[SOLVED] [Solved] Help needed with X58 Classified 3 straight to FF Post code (CPU socket bent pins)

[gt_33]

I built this system in early 2010 and since then I've just upgraded the GPUs. Every year I used to take everything apart (except CPU) and clean it. This time I wanted to change my 950 for a Xeon and I removed the CPU as well. I was careless and did a noobish mistake and ended up with several bent pins on the socket. One thing led to another, tried to restore them in a hurry, but no joy.

So, the machine doesn't post and goes straight to FF. At the moment I am testing with a PSU, the mobo, the CPU and a fanless heatsink, nothing else connected. Yesterday I took my time to restore the pins to the best of my ability, using a magnifying glass and a suitable precision screwdriver, but still no luck. I've tried clearing the CMOS (it has a button for this), I've tried removing the battery, I've tried with the mobo on a non conductive surface to make sure nothing is "shorting" on the back...

At the moment I am at a place where I am not sure if it's worth trying any more, because the machine is stuck at FF even with no CPU installed. Is this normal or is this an indication that I have fried the mobo? I mean, is there not another error code for a missing CPU?

Today I will test my CPUs (i7 950 and a Xeon) and PSU with another mobo to make sure they are ok, but other than that I don't know what to do. Maybe look at the pins carefully one more time and make sure they are nicely aligned and that none of them are in contact with another. I would really appreciate any help.

P.S. I am sure I haven't done a perfect job with the bent pins, considering I am 41yo and my vision is not 20/20 as it used to be, but I am starting to doubt they are the only problem at this point. Thanks in advance.

 

[Darkbreeze]

Fanless heatsink is a problem, unless you have already configured the BIOS to ignore the lack of a signal on the CPU_FAN header. Otherwise, it will not stay running. It will throw a code and shut down to prevent overheating. If you have previously configured that setting, if your BIOS even HAS that setting (Some do, some don't) then obviously that's not a problem. I'd probably connect a fan to the CPU_FAN header just to be sure though, even sitting off to the side should be fine so long as it gets a reference signal.

Also, make sure the EPS power cable is connected to the motherboard as well, in addition to the 24 pin ATX cable.

It is probably also a good idea to pull the CMOS battery for five minutes, after changing CPUs, to hard reset the CMOS BIOS settings.

 

[gt_33]

I had the CPU_FAN header thing in mind and I don't think this is an issue, because I believe this mobo doesn't have such predefined behavior.

The CPU_FAN header is the 3rd pin if I am not mistaken. Meaning two pins for power and a 3rd for status. The fan I was using on this was a 2pin fan and the machine was working just fine. So I believe that it doesn't check for a fan presence in order to boot.

Still, I will try with a fan on, just in case it checks in some other manner. Concerning the battery, I've tried with leaving it out for 10min, but no joy. At the moment I am trying with the battery out.

 

[Darkbreeze]

Do any of the pins in the socket feel "loose" at all. Sometimes they don't just bend, they break, and they break down below where you can see them or do anything about it. You can usually tell by barely nudging the pin with a pair of tweezers. Not enough to bend it again or more, but just enough to tell if it's just being kept in the board by the solder beneath the surface or some such.

 

[gt_33]

I don't think there are any broken pins and they shouldn't, because there wasn't any real forceful action involved. Still, something to keep in mind I guess when I'll recheck the pins in a few hours.

What about the FF post code even with the CPU unplugged? Is this normal or is it one of the following? a) bent pins shorting, b) fried mobo?

 

[Darkbreeze]

This post seems to indicate that your board DOES require a cooler to be connected, like I was saying. Also, what kind of fan are you using that only has two pins? What is your cooler model? You should be using a four pin PWM fan for that system regardless of what cooler is used.

https://forums.evga.com/FF-error-code-m1998166.aspx

 

[gt_33]

I got confused then, 3 pins instead of 4. I just remember that the CPU fan header has an extra pin, which I wasn't utilizing, because the fans I am using on my heatsink don't have that extra pin, but the system was working just fine. My heatsink is a Prolimatech Megahalems Rev.B. I was using 2 standard fans on it, the ones we usually use as system fans (3pin ones).

Anyhow, I'll definitely give it a go with a fan installed, with a 4 pin one if possible (I think I have one), but it should work just fine with 3pin ones too, 'cause it used to.

P.S. I had read that post and did the things it says with no luck. Except CPU fan plugged in, so I shall try.

 

[Darkbreeze]

Yeah, if your fan has three pins, then it will WORK, just not properly. It will be at high speed at all times because SFAIK none of those older boards has the ability to switch from PWM to DC mode on it's four pin fan headers. That means that all three pin fans connected to them will operate at full speed at all times, which will still give an RPM signal to the controller, it just doesn't allow for speed control. There might be a few exceptions to that, and newer motherboards certainly have the option to switch from DC to PWM and back, based on what kind of fan you have attached, in the BIOS settings. Pretty much started around the time Z170 was released. Might be a FEW high end boards that had that feature prior to that, but if so, I haven't seen one.

So in your case I'd think that if it worked before with a three pin fan, even having that connected to that header while trying to POST should be good enough to ensure it's not the issue at hand. If you HAVE a four pin fan though, then at least temporarily that would probably be better.

I agree, it's likely not the issue, but it's imperative to remove potential issues in order to find what IS the issue, by process of elimination.

 

[gt_33]

Update: The pins are bent in more than one ways, which makes it hard. So, even when aligning the pins was not that hard, some still stick out a bit. I was constantly testing with the i7-950 and kept getting FF (straight to FF), so in a last resort attempt I thought to give the Xeon (X5675) a go. So I did and it posted. What a pleasant surprise.



So, I tested for a bit with no overclock and it booted to windows, run a game fine. Temps were ok, even with bad thermal paste coating and a heatsink that can't be tight. The heatsink needs to be relatively loose, ot it won't post (the usual straight to FF) and I have to take the heatsink apart, re-seat the CPU and then it works. Just re-seating the CPU makes it work, I tried it 6-7 times and it reliably works every time.



I guess that this means that 1 or more pins are shorting or not making good contact when there's a given amount of pressure on the CPU or something. I am pretty sure nothing is shorting on the back of the motherboard.



The problem atm is that when I am attaching the 2 fans on my bulky heatsink, I might apply extra pressure to the socket or the heatsink might move a bit and this means no POST (straight to FF) and then I have to take it apart and re-seat the CPU. So, I am thinking of buying a less bulky, easier to mount heatsink, like a hybrid cooling system and try to be gentle with it and see what happens.



This, or I should get into the Socket again and see if I can align the pins a bit better. Still, since it seems to work well, when it works, I am a bit skeptical about tampering with the pins again.

 

[Darkbreeze]

You need to get a new board is what you need to do. With multiple bent pins, you are rolling the dice on the idea that you will not encounter a problem that results in your memory, or cpu, or graphics card, or anything else connected to the PCI bus that runs through the CPU, being damaged or ruined if one of those pins shorts out on something.

Personally, I don't think it's worth the risk if you value the rest of your hardware. If you have the run the system with heatsink that is barely attached, that's a fools game if you ask me. I think it's going to turn out badly at some point.

There are plenty of used boards available at somewhat reasonable prices considering the age of the platform.

https://www.ebay.com/sch/i.html?_fr...board.TRS0&_nkw=LGA+1366+motherboard&_sacat=0

 

[gt_33]

You make sense, but I am a stubborn person and I want to make this work. Also, I love my current board.

Buying a used board means I'd have to buy from EU to avoid a customs' tax. I checked the prices, I'd have to give like 140 Euros for an inferior board.

 

[Darkbreeze]

I understand the complexities of the dilemma, however, that concern regarding prices is going to increase dramatically if you start adding other parts to the list because your stubborness instigates other failures. If you cannot get the board to work with any standard or aftermarket heatsink correctly mounted on it, then it is for all intents and purposes done. There is really no way to argue about that part.

I get not wanting to spend money on something that is mostly still ok, but much like a car that has the frame bent by a front wheel where it cannot be trued in order to bring the alignment and tracking back into line, the vehicle itself is still 95% intact and functional but the cost of having the frame straightened (IF even possible) generally exceeds the worth of the vehicle, IF like your board, the vehicle was 10 years old and had a significant number of miles on it already.

But hey, I'm with you. It's your hardware and that's the way you want to play it, so all I can say is that given your current situation the only thing you can really try that would be both acceptable and make any kind of sense, would be using the stock cooler on it. If the stock cooler is too heavy, then we are done here. There is no cooler lighter than the stock cooler that can or should be used with the CPU. If it requires something less substantial than the stock cooler, you simply cannot resolve the situation in this way, and probably can't anyhow, but there it is.

 

[gt_33]

I haven't updated because I was away of home for 10 days. I repaired the pins to the best of my ability and the CPU works when it's not pressed too tight on the socket. Restoring the pins on the Y-axis proved very difficult for me, so I did my best and then gave it a rest, since it posts, in fear of breaking them down by fiddling too much with them.



The CPU posts when the heatsink is not really tight and it works fine as long as I am careful with the socket (no extra force on it on any axis). I gave it a few quick tries and it overclocked easy to 4.3 Ghz, at 1.34v, so I guess all pins are fine and it draws enough power. I have 2 GPUs and they work just fine. The only thing I am not 100% sure is the RAM, because BIOS is displaying 8gb, while cpuz, aida64, etc are displaying the correct amount (12gb). Maybe it's a BIOS display error, but I'll test thoroughly and see what I get, because I think that while fiddling with the overclock, I saw "dual channel ddr" with some temp bios settings and then triple channel.



Anyway, the heatsink wont budge and temps are fine (on air), so all this ordeal was an acceptable trade off for my noobish mistake. Thanks all for the help and support to keep on trying.

 

[Darkbreeze]

CPU/motherboard could be related to the memory problem of course. Make sure they are installed in the correct slots though.