M6809 on battery sometimes just shuts down - Page 16

It's been about 5 months or so since my last post (I think). Back in Feb 2007, I wrote about how I put a copper shim between the cpu die and the heatsink (I had done that several months before I posted), and my oem A/C adapter finally died, so I bought a cheap (cheesy) A/C adapter from eBay (Brand name is Coonix). Back about 4 months ago, I had one spontaneous reboot, and thought, "here we go again!". But it has not happened since!!

I can't remember who posted the possibility of a bad A/C adapter being a possible cause (I think maybe Eunix or ghn). Anyway, my eMachines m6805 has ran like a top since (other than that one isolated shutdown). So I am inclined to think the possibility of the "space" between the cpu and heatsink, *OR* the A/C adapter may be the problem for many (but not all) experiencing this problem.

I am extremely disappointed in eMachines and Gateway for not officially recognizing this problem, but in their defense, I think there are several problems (including a bad temperature sensor) that makes this problem difficult to diagnose. If your warranty is up, and you have nothing to lose, please try these 2 fixes.

Since these 2 fixes, this 3 year old laptop works rings around my spouse's 3.2 GHz Pentium 4 Toshiba both in performance and in how cool it runs (my fan only runs when scanning for viruses [virii?] or encoding video). But I do not play video-intensive games, so YMMV.

i would vote for it also, drop it down to 1.5gb at pc2700 (333) - it would probably be more stable

cheers ...

Cool. I would like to see some instructions and pics on your mods! Believe it or not I am still running my m6811 and my 7422GX. I think I am going to get me an Alienware Monster machine but I the m6811 and 7422GX are still running strong.

I don't know about many of you but I am still waiting for the promise of X64

Eunix

Well, it's been about 3 years since I've been here (about the time I bought my M6810 from BB), and started having this same problem. The computer would shut down randomly, though always on power. So I went back to Best Buy last night, bought their Dynex Thermal Paste, and it worked! No shutting down, and I still have a decent amount left. Not sure how much to use, but I know it wasn't much.

Also, I read through this post a couple (long) weeks ago, but didn't really find an answer that worked for me. How do you remove that ground wire on the heaksink? I ended up coiling that wire around my screwdriver just to get the heat sink out, but if I could remove it, that would help the next time I have to do this. (I convinced a friend to buy this same laptop 3 years ago, and now she's having the same problems I was having, so I'm going to fix hers too.)

it is a trick, a screw driver and a needle nose plier. it is just stuck, that's all. be careful not to push down on the board itself

it is an amazing piece of machinery, is it not? and it is totally vista compatible

cheers ...

just saw ur sig. stay with pc2700 ram module - it is much more stable, gain in pc3200 is very minimal

cheers ...

Hehe, yeah, that was 3 years ago. A LOT has changed since then. Two hard drives (in fact both of the ones listed there) died.

And I swear I did that (needle nose pliers and screwdriver). That @*$!er was stuck. It was really frustrating me too...thought maybe there was another, easier way.

And it's VISTA compatible? Even 3 years old?

I knew this computer was the best buy of the decade. LOL

I had the random shutdown issue on my M6811... My solution that works the best is to boot up to the bios screen and wait the 20 sec's until the fans kick on....

A few months ago, I started having other motherboard issues... The ethernet port stopped working and the sound chip would stop functioning. I diagnosed it as a bad solder/board contact problem on the motherboard. It was created from picking the laptop up unevenly. e.g. like reaching over and picking it up by one corner. The stress on the motherboard caused it to fail over time.

I bought a new motherboard from ebay ($150) and it works PERFECT. It has never shutdown on me yet... Never! Also, I'm still using my old CPU, heatsink, and fan.

Mark

i m not the only one that would swear by it. go figure; new laptops and users throw vista out to put xp on due to "whatever"; and this workhoorse just plows ahead



cheers ...

definitively a tip here as how to take care of one's laptop

cheers ...

Well, I hate to admit it, but after four days of flawless running, my computer is back to hiccuping. I downloaded some software (SpeedFan and RMClock) to help control the processor speed and voltage, so I'm running at about 1.2V/8x right now, hoping it'll be stable. I'm planning on trying to AC Adapter fix once I get a little more money, but in the mean time, does anyone know of anything else that could be the culprit? It's really irritating me.

My experience

Hi all, I registered just to give you everything I've learned and the countless hours I put into this laptop. Unfortunately, I have no pictures to go with some of the procedures I did.

I purchased the m6805 in August of 2004, for about 1400 dollars. I thought it was a steal at the time.

It worked flawlessly until around July of 2006 when I experienced my first hard shutdown. Being that the intervals were hours apart, I just thought the computer was simply overheating and that the components are not as effecient as from when I first bought it.

But as the intervals started to shorten to about 30 seconds between shutdowns, the laptop was becoming a really expensive paperweight.
Luckily, I found rmclock which solves the problem - IF you make it into windows.

Then the worst thing that can happen - windows stopped working. I figured there was no way to reformat the computer without it shutting down before it can complete. So I popped open the hood.. scraped off the fossilized thermal pad/paste or whatever it came with and put some artic silver on. That was not the end of it.. I even installed a full sized Thermaltake 120mm heatsink on the laptop cpu to keep it as cool as possible. This turned out even better than I thought it would. Even though the latop looked like the hood of some muscle car, it kept the cpu at 20c even under full load!! rmclock wasnt even necessary. I used this everytime I needed to format.. and everything was dandy.. UNTIL

It started shutting down in the middle of a quickrestore..even WITH the full sized heatsink. I even tried my a water cooled heatsink..to no avail. Days and nights of toiling on this thing finally pushed me over the edge. No solution could get windows back onto the laptop successfully. I was not about to invest in another cpu for the laptop. So I did the unthinkable. I took a flat-head screwdriver, took off the CPU, and completely snapped the heat sensor diode right off. Along with it, I snapped off a couple of other things around the diode to make sure NOTHING was going to shut it down. Put everything back together, and pressed the power.. Amazingly, it turned on. It even booted faster! haha. But the problem was NOT fixed at all. So I can tell you right now it is not a heat problem!

So the solution to this was to get into the restore dvd, copy the ghost file images to another computer. Take out your notebook harddrive, and put it into an external enclosure. Use Ghost dosboot (I used version 8.2) and restore the image files onto the harddrive. (This method just mimmicks the quickrestore, just on another computer.) Get a copy of rmclock or amdcpu on a cd. Then pop the harddrive back into the laptop, along with the cd. When it gets to desktop, quickly execute rmclock or amdcpu.

I am back on the m6805 typing this reply. Hope this helps!

Not to sound too optimistic, but I think I've finally found some stable settings! Using RMClock, I'm running at 1.3 V, with a 10x multiplier, and I've yet to see my machine shut down, even after repeated games. The fan even shuts DOWN while playing games, and stays down for a while. The temp is a more-or-less constant 51 C, while at 1.45V it was just around 60 C before rudely shutting down (with no fan kick on warning!)

Anyway, I hope I'm done with this problem for a good long while.

I've got the German version of the arima w730-k8 with a MobRad 9700 128MB and I' ve never heard that this model would have a shutdown problem. Curently I'm running an Athlon 3400+ with 1.150V at 2200 MHz, which shuts off the fan for MOST of the time even during gaming. Every 10 mins, or so i got the fan going for like 1 or 2 minutes. So max. temp is 50 without fan quickly dropping down to 40 when the fan kicks in. Awesome. I also applied a special thermal conductor called Coollaboratory Liquid Pro. It's the best one around, and prolly you want to find out about it:
http://www.coollaboratory.com/en/index.shtml.
Cuts down the temp another 4-6° in comparison to other pastes.

I started having the random shutdowns a few months ago, increasing in frequency until I couldn't take it any more. After a clean install of XP, I determined that it was a hardware fault, not some random software/configuration setting.

I noticed that it was happening most often when the CPU was working at 100%, so I made a quick little app that would chew up the processer. It would shut down a couple of times really quickly, then it would run for a long time after everything warmed up.

With the intention of greater cooling efficiency, I took it apart to clean out all of the cooling components and apply some silver thermal paste. When I took the cooler off of the CPU, there was NO paste at all between the CPU and the cooler. I cleaned out the cooler parts, applied the paste, and put it back together.

Initially, it seemed to solve the problem. Later that night, though, I decided to run the CPU stress program for a long time and monitor the CPU temp. It ran for about an hour, then shut down, along with a burning smell. I happened to be looking at the temp right when it turned off, and it was only about 68 degrees.

The notebook wouldn't turn on after that last shutdown, so I started dismantling it piece by piece to see if there was a fault in an easily replacable component. Eventually, I got down to a nearly bare motherboard and still no power. I noticed that the power brick was "ticking", and from experiance I knew that it made that sound when the output was shorted.

I grabbed my mulitmeter from the garage, and sure enough, there was a direct short circuit between the CPU power supply and ground. The short only seemed to be on the CPU supply, but it was dragging enough current that the entire motherboard couldn't power up.

I saw all of the tantalum capacitors near the CPU socket, and I have seen many a tantalum capacitor short out. I brought it in to work the next day and started pulling out the capacitors one-by-one, and the second one that I removed was the short.

I replaced it with a new one, re-assembled enough of the notebook to test it, and attempted to power it up. No lights, no sounds, excpet for a faint sizzling sound and a little bit of smoke. I opened it back up, and found that the capacitor next to the original shorted cap was now shorted, too.

Those capacitors are rated for 6V, so I think that the CPU power supply regulator has gone bad, sending enough voltage out to the supply line that it fries the most vulnerable component - the tantalum caps.

I decided that I didn't want to spend any more time on it, so I gave it to one of the other engineers at work, keeping the HD to pull my data off.

From the consistancy of the behavior noted in this thread, I think that this is a common problem that starts happening with these things as they age. If yours is doing this and you can keep it stable by clocking it down, I think that you are only delaying the meltdown for a while. I woudn't trust it very much after the symptoms start showing up. Start saving for a new rig...

The QC specs on the German version may be a lot tighter than the ones of the first versions. These boards were originally slotted for the general user (consumer grade) to be sold in retail stores and not the "corporate" business line of machine. Therefore they were a lot more lenient with the tolerence specs and defects.

Eunix

I'm just speculating, but could it be the case that the early emachines models, or the US units in general were fitted out with 81 Watts TDPs Athlon 64s rather than the 62 Watts DTRs?
I can imagine that the "giant" cooling unit of the Arima can handle such high-wattage CPUs with ease, but the voltage and current strain on the capacitors might be the culprit over time, as tanbam pointed out.
I mean, from time to time I've read a few posts in this forum here reporting temperatures as high as 70°C on this laptop and I've never gotten even close to this figure.
Even when I had the Clawhammer 3000+ running at spec's voltages doing stress testing with old thermal goo, I never hit more than 58° or so. I strongly assume that your Athlons draw a bit more juice than our European counterparts.
Also, when u got your machines doing full load/speed Large In-place FFTs under prime95 whats the max. amount of discharge rate you have on your batteries (undervolted or not), with full screen brightness and wifi on?
You can check that figure with RM Clock Utility under battery info- device tab (my fav. tool ever).
I got like 48000mW in a.m. scenario 2200MHz@1.150V temps ranging from 40-50°, depending on fan on/off.

cheers...

I found this thread and registered to post my experiences with a Gateway 7422gx. I read through every post hoping, hoping, hoping that in 22 pages, somebody would have found a solution. Nope.

Unfortunately I also don't have anything positive to report. This machine was my father-in-law's, and as others have reported, it worked fine for quite a while. Then the dreaded high speed fan shutdowns started. He gave the machine to me after getting tired of messing with it.

When he had it, he had a repair shop check the fans. Supposedly they replaced one, but the shutdowns continued. I have taken the laptop apart and I can tell that the HSF was removed and new thermal compound was installed, so the old compound wasn't the culprit after the machine came back from repair. I also verified that both fans do work. When I took it apart, I put the thing back together without hooking up the fans. Stupid, huh? Well, it booted fine and ran several applications. It wasn't until I ran Prime95 and noticed that the CPU temp was 165 and climbing when I realized my mistake. Leads me to believe this isn't a heat issue.

This machine will also run Prime95 for quite a while with no problems and without the fans kicking onto high. But when doing other tasks and the fans kick into high speed, you know it is going to die in a few seconds.

I will spare you the details, but from a software standpoint I have tried updating/changing everything I can think of: chipset drivers, CPU driver, complete reinstall from recovery disc, and even trying Ubuntu Linux. I can't remember what else I have tried, but believe me, I've probably done it. I have messed with the Windows power settings and ran the machine on full power and by letting the CPU driver throttle the power. It still shuts down.

I tend to agree that this is a hardware or design problem, probably involving the power the computer is drawing. This machine has been run on AC power 95% of the time. When on AC power, even with the battery removed, it shuts down.

I've resigned to accept that this thing is simply going to be flaky and unreliable. It is a shame because even after 2+ years, it is still a decent laptop as far as performance and features are concerned.

try using RMClcok and manage the fan/heat urself. We have good threads/posts regarding using this utility to cool down these models

cheers ...

7405 Still Here!

Hi All.

I am still using the Gateway 7405. I just wanted to come back here and post a link to this document that I found, written by one titosfuneral, this has kept the machine going the last few months.

http://discussions.hardwarecentral.c...d.php?t=173823

The main part here is:

"This is NOT an overheating problem. It just reacts like one. It happens well below the maximum temperature for the chip. I can happen at 25C. It happens when the chip spikes from 800MHz to 2200MHz and back, causing a rapid increase in chip temperature. The main fan them comes on full, with the chip at low voltage. The system then halts 4 seconds later. It seems to be a power supply problem."

This describes perfectly how my machine behaves, and for me, the solution is using RMClock to stop going to either 4x or 10x, although I have had a few shutdowns before RMClock gets a chance to control the CPU.

titosfuneral also speculates:

"I have a theory that there is a power short on the AC connection detection circuitry that drains available wattage, but I really don’t have much to back this up. It could just be that eMachines screwed up on AMD’s “cool’n’quiet” technology. There certainly is a problem when the machine goes into its maximum power use mode at the low and high ends."

This grabs my attention when reading the poster above who writes about the tantalum caps getting stressed over time and also Mark K who writes about their experience with the dead ethernet port and picking up the laptop unevenly.

I wonder about whether the first 15 pages or so of this thread may be about people opening the laptop to adjust the cooling mechanism and having results based on moving some bad solder joint, inadvertantly flexing the MB.

It is also possible that their cooling efforts were sufficient to prevent the resultant increase in temperature from this "spike" kicking in the thermal panic.

I read some weeks ago about the Athlon Mobile, I can't find the link now, but that the manufacturing process of the Mobile and Desktop is the same, its just that they test which CPUs can function at low voltage and which cannot.

Those that can become mobiles. I suppose that means that each chip will have it's tolerance. Bear that in mind when trying to find voltages for your CPU.

Thinking about titosfuneral's theory, I thought maybe I would play with RMClocks ability to time the clock changes and also the parameter P-states transition method, thinking that if I don't let it go directly from 4x to 10x, i should be ok, but i got shutdowns.

I had also tried dropping the voltage a little, but I got blue screens (XP) and reboots at 0.800 v at 5x (1Ghz) I got the shutdowns as RMClock started up, which seems to be a reason not to allow 10x 2.0Ghz. Once RmClock is up, I can add 2.0Ghz (so far this has worked), although I wonder really about the benefit.

Currently (at startup) I have 5x at .875v, going up to 9x at 1.25v
Up transition time of 110ms and a down transition time of 400ms
P-states transition method is perform single step only. I've been powered on for more than a week like this. Hibernating instead of shutdown decreases the risk of a shutdown before RMClock gets control.

So more or less I'm ok. At least I can still work and hopefully make enough cash to get a new laptop. It is a good laptop though, yet I wonder why anyone cares about vista compatability. I'd like to run Linux on it, but I guess i'd need RMclock type control.

My heartrate still increases sometimes when I hear the fan kick in at certain moments