As I am an electronics technician with a local (Western Canada) electronics manufacturer, I am quite familiar with the use of thermistors. <BR> <BR>The computer PSU design calls for several DC outputs, I.E. 5VDC @ 15-amperes, 12VDC @ 5-amperes, etc. To acheive these requirements <BR>and remain physically small, the manufacturers use switching designs. Basically, the 120VAC power supply is filtered for internal/external <BR>powerline noise, rectified into DC, and filtered with DC capacitors. This raw DC voltage is the DC supply for the main switching transformer and the <BR>switcher control circuitry. The switcher frequency varies, but will be around 40 kHz (40,000 Hertz). <BR> <BR>The output windings are generally fed into air-core wound coils. These coils are pulsed by the control circuitry and thereby "store" the energy. <BR>Thus, these coils will get hot with the loading of the MoBo, CPU, HDD, RAM, etc. The thermistors add a level of safety protection for the PSU, in that if <BR>the coils get too hot, the thermistor resitance will change, the cooling fans will run faster, thereby cooling off the coils and the rest of the PSU. <BR> <BR>Adding an variable potentiometer to manually adjust the voltage to the fan is an interesting idea, but has one drawback: the PSU now has no automatic <BR>protection against generated heat within the PSU design. If the PSU fails, the designers hope the switcher will fail in the OFF state (not always). However, the <BR>PSU could always fail in the ON state, provide high voltages and wreck havoc on the MoBo, CPU and the rest of your computer. Most manufacturers <BR>(if they are smart) have proection on the DC powerlines to guard against high voltages. These devices are called zener diodes and/or transorbs. Both provide <BR>protection against over-voltages. At best, these devices should be in the PSU design and at least, be on the MoBo. <BR> <BR>In my 30-years of experiences with power supplies, both linear and switching, I would recommend against fooling with the thermistors. The <BR>thermistor is a protection device, like a fuse. You wouldn't bypass a fuse, would you?
[addsig]

Hello TerryW, <br> <br>Thanks for sharing your knowledge and experience with us in that summary of how a switching PSU works. I'm sure lots of folk appreciate the info. <IMG SRC="modules/phpBB_14/images/smiles/icon_smile.gif"> <br> <br>Re -- the thermistor-removal warning, I gather you refer to the Postscript added to the 2nd page of the article "Quieting the Enermax & other thermistor fan PSUs" <!-- BBCode Start --><A HREF="http://www.silentpcreview.com/modules.php?op=modload&name=Sections&file=index&req=viewarticle&artid=9&page=2" TARGET="_blank">here.</A><!-- BBCode End --> <br> <br>I agree with you that the thermistor is a protection device. I think I may have written that it is cheap insurance for the manufacturer. It's is not cheap in the noise price you pay however. <br> <br>My quibble with the Enermax (any many other similar PSUs) is that the protection device is is poorly calibrated. To extend your analogy of a fuse, instead of using a 2 amp fuse which would blow only when the system is really in danger of overload, they've chosen a 0.8 amp fuse that blows all the time. - pun intended <IMG SRC="modules/phpBB_14/images/smiles/icon_wink.gif"> <br> <br>Even with just one fan running at minimal speed, the internal temperature of the unit does not go very high. Yet, the thermistor circuit ensures that the voltage fed to the fan(s) goes up to 8-9-10V with hardly any load. <br> <br>My original modification of simply moving the thermistor away from the hottest spot in the PSU has the effect of locking the fan speed at minimum. Trying to position the thermistor so that the fan voltage rises more gradually and intelligently in response to a load is extremely difficult, a hit and miss affair. Replacing the thermistor with a pot allows the fan to always run at least at minimal speed and unlike the original thermistor reposition mod, does allow the fan speed to be increased. With some intelligence on the part of the user to monitor the exhaust air under high loads and increase the fan speed when necessary, I believe this scheme is reliable. <br> <br>I have personally run many thermistor controlled PSUs with the fan set to contant minimal speed & not seen any failures that can be attributed to this. While any component in stock form is usually safer, replacing the thermistor with the correct value pot does NOT represent the same effect as bypassing a fuse. The fuse bypass means you have no protection whatsoever, while as the POT replacement does allow for at least minimal airflow all the time, which is usually protection enough. (see the article about whether a normal PSU can be run without a fan <!-- BBCode Start --><A HREF="http://www.silentpcreview.com/modules.php?op=modload&name=Sections&file=index&req=viewarticle&artid=15&page=1" TARGET="_blank">here</A><!-- BBCode End -->) <br> <br>The solution represented by the Seasonic and new Zalman PSUs is superior to either the Enermax too-fast- too-soon method or my modification: it provides minimal quiet airflow most of the time and speeds up the fan to cool the components only when really necessarily. <br>

Hi MikeC, <br> <br>Thanks for the comments. Yes, my reply was to the Postscript added to the article of "Quieting the Enermax & other thermistor fan PSUs". <br> <br>In response to your statement of "My quibble with the Enermax (and many other similar PSUs) is that the protection device is is poorly calibrated". I doubt whether the thermistor is strickly used for the fan. Most (smart) manufacturers design their equipment so the product "stands alone" without ancilliary devices, like a fan. I would hazard to guess the fan is an afterthought, to get the heat out of the PSU and away from critical components. <br> <br>One thing I neglected to mention about thermistors: they are usually used in the PSU design to provide protection to ensure the PSU does not go into thermal runaway (meltdown). The thermistor is usually attached to a heatsink or other heat generating device. The thermistors provides a certain amount of control bias (small amount of voltage) to the switcher circuitry to bring the PSU within design limits. The fan may be also tied into that control circuitry. <br> <br>The use thermistors are common practise for stereo receivers, audio amplifiers, TV, computer monitors as used in your offices, homes, vehicles, etc. The thermistor is always attached to the heatsink, as the components that get hot are (most) always attached to a common heatsink. <br> <br>To find out whether the thermistor is for the fan only or part of the control circuitry, one will need to get a copy of the PSU schematic, and/or perform further testing. Getting a PSU schematic is like pulling hens teeth. <br> <br>I don't want to lead anyone down the "garden path". I am here to offer technical assistance and to warn users that such modifications could cause unrecoverable damage. I would shudder to think that someone modifies parts of the PSU, to find they 'cook' the PSU, MoBo, CPU, etc. That would be one hard and expensive lesson.
[addsig]

>> Most (smart) manufacturers design their equipment so the product "stands alone" without ancilliary devices, like a fan. I would hazard to guess the fan is an afterthought, to get the heat out of the PSU and away from critical components. <br> <br>For computer PSUs, that would surprise me a lot. They've all had fans since the days of the 286, and Intel's ATX spec specifically indentifies the PSU fan's key role in exhausting hot air out of the case. Also, this comment doesn't really jibe with the idea of the thermistor as a fail safe device. If the PSU can run without a fan, why bother with one at all & why make it blow SO much air? Another point is that many PSU makers also specify a certain level of forced airflow for safe operation of so-called fanless models at full output (TKPower's 300, for example). <br> <br>>>One thing I neglected to mention about thermistors: they are usually used in the PSU design to provide protection to ensure the PSU does not go into thermal runaway (meltdown). The thermistor is usually attached to a heatsink or other heat generating device. The thermistors provides a certain amount of control bias (small amount of voltage) to the switcher circuitry to bring the PSU within design limits. The fan may be also tied into that control circuitry. <br> <br>>>To find out whether the thermistor is for the fan only or part of the control circuitry, one will need to get a copy of the PSU schematic, and/or perform further testing. Getting a PSU schematic is like pulling hens teeth. <br> <br>Is that like milking a bull? <IMG SRC="modules/phpBB_14/images/smiles/icon_lol.gif"> Surely someone somewhere has one for us to examine. Please post here if anyone has a schematic for a thermistor-quipped ATX PSU they can share! <br> <br>>>I don't want to lead anyone down the "garden path". I am here to offer technical assistance and to warn users that such modifications could cause unrecoverable damage. I would shudder to think that someone modifies parts of the PSU, to find they 'cook' the PSU, MoBo, CPU, etc. That would be one hard and expensive lesson. <br> <br>There are lots of folks here who mod their PSUs -- many of them long before they encountered SPCR. Very few report that kind of overcooking. But it could be they're embarassed to? Anyone with such stories, please speak up!
[addsig]

The Safe Side:
Assuming that the thermistor does more than just control the fans in the PSU, should I then wire up a Zalman Fanmate (or similar device) to control the speed of the fan inside the PSU? (In this setup, I'd plug the fan into a normal 12V plug from the PSU. I wouldn't use the outlet inside the unit.)

The Wide Side:
MikeC's pot solution is so simple. Am I crazy to just assume that the thermistor only controls fan speed? Has anyone had any fallout from slapping a 100K ohm pot on Enermax's thermistor?


I'm so close to having a silent PC. I just don't wanna start a house fire trying to reach my dream..

Why would you not just insert a resistor between the fan and the PSU thus lowering the voltage of the fan without even touching thermistor.

PC Power & Cooling said that their thermistors aren't involved in thermal shutdown and that you can't even order thermal shutdown as an option.

A couple of Enlight PSUs I looked at used the thermistor only for fan control, and I'm pretty sure the same is true with my 300W Antec. I don't know about m Fortron/Sparkle PSUs, but they each have a missing second thermistor and some other parts connected to a small board (not the fan controller board but the one with the KA3511 regulator controller) and I wouldn't mind knowing how this works. I'm afraid of unsoldering this tiny board to trace it out because right near its pins are several surface mount capacitors that the heat would probably unsolder.