Ok, this isn't strictly an "audio" topic, but it is audio-related 
I want to upgrade the cooling system of my HT PC to water cooling, to make it completely silent for the movie/gaming sessions. But I'm running it heavily overclocked and some elements that cannot be easily watercooled, such as voltage regulators, will have to be cooled in a more traditional manner and some fans will still be necessary. 
From my experience, a good quality fan spinning at less than 800rpm is enough to provide this "auxiliary" cooling while being virtually silent (barely audible from 10-20cm away). So I will need a couple of these, BUT because of my PC case layout I will have to use a few different sized fans. So finally, getting to the point:
I need a device that will:
- be able to feed off 12V DC (PC power supply)
- provide independent voltage (speed) regulation for each fan
- regulation range to be at least 3V to 10V (wider range would of course be ok too)
- have one "master" regulation for all channels [so that I can easily turn ALL fans up at once when I play games, because games + overclock = lots of heat]
- ideally also have a "kick-start" feature - provide full voltage for 1/2s to all fans at start-up, to make sure that all fans will actually be able to start spinning at "terminally low" voltages.
Unfortunately, I have too little knowledge to be able to design it myself - I know only a handful of active parts and have very little experience in designing stuff... Here are my "ideas" - I'm not sure if any of this is practical or even possible (and I can't do proper schematics, sorry! it's just a "conceptual" chart):
Any advice will be greatly appreciated - especially if someone could tell me how to keep all this from oscillating... well, I'm not even sure what is the nature of oscillations, I just know it's something bad that happens to parts occasionally when the designer messes something up
Also, if someone could tell me what values of kick-start capacitors to try (is it even a good idea to use capacitor as a bypass for the regulators?)
And if someone would be so kind so as to translate this into an actual "proper" schematic for me, that would be wonderful ...if it makes any sense at all, of course
I want to upgrade the cooling system of my HT PC to water cooling
, to make it completely silent for the movie/gaming sessions. But I'm running it heavily overclocked and some elements that cannot be easily watercooled, such as voltage regulators, will have to be cooled in a more traditional manner and some fans will still be necessary. From my experience, a good quality fan spinning at less than 800rpm is enough to provide this "auxiliary" cooling while being virtually silent (barely audible from 10-20cm away). So I will need a couple of these, BUT because of my PC case layout I will have to use a few different sized fans. So finally, getting to the point:
I need a device that will:
- be able to feed off 12V DC (PC power supply)
- provide independent voltage (speed) regulation for each fan
- regulation range to be at least 3V to 10V (wider range would of course be ok too)
- have one "master" regulation for all channels [so that I can easily turn ALL fans up at once when I play games, because games + overclock = lots of heat
]- ideally also have a "kick-start" feature - provide full voltage for 1/2s to all fans at start-up, to make sure that all fans will actually be able to start spinning at "terminally low" voltages.
Unfortunately, I have too little knowledge to be able to design it myself - I know only a handful of active parts and have very little experience in designing stuff... Here are my "ideas" - I'm not sure if any of this is practical or even possible (and I can't do proper schematics, sorry! it's just a "conceptual" chart):
An externally hosted image should be here but it was not working when we last tested it.
Any advice will be greatly appreciated - especially if someone could tell me how to keep all this from oscillating... well, I'm not even sure what is the nature of oscillations, I just know it's something bad that happens to parts occasionally when the designer messes something up
Also, if someone could tell me what values of kick-start capacitors to try (is it even a good idea to use capacitor as a bypass for the regulators?)
And if someone would be so kind so as to translate this into an actual "proper" schematic for me, that would be wonderful ...if it makes any sense at all, of course
Last edited:
Forget about manual control. Go automatic.
NoiseMagic ThemoControl NMT-3 w/ 3-pin Connector - FrozenCPU.com
Simply select low noise fans which provide the required worst case airflow at 12vdc nominal voltage, then use a couple of these to provide variable speed control based on internal air temperature. If desired, the thermistor can also be remote mounted in order to monitor various zones, though at the loss of closed loop regulation. A kick start function is also provided.
Over the last few years, I have used these in computers, audio amplifiers, network media players, etc. with excellent results. Quiet, reasonably priced, very compact and best of all, fully automatic.
However, if you really must have manual control, there are a myriad of off the shelf products such as these:
Fan Controllers | Page 1 | Sort By: Product Title A-Z - FrozenCPU.com *
Regardless of the control method chosen, I think you will find that you don't actually need all those fans. Most PC cases are very poorly designed from a cooling perspective, using multiple fans in an attempt to move as much air as possible, with very little thought given to airflow optimization.
Try removing and closing off all, except one or two, of the largest rear panel fans. Also, close off all case openings which do not contribute to front to back - bottom to top airflow. Sometimes it will be necessary to create an opening in order to get air where you need it. Use ducts or deflectors to direct air to hot spots. With a bit of experimentation you should be able to achieve and maintain stable internal temperatures over a wide range of ambient conditions. When you do, simply put a thermal speed controller on the remaining fan(s) and your done.
As an example of what is possible, my home server/media server with 12 hard drives, dual core 2.8 processor, and 650 watt power supply in a medium tower case, remains acceptably cool even when room temps are over 30C, using only one 8cm case fan. And, it is barely audible from more than a few feet away. Controlled airflow is the answer.
Good luck!
P.S. It is also possible to control most motherboard connected fans via the internal chipset control bus. It's very possible you don't need any hardware whatsoever to accomplish what you need. Google SpeedFan.
NoiseMagic ThemoControl NMT-3 w/ 3-pin Connector - FrozenCPU.com
Simply select low noise fans which provide the required worst case airflow at 12vdc nominal voltage, then use a couple of these to provide variable speed control based on internal air temperature. If desired, the thermistor can also be remote mounted in order to monitor various zones, though at the loss of closed loop regulation. A kick start function is also provided.
Over the last few years, I have used these in computers, audio amplifiers, network media players, etc. with excellent results. Quiet, reasonably priced, very compact and best of all, fully automatic.
However, if you really must have manual control, there are a myriad of off the shelf products such as these:
Fan Controllers | Page 1 | Sort By: Product Title A-Z - FrozenCPU.com *
Regardless of the control method chosen, I think you will find that you don't actually need all those fans. Most PC cases are very poorly designed from a cooling perspective, using multiple fans in an attempt to move as much air as possible, with very little thought given to airflow optimization.
Try removing and closing off all, except one or two, of the largest rear panel fans. Also, close off all case openings which do not contribute to front to back - bottom to top airflow. Sometimes it will be necessary to create an opening in order to get air where you need it. Use ducts or deflectors to direct air to hot spots. With a bit of experimentation you should be able to achieve and maintain stable internal temperatures over a wide range of ambient conditions. When you do, simply put a thermal speed controller on the remaining fan(s) and your done.
As an example of what is possible, my home server/media server with 12 hard drives, dual core 2.8 processor, and 650 watt power supply in a medium tower case, remains acceptably cool even when room temps are over 30C, using only one 8cm case fan. And, it is barely audible from more than a few feet away. Controlled airflow is the answer.
Good luck!
P.S. It is also possible to control most motherboard connected fans via the internal chipset control bus. It's very possible you don't need any hardware whatsoever to accomplish what you need. Google SpeedFan.
@nigel - I should have thought of that, thanks; the datasheet is actually very helpful.
@kcollins - I am well aware of the existence of fan controllers / motherboard fan control / SpeedFan / etc. I've been overclocking and modding computers since the first Pentium came out in 1993...
There are three main issues with the "ready made" controllers:
- limited regulation range (most of them can only go as low as 5-6V)
- none of them seem to have a "master" dial, which I really want my controller to have
- and lastly, they look stupid with all their "cool" LEDs, UV-reactive cables, LCD displays and whatnots... I just want a simple device that will quietly sit out of the way until I need it.
As for the automatic fan control, I think you misunderstood my intentions - I want my PC to be completely inaudible. I mean so that in order to tell if its on or off you'd have to look at the case diodes or switch on the monitor.
I know it sounds impossible on a heavily overclocked system, but I have done it before using a passive water cooling system in conjunction with semi-passive air coolers I built myself and a semi-passive PSU. Some fans were necessary, but they were running at 3-4V / 600-800rpm - i.e. barely ticking over.
So now I want to do a similar thing, but a bit more flexible this time - so that it CAN be inaudible, but with an option of being easily adjusted for extra cooling (trading off quietness for cooling power).
I managed to draw a schematic of what I have in mind (it's mostly just ripped off the LM317 datasheet but hey):
What do you think?
@kcollins - I am well aware of the existence of fan controllers / motherboard fan control / SpeedFan / etc.
I've been overclocking and modding computers since the first Pentium came out in 1993...There are three main issues with the "ready made" controllers:
- limited regulation range (most of them can only go as low as 5-6V)
- none of them seem to have a "master" dial, which I really want my controller to have
- and lastly, they look stupid with all their "cool" LEDs, UV-reactive cables, LCD displays and whatnots... I just want a simple device that will quietly sit out of the way until I need it.
As for the automatic fan control, I think you misunderstood my intentions - I want my PC to be completely inaudible. I mean so that in order to tell if its on or off you'd have to look at the case diodes or switch on the monitor.
I know it sounds impossible on a heavily overclocked system, but I have done it before using a passive water cooling system in conjunction with semi-passive air coolers I built myself and a semi-passive PSU. Some fans were necessary, but they were running at 3-4V / 600-800rpm - i.e. barely ticking over.
So now I want to do a similar thing, but a bit more flexible this time - so that it CAN be inaudible, but with an option of being easily adjusted for extra cooling (trading off quietness for cooling power).
I managed to draw a schematic of what I have in mind (it's mostly just ripped off the LM317 datasheet but hey):
An externally hosted image should be here but it was not working when we last tested it.
What do you think?
I think that you maybe making it over complicated. Your 12 volts is already regulated so all you need is some npn transistors and some pots. (See attached)
The first cap (connected to the 12 volt rail) should really be equal the sum of the capacitors of the individual fan controllers.
The first cap (connected to the 12 volt rail) should really be equal the sum of the capacitors of the individual fan controllers.
Attachments
Last edited:
Thanks for the schematic RJM1. The main reason I used LM317 in my design is that it is among the few active components that I actually know You're suggesting to use npn transistors, but I still don't know what to actually look for when I go shopping... I mean they come in hundreds of varieties and I'm not really sure what parameters to look for. Do you think you could give me an example of a suitable part?
Also, why is my design more complicated? To my eye it looks simple, and actually very similar to yours (only it replaces NPN transistors with LM317)...
Unless... hang on, what is the role of the 1uF caps in your schematic? They do the "kick-start" function, right? Ah, well in that case, your design is indeed simpler, and in fact exactly what I need.
You're suggesting to use npn transistors, but I still don't know what to actually look for when I go shopping... I mean they come in hundreds of varieties and I'm not really sure what parameters to look for. Do you think you could give me an example of a suitable part?Also, why is my design more complicated?
To my eye it looks simple, and actually very similar to yours (only it replaces NPN transistors with LM317)...Unless... hang on, what is the role of the 1uF caps in your schematic? They do the "kick-start" function, right? Ah, well in that case, your design is indeed simpler, and in fact exactly what I need.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.