Hi everybody!
I got a small problem concerning a laptop fan. During the first 30 minutes (more or less) it didn't work properly: it only started spinning immediately after having turned my laptop on after a system shutdown caused by the excess of CPU temperature (excess of temperature obviously due to the "absence" of the CPU fan). But then once on, the fan kept spinning for about 10-20 seconds before stopping again. So, once again the CPU temperature started rising until the forced shutdown. This kind of process occurred 3-4 times in 30 minutes. In this time span, the CPU fan seemed to be in oblivion and only worked at very high temperatures.
After these 30 minutes the fan started working regularly, even for hours!

Explained this problem, someone told me that it could easily depend on cold joints and a reflow of some components would have been the best solution in these cases. But I don't have a hot gun and neither do people I know. So I preferred to do some tests that someone advised.
Then I managed to reset the bios but nothing changed. I forgot to say that in the first 30 minutes my motherboard often blinked (caps lock and num lock).
At last, I found a quicker solution: to short connector and emitter of a darlington transistor that controls the CPU fan:

Now, it would be very intesting to know any possible disadvantage of this choice, such as energy consumption, transistor lifetime (as well as the ones of the other components that might be affected by the permanent spinning of the CPU fan) and other issues that would arise.
Anyway, thanks to a very small piece of copper (the end of a wire) put between collector and emitter, I tested my laptop for a few hours and it seemed that the full-speed fan didn't cause any problem and CPU temperature was well below acceptable limits.
Strange as it may seem, the DDR fan didn't even turn on, perhaps because CPU was always "mild" and so processor didn't give any signal to let the DDR fan start spinning. But I did want even the DDR fan running, so I did the same trick of joining connector and emitter. Now, I have 2 fans at full speed but I need to remove the temporary piece of copper and make this modification definitive.
There is just 1 mm between the 2 pins, so someone suggested to add a drop of solder in-between (without copper or any other material). Is the best solution if I don't want to waste time to unsolder the transistor? Does it guarantee a good electrical conductivity (and then a constant short between collector and emitter)? It should be a micro solder. I got a WS-70DA soldering iron that is not very precise for this kind of tasks. But I am going to practice a little bit before soldering on my motherboard.

Other issue concerns the noise. With 2 fans spinning at full speed I got to wear something to protect my ears... For this reason, someone suggested to put a resistor in order to decrease a bit the speed of the fan. For a DC 5V 0.30A fan, how many ohms should I need? Consider that often the CPU fan spinned at high speed (though not full speed), when my laptop worked in normal conditions. So I cannot exagerate with the ohms. But where and how putting the resistor in such a small available space? And what kind of resistor? Only SMD resistors?

 

You are going to cause premature failure of brushless motor,bearings will fail and cause a whining noise.

 

My suggestion are:

• reflow the solder on the transistors to make sure bad joints aren't the cause of the problem
• if that isn't the problem, make sure transistors are working

If transistors are working:

• see if you can change fan speed in BIOS
• refresh or update your BIOS
• look for a fan speed control app (I've heard of tpfancontrol for ThinkPads)

 

There is a freeware program called SpeedFan that can monitor CPU temperature and, possibly, adjust fan speed. The computer I have it installed on is very old and isn't fully supported; it monitors CPU temperature, but can't detect fan speed so I can't try to change it...

If your laptop is newer or you're willing to work with the developer to get it supported, you could potentially use it to adjust fan speed; assuming the drive transistors and logic are functioning correctly.

 

My suggestion are:

• reflow the solder on the transistors to make sure bad joints aren't the cause of the problem
• if that isn't the problem, make sure transistors are working

If transistors are working:

• see if you can change fan speed in BIOS
• refresh or update your BIOS
• look for a fan speed control app (I've heard of tpfancontrol for ThinkPads)

I would exclude the BIOS because I can not change fan speed, I did a reset but nothing changed and, as obvious, no software can detect the fan speed because BIOS doesn't provide any option regarding fans.
Well, according to the last test I did, after having shorted the 2 pins, my motherboard detected no problem anymore, no more blinking. I'm led to believe that the real problem is the transistor because, once modified its normal operation and forced it to activate the fan, the motherboard cannot detect any problem that was used to detect. I think I've hidden the symptom from the "eyesight" of the motherboard that is currently unable to recognize that problem. Now CPU temperature is quite cool, so the motherboard wouldn't have any reason to start blinking.
I think this trick is working and would be very quick and cheap to make it permanent (by applying a drop of solder between pins).
On the other hand, an alternative, that most people advised (like you), would be a reflow of the electrical components (in particular, of the transistor). Perhaps, at very last I should try this before the permanent modification I'm willing to do. Anyway, I would run some risks by using hot gun (I just have a WS-70DA soldering iron and I don't know if it's suitable for this kind of task). If I'm not enough careful, I'm going to unsolder some of the closest components (I'm worried about the electrolytic capacitor that can be damaged by the hot gun, as well as the other SMD components such as resistors and diodes), closest to the ones directly affected by the air flow. And what about the components on the same area but of the opposite side of the motherboard? Would they run the risk to be unsoldered and fall to the ground (for the law of gravity)?

This is my reasoning. If the problem is the transistor I should replace it, so unsoldering and then soldering the new one. On PCB I've seen doing this task by using hot gun (firstly, heat the component to remove, then remove it, place the new one and finally heat again the new transistor so that the solder below it can melt properly).
Costs? A new transistor and a brand new hot gun. Results? It could be the right solution. It could, not it is. What if it wasn't and even the hot gun reflow wasn't enough to solve the problem (assuming that the reflow wasn't harmful for motherboard and its components but no one can guarantee this)? Then I would have to definitely finalize the short between collector and emitter, knowing for sure that this remedy already works.
So why not avoiding the purchase of a transistor and a hot gun when there is already a much quicker, cheaper and perhaps even safer solution?
At last, could I try to heat the transistor with my soldering iron, placing its tip on the transistor pins for a few seconds? Would I damage the transistor or would I get the same result as the hot gun?

 

You can try unsoldering it with soldering iron.The key is to get solder between all the 3 pins so that you can lift the transistor up, question is would this be enough to unsolder the GND pad from the PCB.

What is the name of transistor?

 

On the other hand, an alternative, that most people advised (like you), would be a reflow of the electrical components (in particular, of the transistor). Perhaps, at very last I should try this before the permanent modification I'm willing to do. Anyway, I would run some risks by using hot gun (I just have a WS-70DA soldering iron and I don't know if it's suitable for this kind of task).

A soldering iron will be sufficient to reflow solder.

If I'm not enough careful, I'm going to unsolder some of the closest components (I'm worried about the electrolytic capacitor that can be damaged by the hot gun, as well as the other SMD components such as resistors and diodes), closest to the ones directly affected by the air flow. And what about the components on the same area but of the opposite side of the motherboard? Would they run the risk to be unsoldered and fall to the ground (for the law of gravity)?

Using a hot air tool in tight spaces can be tricky. Less so if you use an appropriate tip, air flow, and temperature. If you happened to melt the solder of nearby parts, surface tension from the molten solder should hold components in place as long as you don't jar the board or have air pressure too high. Surface tension is used to center components as they're being soldered...

could I try to heat the transistor with my soldering iron, placing its tip on the transistor pins for a few seconds? Would I damage the transistor or would I get the same result as the hot gun?

Most components will take 10 seconds or so of soldering temperatures.

The best solution to find and correct the problem so you don't substitute a fan speed problem with a dead fan, more noise, and reduced battery life.

 

My suggestion are:

• reflow the solder on the transistors to make sure bad joints aren't the cause of the problem
• if that isn't the problem, make sure transistors are working

If transistors are working:

• see if you can change fan speed in BIOS
• refresh or update your BIOS
• look for a fan speed control app (I've heard of tpfancontrol for ThinkPads)

I would exclude the BIOS because I can not change fan speed, I did a reset but nothing changed and, as obvious, no software can detect the fan speed because BIOS doesn't provide any option regarding fans.
Well, according to the last test I did, after having shorted the 2 pins, my motherboard detected no problem anymore, no more blinking. I'm led to believe that the real problem is the transistor because, once modified its normal operation and forced it to activate the fan, the motherboard cannot detect any problem that was used to detect. I think I've hidden the symptom from the "eyesight" of the motherboard that is currently unable to recognize that problem. Now CPU temperature is quite cool, so the motherboard wouldn't have any reason to start blinking.
I think this trick is working and would be very quick and cheap to make it permanent (by applying a drop of solder between pins).
On the other hand, an alternative, that most people advised (like you), would be a reflow of the electrical components (in particular, of the transistor). Perhaps, at very last I should try this before the permanent modification I'm willing to do. Anyway, I would run some risks by using hot gun (I just have a WS-70DA soldering iron and I don't know if it's suitable for this kind of task). If I'm not enough careful, I'm going to unsolder some of the closest components (I'm worried about the electrolytic capacitor that can be damaged by the hot gun, as well as the other SMD components such as resistors and diodes), closest to the ones directly affected by the air flow. And what about the components on the same area but of the opposite side of the motherboard? Would they run the risk to be unsoldered and fall to the ground (for the law of gravity)?

This is my reasoning. If the problem is the transistor I should replace it, so unsoldering and then soldering the new one. On PCB I've seen doing this task by using hot gun (firstly, heat the component to remove, then remove it, place the new one and finally heat again the new transistor so that the solder below it can melt properly).
Costs? A new transistor and a brand new hot gun. Results? It could be the right solution. It could, not it is. What if it wasn't and even the hot gun reflow wasn't enough to solve the problem (assuming that the reflow wasn't harmful for motherboard and its components but no one can guarantee this)? Then I would have to definitely finalize the short between collector and emitter, knowing for sure that this remedy already works.
So why not avoiding the purchase of a transistor and a hot gun when there is already a much quicker, cheaper and perhaps even safer solution?
At last, could I try to heat the transistor with my soldering iron, placing its tip on the transistor pins for a few seconds? Would I damage the transistor or would I get the same result as the hot gun?

 

I would exclude the BIOS because I can not change fan speed, I did a reset but nothing changed and, as obvious, no software can detect the fan speed because BIOS doesn't provide any option regarding fans.

There may not be any options for you to change fan speed, but that's where the manufacturer stores fan control parameters. If your BIOS has become corrupt, reflashing would correct. Fan control could have had a bug that a BIOS update fixed.

Just because there's nothing you can change in setup to affect fan speed, that doesn't mean there aren't ways to do it. The program I mentioned can detect fan speed from certain motherboards and potentially change it.

So why not avoiding the purchase of a transistor and a hot gun when there is already a much quicker, cheaper and perhaps even safer solution?

If you're determined to ignore advice offered and pursue your hack of shorting the transistor regardless of the consequences; please do it and let this thread die.

 

You can try unsoldering it with soldering iron.The key is to get solder between all the 3 pins so that you can lift the transistor up, question is would this be enough to unsolder the GND pad from the PCB.

What is the name of transistor?

I only know that it's a darlington transistor and I can see an alphabetical code on top of it: D E and W (below the first 2 letters).
I've used my soldering iron to reflow solder (at least trying to) between all the 3 pins of the transistor. My aim was just a reflow, not a transistor lifting. I noticed a nearby diode not horizontally soldered (slightly oblique but always inside its area). So I tried to heat it up some seconds to let the solder melt but nothing happened. I even couldn't move it minimally. Perhaps, my soldering iron isn't enough powerful.

There may not be any options for you to change fan speed, but that's where the manufacturer stores fan control parameters. If your BIOS has become corrupt, reflashing would correct. Fan control could have had a bug that a BIOS update fixed.

I absolutely agree when you talk about BIOS corruption.
Last August I took a note of my BIOS version:
PHOENIX 4.06 30/09/2003
Ver. SMbios 2.3

HwMonitor detected my motherboard as a Clevo 4xx series

So I visited Phoenix website to update but I was redirected to esupport.com, a website providing bios updates with fees. Fees? I don't have any intention to spend even half a cent to update my BIOS!
Some years ago I freely updated BIOS but I remember that my system didn't like that update, so I restored the previous one. Evidently, my hardware really didn't need it.
After the reset of the last week I saw on the POST that the current BIOS version is the 4.0 release 6.0 (then a few lines below: bios revision 1.05).
Is there a way to get access to BIOS updates without paying, as it's always happened (as far as I know)?

Anyway, now I have my laptop completely disassembled and I just want to make some measurements to possibly detect a faulty SMD component.

Just because there's nothing you can change in setup to affect fan speed, that doesn't mean there aren't ways to do it. The program I mentioned can detect fan speed from certain motherboards and potentially change it.

Speedfan and similar software can not detect any fan and relative speed: my motherboard isn't so recent.

If you're determined to ignore advice offered and pursue your hack of shorting the transistor regardless of the consequences; please do it and let this thread die.

OK! Before condemning the transistor to make the fan overwork for the rest of its life, let me see if it's possible to determine the origin of the problem by making some measurements.
Having a simple digital multimeter (GBC KDM-120) and without unsoldering the suspicious transistor, what kind of tests would you recommend to do while the fan is spinning (and even immediately after the turning on of the motherboard) and while the motherboard is off?
I found this schematic diagram on the internet:

Which components should I check out? And how?

 

So I tried to heat it up some seconds to let the solder melt but nothing happened. I even couldn't move it minimally. Perhaps, my soldering iron isn't enough powerful.

Is the soldering iron hot enough to melt solder? Is the tip tinned? Did you try applying a little new solder after heating the joint?

Having a simple digital multimeter (GBC KDM-120) and without unsoldering the suspicious transistor, what kind of tests would you recommend to do while the fan is spinning (and even immediately after the turning on of the motherboard) and while the motherboard is off?

Since shorting the transistor turns on the fan, you only need to measure the voltage on it's base and collector when the fan is, or should be, on. If it's really a darlington, the base will be around 1.4V when the fan should be on and the collector should be around a volt.

The transistor may be switched on/off too rapidly for your multimeter to give meaningful measurements.

 

Is the soldering iron hot enough to melt solder? Is the tip tinned? Did you try applying a little new solder after heating the joint?

Yes to the first 2 questions. No to the last one because applying a little new solder with the not so small tip I have would mean making a mess. I certainly would have to remove the solder in excess between the pins and I tell you this is not an easy task, especially when there is just 1 mm between the pins. I'd make a mess, very likely.

Since shorting the transistor turns on the fan, you only need to measure the voltage on it's base and collector when the fan is, or should be, on. If it's really a darlington, the base will be around 1.4V when the fan should be on and the collector should be around a volt.

The transistor may be switched on/off too rapidly for your multimeter to give meaningful measurements.

If I'm not wrong, the signal for temperature, coming from CPU, is converted in serial form by U27. Look at this picture taken from the service manual (http://www.datasheetarchive.com/dl/Datasheets-USER/DSAUPLD00007992.pdf) on page 88 (Socket 478 & ITP):

Someone told me that the origins of the problem can be traced in a fault of U30. Have a look at this picture from the same manual (page 109, LPC H8):

So, as you suggested, I connected the black probe to Q21 emitter (ground) and the red one to base and then collector. At the very beginning with motherboard on, at the base I measured 0 V (when even the other smaller fan wasn't active), then 0.18 V till 0.37 V (this when the other fan was at maximum speed and the main fan still didn't spin, even if moving in small jerks), finally it reached 1.24 V (while regurlarly spinning). Then when the main fan stopped I measured 0.37 V and I noticed that the voltage got back to 1.24 when the other fan began to spin faster. Perhaps, CPU sends signals to both fans but while the smaller fan increases the speed, the main one is only able to move in jerks, even if there is the same voltage as when it's spinning. With a screw driver when I heard the other smaller fan increasing its speed, I manually helped the main fan to go and I succeeded. But then, why only after the shutdown (and an excessive temperature) does it start spinning (and at full speed)? Does this mean that the fan is faulty?

Let's talk about the collector. At the very beginning with Mobo on, at the collector I measured 4.41 V, then when the other fan started spinning the voltage dropped to 0.81 V and when the other fan increased its speed I measured even less volts (from 0.58 to 0.64) and at that point I noticed that when the main fan showed signs of movement, the volts increased to 0.81 (for a fraction of second, i.e. for the duration of those signs). Finally, when regularly working it reached 0.60 V.
I even measured voltage immeditely after the shutdown: 0.93 V at the base and 0.24 V at the collector.
What does all this mean?
Now, other components to test?
Could I test FAN_SEN with my DMM? There are a diode and a resistor for FAN_SEN. Should I test them? If yes, what measurements should I take?

Anyway, I noticed that during the time of missed spinning (about 30 minutes), the fan moved in small jerks but without starting spinning as it should have had to. Only after the forced shutdown (caused by the excessive temperature), when I turn the Mobo on a few seconds later, the fan starts spinning (at full speed but for some seconds because then it sinks into oblivion again!), otherwise I've to manually let it go. Isn't it a kind of thermal problem? Isn't it a symptom of the so-called cold joints? Or is there a faulty SMD component (transistor or something else closer to processor, unable to trasmit the signal unless the temperature gets excessive)? Or, last, do these symptoms have the characteristics for asserting the corruption of the BIOS code? Or more simply a faulty fan?

 

Yes to the first 2 questions. No to the last one because applying a little new solder with the not so small tip I have would mean making a mess. I certainly would have to remove the solder in excess between the pins and I tell you this is not an easy task, especially when there is just 1 mm between the pins. I'd make a mess, very likely.

What type of tip does your soldering iron have, what is it's wattage and/or temperature?

With a screw driver when I heard the other smaller fan increasing its speed, I manually helped the main fan to go and I succeeded. But then, why only after the shutdown (and an excessive temperature) does it start spinning (and at full speed)? Does this mean that the fan is faulty?

This is a possibility.

Let's talk about the collector. At the very beginning with Mobo on, at the collector I measured 4.41 V, then when the other fan started spinning the voltage dropped to 0.81 V and when the other fan increased its speed I measured even less volts (from 0.58 to 0.64) and at that point I noticed that when the main fan showed signs of movement, the volts increased to 0.81 (for a fraction of second, i.e. for the duration of those signs). Finally, when regularly working it reached 0.60 V.
I even measured voltage immeditely after the shutdown: 0.93 V at the base and 0.24 V at the collector.
What does all this mean?

There are a couple ways to control fan speed. One is to change the voltage, but the more common method is to use a switch to turn the fan on and off as needed to get the desired speed. Your motherboard must be doing the latter.

Could I test FAN_SEN with my DMM? There are a diode and a resistor for FAN_SEN. Should I test them? If yes, what measurements should I take?

Fan sense is a pulse train that represents fan speed. If this were defective, speed control would be ineffective.

Anyway, I noticed that during the time of missed spinning (about 30 minutes), the fan moved in small jerks but without starting spinning as it should have had to. Only after the forced shutdown (caused by the excessive temperature), when I turn the Mobo on a few seconds later, the fan starts spinning (at full speed but for some seconds because then it sinks into oblivion again!), otherwise I've to manually let it go. Isn't it a kind of thermal problem? Isn't it a symptom of the so-called cold joints? Or is there a faulty SMD component (transistor or something else closer to processor, unable to trasmit the signal unless the temperature gets excessive)? Or, last, do these symptoms have the characteristics for asserting the corruption of the BIOS code? Or more simply a faulty fan?

Could be bad solder joints preventing the transistor from switching the fan on/off, could be a defective fan (fan sense or bad bearing), could be the logic generating the PWM control for the transistor. Hard to narrow it down until things can be eliminated.

Do you have a spare fan to substitute for the one that jerks? Any chance you can get an oscilloscope to look at the PWM and fan sense signals? Any chance you can find someone to reflow the solder on the transistors for you? If they're in SOT-23, they shouldn't be difficult to reflow with the correct soldering iron and technique.

 

What type of tip does your soldering iron have, what is it's wattage and/or temperature?

The tip is conical and its width is about 1 mm. There is 1 mm between base-collector-emitter... I'd need something thinner, I suppose. A shot of my soldering iron:

Actually, I don't know if it's an interchangeable tip... I should ask where I bought this item. It's the cheapest soldering iron, perhaps it's not possible to apply another thinner tip.
Temperature is not available on the product, nor on the package, but I can tell you the wattage: 20W/130W (a microswitch allows to select the wattage).

Do you have a spare fan to substitute for the one that jerks?

No, I just have a couple of old PCs with bigger fans (12V). I even tried to ask for a small fan to small shops involved in reparing activities. The result is that they do have old motherboard with fans, old irremediably damaged laptops destined to be brought to a collection center (for waste disposal) but they don't want to sell any possible spare part that can be recycled. I saw with my eyes a big box full of motherboards with fans. Who knows how many of them would have been suitable for my purpose! They'll never use all that stuff. When I asked for an old fan, they began to shilly-shally... saying, we don't have the exact fan model you are looking for, just search on the internet... even if I told them that any could be suitable (voltage is always the same, only connectors may differ) because I'd have managed to adapt them inside my laptop. The truth is that they wanted me to let them replace the fan, recycling an old one that they did have. They wanted to be paid not for a ridicolous fan but for labor!

Any chance you can get an oscilloscope to look at the PWM and fan sense signals?

No, unfortunately.

Any chance you can find someone to reflow the solder on the transistors for you? If they're in SOT-23, they shouldn't be difficult to reflow with the correct soldering iron and technique.

No friends working as electricians or having this particular hobby.
The only electrician I know is all but available to waste his "precious" time with reflowing solder. This man runs a small shop/lab. He's not so kind and he only smiles when someone involves him in gainful projects. Last time he dismissed me in less than a minute...

Anyway, I got some news for my problem.
Last days I found out that there is an old BIOS revision that supports fan control, I think it's the last revision release. I didn't know that!
Now, I think it's worth trying an update. Main difficult: keeping system on during the update. As the transistor is hardly accessible with laptop completely assembled, I should find a way to have the fan working all the time (at least until the end of the update). If I want to short transistor's pin, I need to keep the motherboard in vertical position and without the bottom case, as the transistor is located on the bottom of the motherboard. Otherwise, I couldn't reach it with a screw driver.
Instead, I can get much easier access to the connector of the fan which is on motherboard top. Lifting the keyboard is enough to reach that connector. Is there a way to do something on the connector? Perhaps something on the tachimetric pin (corresponding to the yellow wire)... is there any trick to do?
Or could I easily opt for externally powering the fan? But then the fan should be disconnected from the motherboard. Will the processor react badly impeding even the start up?

 

The tip is conical and its width is about 1 mm. There is 1 mm between base-collector-emitter... I'd need something thinner, I suppose. A shot of my soldering iron: ...snip... perhaps it's not possible to apply another thinner tip.
Temperature is not available on the product, nor on the package, but I can tell you the wattage: 20W/130W (a microswitch allows to select the wattage).

The tip on your iron isn't tinned. It should have a "bright" layer of solder on it to facilitate good heat transfer. 1mm will be tight for the center terminal, but should be doable. I wouldn't use the 130W setting on SMT parts.

I use the flux in the core of my solder to remove any tip oxidation, but I've never had a tip that was badly oxidized. You might need more flux. I wouldn't sand the tip unless you have a spare handy.

Your tip is attached similar to my (old) Weller WTCP. It has tips of various shapes, sizes, and temperatures; though I use 700F for everything.

No, I just have a couple of old PCs with bigger fans (12V).

Without a known good fan or a scope to look at the speed sense signal, you can't eliminate the fan as the problem.

The only electrician I know is all but available to waste his "precious" time with reflowing solder. This man runs a small shop/lab. He's not so kind and he only smiles when someone involves him in gainful projects.

Can't really blame him for not wanting to waste his time. But in little more than the time it took him to dismiss you, he could have reflowed the solder.

Now, I think it's worth trying an update. Main difficult: keeping system on during the update.

Support your laptop on some non-conducting surface (like some books or boxes) and use an external fan to keep things cool.

 

Finally, I managed to update the BIOS but fan control only provides 2 options: "continuos" and "automatics". I didn't even notice the difference. None of them works.
In the BIOS there is no possibility to set voltage or Pwm. Could I set these parameters in a software like speedfan?
If I set "continuos", will I have the possibility via speedfan or similar software to manage the fan in order to let it start at full voltage?
I asked you these questions because I wish to avoid useless trials, as my motherboard is now disconnected and I should lose time to reassemble hardware and case. And then? Realizing that speedfan or similar is unable to better manage fan control?