Is this much heat normal for an FX-55 San Diego?

Guys, this has been bugging me since I got my FX-55 San Diego.

I'm happy with its performance so far as it will do 2.97Ghz on reasonable volts (1.55v), but I'm not comfortable with the temp readings I'm getting from it...they seem way to high for the voltage. If set 1.55v in the bios, smart guargian will read my idle temp at around 42 degrees, but the load temp jumps really high to around 60-62 degrees. I think this chip could do 3ghz with more volts but the temps are holding me back.

I am on water and have a triple fan radiator, will all the other venice and san diego chips I've had I could put way more voltage than this through them with lower temps. I had a 3800+ Venice and 4000+ Sandy at 1.65-1.68v, and the temp never got near this, at first I thought it could be a bad mount so I remounted the waterblock but its exactly the same.

What do you think?

If it was a cheaper CPU I would consider taking the IHS off, but on a new FX-55 it sends a shiver down my spine :p:

Anyone else running FX-55 SD on water or air? What temps do you get at what volts?

Cheers, MadPete.

Load seems too high. Maybe you need to re-seat the block to be sure.

I agree

Thats exactly what my first thought was, which is why I tried reseating the block but it's still the same. Worth another go?

If its still the same after a third seating, do you think maybe the IHS has a bad contact?

Can't say for the SD, but the old clawhammer is a super heater ... I had to go water to keep under 60c at load (50C idle)

edit,

that was not clear,

with water, I'm 42C loaded, but even with an XP-120, I could not stay under 60C loaded with an OC (1.6 volts)

Quote:

---

Originally Posted by madpete

Thats exactly what my first thought was, which is why I tried reseating the block but it's still the same. Worth another go?

If its still the same after a third seating, do you think maybe the IHS has a bad contact?

---

Looks like it :fact:

Is the pump running like it should be? No crap in the water blocking something?

Quote:

---

Originally Posted by timpanogos

Can't say for the SD, but the old clawhammer is a super heater ... I had to go water to keep under 60c at load (50C idle)

edit,

that was not clear,

with water, I'm 42C loaded, but even with an XP-120, I could not stay under 60C loaded with an OC (1.6 volts)

---

So your using an FX-55 and 1.6v on water and your 42 degrees under load right?

Quote:

---

Originally Posted by alexio

Looks like it :fact:

Is the pump running like it should be? No crap in the water blocking something?

---

I got no reason to doubt the pump, flow seems good and the systems been running for a while so worked out any air etc. The temps just seems to rocket right after I installed the FX, its really frustrating.

Temps are too high. My Sandy never went above 49C on air full load overclocked. Do you have antarctica cpu block?


I remember I had Asetek waterchill and my temps were initially really high because water somehow was blocked in radiator. I reassembled the water system, moved radiator up and down and it was fine after that. Not sure if that is in your case. Also you using AS5?

Quote:

---

Originally Posted by kiwi

Temps are too high. My Sandy never went above 49C on air full load overclocked. Do you have antarctica cpu block?


I remember I had Asetek waterchill and my temps were initially really high because water somehow was blocked in radiator. I reassembled the water system, moved radiator up and down and it was fine after that. Not sure if that is in your case. Also you using AS5?

---

Hi yeah using Antarctica waterblock, I will try that with the rad. Where you getting any noise when you had the block cause I'm silent?

Oh and I always use AS5.

yes madpete, I'm actally about 1.66 running 250@2.2.2.5 * 11. Heavy prime95 load will peak at about 45C typically oc load at 42C

idle 34C

Could be one of fluke cpu. My very good clocking venice is 10c higher under all else equal vs the other venices are. I think its just a fluke... (actual temp probably is same)

one other note ... does not really make any sense, but sometime back, I went to the 414-2 bios and that bios showed my cpu 10c higher than the 3-10 bios. I went back to the 310 until recently now on 5-10 - which give the current (and same readings as 3-10) ...

Wow,I thought my temps where high at 52c at full load(prime) on air. I know when I went from 3-10 bios to the 6-18 bios my temps dropped. I haven't tried the new bios yet.

My FX-55 BN usually idles between 39-42 degrees and goes to about 50-52 under full load. It does seem that the FX-55's do tend to put out a bit more heat than the regular San Diego's. Those temps are with both of my 6800GT's in the loop as well running the speeds in my sig.

Well guys, it looks like this chip just runs hot.

I've spent a couple of hours this morning reseating the the waterblock twice, checking the rad and flow, tinkering with fan speed etc but nothing helped.

I have it @ 2.92ghz with 1.450v x 104% in the bios (which is 1.508v by my calculations) and I'm idling at 36-38 degrees, load at 50-52 degrees which I personally think is a bit poor really for highend water. :(

wow, I can not get mine to 2.92 with 1.7v I can run 2880 but no post at about 2915 ... so at least you have that going for you

Yeah your right man, I am please with how she clocks.

Since the re-seat, I seem to be holdin up a little better with...still think she runs hot tho. Just passed OCCT @ 2.95ghz using 1.475v x 104% = 1.53v - 52-53 degrees load.

that does sound really high temps especially with a tripple rad i had the antarctica kit with a double rad on my old prescott system and i used to idle around 29c load was never over 40c fully overclocked .

When ocd, 55BN run hotter than 55AS even with less vcore. Typicaly, if my AS/130nm @1.8V than BN will be equivalent of 1.68V and the load temp. on my vapo LS will be the same @-43C.....

Yeah, remember that the core is smaller now on the new FX-55's so less surface area to conduct heat underneath the IHS. So the core is going to retain a little more heat I would think.

Have you ever taken off a IHS ? Just be careful and won't damage your chip. Trust me it's really easy, just be sure your razorblade is REALLY sharp!

Quote:

---

Originally Posted by sai

Have you ever taken off a IHS ? Just be careful and won't damage your chip. Trust me it's really easy, just be sure your razorblade is REALLY sharp!

---

Sai, no never took one off...wish I had a cheap CPU to practice on

Take one of these

http://www.seilnacht.com/Lexikon/rasier.JPG

and take a marker and paint a mark on the blade to limit how deep you cut under the IHS. This way you can be sure not to destroy anything. JUST TAKE YOUR TIME. Don't apply much pressure while cutting. It's better to cut 100 times than 1 time to deep. ;)

Edit: Just use one corner of the blade, not the whole side, that makes it MUCH easier
BTW I wish I had the money to buy a FX55 :slobber:

This is my friends Sandiego 3700+ get at 2.9GHz with 1.7V vcore :slap:
45C idle.... not sure about load temp... :(
btw on air cooling
just for comparison.. oh its on abit board

http://img45.echo.cx/img45/1268/gkinn13ap.jpg

Is the waterblock warm to the touch? You're sure its well mounted?

saratoga, definatley mounted correctly, I have reseated it four times in total and everytime the temps are the same. I think either all these chips run very hot or it has a bad connection between IHS and the chip, thats why I am considering Sai's suggestion on removing the IHS.

I think that if I removed it, it might only be to reseat it with a good connection. I've heard that if you do this right you can't tell its been done from the outside, so warranty wise you might still be ok?

Whats the general concensus on how dangerous this is? Do you guys think its worth it?

It isn't that dangerous but offcourse you will always lose your warranty. It can't be done without some markers indicating you've done it.

The core itself is pretty strong so you wouldn't crush your chip easily.

First you should try mounting another CPU and check if the temps go down again (confirming that it is indeed the CPU that's causing the problem).

Also try and measure the temp with some kind of probe, never trust software reading of tempratures.

If you apply more volts does the heat go up and does the CPU crash when stressing it for like 5 mins (indicating that temprature is the problem)

Man if you want an adrenaline trip, try dissecting a 600 quid CPU with a razor blade :p:

I did it though, so far it seems to have made around an 7-8 degrees difference, which might not sound a lot but I think that’s a fair drop. I still think these chips run hot, but I must have had a bad connection too.

I also had my Mushkin Redline XP4000 back from RMA today :banana:

Lots of testing and benching ahead :cool:

Quote:

---

Originally Posted by madpete

Man if you want an adrenaline trip, try dissecting a 600 quid CPU with a razor blade :p:

---

LOL, no doubt. That's a nice temp drop you got however....always makes me re-think it when I see these reuults.

I have a tricky SS mounting that keeps me from trying...

good job pete! giving me the courage to try the same!

Well guys, just to wrap this up I thought I'd give my results now that the AS5 has settled...

Before operation: max prime/occt stable speed 2.97ghz with 1.55v - temps @ around 38-40 degrees idle / 58-60 load.

After operation: max prime /occt stable speed 2.98ghz with 1.5v - temps at 34-36 degrees idle 46-48 load. I tried 1.55v again just to see how it compared, load temps never went over 50 degrees, though the CPU will not clock any higher without becoming unstable.

So as you can see I have gained some speed and stability, using slightly less voltage and my temps have dropped a fair bit. A good result.

Just wanted to say thanks to everyone for their advice and encouragement :toast: