7805U fan problems

Seen a lot of these, no?:hi2:

My son complained that his FX started shutting down. I finally got some time to look at it yesterday. It’s a bone stock system all original running BIOS 9C.17.00.

I’m way behind the FX curve. My second son has my old 6831; which won’t die.

All day yesterday I pounded it with Kombustor 2.5 and IntelBurnTestV2 to max it out and check temps, never OC’ed. All heat pipes are clean with original factory TIM.
Temps on the GPU never went over 73C and the Intel Core 2 Duo P8400 ran about 65C max with IBT.

The first thing I noticed is the 120W brick is not enough. Using a Kill a Watt, the GPU pulled 100 to 110 watts using Kombustor. The CPU was pulling 70 watts, total, using IBT. Together the brick was pegged at 120 to 125 watts. I would venture these systems could use a 180watt brick? Not sure if the BIOS has a max current setting. This unit idles around 37 to 40 watts. The readings are at the wall so brick losses are not taken into account.

Day two, trouble. I ran Kombuster again. The temps marched right up to 105C before I shut down the program. I noticed the GPU fan did not ramp up but stayed at slow speeds. Ran a second time, same results.

I did a shut down and re-start, fan ramped up and hung around 73C for 30 minutes as previous day. So, intermittent fan, tachometer, sensor failure, other. Take my pick? I counted at least a hundred similar threads so unless anyone has BTDT, I have some reading to do before opening the thing up. The fan should be easy to check if the feedback is a dc signal or something that can be read with a DVM?

Thx for reading.

 

Is hibernate enabled? I noticed that my fans would not ramp up after returning from sleep or hibernation. To see whether brick is adequate run without battery. Brick will hum and/or shut down if defective.

 

I'll check the hibernation thing but the trouble day started after cold boot. Battery was out of machine. Shut down never happened as I shut it down when temp hit 100C. All i know is that the brick was running full boar with no overhead during tests today but not sure what was being run when son was gaming on it. BTW, all my bricks hum to some degree..I thoght it was nothing but switching noise?

The 7805 has been going for 3.5 years, without any problems, so suspect something on mainboard is about to give it up?

Thx

 

Not talking about switching noise but a loud buzz audible from 3 to 4 feet away. Most bricks will thermally shut down before this though. Your brick seems fine. Never had a problem until I installed a X9100 cpu.

I had fan issue is with sleep/hibernation and had fans give up on my P7805 but thay just quit. Yours might be slip sticking. Can't really see a mainboard issue since it still runs.

Only problem is a complete tear down is required to get at GPU fan and heatsink.

 

Well, hoping, I tore her down. The fans are fine. I hooked the GPU fan up to my B&K supply and it runs fine, 5vdc @ 300ma. The fan is stamped at 5vdc @ 1.8w so close enough. At this point I wonder if it is worth it to replace the TIM. My thinking is if the TIM was bad, temps would be bad 100% of the time?

4 tests today. Twice the temps went up to 85C, then fan ramped up, then back to 73C. The other two times, the fan ramped up and never went higher than 75C using Kombuster. Oh well, I have no spare parts so I might as well change the TIM and go from there.

Cheers

 

Do you have thermal pads for the NB and video memory? I thermal epoxied a .9mm copper shim to my NB heatsink and used IC Diamond TIM. Also used it for the video memory since it doesn't run out. The ICD can be used on the NB instead of the thermal pad or copper shim..

 

I just came back to look at the posted pictures. I do not have any TIM on the GPU memory chips, just a metal roof over them with a clean air gap. All I have is AS5, I used all my ICD on my UL30vt so will have to order more.

Any way gotta run for an hour.

Thx.

 

LOL, well, there is no TIM on the GPU memory! I guess the GPU is running hotter than the Samsung memory chips? Big "OPPS" on Acer's CAD part?

Now, do we second guess Acer and just put a gob of goop on there or just let it ride for another 4 years? :thumbsup:

Any thoughts?

 

Better get the big tube. It's a long way to the aluminum on the keyboard side and I used alot but it really helped. I hope you mean ICD and not AS5 since it runs badly. I've had AS5 short things out even though they claim it's nonconductive.

 

only running the max of 85c will not cause an issue. You are well within tolerances of the chips. Now I am sure the TIM is old and could bennefit from replacement but it does not sound like you NEED it as of yet.

 

I've got more TIM on order @ Amazon.

You had AS5 short things out? Oh really? Have you sent this fact to Artic Silver? Maybe you should use an Ohm meter and show them suckers the readout! You must get the truth out before one of their slick lawers shuts this site down for libel..

I used ICD7 to gap fill on my ULVT30 and it's holding up well.

The problem I have now is should I use that expensive stuff to gap fill the memory chip to heat sink on the 7805? I was shocked to see the photos posted WAY back on this site as I never tore one down completely.

When I see stuff like that I want to know why, not second guess them. If there is a thermal problem with the Samsung chips, the screens would artifact when new?

Anyone NOT gap fill?

 

I agree. I have nothing to lose so will try the new TIM. The intermittentt runaway 100C temps has me worried. Oh well if it's the TIM then that was easy. Also the screen was loose and I may help things by chem cleaning the connectors if something like a high resistant connection was a problem. I'll know next week.

Thx

 

now the runaway 100c is that CPU, GPU or TSZ? I ask as the TSZx can be a false reading...............

 

There was a very involved discussion in this forum regarding TSZx temps and where they come from. As I recall it is a super secret weighted combination of the CPU, NB and others.

I experienced a drastic improvement when I got my NB under control. ICD was a big improvement over thermal pad (expensive replacement material) and an additional small improvement with an epoxied (Artic Silver) copper shim and ICD.

TSZx is certainly a mystery but still will shut you down, false or not, and the algorithm is and has been proven flawed with no way we could figure out to fix it (ACPI dsdt tables).

I noticed that TSZx runs 8 to 10c above cpu and if the NB is not cooled then the gap is wider and shuts down sooner.

IMHO temp runaways are usually caused by a faulty fan.

 

Yes I let them know. Libel, get a grip. Take a little and spread on you CPU leads then get back to me.

In my discussions with AS, all they could come up with is the material created capacitence which caused problems. Have you ever measured the resistance of tap water? By your standards it's also nonconductive.
That was a long time ago and now I avoid getting it on anything but the interface since anywhere else I can expect problems.

Expensive? A new laptop is expensive.

Not gap fill, of course. Since there are KNOWN instances of premature failures with this memory I want to extend it's life, if possible, and I KNOW that lowering it's temp will do that.

You seem to be pretty sure yourself, good, so am I. I throw things out here for consideration in fixing a problem not as absolutes. You go on and on about the manufacturers and why they did this and didn't do that, well, ahem, costs and design constraints.

I think they did and excellent job with the P7805u considering it's price point. Perfect, no, nothing is.

Love your avatar. I guess mine would be a PDP-8/E.

 

Easy now, don't get in a wad.

Yes, I'm very sure of myself. As an ET by trade and having a meter, I check things, always will. You can even Meg it at 250v if you want.

You said it "shorted your things out". Sorry, I had a problem with that, if you said i had a strange capacitance problem, I would have taken note.

My apologizes by caling you on the "shorting out stuff" just pointing out a thorn in my side.

Shee, I wonder why I don'y make friends, no?

Anyway computer all fixed and working great. It was the paste that had failed in a way I never seen with the intermittent temps.

Thanks for your comments.

 

It was the GPU reading. Anyway, all fixed. I was all hardheaded about the TIM due the the odd intermittent readings. I had enough diamond and silver stuff left to get me through.

@ smlindahl, I did check the fans with external B&K supply that gives me current read out. @5v the current was below specs, so good there. The little bugger is fast at full 5 volt.

 

It was strange that the TIM was causing intermitent runaway GPU temps. It's like the heat was breaking the interface, sometimes. I have a P79 GPU heatsink/fan that I am waiting for some reason to install.

I check my fans with the frame right against my ear and usually can hear bad balls or a sleeve starting to seize. The marginal sleeve bearings I'll peal the label back and add a little light instrument oil the bad ball bearing ones I'll toss.

 

Yea, I've only pasted 4 laptops, total, so always something new? This time I only had 1ea, 4x1mm screw left over.

I was considering ordering the fan Tanware ordered for back up, but moving on now. Oiling the sleeve fans is a good and will usally buy you more life.

Last 2 hour bench ran today. After slow build up, temps are holding fast at 70C (ambiant 21C) so we should be good. (using Kombustor)

Update:

LOL, big fail. Almost back to where I was. On third run to recheck CPU, GPU ran up to 90C before I shut it down. Last two days, no problem. As before GPU fan running slow and quiet as if are not being commanded to run fast, or like smlindahl wrote, a bad bearing?

Funny I ran this fan up on external suppy, no vibration, sticktion, and aways free turning. Oh well, when I get time I plan to pull the yellow tach wire, cut and see if the fan will run full speed at all times and to monitor TZ voltage commands to fan.

...and I get to look for where the extra screw went..

cheers

 

The fan, running at full speed, keeps the temps around 79C after 90 min. of Kombuster. I want to find a GPU fan to rule out a bad tach and to insure fan at full RPM. I could just lube this fan and the heck with it as son only games on it but I want to test a fresh tach circuit.

Anyone have a good source of fans? I found the forcecon? version, around $20 new. Some sites are asking $38 bucks for a used fan. No thanks. Most of the ones seen are what I see on the CPU but they may work fine on the GPU.

Thx

 

If you are replacing the GPU fan it is best to actually find a HS from a P79xx s these were integrated as aq single metal enclosure for the FAN and HS and are more efficient than the old stock P68xx and P78xx units................

 

Thx, do you have any suppliers? My quick search failed...

 

ADEAL used to have them from broken up P7901's. It doesn't look like they have them anymore. I still have not installed mine yet, but I'm gladd I got it while they were still available.

 

OK, thanks anyway..

 

all you need is google................
ACER HEATSINK FAN FOR LAPTOPS 60.4FE04.001

 

LOL, +1

I must have been holding my tongue wrong..I never did get a hit on that site.

cheers