I've been working with a Fortron FSP300 (non-PFC) PSU and I've been trying to figure out how much voltage it feeds to the 120mm fan (why, when, and how much.)

I tapped into the fan wires and ran a lead to the outside of the PSU, where I hooked up a multimeter to measure the voltage being fed to the fan. I've determined that the fan speed is determined by intake air temperature, more than actual load and having another exhaust fan makes the problem worse. I think this is because of the placement and type of thermistor it uses.

With this in mind, I'd like to modify the PSU to make it always run the same speed, instead of being dynamic. I've noticed that at 4 or 5 volts (initially starts at 4V), the PSU is very quiet.

I'm interested in just unhooking it from the fan controller and hooking it up to 5V from a molex.

Any reason to not do this? Any problems I might encounter? I appreciate any advice on this. I love this PSU and with some systems it runs really cool, slow, and quiet. On other systems (like this AMD dually), it runs hot, fast, and loud;) (Still quieter than MANY power supplies, though.)

The limit I would recommend you watch is that the intake air to the psu remains below 35C, because the capacity of all psu is reduced as their ambient temps rise. e.g. SeaSonic say "The rated power will derate from 100% to 80% from 40oC to 50 oC Linearly"

Which means (especially with a dually) you probably need some additional case ventilation in addition to the psu itself. Once you have this under control then I think there is every reason you can safely drop the psu fan speed: as you say fix it at 5V. That Yate Loon D12SM-12 pushes 23cfm in free air at 5V - even allowing for static pressure to reduce this to say 10-12cfm (and FSP psu are fairly dense) you should be OK. While you are at it removing the grill in front of the fan is also my recommendation - I have done that to both my SeaSonic 120mm psus.

This thread had a similar discussion but in your case (with a dually) managing the case temps is the big deal. Obviously 10-12cfm is not enough case ventilation.

I have a 120mm exhaust fan, in addition to the PSU, so my case temps stay around 30-35C (according to the motherboard sensor, which errs on the high side, I think.)

I have a indoor/outdoor thermometer that I can jam into the PSU exhaust grill to measure PSU exhaust air temps. Any idea what temps should I be shooting for?

Here's some data:

With the case side off, the intake and exhaust fans off, the PSU fan runs 7.19V and the PSU exhaust air measures 35.1C.

With the case side on, the intake and exhaust fans still off, the PSU exhaust fan runs 8.79V and the PSU exhaust air measures 37.2C.

With the case side on, the exhaust fan running it's usual 6V and the intake fan running it's usual 7V (for hard drive cooling) the PSU fan runs 9.24V and the PSU exhaust air temp is 38.8C.

Ok, now with the PSU fan modded for 5V:

With the case side off, the intake and exhaust fans off, the PSU exhaust air measures 36.7C.

With the case side on, the intake and exhaust fans still off, the PSU exhaust air measures 43.9C.

With the case side on, the exhaust fan running it's usual 6V and the intake fan running it's usual 7V (for hard drive cooling) the PSU exhaust air temp is 47.8C.

What do you think about these temps, as compared to the old temps?

Look OK to me.

Personally I am not sure the psu exhaust temps are something to bother with - they are a direct function of the intake temp, cfm through the psu and heat being disipated in the psu. To me 50 or even 60 is OK (although 40 feels more comfortable). As long as the intake temp is below 40 (and preferably a lot below 40) the exhaust is academic IMHO.

Both my cases are exhaust only (now): the hdds hang at the bottom front with big open intakes (but no fans) and keep plenty cool enough as I tend to block most other case holes - especially on the sides and rear by the PCI cards. I mention this as your last psu exhaust temp (fan 5V, both other fans running, case closed) seems significantly higher than other configs. I suspect the psu fan is struggling with static pressure and not getting as much air as it would like (compared to case open).

Would be interested to see if you could run with no intake fan, and 5V the other case exhaust as well. (Should reduce static pressure on psu fan, drop psu exhaust temp and as long as the main case intake is by the hdd, their temps should not rise more than 2 or 3 degress.)

My front intake fan isn't really an intake fan, per se'.

I have a 120mm hole in the front of the case and I have an 80mm fan suspended in this hole, using elastic cord. The fan is only there to blow cool air across the hard drives. I rely on the exhaust fan and PSU fan to draw into the case, passively, for the most part. (through the front 120mm hole.)

My case, just so you know, is a Compucase LX-6A19. It is VERY similar to the Antec SLK3700 series cases. except it has a much better front bezel design. I cut out the 120mm fan grill in front and back.

5V modding the PSU fan did make an audible difference in overall noise level, by the way. The reason I measured PSU exhaust air temp is because that was easy. Ideally, I'd like to measure the temp of the PSU heatsinks. I do have some thermistors I could put on them, but I really don't want to get in there and do it. I'm a little afraid that my thermistors aren't rated to handle the kind of heat that I suspect those PSU heatsinks have and I'm not sure how to attach them and keep them from being overly affected by the PSU intake air temp (as the PSU thermistor is.)

You might want to rig a small baffle seperating the exhaust fan "space" from the PSU fan. I think it is stealing air from the PSU fan. I have the same case. I was using the 350W version of the same PS. Im about to get a 550W version and will put a small baffle in.

I have an Alpha 8045, and use a duct from the front to the 80mm fan, pulling fresh air in and blowing it on the CPU HSink.

The air temp you showed shows the affect air stealing has on the PS's fan, hence the rise when the 2 case fans are on.

Thought I'd post an update.

I've now modded four FSP300's, total, and I have one in my AMD dually and one in my wife's AMD dually. Both have been running this way for over a week, now, with no problems.

More interesting, however, is two other computers I'm working on. One is a 3.2 GHz P4 system, in a LX-6A19 case, with an Evercool 120mm fan as an exhaust fan. On this computer, I'm using another 5V modded FSP300 and a Zalman 7000A heatsink. This system is extremely quiet. I opted for no intake fan as the PSU and exhaust fan are drawing in enough air through the front to keep the hard drive quite cool. For the heatsink and exhaust fan, I tried to use the Zalman fan controller, but even at 5V, the heatsink and exhaust fan were a little louder than needed. I wanted to try to use Speedfan to further lower the fan speeds, when the CPU was running cool, but Speedfan wasn't able to vary the fan voltage, with the Zalman fan controller, for some reason (has something to do with the way the Zalman fan controller is made, I think.) I made my own fan controller, sort of. I simply daisy chained 9 diodes, together, on a fan extension cable plugged into the CPU fan header. This drops the supplied voltage down to 5.25V, but the RPM wire still works and is able to report RPM's to the board, which is necessary for Speedfan to do its work. Now, the heatsink and exhaust fan run 5.25V, when the CPU is hot, but run 3.75V when the CPU is below 50C (which is 99% of the time.) Overall, I'm really pleased with the low noise of this system, considering it is runnning the hottest CPU Intel makes, and one of the hottest video cards you can get (ATI Radeon 9800 Pro, with Arctic Cooling VGA Silencer.)

On another P4 system, this one a 3.0C, I used another 5V modded Fortron FSP300, but I taped over all the exhaust and intake fan holes on the case, so the PSU is the only exhaust and the only intake is directly in front of the hard drive (no fan, just a hole.) This is working really well, too. I made another diode fan controller to work with the Zalman 7000A heatsink. This motherboard is an Asus PC800 Deluxe, which is a BIOS function which slows the CPU fan down when the CPU is cool. It drops the voltage down to about 3.5-3.75V. This system is amazingly different from when I got it. The owner sent it to me to make it quiet. It is a huge, ugly Aluminum Thermaltake case with 7 case fans and sounded like a vacuum cleaner! Now, it has 0 case fans, a different heatsink and PSU, and it is very quiet.

Sounds like you are having fun. Be careful you aren't taking things too low for the boxen you cant monitor. When summer comes there may be surprises, also if the use changes and the owner starts Folding 24/7!

Also while FSP are damn fine psu, they are 'weighted' towards the +5V instead of the +12V rail. The big P4Cs with 9800 vid have all their load on the +12V and with 2 or 3 hdd, 2 optical will be close to the limit. (And OK, so they typically run good overload as well.) The psu heatsinks shouldn't be getting much above 60/70C from my experience measuring at the top of the fins: my indoor/outdoors max at 70. The transistors at the bottom of the heatsink are no doubt a lot hotter but they are rated well above 100C anyway.

Speedfan doesn't use the rpm sense except for reporting: I have run it fine on several 2 wire fans like the stock 'flo L1A etc.

Really? I tried it with Panaflo L1A's on a different board and it would do anything for me. Maybe it was just that board, or the way I had it configured.

In my experience with the Zalman fan controller, though, the fan controller somehow defeats Speedfan. [EDIT] The Zalman fan controller also defeated Asus Q-fan control on a P4C800 Deluxe motherboard. [/EDIT]

By the way, I measured the intake and exhaust air temp on the 5V modded FSP300 with the P4 3.2C. The intake air temp is about 37-38C at full load and the exhaust air temp is about 47-50C at full load. Those temps seems ok, to me. What do you think?

I've got another update for this thread.

I just bought two FSP350 PSU's from Newegg. This PSU has a fan controller on the back, but this controller is still connected to a temperature controller. You can find Mike C's review here.

My PSU's are a bit different from the one Mike reviewed, though. For one, mine don't have an LED fan. It is a simple black plastic fan. Secondly, I also noticed that this PSU uses a different fan from the FSP300 and it is different from the fan in Mike C's reviewed PSU, too (5 fan vanes, as opposed to 7, as Mike's sample had and my FSP300's have all had.)

The FSP300 uses a Yate Loon D12SM-12 fan, rated for 56 CFM & 39 dB @ 12V. This fan is failry quiet and transparent at 5V. The FSP350, however, uses a Yate Loon D12SH-12 fan, rated for 78 CFM & 43 dB @ 12V. Logic would dictate that the FSP350 will be louder than the FSP300, if running at the same voltage. The FSP350 does, however, have a different location and mounting for its temperature control circuit, however, so if the PSU is left unmodded, I thought it may be quieter than an unmodded FSP300, if the temperature control circuit works better than the FSP300's temp control circuit.

You can find the Yate Loon fan specs here.

I opened up the FSP350 and did a little splicing. I spliced a wire into the fan leads and ran it out of the PSU, to measure fan voltage. I also hooked up a 5V bypass for the fan, so I can switch it off the onboard controller and directly into 5V, if I want to.

I took my 5V modded FSP300 and the FSP350 (modded so I could measure fan voltage) and I tested their noise level, side by side. I found a couple interesting points.

The FSP300, in stock condition, has a starting voltage of about 4.25 volts. The FSP350, in stock condition, has a starting voltage of about 3.6 volts. The FSP350's fan is faster and louder than the FSP300 and this is immediately noticable, when comparing the two PSU's running the same fan voltage. The FSP350, at 3.6 volts (its lowest voltage level) is very close, in noise level to the FSP300, modded to run at 5V. When they are both run at 5V, the FSP350 is quite noticably louder.

I wanted to see how the temp controller worked on the FSP350, so I threw it in my dual AMD system and started it up. I let the system run for quite a while (at full load, since I fold proteins) and I was dissapointed that the fan sped up to a quite loud level. It is now running at a noisy 8.36 volts with an exhaust air temp of 37.5C. [EDIT] After modding the FSP350 fan for 5V (because it was easy), the exhaust temp only raised a bit. It is now 41.8C, which is still very safe, I believe. I will later do the diode mod to drop the voltage down some more, and then post more findings. The PSU exhaust temp, of course, is only a baseline for comparison. I've found that it is generally proportional to the PSU intake air temp. With this PSU, the exhaust temp is generally about 10C higher than the intake air temp and I am striving to keep the intake temp below 40C, with an exhaust temp not exceeding 50C.[/EDIT] [EDIT 2] I've now modded the PSU fan with two diodes, running series, between the 5V line and the fan. The measure voltage is 3.55V. The PSU exhaust temperature has raised to 45.4C. To get an idea of my case airflow, you can see pictures of my case, here.[/EDIT 2]

Like everyone else, around here, I strive for low noise levels. This PSU is not low noise. It can be modded for low noise, just like the FSP300, but it is a bit more difficult. This PSU fan would need to be modded for around 3.6V, to get the noise level achievable by the FSP300, modded for 5V. The 5V mod is really easy. The 3.6V mod would take a little more work. The easiest way to achieve about 3.6V, would be to tap into the 5V line, then run two IN4001 (or similar) diodes, in series, before going to the fan. This would drop voltage down to about 3.5V (the diodes drop the voltage by about .75V each). This mod wouldn't be terribly hard, but it would mean an additional step, which you wouldn't have to take, if using the FSP300. For more information about using diodes for voltage control, please see this website.

In my experience, so far, I haven't found a single system that wouldn't run rock-solid stable, with a modded FSP300. So far, I have used a 5V modded FSP300 in two AMD duallies (running dual 2GHz Athlon XP's), a 3.0C P4 system with no case airflow except the 5V FSP300, and a 3.2C P4 system with a 9800 Pro and a 120mm Evercool exhaust fan (modded for low voltage.) All these systems are high wattage systems. The P4 systems, although running the hottest and fastest CPU's and video cards, used far less power than either of the AMD duallies, as tested with a Kill-A-Watt meter. Neither of them used more than 150 watts (from the wall), if memory serves. Both AMD duallies use a little over 200 watts.

Hope this thread helps anyone interested in the FSP300 or FSP350.

I was wondering if you could post some pictues because im not too clear on how u did the cutting, because my 350 aurora has 4 wires running from it, probably 2 for the fan and 2 for the leds... anyway, just wondering if your psu has the "noise killer" sticker on it, because mine does.

~EO

Let's think about what a fan controller does for a minute. It takes power straight from the PSU, regulates it to a user setting, and passes it on to the fan. No motherboard needed! And since Speedfan is not connected to the Zalman, well, you see the problem. I suspect that you're using the Zalman's "split" 3-pin extensions which do have a plug for the mobo, but that carries only the RPM signal. But Speedfan's ability to read RPM doesn't imply the ability to change it. Kind of the converse to using Speedfan on an L1A, now that I think about it...

I'm talking about a Zalman fan controller that plugs into the motherboard fan header, then the fan plugs into it. It is NOT plugged into a molex. It is fed power by the motherboard and it reports RPMs to the motherboard. Speedfan can't vary the speed, though, and Asus Q-fan can't vary the speed, either. I'm not sure exactly why, because the Intel motherboard that ARM systems uses is able to lower fan speed, through a Zalman fan controller, according to Mike C's review.

[EDIT] Here is a picture of the Zalman FAN MATE I'm using. [/EDIT]

My own voltage controller, built with diodes, in series, was able to drop the voltage going to the fan and Speedfan and Asus Q-fan were able to vary the fan speed with my controller.

Sorry, Chris, I made a bad assumption there. I call the single-channel unit the Fan Mate and think of the ZM-MFC1 when I hear "fan controller." As for why it's not working, I can't help you.

some fans just stop reporting rpm when undervolted... one blower fan that i had on my cpu stopped reporting rpm with anything under 12v! Probably something to do with the circuitry of the fan or something... anyway, pix? and could your respond to my post?

~EO

ps: asus q-fan kinda sux

Generally, it is the fan monitoring circuits in the motherboards that can't track low rpm. Those engineers are probably told that 1000 rpm min speed is all they need to design for.

cmcquistion --

Since you're using 120mm Fortrons all the time, why don't you try one upside down with a top CDbay fresh air intake duct? As per Lilla's post. As long as you have a case that gives ~1" or more space above the PSU, you should not hear turbulence if the fan is spinning at under 5V...

Just a thought.

My Panaflo 80L modded PSU systems all feature PSU intake ducts and none of these PSUs run hot or ramp up in fan speed.

Sorry, hadn't noticed your earlier post.

My FSP350 does not have an LED fan. It is just a plain-jane black plastic fan, so it only has two leads. I can't help you there, except to suggest you cut one of the red wires, then fire up the PSU. If the fan spins, but the LEDs aren't lit, then the wire you cut is for the LEDs. If the LEDs light up, but the fan doesn't spin, then the wire you cut is for the fan.

I'm afraid I don't own a decent camera, so I don't have any pictures. If I think of it, later this week, I'll take some pictures, when I mod the other FSP350, sitting in my office (if I can borrow a camera).

P.S. I've always had really good experience with Q-fan, but my experience is limited to the A7N8X Deluxe and the P4C800 Deluxe. On those two systems, it worked really well, as long as I was using a fan with RPM reporting (and not using the Zalman Fan Mate.)

That is a very good idea, but I kinda rely on the PSU fan for case airflow. I've been building several systems, lately where the FSP300's 120mm fan @ 5V, is the only case fan. The results have been remarkable good. I even built a system for a P4 3.0C and Geforce FX5600 Ultra, in this configuration. The temps were great and the noise was very low.

On my AMD duallies, I rely on the PSU fan for additional case airflow, too. I think it makes a positive difference in CPU temps.

cmcquistion -- I guess if the PSU is not subject a lot of heat, then the duct is not worth trying. I just have a sneaking suspicion that -- all things being equal -- those bottom feeders (the PSUs with a fan facing the CPU) cannot survive that long with much heat running through them.

Your P4 3.0C system in an average config will probably draw ~200W AC max. (The recently reviewed ARM Systems StealthPC P4-3.2 draws ~200W running [email protected], which is about 10-15% less than the max possible.) If the 120mm PSU fan is the only exhaust in the system, then ALL 200W of this heat is passing through the PSU.

In contrast, in a 80mm fan PSU intake ducted system with a rear case fan at ~6V, the PSU passes little more than the heat that it generates -- ie, for the system above, if it's 70% efficiency, then it would generate & pass ~60W of heat. Well, maybe a bit more because even with a duct, some of that heat rises up... so call it 80W. The remainder is dissipated via the case fan.

Now if the 80mm fan and the 120mm fan have the SAME airflow, then the former would definitely last longer. BUT they don't have the same CFM, so it gets a more complex. A typical 120mm fan at 5V is probably still pushing close to 20 cfm while a Panaflo or similar at 5~6V is at ~10 cfm. Still your the 120mm fan PSU has to move more than double the heat.

I just wonder if it is not possible, with a duct + rear case fan, to have the same or even lower noise while keeping the PSU running cooler, that's all. It is generally true that even 2-3 Panaflos at 6V are quieter than a single noisier fan.

I think you definitely have a good case.

With the PSU flipped upside down and ducted, the fan could be run slower. An additional exhaust fan would be needed, but it could probably be run slow and quiet.

The net effect may be quieter, but it would be more work, and it would require cutting a hole in the top of the case, which I'm a little hesitant about, since my hole cutting skills are pretty poor.

Definitely worth considering, though, and I may try it some day.

cmcquistion --

No hole in top of case, that's potentially a direct path to users ears! Just need a case that gives you 1" clearance above the PSU -- and then create a duct to front top 1 or 2 CD drive bays which remain empty. PSU fan simply draws air through the front panel CD drive opening(S), through duct, then through the PSU & out.

Oh... I misunderstood what you meant, before.

In that case, I may try this a little sooner.

[EDIT] I just took a look at my two favorite and most widely used cases, the Evercase 4252 and the Compucase LX-6A19. Neither case has much of any room above the power supply, so I don't think I could do this in one of those cases, unless using a typical 80mm PSU. [/EDIT]

maybe you could also take the fan outside of the PSU, to reduce static pressure?

I did it like this with my Seasonic SS-300FT. I replaced the Yate Loon fan with a Adda 120mm "16dB" 1000rpm fan... a bit noisy at 12V so Im now running it at ~7V with a Zalman Fanmate.
http://aegd.cjb.net/dexton/osorterat/Se ... 300-ft.jpg

Since the fan tended to vibrate some, it was fastened with these rubber things:
http://www.gtek.se/img_prod/efs.jpg

Im using this PSU as only exhaust, Ill see if it will work during the summer also

// Johan

I thought about doing that, but I wasn't sure if it would actually make any difference in noise levels.

Have you noticed a definite difference?

Do you guys think an FSP300-60PN with a Globe 120 mm fan at a fixed 7V input is safe in an airflow restricted case such as the Antec Sonata? (my idle temps run about 30 - 31 C during the summer. With Prime95, peak case tempts are ~ 34 - 35 C).

I've run nearly that exact setup with a Globe fan at 5V, so yes, that's safe.

Hi cmcquistion:

My only concern is that airflow in my Antec Sonata is much lower than your Compucase LX-6A19.

Also, what type of peak ambient temps do you have during summer months?

I think I posted this in one of your other threads mshan, but I'll put it up again.

I have a Coolmax Taurus 400w. I can only assume this is a Fortron model as its characteristics are pretty much the same and look the same from what I CAN see internally.

This model comes with a switch to adjust from Low/High/Auto on the fan. I have NEVER and I stress NEVER heard this fan ramp up when on Auto. My temps as of right now are 45C Cpu/30C Case and after 2xPrime95 they get up to about 54C/33-34C. Still no apparent noise change. This is also inside an Antec Sonata, only bezel modded and the Antec holes taped shut.

I can only guess that this fan is never seeing any higher than ~6-6.5volts even when on Auto. Depending on what fan you're working with, I assume 5v to be perfectly safe for this Fortron model. My exhaust temps feel warm to the touch, however this PSU (my first experience with a Fortron so someone correct me if I'm wrong) seems to actually use the casing as a heat dissipation device. It gets very warm on the outside. I think this is helping it rather than hurting it. (Again, no fan ramping whatsoever even on Auto)

Hi interitus!

Can anyone else confirm that this is indeed a rebadged Fortron?