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

[Dark Mantis]

Yes apparently it connects to the CPU_FAN connector.

[forumjoe]

Away from the computer, I've now had a good think about this matter, and I reckon I've sussed out basically what's meant to happen. Possibly, this may have already been blindingly obvious to you and others, DM, but not to me. This still leaves a few unknowns, though. Feel free to pick holes in this.

Here's what I think happens. Consider the labeled four pins of the CPU fan header. In the BIOS, when you put the Fan Mode to Voltage, the CPU controls the speed of the CPU fan by varying the voltage on Pin 2 of the header, and the CPU will do this according to its core temperature. Conversely, if you put the Fan Mode to PWM, the voltage on Pin 2 will revert to a fixed value of around +12v and instead a PWM signal will function on Pin 4, controlling the speed instead by that method. Again, that PWM signal will be actioned by the CPU and will be a function of the CPU's core temperature. If you completely disable Smart Fan Control, then PWM is curtailed, as is also Voltage control, and the voltage on Pin 2 will revert to around +12v, making the fan run at full speed all the time.

I suspect that the Sys 2 Fan header works in much the same way. That is, it uses Voltage control on its Pin 2 in order to determine its fan's speed. However, Sys 2 fan header doesn't have any PWM control. And if you disable Smart Fan Control, I reckon the voltage on Pin 2 will again revert to a fixed value of around +12v, making the fan run at maximum speed.

lf my conclusions are correct, it should mean that, if I disable Smart Fan Control and then connect the CPU PWM fan via a Fanmate2 (no, not a PWM Mate, a Fanmate), the Fanmate should allow me to manually control the speed of the CPU fan. There's of course a physical mismatch with the connectors (3-pin to 4-pin) but this can easily be overcome by just snipping away one outer edge of the Fanmate's 3-pin connector. Once the speed is set manually in this way, of course, it won't then vary with temperature but if I'm confident about the likely heat increase and I gauge the required fan speed correctly, I should be able to set the fan's speed somewhere between its minimum and maximum and adequate cooling will result, consistent with achieving minimum noise. This method will avoid having to later install the Easytune utility just to be able to set the speed.

The Sys 2 fan should similarly be manually controllable by a Fanmate but, without experimenting, I'm unsure as to whether disabling Smart Fan Control will also disable the Voltage control of the Sys 2 fan. I suspect that, with Smart Fan Control enabled, the speed of the Sys 2 fan will be determined by the CPU and its sensing of the air temperature in the case. It could well be that, in my situation, the case temperature has not, as yet, been high enough to warrant the Sys 2 fan, ie. the inlet fan, being on. So, maybe that's why the voltage on its Pin 2 is only around +5.4v. (Earlier, DM, you hinted that this might have been the situation). This would be a reasonable way of automatically operating the inlet fan but, frankly, wouldn't be my first choice. As with the CPU fan, I'd rather have complete manual control of it, making it run at a fixed low speed all the time. Again, the manual control solution would seem to be to operate the Sys 2 fan via a Fanmate.
 

[Lsdmeasap]

From what I've noticed, SYS_Fan2 on many boards only runs at half speed (Or voltage) all the time no matter what.

That or it fails to work at all, but that may just be with certain fans or users not knowing about the issues with it.

[Dark Mantis]

I suspect that the Sys 2 Fan header works in much the same way. That is, it uses Voltage control on its Pin 2 in order to determine its fan's speed. However, Sys 2 fan header doesn't have any PWM control. And if you disable Smart Fan Control, I reckon the voltage on Pin 2 will again revert to a fixed value of around +12v, making the fan run at maximum speed.

lf my conclusions are correct, it should mean that, if I disable Smart Fan Control and then connect the CPU PWM fan via a Fanmate2 (no, not a PWM Mate, a Fanmate), the Fanmate should allow me to manually control the speed of the CPU fan. There's of course a physical mismatch with the connectors (3-pin to 4-pin) but this can easily be overcome by just snipping away one outer edge of the Fanmate's 3-pin connector. Once the speed is set manually in this way, of course, it won't then vary with temperature but if I'm confident about the likely heat increase and I gauge the required fan speed correctly, I should be able to set the fan's speed somewhere between its minimum and maximum and adequate cooling will result, consistent with achieving minimum noise. This method will avoid having to later install the Easytune utility just to be able to set the speed.

The Sys 2 fan should similarly be manually controllable by a Fanmate but, without experimenting, I'm unsure as to whether disabling Smart Fan Control will also disable the Voltage control of the Sys 2 fan. I suspect that, with Smart Fan Control enabled, the speed of the Sys 2 fan will be determined by the CPU and its sensing of the air temperature in the case. It could well be that, in my situation, the case temperature has not, as yet, been high enough to warrant the Sys 2 fan, ie. the inlet fan, being on. So, maybe that's why the voltage on its Pin 2 is only around +5.4v. (Earlier, DM, you hinted that this might have been the situation). This would be a reasonable way of automatically operating the inlet fan but, frankly, wouldn't be my first choice. As with the CPU fan, I'd rather have complete manual control of it, making it run at a fixed low speed all the time. Again, the manual control solution would seem to be to operate the Sys 2 fan via a Fanmate.

I think that you are fairly much in the right area and I use a similar  setup on my system for controlling the fans even though  have a watercooling setup. I find the Fanmate2 is a good controller for this situation, and cheap which is a bonus. You shouldn't need to alter the conections at all when installing them as the 3pin connector fits onto the 4 pin header anyway.

[forumjoe]

Yes, what you say is true, DM, but you've forgotten about the connector on the fan's lead. It has to mate with the shrouded, thee male pins of the connector on the lead from the Fanmate. Normally, the two cannot come together because of the keying of the respective connectors. However, if you just snip away the side of the aforementioned shroud with a pair of sidecutters, that problem is dealt with. There's no problem of that kind with the connection into the mobo; you just have to be careful you slide the connector on using only the first three pins of the header.

Re the PWM Mate device, QuietPC have confirmed that it can be run off the mobo header. They've pointed me to this:

http://www.zalman.co.kr/Eng/product/Product_Read.asp?idx=375

Scroll down to the How To Install diagram. The "4-Pin CPU Fan Connector" is the header in the mobo. Zalman stress that this device will only work with a 4-pin PWM fan, not a 3-pin fan.

[forumjoe]

Lsdmeasap,

Thanks for your comment. I'd hitherto thought I was all but alone with the Sys 2 fan problem.

As I stated earlier, I think the stop-gap solution is to run the appropriate fan (in this case, the inlet fan) off the Pwr Fan header instead, if there's one on the board (which fortunately there is, in my case). There's no truck with that one, as it's a straightforward 3-pin affair. What has been concerning me has been the unusual voltages I found on the 4-pin Sys 2 fan's header. They might have indicated something seriously wrong with the bit of circuitry on the board that deals with the control of that header, in which case it would have been important to get to the bottom of it. Ignoring a hardware problem with the mobo could have more serious consequences later. However, I'm inclined to think that maybe those voltages are correct for the circumstances. It's just so annoying that, in the user manual, Gigabyte have not explained in anything like enough detail how these various headers should work. Of course, your suspicion could be well-founded; it might transpire that Gigabyte have indeed dropped a design clanger with regard to the Sys 2 fan header and that the error has been perpetrated through P55 boards other than my US3L.

[Dark Mantis]

Actually I have been trying to gain access to detailed specs and documents for just this kind of problem solving but so far have been unable to obtain them. Hopefully things might change in the future which will make it much easier to pinpoint  the cause.

[Lsdmeasap]

Lsdmeasap,

Thanks for your comment. I'd hitherto thought I was all but alone with the Sys 2 fan problem.

As I stated earlier, I think the stop-gap solution is to run the appropriate fan (in this case, the inlet fan) off the Pwr Fan header instead, if there's one on the board (which fortunately there is, in my case). There's no truck with that one, as it's a straightforward 3-pin affair. What has been concerning me has been the unusual voltages I found on the 4-pin Sys 2 fan's header. They might have indicated something seriously wrong with the bit of circuitry on the board that deals with the control of that header, in which case it would have been important to get to the bottom of it. Ignoring a hardware problem with the mobo could have more serious consequences later. However, I'm inclined to think that maybe those voltages are correct for the circumstances. It's just so annoying that, in the user manual, Gigabyte have not explained in anything like enough detail how these various headers should work. Of course, your suspicion could be well-founded; it might transpire that Gigabyte have indeed dropped a design clanger with regard to the Sys 2 fan header and that the error has been perpetrated through P55 boards other than my US3L.

It's been that way since P35 boards, that header is always the same and I'm not sure why but I do know they do it by design or fault on all of them that I've seen.

I help 100's of Gigabyte users a week, so I see this all the time

[forumjoe]

Really? That seems almost unbelievable. Why on earth don't Gigabyte fix it? They've had more than enough opportunity to do so, as they've produced several revisions of boards. As it stands at present, that particular header is all but unusable and is obviously causing considerable annoyance to users of those series of boards. It can't be doing Gigabyte's reputation much good.

[Lsdmeasap]

I believe it is by design meant to be slower than the normal headers, not sure why?

Maybe Runn3r could comment on this or find out why for us?

[runn3R]

Hi forumjoe

I will check this out for you.
But first I need to know Rev of your MB and results of tests with the latest bios FH.

[forumjoe]

Hello runn3R,

You're lucky to have caught my attention, as I thought this topic was well 'done-and-dusted'.

I'm afraid I can't tell you whether BIOS FH has made any difference to the voltages at the Sys2 fan header, as I'm simply not using that header any more and am using Pwr Fan header instead. It's a pity I didn't think to test Sys2 header shortly after re-flashing the BIOS. I kinda got it stuck in my mind that the voltages at that header would have been set in hardware and therefore unchangeable by a BIOS update but, of course, it's possible that that's not the case.

Unfortunately, I'm now not going to be able to re-test my Sys2 header, as I altered the fan controller wiring in my setup and now it's not going to stretch as far as the Sys2 header. I may possibly be able to do it at a later time, when I've obtained a lead-extender, but who knows when that might be. I'm about to install the OS and a large amount of software.

Maybe some other kind soul on this forum who's installed BIOS FH and who's skilled with a meter will be able to take the measurements instead, or be able to nevertheless say whether FH has made any difference to that header? I'd advise that anyone who measures the voltages at that header does so with the utmost care, as there's very little space for probes and it'd be all too easy to short one pin, and hence voltage, to another. You need to fashion a fine-wire probe (and have a very steady hand) or use an extension lead and suitable probe. And very importantly, you need to take good precautions against static.

My board revn. no. is 2.1.

[Dark Mantis]

I don't know in this particular instance with the FH BIOS version but as far as I have seen this half speed on SYS_FAN2 header is the norm.

[forumjoe]

Yup, so the likelihood is that it's set in hardware, in stone.

[Dark Mantis]

I would expect so. I don't know if it is by design or accident but all the boards I've tested or seen the specs of have been the same.