can undervolt that Papst fan, it might mean going to a Nexus

. This might not get rid of the erratic fan controller behaviour, though. Your options are then modding the PSU (putting fan on a fixed voltage, will void warranty) or getting a new PSU.

I swapped the fan for a Nexus and put it on a fixed voltage. This is how I chose that voltage (warning, calculations ahead!):

Ballparked the typical DC power draw of my system under load. Was 90W for my fairly low power system.
Went to the review of the Seasonic SuperTornado, Table B, to figure out how much heat is generated inside the PSU (I'd assume the 300W and 400W SuperTornado models, RevA1 or RevA2 to have the same efficiency): 27W for me.
Took that number and went to the review of the RevA3. The RevA3 is more efficient, so the same internal loss will correspond to a higher DC power, somewhere between 90W (=23W loss, 4.4V to the fan) and 150W(=36W loss, 5.25V to the fan) numbers in that table. Figured out (by linear interpolation) what voltage the good fan controller would output at this internal loss: U = 4.4V + (5.25V-4.4V) * (27W-23W)/(36W-23W) = 4.66V
Next I needed to know what airflow this corresponds to (this is necessary, even if you don't swap fans, as RevA3 and RevA2 don't have the same fan): The NX3500 Review states that at 12V the D12SM-12 moves 56 CFM. Assuming a linear CFM-to-voltage relation, which seems about accurate for these (don't remember the link, I think MikeC measured it, at least for the low speed YateLoon), I got: F = 56 CFM * (4.66V/12V) = 21.7 CFM.
Since I swapped the Fan for a Nexus (spec'ed for 37CFM at 12V), I had to run that fan at: U = 12V * (21.7CFM/37CFM) = 7.0V. How convenient.

~El~Jefe~ · Post by **~El~Jefe~** » Wed Mar 02, 2005 2:45 pm

Hm, I can't really figure on this case of mine. i see a slot on the bottom in the front, but the slot is really kinda small. some coolermaster cheaper ones have more front grill going on, this has a big grill inside it but the port for it is a slit on the bottom, kinda small. I have one, bottom pci slot open, i thought this would be good, but I am considering doing a cut job somewhere else on the case.