Hello! This is my first post on this forum, and I'm putting a call out to any electrical and computer engineering gurus or wannabes (like myself).
As many of you surely know, the monitoring chips on any given motherboard often have more capabilities than most motherboards utilize. For example, the IT8712F can control and monitor up to 5 fans via PWM depending on the chips configuration. Most motherboards though have fewer fan headers than this, and often fewer still that are controllable. My ignorantly optimistic goal is to access these additional capabilities.
I have two systems: a test system, and my main rig. I'm working on the test rig first to determine if my goal is remotely viable. Specs as follows:
-AMD Athlon 64 3200+
-1GB generic DDR RAM
-Unknown nVidia 256MB Graphics Card
-Generic 480W PSU
-Windows XP x86 Professional SP2
*-Monitoring chip is an IT8712
*-Two fan headers: CPU_FAN and SYS_FAN; both are 3 pin, and only CPU_FAN is controllable.
-Asus M5A97 Evo
-AMD Phenom II x4 945
-8GB DDR3 Corsair RAM
-Sapphire Radeon HD 6870 1GB
-500GB WD Green HDD (System)
-Corsair 750W PSU
-Windows 7 x64 Home Premium SP1
*-Monitoring chip is an IT8721
*-Three fan headers: CPU_FAN, CHA_FAN1 and CHA_FAN2. The first two are 4 pin, the last is 3 pin. Only CPU_FAN is controllable (CHA_FAN1 outputs a steady 5 volts on pin 4; odd)
I use Alfredo Comparetti's excellent SpeedFan software
for fan control and monitoring.
First, about the IT8712F and the IT8721F. ITE seems to be secretive about their products, and information on either chip is not easily found. I have found a few revisions of the datasheet for the IT8712, but not for the IT8721. All revisions of the IT8712 datasheets show the same pinout. The IT8721 is not even listed on ITE's product page. So I know I could be working off of outdated information. Here is the datasheet I used.
Based on the datasheet, only 1 set of pins is DEDICATED to fan control and monitoring: pins 7 and 8 for tach and PWM respectively. The other four sets of pins (9-12 and 20-23)can be configured as general purpose I/O pins or joystick buttons or axis. I don't have any way of knowing for sure whether these other pins have been used for other purposes, but SeedFan does pull multiple readings from both rigs, so I'm sure that not all of them have been used for other things.
My plan was to solder a wire to each pin 7-12, and then use a multimeter to check the voltage and frequency of each. In a perfect world, I would have been able to read the incoming tach pulses and the outgoing PWM pulses for each channel. Then, assemble some external circuitry to take advantage of this new found functionality! But alas, this is not a perfect world, and I have a very imperfect understanding of motherboard design.
DISCLAIMER: before anyone tries this at home, keep in mind that the lead spacing on these chips is 0.5mm, and it takes some skill and know-how to solder to them. Messing this up could very well permanently damage the motherboard in unpredictable and/or catastrophic ways. I will link to some pics, but I didn't include a penny or anything for reference. This stuff is SMALL!
After clicking the links, click File>Download for the full 1619x1079 image.Close up of solder job.View from further out.Splayed wires.
After soldering and cleaning, I fired up the rig and was pleased to see that everything still worked correctly. SpeedFan reported CPU_FAN at around 900RPM and SYS_FAN at 2900RPM. CPU_FAN control still worked normally.
I pulled out my meter, a Fluke 287, and began measuring. First, I measured the voltage at the CPU_FAN header on pin 2 (pin 1 is ground). As expected, it was a steady DC voltage that varied with SpeedFan control. Pin 4 was around 28hz (that's 28 pulses per second, times 60 to convert to minutes, and divided by 2 since there are two pulses for each rotation: 29*60/2=870). That's also normal. SYS_FAN pin 2: 12V, pin 3: 96hz. Also Normal. Now onto the IT8712. Pin 7 (FAN_TAC1) measures 0.024V either steady or pulsing at a frequency beyond the range of my meter (999.99khz). Pins 8-12 measure 4.34V either steady or beyond the range of my meter. changing settings in SpeedFan do not affect anything.
What's going on here?
I realize that conductor routing is critical in digital board design, and my wires do not conform to Gigabyte's design plan, but they carry virtually zero current, and should at least provide reasonable voltage and frequency readings. Also, as far as I can tell, I did no damage to the board while soldering, and didn't short anything.
I have verified that the readings and speed controls listed in SpeedFan are from the IT8712, and not from some other device. Do I just have the pins wrong? It seems doubtful that ITE would change the pinout for a device that has been the same for several generations of chips, and I have triple checked that I have soldered onto pins 7-12. Probing each lead without soldering wires to them is risky (again, they're SMALL), and I really don't want to solder a wire to each lead because it's a royal PITA.
Wouldn't the tach inputs from both fan headers be passed straight through to the chip? I can't see any reason to modify the signal in any way before passing them to the chip.
On my main rig, I can measure the PWM pulses on pin 4 of the CPU_FAN header, and so I was expecting to be able to do it somewhere on the IT8712 too.
Do I have the pinout wrong?
I look forward to any comments and questions, and also to suggestions for other places to look for help.
EDIT: I don't know how I overlooked this before, but on my main rig, CHA_FAN1 and CHA_FAN2 are both controllable via IT8721's PWM3 output converted to a linear voltage. The 4th pin on CHA_FAN1 always remains at 5V steady.