Dell Latitude E6400 - Overheating issues with minute tasks such as playing flash videos for extended periods?

I just found out, that it is no more needed to shutdown the system to restore the speed. I just need to stop / close resources intensive tasks for few minutes, wait for the hw to cool down and the system is again running at full speed. I can't say it is new in bios ver 14, because I didn't tested 13 this way.

I know, it's not a solution, but I was surprised.

to JoeB7: I'm also thinking it's the chipset what is overheating. The chipset is on the inner side of the cooler plate (lets say far-far away from the fan) and the only contact with the cooler plate itself is managed by small foam pad. I'm thinking of replacing this with copper plate - but really, am I supposed to do this ??? Or replace the fan with more powerful one (but I don't know how).

btw: I bought a cooling pad. This is really help for this notebook.

DELL escalation expert Europe just called me. The are aware due to forums and complaints that there must be a problem (most likely BIOS related) in the E-series. Similar problem was solved in the M Series with an BIOS update. Problem was again escalated to the product management of the Latitude E series. Hopefully there will be news sooner or later. Some of you are waiting much longer than me - I saw.

@ivka57: I still believe it's a wrong temperature information or interpretation of the chipset (northbridge) that hosts (imho) the GPU as well. I won't touch the system during the warranty period. And a cooling pad is a good solution - however, from a quality manufacturer I'd like to expect that the system just works fine straight out of the box. It's not a no-name thingy.

So - lets hope for a cure some day involving a few clicks updating software/drivers/bios...

I have this problem on my Ubuntu setup as well. I am using the E Dock with dual monitors, NVidia graphics and have a 2.94GHz duo processor.

I am working on mission critical applications and started having these performance problems under Vista so I changed to Ubuntu hoping that it was a Vista problem.

I noticed that the processor cores are running at 800MHz when this happens so I installed a utility to change the speed to 2.94GHz again. That helps a little but it seems like the bus is also down to 800MHz.

I ran some benchmarks (hardinfo) that indicate that the CPU is 50%-75% slower than a Centrino 1.5GHz (even though the CPU speed is 2.94GHz).

Before the problem starts (it takes about 2 hours) it is performing about 120% faster than a Centrino.

I got my motherboard replaced yesterday but that didn't change anything.

I have the Dell Latitude E6500 connected to two external monitors through E-port replicator, although these problems also occurred with external monitors plugged directly into the laptop.   I had the same problems with CPU gradually increasing to the point that the system was unusable.  (This does NOT happen when using the laptop monitor only.)  This occurred when using the Slingbox, so I thought it was a GPU problem.

Called Dell and they agreed to send out a new motherboard.  After reading this forum, I did more troubleshooting and noticed that under the Resource Overview, (Right click - Computer\Manage\Reliability and Performance) when the graphics began to act up, the Maximum Frequency for the CPU dropped to 31%.  This was after running for about 10 minutes, and the cooling fan running on high.  It's like a switch was tripped.

The good news is that  I found the fix for me.  First, I changed the Bios back to A03.  Also using the original Nvidia video driver.  Newer drivers have trouble recognizing the twin external monitors properly.  Second, I removed the back and using compressed air, blew out a TON of dust.  I could not believe how much dirt was in there since the system is only several months old.

After that, I connected the laptop to the replicator with the back panel removed.  No problems; even with Sling, backup software AND Google Earth running together.  The fan turned up to high, but later than before, and the CPU did not cycle down like before.  Then I restarted with the back panel on - similar results.  Problem solved.

I tried the same test the next morning with no problems.  Called Dell and cancelled the motherboard replacement.  I hope this helps.  Cheers! 

Thanks for the dust spraying info. It might be related, yes, however, I have two of that E6400 here. One is just 8 weeks old, got new MOBO and heatsink(s) some weeks ago, the other one is brand new. Both have that issue.

I've investigated a bit more and hope some hardware experts can confirm of what I think:

The problems happen more often when connected to the E-Dock and/or heavy GPU use undocked. DELL has redesigned the docking station connection since the E-series. The D-series was a "real" docking scenario, where you had even some different hardware profiles for docked and undocked.

Now, the docking is like a USB connection. When you dock/undock the device it doesn't change hw profiles nor do you see the undock button in the Windows start menu, nor do you have an eject button to prepare the undock like you had with the D-series docking stations/port replicators.

I assume that a lot of data traffic is now handled via USB when the E6400 is connected to the E-dock. Therefore the northbridge chipset, that handles (imho) the traffic for the USB ports is already in heavy use (likely).

That might explain, why I managed to slow down the system when copying a very large file archive to an external USB2.0 hdd (connected to the E-Dock USB port). No GPU use at all but loads of traffic on the chipset. The processors stood cool, no downclocking of them, but most likely downclocking of the FSB. The system was dead slow and 100% usage after 10 mins of copying files to my external hdd.

Notice that the Intel GMA GPUs are no real separate GPUs and are built into the northbridge / chipset as well. So GPU usage high traffic on the northbridge might lead to similar effects?

Therefore, I assume (and hope experts can join) the problem lies in the chipset and the temperature there. Either wrong temp or wrong handling of the temp info from the sensors (thermal table -> bios).

Hope that makes sense. I'm no HW designer at all. I'm repeating myself: hopefully dell will find a solution soon.

PS: somewhere I've read that all of the trouble is worse when the system has been waken up from standby/hibernate rather than a new booted item. So I suggest DELL to test it that way either.

I downgraded to A03. I will report later.

mk

The dust is not the issue, as I took out brand new E6400 out of the box and "achieved" "overheating". Both quoted, as we all know it does not overheat in fact. The air comming out of the notebook is cool. The hypothesis about northbridge/chipset/bus looks promissing. As there are problems with separate GPUs and integrated as well. Any ideas how to check all sensors in the notebook? More than SpeedFan could? Maybe some linux app... Maybe there's some sensor we don't know.

Ivka: Any progress with A03? Thx.

Sidenote: does any one of us use the DCP bloatware? (I know it's win only, that it can somehow interact with Bios from OS and override some stuff.)

I8kfangui will read the CPU, GPU RAM and Chipset temps for the M1530, and probably the M1330. Initially, I was also able to read fan speed, but after some time, the fan disappeared. Probably after a BIOS update.

Hi again.

I am now using my laptop out of the dock connected to one external monitor. The fan goes crazy after 30 mins, but downclocking is rare and fixed with shutting down for 15 minutes. It will then last for the remainder of the day.

I have contacted Dell support but they insist that I reinstall Vista and download some process explorer program from Microsoft (I am running Ubuntu now) in order for them to debug the problem!!!

Has anyone here gotten a sensible reply from Dell yet?

BTW. It seems that they still believe that this is a software problem even though I have exactly the same issues in Vista and Ubuntu. Also when I enter the Bios utility when the downclocking occurs, it behaves exactly like the OSs. Slow and unresponsive.

With A03 so far so good. I didn't test with E-Dock, I'm out of my office for whole week, but I installed BOINC running 100% of CPUs, with 10x firefox and flash intensive webs, virtualbox (virtual windows XP running IE and office apps). During night, I closed the lid. Until now, I experienced no slowdown.

Regarding the DELL support, they also forced me to test with DELL diag utility and the behavior was the same as JONATANEINARSSON explained. After slowdown, reboot to diag util, no errors, but very slow and unresponsive. This means, this is HW problem, now SW. Even after the system cooled down, the speed wasn't restored. But this changed with A14, the speed is actually restored after system cools down, without shutdown / start cycle.

mk

Going with A03, my current mobo in system was stock A12, hope I won't ruin it. =)) Will report!

UPDATE: A03 did not solve anything for me. So A03 BIOS is a no go.

fantomex: I'm sorry to hear that.

today I was all day working on E-Dock with two monitors, using my usual apps (the same, I experienced the slowdowns with in the past). at the evening, before shutdown I performed the benchmark with BOINC. I got approx 2300 / 6900 (floating point speed / integer speed). this is exactly the same value as I get after start in the morning.

I hope, that was not just a lucky day and someone will also confirm the A03 approach?

mk

BOINC is a cpu-intensive application.  I did the same with Prime95 and never experienced any slowdowns, using A14 BIOS.    It looks like these slowdowns are happening when CPU load is relatively low (i.e. CPU temp is low) but chipset/GPU load is high. 

You are right, but I experienced the slowdowns before even with just BOINC running on E-Dock. Because you have just one cooling system in the laptop, which connects CPU, GPU and CHIPSET, the rising temperature on CPU will cause the chipset and GPU to overheat very quickly.

Again, I'm running the same apps and the same way as before, don't take the BOINC as only test solution. But at the end, I can check the current speed with BOINC. After slowdowns, the benchmark was down to 700/2500 (and even less).

I can just hope, that downgrade to A03 and cleaning of the fan was interim solution for me and I don't have to shutdown the PC during the day anymore (what is actually the case).

mk