My old Compaq has a CPU VRM based on an SC1101 controller and is similar to the test circuit in the 2000 datasheet (the current datasheet has a different test circuit). The actual circuit uses an FDB603AL MOSFET, an S10S40C diode, less capacitance on the input (one 1000uF cap and one 1500uF cap), and different resistor values. The diode runs very warm and is almost too hot to touch, and the MOSFET and the controller IC are only slightly cooler. Why do these components run so warm? I don't know anything about switching power supplies and DC-DC converters.
-
-
Re: Warm CPU VRM components
Because they cheaped out and used components with high voltage drop (or high on resistance - mosfets). Or they simply didn't use enough of them. Or the board layout is tight and they didn't leave enough cooling area (D2PAK parts rely on the PCB area they are mounted on to cool them).
Btw - the diode is rated 125C, the mosfet is rated 175C. If the manufacturer says the parts can take high temperature, there aren't too many efforts made to keep it low.A working TV? How boring!Originally posted by PeteS in CA -
Re: Warm CPU VRM components
Why would the diode get so warm? It doesn't seem like much current should be flowing through it. I have added the SC1101 datasheet. Also, I measured 2.30V on the output with the output voltage set to 2.2V. The values of the resistors must not perfectly match their marked rating. There could also be a flaw in the circuit design (this is a Compaq
).
Comment
-
Re: Warm CPU VRM components
It's not current alone that gives temperature of a diode. It's current times voltage drop over package plus heatsink area. Like i said, if the board layout is tight, then the diode will get hot.
If the package is not over 100C (simple test - lick your finger then touch the diode and see if it fizzles) then it's operating under normal conditions.A working TV? How boring!Originally posted by PeteS in CAComment
-
Re: Warm CPU VRM components
It isn't running that warm. It is just hard to keep your finger on it for more than a couple seconds. It isn't as bad now that I have a fan blowing over the VRM. The board is cramped, so there isn't much room to cool these components.Comment
-
Re: Warm CPU VRM components
A fan won't do much by itself but it's still an improvement. You can try sticking some ramsinks on top of them if you have spare ones, they will help more than the fan alone does.
Reminds me of my Asus P55T2P4. Pentium MMX 200 @ 292MHz. Took quite a bit of heatsinking to pull that one off. For the CPU it was simple - a Socket A heatsink, for the VRM, not much so.A working TV? How boring!Originally posted by PeteS in CAComment
-
Re: Warm CPU VRM components
That's not an old compaq!
At least it has a proper PWM DC-DC converter. These old Presarios (and others) used linear regs for the CPU. Both the CPU and VRM were cooled by air entering the case through the holes on the left side. Just oversized heatsinks, the only fan was in the power supply.
I have personally seen three out of a set of eight overheat and freeze when the 'genious' had them open to reinstall Win95.
Don't even ask.
I would concur with the ramsinks. LTI, are those semi soldered to the board or is there a screw, washer and nut? If it's the latter, you may be able to get some small TO-220 heatsinks in there.
-PaulAttached FilesLast edited by kaboom; 06-18-2011, 12:26 AM."pokemon go... to hell!"
EOL it...
Originally posted by shango066Originally posted by smashstuff30guilty,guilty,guilty,guilty!
guilty of being cheap-made!
Comment
-
-
Re: Warm CPU VRM components
If you are really concerned you can put another same size diode in parallel with it.
.Mann-Made Global Warming.
- We should be more concerned about the Intellectual Climate.
-
Be who you are and say what you feel, because those who mind don't matter and those who matter don't mind.
- Dr Seuss
-
You can teach a man to fish and feed him for life, but if he can't handle sushi you must also teach him to cook.
-
Comment
-
Re: Warm CPU VRM components
This computer is a Presario 2286. It is cooled in the manner you mentioned, but the VRM is not in the airflow path. The original power supply fan spun so slowly that there was no air coming out of the exhaust vent. I put a better fan in the power supply and the original power supply fan is now on the CPU heatsink and connected to a homemade fan controller. This computer uses the same case, motherboard, and power supply. The components are soldered to the board.Originally posted by kaboom View PostThat's not an old compaq!
At least it has a proper PWM DC-DC converter. These old Presarios (and others) used linear regs for the CPU. Both the CPU and VRM were cooled by air entering the case through the holes on the left side. Just oversized heatsinks, the only fan was in the power supply.
I have personally seen three out of a set of eight overheat and freeze when the 'genious' had them open to reinstall Win95. Don't even ask.
I would concur with the ramsinks. LTI, are those semi soldered to the board or is there a screw, washer and nut? If it's the latter, you may be able to get some small TO-220 heatsinks in there.
-Paul
If those computers overheated with the top cover off, they must have been close to overheating with the cover on. This one has never overheated. If it does, my fan controller has an alarm.
I would like to know what purpose the diode has.Last edited by lti; 06-18-2011, 07:45 PM.Comment
-
Re: Warm CPU VRM components
100% for persistence https://www.badcaps.net/forum/showthread.php?t=14686better to keep quiet and be thought a fool, than to open your mouth and remove all doubtComment
-
Re: Warm CPU VRM components
When the MOSFET is turned on, current flows through the inductor to the load. Due to the inductance, the current increases with time, as a magnetic field builds up. When the MOSFET turns off, the diode commutates the current still flowing through the inductor to the load. The inductor stores energy in its magnetic field so as to make the PWM possible. The diode, or bottom MOSFET if synchronous rectification is used, is the current path when the top MOSFET is off.Originally posted by lti View Post
Without the diode, the nature of the inductor to 'keep current flowing' would generate nasty spikes.
These pdfs may help.
-Paul"pokemon go... to hell!"
EOL it...
Originally posted by shango066Originally posted by smashstuff30guilty,guilty,guilty,guilty!
guilty of being cheap-made!Comment
-
Re: Warm CPU VRM components
What maximum voltage and current should the diode be rated for? Is the current 10A, 40V diode good enough? This voltage regulator currently powers an AMD K6-2 running at 500MHz, but the diode ran just as warm with the original Cyrix MII-333GP CPU.Comment
-
Re: Warm CPU VRM components
40V is plenty since that diode is in the Vcore circuit which is below 3.3V (and ~1V these days). I'm not sure of the power output of the CPU but 10A may or may not be sufficient depending on how many diodes there are. (For comparison, modern CPUs now consume around 100A at full load
)
Comment
-
Re: Warm CPU VRM components
If you are going for two diodes in parallel then use a matched pair least one take all the load and the other do nothing.
.Mann-Made Global Warming.
- We should be more concerned about the Intellectual Climate.
-
Be who you are and say what you feel, because those who mind don't matter and those who matter don't mind.
- Dr Seuss
-
You can teach a man to fish and feed him for life, but if he can't handle sushi you must also teach him to cook.
-
Comment
-
Re: Warm CPU VRM components
The diode is connected to ground, so the full output current doesn't pass through it. I don't know how much power the CPU draws. The diode that is currently installed is actually two 5A diodes in a single package.Comment
-
Re: Warm CPU VRM components
http://www.m571.com/m571/m571_amd.htmOriginally posted by lti View Post
21W @ 2.2 volts = 9.55A.A working TV? How boring!Originally posted by PeteS in CAComment
Comment