I just completed a build using a Sunbeam Rheosmart 6 Fan Controller. (I'm about to start overclocking; the assembly is stable.)
I've convinced myself that I'm happy with the Rheosmart for now. There were however some quirks with my ASUS P8P67 Pro motherboard, that may be quite common. I'm posting my setup in hopes I might save others some time, or lead them to a different choice. This is a long read, but not if you're agonizing over this choice or trying to get it to work.
Mountain Mods Plateau 18 case
ASUS P8P67 Pro motherboard
Intel Core i7-2600K cpu
Noctua NH-D14 cooler
G.Skill Ripjaws 16 GB memory
MSI N460GTX Cyclone 1GD5/OC 1GB Graphics Card
9 Scythe S-Flex SFF21E 1200 rpm fans
1 Cooler Master Blade Master PWM 120mm Fan
ASUS DVD burner
Addonics Snap-In Double Drive
Intel SSD drive
Sunbeam Rheosmart 6 Fan Controller
This is a compute server, running Ubuntu, for math computations; the graphics card is for the 336 CUDA cores. Despite my misgivings about the Rheosmart, this build is virtually silent except when working at full load. In particular, only the graphics card fan and one cooler fan runs except when working, and I can't hear either of them over background noise.
I've come to love Mountain Mods cases, despite the cost. They're like renting a large, unfurnished apartment. I resent all the crap everyone else puts in cases, that just gets in my way. I picked a horizontal orientation this time, after seeing what heavy coolers can do to vertical motherboards without additional support. As a bonus, this case is more inviting to work on, either in place or by sliding out the motherboard assembly. The flagship MM cases would be twice as tall, leaving room for water cooling I'm not using. They're clearly aiming the Plateau case at yahoos, as there's no reset button. For now, I cross two wires like I'm stealing a car. The Plateau dimensions are arbitrary, chosen to be consistent with the rest of their product line, and slightly small for extreme cooling: I have to mount my three top fans outside the case to leave room for the Noctua NH-D14 cooler, and if one orders a 3-fan panel for the right side, there's no room for fans to fit inside next to the power supply and bay devices. I guess 10" by 19" by 18" doesn't have the same ring to it as 9" by 18" by 18". I'm quibbling; I've never had this much room in a mass-market case.
The memory is on the motherboard supported lists, and fits under the cooler. Addonics makes great stuff; this 2 x 2.5" trayless drive bay is more convenient than the 4 x 2.5" tray-based drive bay I've used before.
So what's up with the Rheosmart? The basic idea is that the front panel offers six 3-pin standard fan channels, each of which lets one choose between manual control with a dial (green LED), and using voltage control to follow the motherboard's PWM control of one fan (red LED). This makes sense, the motherboard has a better idea of the cpu temperature than any add-on fan controller could possibly have. I've built my share of DIY fan controllers (I'm partial to diode arrays) but I really want the fans to automatically ramp up when the CPU gets hot, and ramp down when the CPU cools off. My ASUS motherboard BIOS offers fine manual control of two 4-pin PWM fan headers, e.g. I'm now running 100% at 70 C CPU temp, 0% at 50 C, scaling in between.
Most of the reviews of the Rheosmart series could have been written without opening the box. They're all on those look-alike 17 page, ad-filled review sites, so no surprise. There are also Amazon and Newegg user reviews along the lines of "I installed it and my fans still work." There's an information gap here.
Motherboards vary in how they modulate fan headers. My ASUS in particular has no voltage modulation, only PWM, and needs to see a real PWM fan to function properly. In particular, the Rheosmart does not provide a sufficiently faithful simulation to fool the motherboard. One symptom (the first symptom I saw) was that the BIOS wouldn't retain custom manual fan control parameters until I attached an actual PWM fan.
The motherboard probably depends on the RPM signal for feedback; the Rheosmart doesn't let you choose which fan should be used for this signal, suggesting that it's not providing any RPM signal to the motherboard. Were it, say, defaulting to fan channel 1 for this signal, a warning to this effect would be nice, as I'm Y-splitting many of my channels, and that always confuses reading the RPM unless you snip one of the two RPM wires.
Sunbeam provides many cables; the most crucial cable splits a 4-pin PWM motherboard fan header into a 4-pin PWM fan out, and a copy of the PWM signals for the Rheosmart to follow. The documentation doesn't assert this, but one needs an actual PWM fan attached to the PWM fan out, for their whole rig to work. Otherwise, in my experience, it just runs full throttle. With the assortment of other cables, one can jury-rig returning the RPM signal from one of the analog fans, while reading the PWM signal from the fan header. At least, one can cable this. But it doesn't work.
After many experiments, I swapped the front 120mm 3-pin Noctua fan that came with the cooler for the Cooler Master Blade Master 4-pin PWM fan, attached it to the PWM splitter cable so the motherboard could directly see it, and the Rheosmart started functioning as advertised. PWM fans click, but I can't hear this one click once it's in place.
The Rheosmart more or less works. I used a voltmeter to check the analog fan output. (Sunbeam provides two 3-pin fan to 4-pin power converter cables, that are the right genders to allow monitoring their output with a voltmeter.) With the motherboard set to 0% (CPU below 50 C) the analog fans see around 1 volt, and don't run. The Cooler Master PWM fan continues to run at around 600 rpm; perhaps it defaults to a minimum speed with no PWM signal, while other fans would stop. In any case, I like this outcome. The Noctua NH-D14 cooler is perhaps overkill for a Sandy Bridge processor, but my rig sure is quiet in this mode.
As soon as one sets the floor PWM to 1% in the BIOS, the analog fans now see 10 to 11 volts, and run only marginally below their nominal 1200 rpm. Not ideal, but one can live with this: The fans come on for computations, turn off when they're done. This could be an artifact of my fan choices; with a different combination of the PWM fan and the 3-pin fans, the 1% voltage might be lower. But the point is, I can't adjust this on the fly.
I corresponded with mCubed over their FanAmp. It amplifies a 3-pin fan signal for many fans, and it can be tuned, unlike the Rheosmart. (Ideally, it takes two controls per channel, but I wish that the Rheosmart would at least have the dial adjust the floor voltage in PWM mode, leaving the ceiling voltage at 12 V. Instead, the dial goes numb in PWM mode.) mCubed claims that one can modify their amp for PWM use by cutting the red wire, and connecting it to the PWM wire. I haven't tried this, but I may.
Or one can go the DIY route. A trivial C-R circuit can smooth the PWM signal to a varying voltage, an op-amp can buffer the fan header from the output current draw, and a power amplifier can boost this to a range to drive all one's fans. A good beginner project. Ideally one has two such circuits, so some fans follow the CPU fan header, and some fans follow the chassis fan header.
What would be very slick would be to have all fans run at exactly the same RPM; there are chips intended for this, which one could smooth individually from PWM to voltage, allowing the use of a mixed collection of 3-pin fans. I wouldn't go to this much trouble for myself, but if I were entering the market in competition with Sunbeam, I'd do this. It probably doesn't matter for case fans, but I have heard out-of-sync push-pull cooler fans beat against each other. And "exactly the same RPM" is a good way to accomplish "roughly the same RPM" when using a mixed collection of fans.
One could also go the all-PWM fan route. The motherboard can't supply enough power, but there are cables that take power from a 4-pin power cable, and the PWM signal from the motherboard, to run many PWM fans.
I'm amazed at how quiet my fan-on-demand build is, using the Rheosmart, even if it was puzzling to get working. I give it a qualified recommendation, until someone does better. This shouldn't be long; adjustable PWM fan control is becoming the norm. There's the usual lag, where the 2600K doesn't need water but we know how to do water, the motherboard doesn't need an independent fan controller but look at all the independent fan controllers on the market, etc. It always pays to forget everything one knows and start over. Eventually various players in the market will do this.
Sunbeam Rheosmart 6 Fan Controller
Noctua NH-D14 120mm & 140mm SSO CPU Cooler
Mountain Mods Plateau 18
mCubed FanAmp Fan Controller
Cooler Master Blade Master PWM 120mm Fan
Addonics Snap-In Double Drive
I've convinced myself that I'm happy with the Rheosmart for now. There were however some quirks with my ASUS P8P67 Pro motherboard, that may be quite common. I'm posting my setup in hopes I might save others some time, or lead them to a different choice. This is a long read, but not if you're agonizing over this choice or trying to get it to work.
Mountain Mods Plateau 18 case
ASUS P8P67 Pro motherboard
Intel Core i7-2600K cpu
Noctua NH-D14 cooler
G.Skill Ripjaws 16 GB memory
MSI N460GTX Cyclone 1GD5/OC 1GB Graphics Card
9 Scythe S-Flex SFF21E 1200 rpm fans
1 Cooler Master Blade Master PWM 120mm Fan
ASUS DVD burner
Addonics Snap-In Double Drive
Intel SSD drive
Sunbeam Rheosmart 6 Fan Controller
This is a compute server, running Ubuntu, for math computations; the graphics card is for the 336 CUDA cores. Despite my misgivings about the Rheosmart, this build is virtually silent except when working at full load. In particular, only the graphics card fan and one cooler fan runs except when working, and I can't hear either of them over background noise.
I've come to love Mountain Mods cases, despite the cost. They're like renting a large, unfurnished apartment. I resent all the crap everyone else puts in cases, that just gets in my way. I picked a horizontal orientation this time, after seeing what heavy coolers can do to vertical motherboards without additional support. As a bonus, this case is more inviting to work on, either in place or by sliding out the motherboard assembly. The flagship MM cases would be twice as tall, leaving room for water cooling I'm not using. They're clearly aiming the Plateau case at yahoos, as there's no reset button. For now, I cross two wires like I'm stealing a car. The Plateau dimensions are arbitrary, chosen to be consistent with the rest of their product line, and slightly small for extreme cooling: I have to mount my three top fans outside the case to leave room for the Noctua NH-D14 cooler, and if one orders a 3-fan panel for the right side, there's no room for fans to fit inside next to the power supply and bay devices. I guess 10" by 19" by 18" doesn't have the same ring to it as 9" by 18" by 18". I'm quibbling; I've never had this much room in a mass-market case.
The memory is on the motherboard supported lists, and fits under the cooler. Addonics makes great stuff; this 2 x 2.5" trayless drive bay is more convenient than the 4 x 2.5" tray-based drive bay I've used before.
So what's up with the Rheosmart? The basic idea is that the front panel offers six 3-pin standard fan channels, each of which lets one choose between manual control with a dial (green LED), and using voltage control to follow the motherboard's PWM control of one fan (red LED). This makes sense, the motherboard has a better idea of the cpu temperature than any add-on fan controller could possibly have. I've built my share of DIY fan controllers (I'm partial to diode arrays) but I really want the fans to automatically ramp up when the CPU gets hot, and ramp down when the CPU cools off. My ASUS motherboard BIOS offers fine manual control of two 4-pin PWM fan headers, e.g. I'm now running 100% at 70 C CPU temp, 0% at 50 C, scaling in between.
Most of the reviews of the Rheosmart series could have been written without opening the box. They're all on those look-alike 17 page, ad-filled review sites, so no surprise. There are also Amazon and Newegg user reviews along the lines of "I installed it and my fans still work." There's an information gap here.
Motherboards vary in how they modulate fan headers. My ASUS in particular has no voltage modulation, only PWM, and needs to see a real PWM fan to function properly. In particular, the Rheosmart does not provide a sufficiently faithful simulation to fool the motherboard. One symptom (the first symptom I saw) was that the BIOS wouldn't retain custom manual fan control parameters until I attached an actual PWM fan.
The motherboard probably depends on the RPM signal for feedback; the Rheosmart doesn't let you choose which fan should be used for this signal, suggesting that it's not providing any RPM signal to the motherboard. Were it, say, defaulting to fan channel 1 for this signal, a warning to this effect would be nice, as I'm Y-splitting many of my channels, and that always confuses reading the RPM unless you snip one of the two RPM wires.
Sunbeam provides many cables; the most crucial cable splits a 4-pin PWM motherboard fan header into a 4-pin PWM fan out, and a copy of the PWM signals for the Rheosmart to follow. The documentation doesn't assert this, but one needs an actual PWM fan attached to the PWM fan out, for their whole rig to work. Otherwise, in my experience, it just runs full throttle. With the assortment of other cables, one can jury-rig returning the RPM signal from one of the analog fans, while reading the PWM signal from the fan header. At least, one can cable this. But it doesn't work.
After many experiments, I swapped the front 120mm 3-pin Noctua fan that came with the cooler for the Cooler Master Blade Master 4-pin PWM fan, attached it to the PWM splitter cable so the motherboard could directly see it, and the Rheosmart started functioning as advertised. PWM fans click, but I can't hear this one click once it's in place.
The Rheosmart more or less works. I used a voltmeter to check the analog fan output. (Sunbeam provides two 3-pin fan to 4-pin power converter cables, that are the right genders to allow monitoring their output with a voltmeter.) With the motherboard set to 0% (CPU below 50 C) the analog fans see around 1 volt, and don't run. The Cooler Master PWM fan continues to run at around 600 rpm; perhaps it defaults to a minimum speed with no PWM signal, while other fans would stop. In any case, I like this outcome. The Noctua NH-D14 cooler is perhaps overkill for a Sandy Bridge processor, but my rig sure is quiet in this mode.
As soon as one sets the floor PWM to 1% in the BIOS, the analog fans now see 10 to 11 volts, and run only marginally below their nominal 1200 rpm. Not ideal, but one can live with this: The fans come on for computations, turn off when they're done. This could be an artifact of my fan choices; with a different combination of the PWM fan and the 3-pin fans, the 1% voltage might be lower. But the point is, I can't adjust this on the fly.
I corresponded with mCubed over their FanAmp. It amplifies a 3-pin fan signal for many fans, and it can be tuned, unlike the Rheosmart. (Ideally, it takes two controls per channel, but I wish that the Rheosmart would at least have the dial adjust the floor voltage in PWM mode, leaving the ceiling voltage at 12 V. Instead, the dial goes numb in PWM mode.) mCubed claims that one can modify their amp for PWM use by cutting the red wire, and connecting it to the PWM wire. I haven't tried this, but I may.
Or one can go the DIY route. A trivial C-R circuit can smooth the PWM signal to a varying voltage, an op-amp can buffer the fan header from the output current draw, and a power amplifier can boost this to a range to drive all one's fans. A good beginner project. Ideally one has two such circuits, so some fans follow the CPU fan header, and some fans follow the chassis fan header.
What would be very slick would be to have all fans run at exactly the same RPM; there are chips intended for this, which one could smooth individually from PWM to voltage, allowing the use of a mixed collection of 3-pin fans. I wouldn't go to this much trouble for myself, but if I were entering the market in competition with Sunbeam, I'd do this. It probably doesn't matter for case fans, but I have heard out-of-sync push-pull cooler fans beat against each other. And "exactly the same RPM" is a good way to accomplish "roughly the same RPM" when using a mixed collection of fans.
One could also go the all-PWM fan route. The motherboard can't supply enough power, but there are cables that take power from a 4-pin power cable, and the PWM signal from the motherboard, to run many PWM fans.
I'm amazed at how quiet my fan-on-demand build is, using the Rheosmart, even if it was puzzling to get working. I give it a qualified recommendation, until someone does better. This shouldn't be long; adjustable PWM fan control is becoming the norm. There's the usual lag, where the 2600K doesn't need water but we know how to do water, the motherboard doesn't need an independent fan controller but look at all the independent fan controllers on the market, etc. It always pays to forget everything one knows and start over. Eventually various players in the market will do this.
Sunbeam Rheosmart 6 Fan Controller
Noctua NH-D14 120mm & 140mm SSO CPU Cooler
Mountain Mods Plateau 18
mCubed FanAmp Fan Controller
Cooler Master Blade Master PWM 120mm Fan
Addonics Snap-In Double Drive
Last edited: