GA-P55-US3L: strange fan header voltages and help needed with some BIOS settings

[forumjoe]

Mobo: GA-P55-US3L.
Bios version: FE.
CPU: Intel i5-650 Clarkdale.

This is a new topic, following on from my earlier one entitled GA-P55-US3L: operating a PWM fan from a 3-pin arrangement. The CPU fan I'd bought turned out to be faulty. I've been delivered a replacement CPU (PWM) fan and have now got it running, albeit at top speed only. Thus, I've finally been able to keep the PC powered on and been able to inspect and configure the BIOS (no OS installed as yet).

More on the unmoveable top speed of the CPU fan later, but the other 4-pin header on this mobo, called the Sys 2 Fan header, is showing some odd open-circuit voltages on its pins. Here's what I've measured with a meter:

Pin 1, labeled GND: 0v
Pin 2, labeled +12v/SPEED CONTROL: +5.38v
Pin 3, labeled SENSE: +10.58v
Pin 4, labeled RESERVED: +5.08v.

The other day I connected my 3-pin inlet fan to it, obviously being careful to use only the first three pins of the header, but the fan hadn't run. Now I know why! But why on earth should the header be producing these voltages? Are they somehow linked to the Smart Fan Control, in the same way that the CPU Fan header is? Or are some 4-pin headers simply unsuited to operating 3-pin fans? I've noticed that Pin 2 of the Sys 2 fan header is labeled as +12v/SPEED CONTROL in the same way that the CPU fan header is, but Sys Fan 1 header and Power Fan header have Pin 2 labelled only as +12v.

As for the CPU Fan which is now running, there's nothing in the Smart Fan Control in the BIOS (version FE) that allows me to manually control the fan's speed. The settings are AUTO, VOLTAGE, and PWM. I've tried all three but to no effect. Is the only way to control the speed to install the Smart 6 utility, once Windows is there? Or am I missing a trick somewhere?

Now to the BIOS itself. Anyone know what QPI/Vtt Voltage is? (currently, 1.1v). Just interested to know the semantics.

Anyone know what PCH Core Voltage is? (currently, 1.050v). And CPU PLL? (1.800v).

Why isn't a Battery Voltage given? Surely that's a serious omission, as at some stage in the future you might need to change the battery and therefore, during the life of the PC, you need to keep an eye on its voltage? How can I suggest to Gigabyte that they add that to a future BIOS revision?

What is PCH Temperature? (currently, around 39 degrees C).

I was surprised to find that the BIOS has put the first of my two optical drives on the Slave of the SATA Channel 0, the single hard drive being the Master on that same channel. How can I reorganise the SATA physical connections so that the hard drive is on a channel on its own? The SATA2 group of ports are SATA2 0 - 5. I've currently got the SATA connections on them as: 0: HDD, 1: CDROM drive, 2: DVD/RW drive. I assumed that because these are SATA connections there was no concept of master and slave. BTW, I'll be using them in IDE Native mode. I'll not be using RAID or AHCI.

Finally, anyone know what the safe, average value for the core voltage of an i5-650 is? According to some CPU tables I've seen, it's somewhere between 0.65v (min core voltage) and 1.4v (max core voltage). The speed of the CPU is currently being reported as 3333 MHz. I don't need it that high, and the nominal speed of this CPU is 3.2GHz, so I'd like to get it down to 3.200 GHz, by adjusting the frequency and voltage downward slightly, to something that will give really safe working, with the CPU dissipating only its nominally-designed-for amount of heat. The core voltage at present is being reported as 1.220v.

Sequel: Using a 3-pin fan extension lead, I've now measured the on-load voltages on Pins 2 and 3 of the Sys 2 Fan header, ie. with the inlet fan attached. They're 5.38v and 5.08v respectively. However, the inlet fan (nothing special about it; it's a 3-pin fan) now runs (and very fast), even though the basic voltage supplied to it is still only around 5v. This is getting annoying.

Meanwhile, as the system boots I'm getting a continuous tone as a warning signal. I think that's indicating that the graphics card is possibly misaligned, though the monitor still shows everything okay. Looks like I'll have to take the graphics card out and reseat it.

Sequel to sequel: Does a continuous tone indicate a problem with the graphics card, or a problem with the PSU or the supplies generated on the mobo? When I google for it, I get all sorts of different answers. The POST sequence gives a single beep, suggesting that the voltages outputted by the PSU are all good. It's after that that the continuous tone starts. 

[Dark Mantis]

Meanwhile, as the system boots I'm getting a continuous tone as a warning signal. I think that's indicating that the graphics card is possibly misaligned, though the monitor still shows everything okay. Looks like I'll have to take the graphics card out and reseat it. 

If you are getting a continuous tone I would think that it is more likely to be a motherboard alarm, probably for the fan speed.I would advise youy to check in the BIOS and see which one it is.

Sequel: Using a 3-pin fan extension lead, I've now measured the on-load voltages on Pins 2 and 3 of the Sys 2 Fan header, ie. with the inlet fan attached. They're 5.38v and 5.08v respectively. However, the inlet fan (nothing special about it; it's a 3-pin fan) now runs (and very fast), even though the basic voltage supplied to it is still only around 5v. This is getting annoying.

That is very unusaul to say the least as most fans are 12V and won't even start up at anything under 7V. They may be coerced into running at less than 7V once turning but even then would be at slow speed. What are the same pins measured at no load ie fan disconnected?

Now to the BIOS itself. Anyone know what QPI/Vtt Voltage is? (currently, 1.1v). Just interested to know the semantics.

QPI is the acronym for Quick Path Interconnect and is what has now taken over from the old Front Side bus. NB to CPU.

I will answer the rest of your questions tomorrow.

[forumjoe]

Before seeing your reply, DM, I realised the reason for the long continous tone. In the BIOS, I'd set the then-used fans to the fail alarmable state. So, your reply above on that matter was obviously along the right lines. I strongly suspect that if I either put the inlet fan back on the Pwr fan header or I disable the Pwr fan alarm, the tone will go away. Mystery solved. Nothing to do with the graphics card.

Yes, those on-load voltages at the Sys 2 Fan header are nothing what you'd expect, are they? The off-load voltages of that header are as I reported above, namely this:

Pin 1, labeled GND: 0v
Pin 2, labeled +12v/SPEED CONTROL: +5.38v
Pin 3, labeled SENSE: +10.58v
Pin 4, labeled RESERVED: +5.08v.

So, they're very much the same as the on-load voltages.

Thanks for letting me know what QPI stands for. The QPI, or FrontSide Bus, voltage on my mobo would appear to be just 1.1v. That seems a low value to me.

I mentioned the battery above. I think it's the case that most mobos actually trickle-charge the battery, once the PSU has been attached. So, initially the battery voltage will be low, but in time will creep up, levelling out. To be able to monitor this happening makes it even more important to have the BIOS tell you what the battery voltage is. This is the first mobo I've ever come across where the battery voltage hasn't been shown in the BIOS.

[forumjoe]

I think I've realised how I can get the CPU frequency down a bit. In the BIOS, I already had CIA2 disabled but I've noticed that there's also a Performance Enhance setting in another part of the BIOS, which is currently set to Turbo. I suspect that if I change that to Standard the frequency will more closely align with 3.2Ghz.

The most worrying problem at present is the non-functioning of the Sys 2 fan header. With what I've found, it's almost as though Gigabyte have put the wrong voltages on the pins of that header - about 11v on Pin 3 instead of it being on Pin2. Well, it looks like it's either that or somehow the bit of circuitry on the mobo that handles the sys 2 header has been damaged, and possibly irrepairably so.

[Dark Mantis]

The most worrying problem at present is the non-functioning of the Sys 2 fan header. With what I've found, it's almost as though Gigabyte have put the wrong voltages on the pins of that header - about 11v on Pin 3 instead of it being on Pin2. Well, it looks like it's either that or somehow the bit of circuitry on the mobo that handles the sys 2 header has been damaged, and possibly irrepairably so.

I have been trying to envisage how the voltages you have given could be found and the only way I came up with is if you are reading them from the wrong side. Please don't take offence at this suggestion but is there any chance you could be looking at the header from the wrong end?

[forumjoe]

A good point, but all the fan header pinniings in the manual show Pin 1 as being the lowermost pin, with the keying of the header then being on the righthand side. Pin 1 is normally Gnd, isn't it, with +12v appearing on pin 2? I think it's a convention with headers that, with the keying piece on the right, pin 1 will always be the lowermost pin. That's vertical orientation, of course.

Theoretically, I don't think it's possible to mis-connect a fan or a fan controller, due to the built-in keying of the respective connectors. On the other hand, there'd be nothing stopping Gigabyte from making a mistake in the design of the board and, as a consequence, feeding the incorrect voltages to the header. They do, after all, produce revisions of their boards, and though this one is merely the latest available in Europe, it's not the very latest revision. Personally, though, I think it's more likely that one of the regulator components in that precise area of the board has blown.

Maybe it's also possible that Gigabyte have set up the Sys fan 2 header on this particular board as a PWM header? Maybe that's why such odd voltages are on the pins? If you, for the sake of argument, assume that something daft like that has been done, then if you just attach a 3-pin fan, as I've done, the chances are that the fan wouldn't run, as Pin 4 is open.

I could ignore using this header and instead use the 3-pin Pwr fan header. That works flawlessly. But if one or more board components have blown I wouldn't want to continue using the board, as it might overheat in time or give rise to other failure conditions.

Here's a thought, though. Why did Gigabyte make this header a 4-pin one? According to the manual, pin 4 of it is merely 'reserved' (but it's actually got 5v on it at present). Before PWM fans came on the scene, why was it necessary to have 4-pin fan headers? I don't see the logic of it.

[Dark Mantis]

Right well I thought that it was time I dived in there with my trusty digital meter and took some readings. Don't forget that my board is different to yours (GA-X58A-UD7) but the fan headers should give the same readouts approximately. These readings were at a no load condition.

Yours
Pin 1, labeled GND: 0v
Pin 2, labeled +12v/SPEED CONTROL: +5.38v
Pin 3, labeled SENSE: +10.58v
Pin 4, labeled RESERVED: +5.08v.

Mine
Pin 1, labeled GND: 0v (Ground)
Pin 2, labeled +12v/SPEED CONTROL: +7.99v
Pin 3, labeled SENSE: +10.46v
Pin 4, labeled RESERVED: +5.01v.

As you can see there is not so much difference at all. I would guess that the speed control (pin2) is the 7v that I said most fans need as a starting voltage and then the sense (pin3) would fluctuate depending on the feedback from the "tacho".
I would suggest from these figures that your Speed control(pin2) voltage is too low and that is why your fan wouldn't start.

Hope this helps somewhat.

[forumjoe]

I've just found a forum question-and-answer elsewhere that relates to this header problem. See pages 1 and 2 of this, especially page 2:

http://forums.tweaktown.com/gigabyte/41371-3-pin-fan-4-pin-fan-header.html

It's considering a somewhat older Gigabyte board but demonstrates the kind of problem I have. My mobo manual simply doesn't explain the necessary configuring of the BIOS for this in sufficient detail and how you subsequently achieve control of the speed of whatever type of fan you're using.

According to one or two of those respondents on that tweaktown forum, if you use a 3-pin fan on a 4-pin header on a Gigabyte mobo, you have to set up the Smart Fan Mode in the BIOS to 'Voltage'. The implication is that if you need to do it for a Sys fan, it will also cause the CPU fan to operate in the self same mode. In effect, those respondents are saying that, on a mobo like mine, the CPU fan control doubles up with that to the Sys 2 fan. You can't run those two fans with independent control. Thus, if you set the mode for the CPU fan to PWM and you connect a 3-pin fan to the Sys 2 fan header, the CPU fan will run but the 3-pin fan on Sys 2 won't (as the 4th-pin connection on that header can't be used by the 3-pin fan and won't be connected through anyway).

Maybe there's something in this. I'll have to explore my machine a bit more. Perhaps, in the end, I'll just have to use the Pwr fan header for that 3-pin inlet fan.

My 4-pin PWM CPU fan, incidentally, is running at just over half its maximum speed, that being about 1440 rpm so, for that fan, clearly some form of speed control is automatically happening. I'd like to manually drop that to about the fan's minimum speed, which is 900 rpm. But I don't want to have to install any software just to do that. Consequently, I've been wondering whether one of the latest devices from Zalman, the so-called PWM Mate, will do that job instead. Anyone had any experience of using a PWM Mate controller with a 4-pin PWM fan?

[Dark Mantis]

Yes interesting . I am at present using a couple of the older versions the Zalman Fanmate2 s but of course they are not PWM controllers.
I take it the PWM MATE Fan Controller is fed from a 4 pin molex connector directly from the PSU.

[forumjoe]

DM,

Many thanks for taking the trouble to measure your own header voltages. The results are interesting. At the time you took those readings, how exactly was your Smart Fan Control set up in the BIOS? Don't forget there's both CPU Smart Fan Control and then CPU Smart Fan Mode. Is it conceivable that the voltage on Pin 2 will be inadequate on the Sys 2 header if the CPU Fan Mode is set to PWM rather than Voltage?

I'm finding this all about as clear as mud, at the moment.

[Dark Mantis]

I will check the settings in the BIOS when I next reboot and then post here.

[forumjoe]

Re the PWM Mate, no I beliieve the controller is fed from the mobo's PWM header. Zalman market a PWM cooler/fan assembly of their own which can accommodate a PWM MAte Controller directly on the assembly. But a short video I've seen at QuietPC.com suggests that the PWM Mate can otherwise be connected by the supplied adaptor lead into the mobo's PWM header. See here:

http://www.quietpc.com/gb-en-gbp/products/caseaccessories/pwmmate

[Dark Mantis]

Yes I had already looked at that but it doesn't actually show where the fan/controller is attached for the power. You may well be right.

[forumjoe]

I don't know whether it's just coincidence but, as we've seen, the voltage on Pin 2 (+12v/SPEED CONTROL) of my Sys 2 fan header is roughly half that of the voltage on Pin 3 (SENSE). It happens that the speed of my CPU fan is currently running at around half its potential.

So, could there actually be a tie-up between the functioning of the CPU fan and that of any Sys 2 fan? In other words, could it be that they are NOT independent and that you have to be careful as to which fan mode you set up in the BIOS?  Perhaps if you try to use a PWM CPU fan and also a non-PWM fan on 4-pin headers, only one of the two fans is ever likely to run, at certain CPU temperatures?

[forumjoe]

DM,

According to this, the PWM Mate gets its power from a lead that plugs into the mobo. It must be some sort of split lead that's supplied with it, if that's the case, as the Mate controller seems to have itself only a 4-pin connection and yet it has to not only marry up with the fan's connector but also to the mobo's connector, be that the PWM mobo header or a spare power lead from the PC's PSU.

http://www.techpowerup.com/107982/Zalman_Releases_PWM_Mate_Fan_Controller.html