Compaq Presario v6000 laptop repair (failed nVidia chipset)

[momaka]

So I found this Compaq Presario v6000 laptop in September of last year and posted about it in the Best Cheap/Free Scores 1.1 thread (along with some other stuff I found). It came complete, with a power adapter, a battery, and even HDD. As mentioned there, as soon as I saw the “Graphics by nVidia” sticker, I knew there would be problems with the thing. And sure enough, that was the case indeed – the laptop would not POST (all of its lights would turn ON, but screen stayed black with backlight ON).

I then did some research on my model, and found that there were many, many complaints online about this laptop along with its equivalent: the HP DV6000. The complaints ranged from dead WIFI, no sound, dead trackpad, no POST (like mine) and sometimes even a combination of these. It wasn't too much of a surprise to me, as I've heard of bad stories regarding the HP DV6000 laptops before (as well as some similar models too, all based on similar motherboards made by Quanta). Heck, we even had a few threads here at BCN:
https://www.badcaps.net/forum/showth...ighlight=v6000
https://www.badcaps.net/forum/showth...ighlight=v6000

While researching, I also found a quick specs brochure for my laptop, which said it came with an AMD Athlon 64 Turion processor of some sort and a GeForce Go 6150 GPU (or dedicated 7150M for higher-end models).

Thus, I was pretty sure what this meant before I even opened my laptop – failed nVidia chipset/GPU. I wasn't sure if I was going to repair it at the time, but I cleaned it nevertheless, as I simply don't like storing dirty stuff in my house. The laptop was pretty terrible when I got it – hair, dust, and some food all over the keyboard. To make things worse, the top case cover above the keyboard had a “dotted/dimpled” texture on it, making it even harder to clean. Even the top case on the LCD was nasty. In the end, it took a lot of work just to clean it, but at least now I had a clean laptop that I could either part out or repair and not worry about getting my bench/carpet dirty. After cleaning, I put the laptop back in the box I found it in and stored it in the garage until I felt like playing with it another day.

That time came a month later. We had some good weather (not too cold or too warm), which meant I could do some reflows (on the kitchen stove, mind you ) and easily air out the fumes afterwards by opening the windows. Thus, I decided to reflow the Northbridge/GPU in this laptop, just for fun to see if that would fix it. After all, I had nothing to loose but a bit of time with that.

I began stripping down the laptop. But after removing the HDD, RAM, wireless card, and a few other small plastics, I quickly got stuck – I just couldn't figure out how to open the damned thing! So I searched online and found the service guide for that laptop. After going through the disassembly steps and finding a few “hidden” screws that I had missed, eventually, I got the laptop fully taken apart. (I suppose for those of you who do laptop repairs for a living, you probably find this amusing. But I've only worked on a few laptops before this, most of them being Dell Latitudes, which are a piece of cake to take apart and put back together).

Then finally I got it all disassembled:
https://www.badcaps.net/forum/attach...1&d=1504815472

Upon seeing the motherboard's cooling, it was obvious why the NB/GPU had failed: it was thermally “coupled” to the CPU heatsink with a thick thermal pad. The thermal pad was at least 2 mm thick . Moreover, the CPU heatsink had a copper pad soldered to it for the NB/GPU, but why didn't they make it just a small bit thicker to deal away with the thick thermal pad?
Here is a picture of the bottom side of the heatsink showing the design:
https://www.badcaps.net/forum/attach...1&d=1504815472

To the manufacturer's credit, at least they used a good quality fan:
https://www.badcaps.net/forum/attach...1&d=1504815472

And here is the board, cleaned and ready for reflow:
https://www.badcaps.net/forum/attach...1&d=1504815472

CPU and NB/GPU up close:
https://www.badcaps.net/forum/attach...1&d=1504815472

So then I did the reflow, as I usually do: place a grille above one of the stove's gas burners, put the board on top of the grille, place a type-K thermocouple near the GPU, and turn the gas burner on LOW. Wait until the board temperature reaches about 110°C and keep it around that temperature for a few minutes to “soak” the board with heat. This also allows any moisture to escape before reflowing, so that I don't get “popcorn” at high temperatures.

After soaking the board, I usually turn up the heat just a slight bit so that the temperature raises about 1-2°C per second. When the area around the GPU hits 150°C, I turn on my heat gun to its high setting (about 1400 Watts) and point it up in the air for a few seconds until it heats up (the heating elements inside starts to glow orange). Then point at the GPU, keeping it about 2.5-3 cm away. This is where it gets tricky, as my thermocouple starts to show the temperature of the hot air from the heat gun, which varies between 250-300°C, depending on where the thermocouple is placed. So to avoid overheating or under-heating the board, I periodically move the heat gun away from the board and let the thermocouple read the temperature of the board. As I continue to heat with the heat gun, I also crank the bottom heat a bit. When the temperature is 210-215°C, I heat for about 10-15 more seconds, then turn off all heat and let the board cool on its own. (Unfortunately, I don't have a picture of the setup)

While the board was cooling, I started to think of a way to better couple the NB to the CPU heatsink. I went in the garage and found several aluminum pieces, which looked like they might have worked. But then getting them filed down to the right thickness seemed like a lot of work and I tried to think of something simpler… and so I did!
https://www.badcaps.net/forum/attach...1&d=1504815472

It was just a simple piece of copper folded like paper in a U-shape. Since copper has a bit of a spring-action to it, I though that placing this piece between the GPU core and the heatsink would make it so that the copper is always keeping a bit of pressure on the GPU core, and thus would sit tightly on there. I test-fitted it on the heatsink and it seemed to work (which is why you see some thermal compound on it in the above picture). Then I had another thought: why not fill in the empty gap in the copper piece with whatever was left of the original thermal pad (the thermal pad tore itself to pieces for some reason when I took it off the heatsink). So the end result was this:
https://www.badcaps.net/forum/attach...1&d=1504815472

Now was time for the real test. However, I wanted to test the board without putting the laptop back together, as it was a pain to take apart. And I also wanted to monitor the temperature of the GPU with my thermocouple, which would have been difficult to do with the laptop in the case. So I did this:
https://www.badcaps.net/forum/attach...1&d=1504815472

Then, I connected and external USB keyboard and monitor.
But did it work? ...... TBC in next post.

[momaka]

Compaq Presario v6000 laptop repair (failed nVidia chipset) pt. 2

... Yup, it worked!

Just had to set the date and time in CMOS (CMOS battery had to be removed before the reflow… and it was dead anyways), after which she booted right to the OS that the owner left on the HDD. No password either.

On that note, I’m not sure who installed the OS for this person (or if they did it themselves), but most drivers were missing and everything ran SLOOOOOOOOOOOOOOOOOW. I actually had to go on another PC and download the drivers there, because this thing was nearly useless. I used a flash drive to transfer the drivers (had to unplug the USB keyboard for this, as this laptop only has two USB ports, and I was using the other for the mouse). So I installed the drivers, along with a few favorite utilities like Firefox 24 Portable, CCleaner, and SpeedFan. Everything ran WAY better now. Hardware-wise, this is what the laptop came with:
CPU: AMD Turion 64 TK36 (2 GHz, single core, 31 Watts TDP)
Motherboard: Quanta model 30B7 with F.42 BIOS
GPU/Northbridge: GeForce Go 6150 (see below)
RAM: 1x 1 GB DDR2
HDD: 80 GB Seagate


I then proceeded to test the laptop more, as I wanted to make sure it works before putting it back together. The first thing I did is check the temperature of the CPU and GPU over time with SpeedFan. With the laptop running “open” like this, the temperatures didn’t look too terrible.
https://www.badcaps.net/forum/attach...1&d=1504816553

In idle, the CPU and GPU ran fairly cool. But under load, the CPU temperature would spike up quite fast before settling to around the 60°C mark. That said, at least my copper shim seemed to be working well, because the GPU would follow the CPU’s temperature to within a 1-2°C range in the steady states (i.e when the CPU temperature was settled either low or high).

It is also interesting to note that my thermocouple was reading much lower temperatures, despite the fact that it was directly touching the GPU core. This most likely happens because CPU and GPU cores typically have their thermal diodes deep inside them, and the internal temperature is always much higher than the core closer to the heatsink interface (where my thermocouple was). So both SpeedFan’s and my thermocouple’s temperatures were likely accurate, despite being different.

Next, I proceeded to do one more thing before calling the laptop fixed: check the wireless function. But how to do that without assembling the laptop and connecting the antennas? - This is where my spare stash of Xbox 360 and PS3 parts came to the rescue again. I took two 2.4 GHz wireless antennas from a PS3 and connected them to the WiFi module (you can see those in the picture above).
Ghetto?
- YES!
Does it work?
- YES! I simply entered my WiFi connection details and we had internet!
Does that mean I could watch YouTube?
- Yes! And YouTube didn’t run too badly with a forced Flash Player add-on in Firefox. Unfortunately, that only worked for one or two videos. After that, I started getting a weird error that would not allow me to watch any videos unless I switched to HTML5 mode… and HTML5 runs like crap on old hardware.

But that’s where the good stuff ended. The bad stuff: audio was NOT working AT ALL.

So my reflow may not have been 100% successful after all. Or maybe it was? I know that for most older desktop motherboards, the audio is almost always controlled by the Southbridge. Thus, maybe the SB in this laptop had problems too? I noticed that it ran hot as well. As such, I thought about reflowing that too… but in the end, I decided to use the laptop as it is with the broken audio and see how long it lasts. (Seems like I underestimated my repair, as the laptop is still working. )

With the laptop still not assembled, I took it upon myself to further clean some of the case plastics. Now that everything was taken apart, it was much easier to clean each individual pieces in detail. Thus, I took some of the plastics and washed them with water + dish detergent and a brush. Even that plastic keyboard top case with the “dotted” pattern could finally get cleaned properly.

After everything dried, I started putting the laptop back together and noticed a few interesting things:
https://www.badcaps.net/forum/attach...1&d=1504816553
Looks like this motherboard came with three (3) USB ports on some models. Also looks like some models had a Line In jack, as opposed to mine only having Line Out and Mic jacks.

Small stuff aside, it looked like everything was coming well together…. Until this:
https://www.badcaps.net/forum/attach...1&d=1504816553
Ugh! I placed a screw on the motherboard that was not meant to go there. So I had a fun time taking the laptop apart again to fix that. JOY! When all was done, this was the finished result:

https://www.badcaps.net/forum/attach...1&d=1504816553
https://www.badcaps.net/forum/attach...1&d=1504816553

I can’t say that I am 100% satisfied, as the top case had quite a few scratches… but for what it’s worth, it looked pretty decent when I was done with it and certainly much better than when I got it (I wish I had taken some “before” pics, but I didn’t care much for this laptop when I got it, so I never did). The only ugly part was the bottom:
https://www.badcaps.net/forum/attach...1&d=1504816553
All but two of the rubber feet were missing. As such, the laptop easily slides across most surfaces. What’s worse is those rubber feet are needed as they raise the bottom of the laptop so that the fan vents can breathe easier.

Without them, this is what the laptop temperatures looked like with it sitting directly on top of my carpet:
https://www.badcaps.net/forum/attach...1&d=1504816553
Unacceptable! And this was with a fairly cool room temperature of around 21°C / 70°F . The laptop will surely die again if I kept running it like this.

Fortunately, I found that with the laptop sitting on top of a book, with one side hanging so the fan intake holes are unobstructed, I was getting temperatures that were much closer to those initial test temperatures: the CPU and GPU were generally sitting around the 47-50°C mark in idle and hovered in the high 50’s under light/moderate load browsing the internet (depending on the website visited).

I tried using the laptop to watch some MP4 movies on it (high bitrate SD). With the CPU load being around 25-50%, its temperature was around 58-60°C. Meanwhile, the GPU never went above 56°C. But again, this was with a fairly cool room temperature of about 21°C / 70°F. Once I started testing the laptop in the summer when my room temperature hit above 27-28°C / 80°F, the GPU managed to go above 60°C under load a few times with the same load after a long session. After doing this several times, eventually I got the laptop to crash once with no video, which required a cool down before I could get it to work again.

So I suppose the truth is that the Compaq Presario v6000 and HP Pavilion DV6000 just don’t have a properly-designed cooling system. Not only does the thermal pad allow the GPU to run hot, but the whole heatsink in general is NOT ADEQUATE for the CPUs that they put in these laptops (generally around 30 Watts TDP). The heatsink should have had at least 1½ times more surface area. And although the latest BIOS for these laptops is meant to crank up the fan speed more, it still wouldn’t be enough to keep the laptop cool IMO. Thus, I am still expecting this laptop to fail as is.

That said, I have a spare Athlon 64 X2 TK-57 CPU that I am planning on installing in it. I’m hoping that with this CPU, my CPU usage would go considerably down due to it being a dual core. It has the same TDP as the MK-36 that is currently in the laptop, so I think the temperatures will drop. But I guess we will see. If the temperatures are decent, I might even open the laptop once more and do a reflow on the SB to try and fix the audio. I guess more to come later.

[piernov]

Re: Compaq Presario v6000 laptop repair (failed nVidia chipset)

Those NVidia chips are garbage, the C51M is from the well-known NVidia fuckups era and merely reflowing it won't fix the problem (or at least, it's only a short term workaround). You would need to replace the chip with a fixed version. They've a datecode of at least 09, I'm not sure if the 10 ones are genuine but here is one:

http://www.ebay.fr/itm/Brand-NEW-Nvi...UAAMXQnYJRBQIN

Though you always have to pray to actually receive a new, genuine chip and not a dead one or rebadged defective one. Sadly that's what happens most of the times, especially for those NVidia ICs were hundreds of thousands of defective ones were sold, thus many sellers take advantage of this and rebadge an older defective chip to have a newer datecode.

Btw you don't need to actually reflow the IC, merely heating it up to 200°C for a couple of minutes should bring it back to life for a short amount of time. If it dies again soon after, try to heat it for a bit longer. Reflowing it isn't needed and could damage the chip even more (they're not designed to get over 260°C or something like that) or the motherboard.

The MCP51M on the other hand is not known to fail so it's unlikely that reflowing it would fix your issue. If you really have an audio issue, first rule out a software issue (you could run Linux to see if it works there), a parts issue (bad speakers/audio jack) and then check if the audio IC is overheating.

Also adding a copper shim is not always the best idea. Having multiple layers adds to the thermal resistance. And especially in the case of multiple chips under the same heatsink, the copper shim can't align properly so it doesn't make good contact with the die. Those heatsinks are often designed in a specific way that they make the best flat contact with the CPU, but not the other chips, and it's not possible to fix that. Also be aware that it would put more stress on the die and you risk breaking it (especially when disassembling it since the imbalance will put more pressure on a specific corner).
On the other hand, if it's on a separate heatsink which has a thermal pad, it should be fine to replace the thermal pad with a copper shim, though it may not help that much.

If you can get away without adding a copper shim, it's better, you just have to find a suitable thermal pad with high thermal conductivity (that's the hard part).
Well that said, I've done it myself (replacing thermal pads with copper shim especially on HPs) and sure it helps in some situation. I've got a CQ61 (IIRC) to loose 10°C on the GPU.

[jiroy]

Re: Compaq Presario v6000 laptop repair (failed nVidia chipset)

I agree with piernov about the reflowing procedure being a waste of time . A hair dryer is more than enough to bring it to life for a short period , usually between 1 hour to one month at maximum .
Pointing a hair dryer on the cooling system exhaust can do the trick , but i saw hundreds of those laptops with totally melted plastics in the bottom case .

Some would simply on those nvidia built in VGA for some time inside their bags , zipped and closed , and when they heat enough , they would turn on .

Some others found a more weird way .. Once the thingy is on , they wouldn't turn him off again , lol .

[piernov]

Re: Compaq Presario v6000 laptop repair (failed nVidia chipset)

That's like putting a PSU under the sun for a few hours to workaround a capacitor problems… Though the capacitors require less heat, and they're actually replacable.
Anyway, better not waste time with those NVidia chips (don't make the same mistake as me), selling them on eBay for parts or keeping the motherboard as a donor is a better idea IMHO.
When it is a GPU (and not a part of the chipset) problem, some motherboards can be converted to UMA only, but I never tried it, and it seems like a pain on boards from that era since at that time the GPU wasn't integrated in the CPU but rather in the northbridge. They may require replacing the MCH with a GMCH for example on Intel platforms.

[Th3_uN1Qu3]

Re: Compaq Presario v6000 laptop repair (failed nVidia chipset)

The 6150 is unfortunately integrated, it is already UMA. For the audio issue, a faulty jack or burmed audio codec is likely the issue. Do you at least see the sound device in Device Manager? Link to the schematic and i'll give you a few things to check.

[jiroy]

Re: Compaq Presario v6000 laptop repair (failed nVidia chipset)

There were many other problems back then , the 6150 Nvidia wasn't compatible with win 7 and maybe impossible for win7 64 bit drivers . So , motives were missing .. No enthusiasm .

Models equipped with Quickplay were the worst case . The Quick play would fail in the first year , and in the second year , the Wireless simply disappear (due to the Nvidia SB And not the wireless module itself ) .
Within one year and a half , Display start showing its problems .

I had the bad luck of purshasing 3 DV2000 for some 3000 $ , and when i tried to sell them , as is , many replied : Worth 0 $ , lol .

[piernov]

Re: Compaq Presario v6000 laptop repair (failed nVidia chipset)

Schematics attached (hoping this is the right one).

[momaka]

Re: Compaq Presario v6000 laptop repair (failed nVidia chipset)

Those NVidia chips are garbage, the C51M is from the well-known NVidia fuckups era and merely reflowing it won't fix the problem (or at least, it's only a short term workaround).

Yeah, I know the reflow is only a temporary fix. But I imagine if the temperature is kept fairly low, this could prolong the life of the repair quite a bit. From what I remember, the bumpgate issue becomes much bigger only after the chip goes above 60C.

You would need to replace the chip with a fixed version. They've a datecode of at least 09, I'm not sure if the 10 ones are genuine but here is one:

http://www.ebay.fr/itm/Brand-NEW-Nvi...UAAMXQnYJRBQIN

Thanks for the link.

I'm not sure if it is worth spending that much money on an old laptop (and that's not even considering the time I would have to put to tack it apart again and reflow it). Heck, for $30, I can usually find a working Dell Latitude D or E series with Core 2 Duo, and those are 100x better built than this Compaq (except the ones with the nVidia Quadro GPUs, which have the same issues).

Btw you don't need to actually reflow the IC, merely heating it up to 200°C for a couple of minutes should bring it back to life for a short amount of time. If it dies again soon after, try to heat it for a bit longer. Reflowing it isn't needed and could damage the chip even more (they're not designed to get over 260°C or something like that) or the motherboard.

Ah, okay. Will keep that in mind if I have to reflow/re-heat it again.

The MCP51M on the other hand is not known to fail so it's unlikely that reflowing it would fix your issue. If you really have an audio issue, first rule out a software issue (you could run Linux to see if it works there), a parts issue (bad speakers/audio jack) and then check if the audio IC is overheating.

Good idea.

Indeed I haven't narrowed down the issue yet. All I know is that Windows XP SP3 has the drivers for the audio chip installed, but I don't have a sound icon in the sys tray and no software will see the audio device, despite Windows XP recognizing the audio device hardware in Device Manager.

That said, it seems that at least some portion of the hardware is working normally, as warning messages will send a "beep" through the system speakers. Likewise, the Compaq volume up/down/mute shortcut keys on the keyboard will also emit their own distinctive beep when being pressed (even though they don't seem to do anything). So I guess the hardware is not completely dead. But I haven't opened the laptop after I "fixed" it last year.

I'm thinking about trying a Linux distro, but I am worried that if it loads the CPU down to 100% for more than a few minutes, it will overheat the GPU as well, as my copper shim is working pretty well. So I might try some driver pack software to see if that picks up something that perhaps wasn't installed that I missed. I have both the Russian driver packs and the other big one (their names escape me now).

Also adding a copper shim is not always the best idea. Having multiple layers adds to the thermal resistance.

It does, but it is still many times better than a thermal pad. Even not-so-thermally-conductive metals like steel have lower thermal resistance than most thermal pastes and thermal pads.

And especially in the case of multiple chips under the same heatsink, the copper shim can't align properly so it doesn't make good contact with the die.

That's why I made mine as a U-shaped piece, because it keeps a pressure between the heatsink and the core. You can see I have decent thermal conduction from the temperature graphs above.

Those heatsinks are often designed in a specific way that they make the best flat contact with the CPU, but not the other chips, and it's not possible to fix that.

Yes, true.

Also be aware that it would put more stress on the die and you risk breaking it (especially when disassembling it since the imbalance will put more pressure on a specific corner).

I've chipped many corners with custom heatsink mods before. Usually it's not an issue, unless you chip way too much, and that typically requires a huge force.

The only scenario where a copper shim may be more dangerous is if the laptop gets dropped.

...you just have to find a suitable thermal pad with high thermal conductivity (that's the hard part).

Yeah, maybe not for this laptop. Not worth it.

Well that said, I've done it myself (replacing thermal pads with copper shim especially on HPs) and sure it helps in some situation. I've got a CQ61 (IIRC) to loose 10°C on the GPU.

Nice!

My sisters had a CQ61 (or was it a CQ62?) that is starting to get flaky due to the ATI NB going on its way out. They already had it serviced once somewhere, and I believe they changed the mobo. This time, while still working, I'm just going to shim/improve the NB cooling and hope for the best, as the laptop still works fine.

Some others found a more weird way .. Once the thingy is on , they wouldn't turn him off again , lol .

Lol, that will work..................... for a little bit.
... Until the power goes or when the PC crashes and you need to restart it (there is not reset button like on desktops, so the laptop will have to be turned off).

Plus, it's a waste of power keeping the computer on. I prefer to turn things off when I don't need them.

That's like putting a PSU under the sun for a few hours to workaround a capacitor problems… Though the capacitors require less heat, and they're actually replacable.

I've actually done that with my first router's power adapter when it was on its last legs and I needed to use it. Once it fired up, it worked fine until next time power went out. Eventually, I recapped it and solved my issue once and for all with the adapter. The router (a DI-614+) is still a flaky piece of crap, despite getting recapped as well.

Anyway, better not waste time with those NVidia chips (don't make the same mistake as me), selling them on eBay for parts or keeping the motherboard as a donor is a better idea IMHO.

That's what I will probably end up doing if it breaks again and a reflow/re-heat doesn't fix it. There are indeed more worthwhile used laptops out on the market that I can get for the money.

For the audio issue, a faulty jack or burmed audio codec is likely the issue. Do you at least see the sound device in Device Manager? Link to the schematic and i'll give you a few things to check.

Jack okay.
Yes, I see the sound device in Device Manager, as mentioned above. Schematic is attached above by Piernov (thanks man! ).

There were many other problems back then , the 6150 Nvidia wasn't compatible with win 7 and maybe impossible for win7 64 bit drivers . So , motives were missing .. No enthusiasm .

Well, given the hardware specs of this laptop, I don't think Windows 7 would be a good idea anyways. Heck, even Vista was a bad enough idea. I'm not sure if it came with 512 MB stock or 1 GB, but it had 1 GB when I got it. That itself is just barely good enough for XP for online browsing nowadays. Of course, as of last June, I increased the RAM to 2 GB, which is plenty for XP SP2/3 32-bit.

Actually, to my surprise, Firefox 4x runs quite well on this laptop, even though it is only a single-core Athlon. Most pages load quite smoothly - or at least much better than compared to my Pentium 4 PCs. Maybe with Flash going out the door and HTML5 being the norm, the long architecture of the P4 is no longer an advantage.

[jiroy]

Re: Compaq Presario v6000 laptop repair (failed nVidia chipset)

Yes, true.


I've chipped many corners with custom heatsink mods before. Usually it's not an issue, unless you chip way too much, and that typically requires a huge force.

The only scenario where a copper shim may be more dangerous is if the laptop gets dropped.


Yeah, maybe not for this laptop. Not worth it.




Lol, that will work..................... for a little bit.
... Until the power goes or when the PC crashes and you need to restart it (there is not reset button like on desktops, so the laptop will have to be turned off).

Plus, it's a waste of power keeping the computer on. I prefer to turn things off when I don't need them.
Jack okay.
Yes, I see the sound device in Device Manager, as mentioned above. Schematic is attached above by Piernov (thanks man! ).

.

About the Audio missing Momaka , those Russian packs can do miracles , i must admit , but they always fail in some old specialized Audio codecs .
Suspect a VIA codec , or even worse , the infamous Sigmatel . But my bet, it's a Conexant AC-Link Audio .
In any case my friend , the very easy way is to install windows 7 , and let the Windows updates catch the difficult ones .

About the Nvidia 6150 GO , and to be positive , i always thought one of the possible solutions is to acquire the original chipset , do a one good reball , and correct the cooling system failures ...Just so to say .

[Th3_uN1Qu3]

Re: Compaq Presario v6000 laptop repair (failed nVidia chipset)

Yeah, I know the reflow is only a temporary fix. But I imagine if the temperature is kept fairly low, this could prolong the life of the repair quite a bit. From what I remember, the bumpgate issue becomes much bigger only after the chip goes above 60C.

True, especially if it's on its first reflow.

Indeed I haven't narrowed down the issue yet. All I know is that Windows XP SP3 has the drivers for the audio chip installed, but I don't have a sound icon in the sys tray and no software will see the audio device, despite Windows XP recognizing the audio device hardware in Device Manager.

That said, it seems that at least some portion of the hardware is working normally, as warning messages will send a "beep" through the system speakers. Likewise, the Compaq volume up/down/mute shortcut keys on the keyboard will also emit their own distinctive beep when being pressed (even though they don't seem to do anything).

Then your audio chip is fine. You have a driver issue. If you're running XP, make sure you have the "Microsoft UAA Bus for High Definition Audio" installed before you install the sound driver.

[momaka]

Re: Compaq Presario v6000 laptop repair (failed nVidia chipset)

But my bet, it's a Conexant AC-Link Audio .

It is, indeed.

Then your audio chip is fine. You have a driver issue. If you're running XP, make sure you have the "Microsoft UAA Bus for High Definition Audio" installed before you install the sound driver.

Just checked - it's installed.

Anything else I might be missing?

Perhaps I will try uninstalling both the audio driver and the UAA Bus and then reinstall each in the correct order.

If not, I might try a light Linux distro to see what it does. Hopefully those Linux distros don't load my CPU down to 100% for too long, because like I said, my NB/GPU temps tend to follow the CPU temps to within a few degrees Centigrade (and the CPU easily goes over 60C under load at stock voltages and clocks).

... which leads me to:

...and correct the cooling system failures ...Just so to say .

Welp, I found a good workaround for that. It is a little program called CrystalCPUID. With CrystalCPUID, I can change the CPU core voltage and multiplier straight from Windows desktop.

At first, I down-clocked to 1.6 GHz (from 2 GHz stock) and reduced the voltage to 1.10V (from 1.20V stock). That instantly made a difference of about 2°C in idle. Under load, the high CPU spikes in temperature decreased by at least 8°C. In fact, the CPU would rarely go over 60°C with that.

Then I decided to experiment a bit more: I increased the core clock to 1.8 Ghz while keeping the core voltage the same at 1.10V. The CPU was stable with that too. Due to the slight increase in the core clock, my temperatures rised a bit, but not that much. This is what the temps looked like after a long average internet browsing session:

As you can see, I still managed to make the NB/GPU core touch 60°C, but this was after my browser (FF 49) got stuck on a web page that made the CPU stay at 100% for about a minute or two (the high spike in the middle of the graph). Also, this was with an ambient room temperature of 25°C / 78°F - so not bad at all. Before, my core temperature would be much higher.

After all of this, I got a little ballsy and decided to try one more experiment: 1.6 GHz core clock at 1.00V.

Guess what? That worked fine too. The CPU was perfectly stable at this voltage as well, and my temperatures dropped even further. In fact, my highest CPU temperature spikes rarely touched 60°C. For the most part, the CPU stayed in the low-mid 50's Centigrade while browsing online, with the NB/GPU following to within 1°C of that.

As for the under-clock itself: well, it did make the computer slightly slower, particularly when watching 720p MP4/AVI videos. But online, there was hardly a noticeable difference. In fact, I was again surprised how smooth this PC runs with FF 49.

[piernov]

Re: Compaq Presario v6000 laptop repair (failed nVidia chipset)

If not, I might try a light Linux distro to see what it does. Hopefully those Linux distros don't load my CPU down to 100% for too long, because like I said, my NB/GPU temps tend to follow the CPU temps to within a few degrees Centigrade (and the CPU easily goes over 60C under load at stock voltages and clocks).

No reason for a Linux system to use more the CPU than Windows, that's usually the other way around. Obviously for a machine like this, don't take a heavy graphics environment (for example stock Ubuntu or Kubuntu), use a lighter one without graphical effects (for example you could use Xubuntu or Lubuntu).

Welp, I found a good workaround for that. It is a little program called CrystalCPUID. With CrystalCPUID, I can change the CPU core voltage and multiplier straight from Windows desktop.

I'm not convinced it actually changes the CPU voltage, but if it does it is great.

[jiroy]

Re: Compaq Presario v6000 laptop repair (failed nVidia chipset)

It is, indeed.


Just checked - it's installed.

Anything else I might be missing?

Perhaps I will try uninstalling both the audio driver and the UAA Bus and then reinstall each in the correct order.

If not, I might try a light Linux distro to see what it does. Hopefully those Linux distros don't load my CPU down to 100% for too long, because like I said, my NB/GPU temps tend to follow the CPU temps to within a few degrees Centigrade (and the CPU easily goes over 60C under load at stock voltages and clocks).

... which leads me to:

Welp, I found a good workaround for that. It is a little program called CrystalCPUID. With CrystalCPUID, I can change the CPU core voltage and multiplier straight from Windows desktop.

At first, I down-clocked to 1.6 GHz (from 2 GHz stock) and reduced the voltage to 1.10V (from 1.20V stock). That instantly made a difference of about 2°C in idle. Under load, the high CPU spikes in temperature decreased by at least 8°C. In fact, the CPU would rarely go over 60°C with that.

Then I decided to experiment a bit more: I increased the core clock to 1.8 Ghz while keeping the core voltage the same at 1.10V. The CPU was stable with that too. Due to the slight increase in the core clock, my temperatures rised a bit, but not that much. This is what the temps looked like after a long average internet browsing session:

As you can see, I still managed to make the NB/GPU core touch 60°C, but this was after my browser (FF 49) got stuck on a web page that made the CPU stay at 100% for about a minute or two (the high spike in the middle of the graph). Also, this was with an ambient room temperature of 25°C / 78°F - so not bad at all. Before, my core temperature would be much higher.

After all of this, I got a little ballsy and decided to try one more experiment: 1.6 GHz core clock at 1.00V.

Guess what? That worked fine too. The CPU was perfectly stable at this voltage as well, and my temperatures dropped even further. In fact, my highest CPU temperature spikes rarely touched 60°C. For the most part, the CPU stayed in the low-mid 50's Centigrade while browsing online, with the NB/GPU following to within 1°C of that.

As for the under-clock itself: well, it did make the computer slightly slower, particularly when watching 720p MP4/AVI videos. But online, there was hardly a noticeable difference. In fact, I was again surprised how smooth this PC runs with FF 49.

Nicely done Momaka; Excellent thought i say . Instead of working on the cooling system , just control the CPU . And yes , such programs are indeed made by genius people . There's no need to underestimate excellent programs . An example , I found it very harsh from one member saying Spinrite is a Hoax . Probably never used its efficiency .
Anyway , I'm definitely gonna add this CrystalCPUID to the arsenal .

Keeping Temperature under 60° is a turn hit , because excessive heating is one of two factors destroying those VGAs , of course the other factor was the cheap materials , including solder ..

Keep us informed

[Th3_uN1Qu3]

Re: Compaq Presario v6000 laptop repair (failed nVidia chipset)

It does actually change CPU Voltage, that's because the nForce chipset allows it. Back in my days with the DV9000 i used to use RMClock.

[momaka]

Re: Compaq Presario v6000 laptop repair (failed nVidia chipset)

Okay, I have a few small updates as I've been using this laptop a lot more often in this last month.

First thing I want to say is how AMAZED I am by this laptop's little AMD Turion CPU. Last time I mentioned I was able to run the CPU at 1.6 GHz @ 1.00V core without a hitch. Well, guess what? Turns out I could lower the core voltage even further for that clock. How low? Let's say I got too ballsy at first and tried 0.925V just for fun. That made the laptop bluescreen immediately, so obviously 0.925V was too low. Next I tried 0.95V core. The CPU took it, but crashed after about 30 minutes of heavy browsing online (stress-tested with some YouTube), so that was too low as well. But clearly 0.95V wasn't too far from what the core needed. Thus, I tried 0.975V and BINGO! - CPU 100% stable at 1.6 GHz. So I started using the laptop with that setting almost daily. Even with heavy browsing (YouTube - for this laptop, it is) it NEVER crashed. So 1.6 GHz @ 0.975V core is perfect for my CPU.

As you might imagine, this drop in core voltage also dropped the CPU (and GPU) temperatures further - about 2°C lower for the GPU (with CPU at full load) with an ambient room temperature of 79°F/26°C.

Then, we had some very warm weather here, so I decided to do even more testing as my room was steady at 80°F/27°C. I sat down one night and started playing YouTube videos at 480p (the max this laptop can do without stutter, thanks to the GeForce 6150's H.264 hardware acceleration). After 2 hours of watching, I was finally able to see the absolute maximum temperature for the CPU and GPU: 60°C with the CPU clocked at 1.6 GHz and 0.975V core. Here's a SpeedFan graph, for those who like graphs:
https://www.badcaps.net/forum/attach...1&d=1508366918

While 60°C isn't very good, this is again only after the laptop has been stress-testing for 2 hours. If I was to watch a 10-20 minute video, take a break for 10 miutes, then watch again, my temperatures were much lower: around 54-55°C tops. In fact, here is a graph of that:
https://www.badcaps.net/forum/attach...1&d=1508366918
On the left side side of this graph, you can see the CPU (green line) easily peaks over 60°C - this is with the stock settings of 2 Ghz @ 1.20V core. And I didn't even let it run for too long, because the GPU (red line) was already starting to quickly peak towards 60°C. If I actually ran the laptop for 2 hours of YouTube, I wouldn't be surprised if the GPU and CPU go way above 65°C.
Now, on the right side of that graph above, I changed the CPU frequency and core voltage to 1.6 GHz / 0.975V and ran the same browsing tests. And here, you see that for everyday normal browsing, both the GPU and CPU don't really go above 56°C.

Therefore, I think that 1.6 GHz @ 0.975V core is the perfect balance between CPU speed and heat. After all, 1.6 Ghz is only a 20% loss in performance from the stock 2 GHz clock, so not a big deal, considering I am getting an almost proportional drop in temperatures.

But if you think I would stop here, you're wrong. Just like overclockers who want to squeeze every drop of performance out of their CPUs - well, I went pretty much in the opposite direction: see how far I can bring down the CPU core voltage for each multiplier/CPU speed setting I had available in CrystalCPUID. That said, I didn't go all the way down to the lowest frequency, as my main goal was really to see if I could reach 0.900V core.

And guess what? IT IS POSSIBLE!!! All I had to do was under-clock the CPU to 1.2 GHz. Actually, 1.4 Ghz might have worked too, but based on the other voltages and multipliers I experimented with, I didn't think to try it. Nonetheless, I am still so amazed - after all, who would've thunk that a 90 nm CPU from over a decade ago could actually run on 0.900V core? (Now I actually want to go and dig every AMD Athlon XP and Athlon 64 rig I have and test it with CrystalCPUID, but I think that will be material for another thread. )

With the CPU @ 1.2 GHz and 0.900V core, I could run YouTube all day and my temperatures would not actually touch 60°C at all. Of course, it is worth noting that at frequencies below 1.6 GHz, YouTube actually started getting choppy at both 360p and 480p fullscreen (especially at 1.2 GHz). With the normal non-fullscreen video, there videos were smooth - at least after the video downloaded fully. Actually, this was true even when the CPU is running at stock 2 GHz, because it is only a single-core CPU after all.

Now for those of you who are still suspicious of my findings, I checked the total power draw of the laptop with my Kill-A-Watt meter as I tested the various speeds and core voltages. Here are the results of that:

AMD Turion 64 MK-36
==== stock CPU voltage and speed (2.0 GHz, 1.2V)====
idle: 28.8 Watts
Youtube: 47.5 Watts
Max load: 48.7 Watts

==== CPU @ 1.8 Ghz (9x) and V_core = 1.00V ====
idle: 25.7 Watts
Youtube: 36.5 Watts
Max load: 37.7 Watts

==== CPU @ 1.6 Ghz (8x) and V_core = 1.00V ====
idle: 25.6 Watts
Youtube: 36.2 Watts
Max load: 36.8 Watts

==== CPU @ 1.6 Ghz (8x) and V_core = 0.975V ====
idle: 25.5 Watts
Youtube: 33.8 Watts
Max load: 35.8 Watts

==== CPU @ 1.4 Ghz (7x) and V_core = 0.950V ====
idle: 24.9 Watts
Youtube: 33.4 Watts
Max load: 34.0 Watts

==== CPU @ 1.4 Ghz (7x) and V_core = 0.925V ====
idle: 24.7 Watts
Youtube: 32.9 Watts
Max load: 33.4 Watts

==== CPU @ 1.2 Ghz (6x) and V_core = 0.900V ====
idle: 24.4 Watts
Youtube: 30.9 Watts
Max load: 31.4 Watts

As you can see, going from stock frequency/core voltage to 1.6 Ghz @ 0.975V core made a huge difference and dropped almost 13 Watts from the CPU's TDP. So for 20% loss in performance, I dropped the overall TDP of the CPU by almost 40% - that's amazing! (Of course, the overall temperature did not drop by 40%, as the heatsink is another factor here.)

However, going from 1.6 GHz / 0.975V core to 1.2 GHz / 0.900V core only lowered the total power draw by 4.4 Watts. Thus, it doesn't make sense to lower the CPU frequency and core voltage past a certain point, as there is a point of diminishing returns. Perhaps I might try lowering the core voltage even further below 0.900V, though I just don't see a point to it and I have a feeling the CPU won't be able to go that much further down. After all, this is still a 90 nm CPU.

That said, I can't wait to install my 1.9 GHz Athlon X2 CPU in this thing and play with CrystalCPUID. On that note, BIG THANK YOU to the makers of CrystalCPUID. Not only does it work, but it works perfectly. I can even automate things so that the CPU throttles up and down depending on the load (though I didn't do that, as I don't really see a reason to). The neat thing about CrystalCPUID is that it allow you to create a desktop shortcut with the core speed and voltage you like. So all I have to do is double-click that when the OS loads, and all is good

As for the audio issue - I still haven't attempted to fix it , mainly because I've only been using the laptop for online browsing, so it wasn't that big of a deal. But I haven't forgotten about it and it's still on my TO-DO list. The more I play with this laptop, the more I see it as a possible future upgrade option once I need to move away from (now-ancient) Dell Latitude C600, which is still currently my "main" laptop.

[piernov]

Re: Compaq Presario v6000 laptop repair (failed nVidia chipset)

Great experiments, though you could use a software like OCCT to stress-test the CPU and get nice plots of temperature (and other info depending on what your system reports).
Indeed, 0.9V is quite low for those CPUs. However they are already painfully slow at their max frequency so at a lower frequency it must be even harder to use comfortably.

Also it may be possible to do a hardware voltage mod for the chipset. Though it may have an integrated voltage regulator.

[momaka]

Re: Compaq Presario v6000 laptop repair (failed nVidia chipset)

Great experiments, though you could use a software like OCCT to stress-test the CPU and get nice plots of temperature (and other info depending on what your system reports).

Thanks!

Yes, OCCT is a good idea, as I probably would have gotten more repeatable results and more solid way of testing. But I figured YouTube was good enough, as that is the max I really stress most of my browsing PCs.

Would be pointless to try to run any games on the GF 6150 GO chipset, as it's a pretty weak GPU - even for its time.

However they are already painfully slow at their max frequency so at a lower frequency it must be even harder to use comfortably.

No, not really.

Between 1.6 GHz and 2 GHz, the only difference is on websites with heavy scripting or YouTube. Other than that, it's pretty snappy. I've actually seen stock Windows 10 PCs with new hardware run slower than that even with minimum bloat - at least until the OS loads everything. When it comes to browsing, of course a new gen CPU/GPU runs much faster.

That said, I think Firefox has really made some good improvements in version 47 and up. I thought I'd get better performance with the older versions, but this laptop actually runs much much better with the newer versions. No issues browsing whatsoever. Even dumb news websites with heavy scripts are still bearable.

Also, compared to my Latitude C600 laptop (Pentium 3 @ 700 MHz and 512 MB of RAM), this laptop clearly flies. Heck, even compared to my Pentium 4 PCs, it seems to do better. Though I suspect a big part of that is now due to Adobe Flash getting dumped in favor of HTML5. Seems that HTML5 favors short wide CPU architectures over long ones (Pentium 4 and D), whereas Adobe Flash was is opposite.

Also it may be possible to do a hardware voltage mod for the chipset. Though it may have an integrated voltage regulator.

Yeah, I am going to check when I crack it open again to upgrade the CPU. Small chance, though I really hope the board has one of those "dumb" PWM controllers that I can just set the voltage with just a few resistors. If that's the case, GPU/chipset will get volt-modded too.

[momaka]

Re: Compaq Presario v6000 laptop repair (failed nVidia chipset)

As for the audio issue - I still haven't attempted to fix it...

Well, turns out there is nothing to fix here - the audio hardware is working!

Hence, Mr. Th3_uN1Qu3 was right (as usual ) - it was a driver issue all along. Actually, I wasn't able to get the audio working under the original Windows XP SP3 installation, but I made a test installation of Windows 7 Pro onto another HDD and it picked up the audio drivers all by itself. Not only that, but even the HP/Compaq "quick" buttons driver was picked up by Windows 7, which also made them work fine (the ones that mute and turn up/down the audio).

So we have audio too!

Now here's the kicker, though: Windows 7 has it's own (updated) CPU driver for that Turion CPU, and it throttles the CPU automatically by itself. This might seem like a good thing, but I don't like it.
Why?
Because the throttling is not that optimal. In idle mode, the CPU frequency scales all the way down to 800 MHz with the core voltage held at 0.950V. This is a fine idling speed and voltage. However, when I put the CPU under any considerable load, the frequency jumps up to 2 GHz while the core voltage goes up to the stock 1.200V - and that's what I want to avoid, as running the CPU at 1.200V core for prolonged periods of time eventually makes it run very hot, which eventually overheats the GPU as well.

What's worse is the Windows 7 AMD CPU driver automatically and instantly overrides my CrystalCPUID settings, so I can't control the CPU anymore! :\

However, I noticed something interesting: the Windows 7 AMD CPU driver has an intermediate throttling speed for this CPU at 1.6 GHz. At this frequency, the core voltage is lowered to 1.100V - not quite as optimal as my 0.975V core , but still a whopping 0.1V less than stock.

With that in mind, I started trying stuff and eventually found a somewhat okay method for throttling down the CPU in Windows 7: the Windows Power options/settings in Control Panel. In particular, what I did is I set the power plan to "Balanced", then went in the Advanced settings and set the Maximum CPU frequency to 80%. This throttles the CPU to 1.6 GHz (as 80% of 2 GHz is 1.6 GHz). So with this "tweak", my CPU now runs at a maximum of 1.6 GHz under full load and 1.100V CPU core voltage. I tried other throttling percentages, but none worked - anything lower than 80% would lock the CPU at 800 MHz and 0.950V core... which is nice from a thermal stand point, but also A LOT slower online.

Thus, 1.6 GHz @ 1.100V CPU core it was. Not great, but still better, as the instant spikes in CPU temperature were kept under 60°C for the most part (while at 2 GHz, it's not uncommon to see them go up to and above 65°C) - and all of that with a chilly 18°C room temperature. I can't imagine this to do well in the hot summer temperatures (whereas my CPU core voltages with CrystalCPUID were proven to work fine in warm weather). Thus, if anyone has any ideas how to stop Windows 7 from overriding CrystalCPUID, I'm all ears.

On the positive side of things, the nVidia GeForce 6150 Go driver for Windows 7 (from nVidia, not the stock MS one) also appears to throttle both the onboard GPU core clock (down to 100 MHz, up from 300 or 400 MHz stock) and its voltage, because the GPU stays much cooler than before, even with the above large CPU spikes in temperature. And, as an added bonus, H.264 acceleration functions A LOT smoother under Windows 7 than XP. So YouTube with Firefox appears to run absolutely smoothly in windowed (standard) mode. In fullscreen, the video can still get a bit jittery until it fully downloads/buffers, but after that it's smooth again - all the way up to 480p. Funny thing is, YouTube HTML5 player automatically tries to load in 720p... and in standard mode, it runs okay too, as long as it's a 30 FPS video.

That said, I still prefer Windows XP as my main OS. Therefore, I would still like to get the audio driver issue solved under Windows XP, if possible. On that note, I also bought two cheap USB stereo adapters from eBay for $1 each (I'll do a review on those someday later when I use them a bit more). I tested both of them with the Windows XP installation, and I had the audio work from the USB adapter just fine. Thus, it looks like whatever is wrong with that XP installation does not completely affect all audio devices. So there is some hope.

But if not possible, then I guess Windows 7 it would be. Funny thing is, since I didn't want to bog down the laptop's CPU with the Windows 7 installation, I installed it on a 2.5" HDD through a desktop motherboard first - a Biostar GeForce6100-M9 (socket 939 with onboard GeForce 6100 Northbridge/video and NF-410 Southbridge - i.e. pretty similar to the laptop). After Windows 7 installed and I tested it on the Biostar mobo, I transfered the HDD into the Compaq Presario V6000 laptop and... it worked!

So anyways, I'll come back with more updates on this after I try the dual core Athlon CPU in this laptop (might be a while again). But until then, if anyone has any ideas how to solve the Windows 7 issue with overriding CrystalCPUID CPU settings or solve the issue with the audio drivers in XP, that would be great. Actually, I might try another test install of Window XP onto another HDD, but I will have to clear one of my 2.5" HDDs first. The one I used for the test Windows 7 installation had tons of bad sectors, so on every other boot, Windows would either blue screen (BSOD) during the loading screen or say a file is missing/corrupt. After restarting once or twice, it would load again. (Thus, for those of you who don't think that a bad/marginal HDD can cause BSODs, this is pretty solid proof right there).

[jiroy]

Re: Compaq Presario v6000 laptop repair (failed nVidia chipset)

Hi Momaka; About the Audio problem with XP SP3 , Your Compaq Presario V6000 originally come with a Vista CD as a system recovery , Something very logical since Nvidia didn't make XP drivers for Nvidia 6150GO , (that with taking into consideration of some third parties or programmers like me , who managed to make compatible and successful drivers for this chipset ) , so it is also logical to assume that your audio chipset isn't also compatible with XP or XP SP3 , While it's totally compatible vista and above ..

That's being said , and looking at the Schematics brought by Piernov; , your Audio Chipset is Conexant Venice Amon CX-20549-12 .. So It's a Conexant driver that windows Automatic updates can take it easily under Conexant Audio-link .

If we look at Compaq Presario v6000 series in Hp site , there are only Vista Drivers , one is Conexant High definition Audio driver , and the other is Realtek high definition audio driver . If you're lucky and you have the conexant version
like Piernov schematic , and it's compatible with your board , then next you'll have to search for another kind of laptops that have the same chipset , but have also XP drivers . This link from Hp state that HP 530 have the XP driver for the same chipset Conexant Venice Amon CX-20549-12 :

https://support.hp.com/us-en/drivers...ook-pc/3375953

Hope it works my friend