Recapping Asus PTGD1-LA

[Pentium4]

This thing was handed down to me, I'm the third owner. The previous owners are both now passed away so this is sentimental to me now. I ran it in my file server up until March, 2012 when I saw one of the Ltec caps bulged. Since then, the other Ltec bulged just sitting since then

I want to re-purpose this as a HTPC. It has a lot of caps on it that are from bad manufacturers. There aren't may good caps at all. There are two Rubycon caps next to the NB chip that I will leave, and two Fujitsu caps that will stay. The VRM low caps are good polymers.

Besides that, the rest are OST, Ltec, KZG, and GSC. My question is...has anyone ever had any grief or issues from not replacing caps on motherboards in the 330uF and lower range? I only want to do this job once, but don't want to replace more caps than I need to. Last time I recapped an old board, even the 47uF GSC caps tested perfectly in spec, and they were 13 years old. Any shared experience is appreciated

[ChaosLegionnaire]

Re: Recapping Asus PTGD1-LA

actually, where a cap is located and what filtering its doing is more important than its capacitance value or can size.

those 47uF caps are typically tasked with filtering the onboard sound or fan headers right? those arent stressed much thermally or otherwise so its fine to leave them. if there are any caps near the pci-e, agp or pci slot where there is going to be a hot n' powerful video card around those caps, jap caps or maybe even poly modding have to be used there. any non-jap caps or non-polys that arent jap in such a hot location will pop in a couple of years in a cramped htpc case.

likewise, look for caps near any switching mosfets or chips (NB, SB, LAN etc.)that get very hot during use. those hot mosfets/chips conduct their heat to the caps via the thick copper traces and kill them fast. stick heatsinks on those things and modify the casing airflow over those areas to avoid this from happening.

also try to analyse the airflow pattern in the closed case. look out for caps on board edges or airflow deadzones without moving air. ghetto mount inside the case a small 40-60mm fan to circulate these deadzones. u can also mod the side panel to mount 120-140mm fans to cool the whole board.

moral of the story, its better to understand where the hot zones are for a particular mobo & case setup so u know where its going to get hot so precautions (e.g. recapping in the right places) can be taken...

btw, i see the cpu vrm out uses sanyo sepc polymers right? those are one of the best polys out there. i luv those on hot p4 boards and on my overclocking boards when its going to get toasty there.

[momaka]

Re: Recapping Asus PTGD1-LA

My question is...has anyone ever had any grief or issues from not replacing caps on motherboards in the 330uF and lower range? I only want to do this job once, but don't want to replace more caps than I need to.

No, I haven't.
But I usually change caps based on how close they are located to sources of heat and what type of filtering those caps do, just as ChaosLegionnaire suggested.

In general, crap caps should never be left at the input nor the output of buck-type regulators. You can recognize buck regulators since they have an inductor coil (or two) near them. On your motherboard, there are three (3) of these: one between the ATX connector and Southbrdige (likely for the RAM DDR voltage, so about 2.5-2.6 Volts), another near the PCI-E slot (probably Northbridge Vcc/Vdd, usually around 1.2 - 1.5 V), and the CPU VRM itself (CPU V_core).

Any caps filtering the PSU outputs should also be changed (i.e. anything directly connected to the 3.3 V, 5 V, and 12 V rails on the PSU). You can use your multimeter to find those, as their positive (+) leg will show 0-1 Ohms to one of the pins on the ATX connector.

Most of the caps around the PCI and PCI-E slots are usually used for that ^, so again, as ChaosLegionnaire suggested, you should change those if there will be a hot video card blowing air on them. I normally tend to leave them when my cap stock is low, though, since I often leave the cases in my PCs open, and I rarely use high-power cards. Never had a problem yet.

Everything else is probably behind a linear regulator, so crap caps aren't *usually* a problem there. That said, it's still not a bad idea to replace some of the bigger caps, particularly around the NB, SB, and RAM. But for linear regulator outputs, try *not* to use polymers, unless the cap(s) will be in a very hot spot.

Also worth nothing is that those tiny caps between the RAM slots are probably for the RAM Vtt (which is half of the DDR voltage, or roughly 1.25-1.3 V). For those, if you replace just one or two, that would be enough. However, your replacement caps need not have the same exact capacitance here. So don't feel like you need to purchase tiny caps just for that. Anything up to 1000 uF will actually be fine (but much less if you really have to use a polymer there).

The caps that you can certainly leave alone are those that filter the audio coupling, USB power (two caps by the FP USB headers, for example), and some of the other tiny caps by the fan headers and near the LAN, sound, and LPC/super I/O chip.

This thing was handed down to me, I'm the third owner. The previous owners are both now passed away so this is sentimental to me now.

Sorry to hear that. I guess you will be making it work work again in their honor.

[Pentium4]

Re: Recapping Asus PTGD1-LA

Thanks for the replies! It's just hard to leave GSC caps even if they aren't likely to cause trouble. Even though those caps between the RAM slots aren't be stressed much, those DDR RAM modules run hot! They idle around 55C on this board.

This board came with a Prescott P4, but I will be using a Cedar Mill P4 in place of it to reduce heat. It's in a full sized ATX case so heat shouldn't be a problem.

And we like to view recently taken picture/videos on the big screen so that card reader on the front will come in handy.

As for video card, this is just going to have a Radeon HD 4350 in it. This one...That I recapped almost 2 years ago, and has had 11,000 more running on it so far with no issues

There will be a 92mm exhaust fan right above it.

The caps that you can certainly leave alone are those that filter the audio coupling, USB power

I've actually had trouble with these caps on multiple occasions. I've seen KZG, and GSC fail in this section before, causing plugged in devices to act funny. I usually replace them

btw, i see the cpu vrm out uses sanyo sepc polymers right? those are one of the best polys out there. i luv those on hot p4 boards and on my overclocking boards when its going to get toasty there.

YES. Love those caps for boards like these

However, your replacement caps need not have the same exact capacitance here. So don't feel like you need to purchase tiny caps just for that. Anything up to 1000 uF will actually be fine (but much less if you really have to use a polymer there).

You mean for those tiny caps? How important do you think ESR is there?

[Pentium4]

Re: Recapping Asus PTGD1-LA

momaka, I think you will be very pleased to know that this PSU will be powering the machine (and it's also been powering the current machine for 11 months 24/7)






Original post for reference: https://www.badcaps.net/forum/showpo...postcount=1618

[momaka]

Re: Recapping Asus PTGD1-LA

It's just hard to leave GSC caps even if they aren't likely to cause trouble. Even though those caps between the RAM slots aren't be stressed much, those DDR RAM modules run hot! They idle around 55C on this board.

Was that after you recapped the board or before? Because those bulged caps on the RAM Vdd buck regulator could sure have caused that.

As for video card, this is just going to have a Radeon HD 4350 in it. This one...That I recapped almost 2 years ago, and has had 11,000 more running on it so far with no issues
https://www.badcaps.net/forum/attach...9&d=1391966498

That's a good choice for a HTPC.

And I love that PSU heatsink wedged in there . Probably helps dissipate an extra Watt or two. With a 92 mm fan blowing on it, however, I doubt it will run hot at all.

I've actually had trouble with these caps on multiple occasions. I've seen KZG, and GSC fail in this section before, causing plugged in devices to act funny. I usually replace them

I guess I better take that suggestion back, lol.
Maybe it's because I run most of my PCs with the case open that the caps never bulged?
IDK, I still haven't had any of my own motherboards develop puffy caps on me. With PSUs, it's a different story.

YES. Love those caps for boards like these

Me three. In fact, I replaced those short UCC TMZ caps on the ASUS P5GC-MX you gave me exactly with Sanyo SEPC. I had other polymer cap choices too (Nichicon LF, UCC PSC, and Lelon OCRZ), but I settled on the SEPC, because I though the purple strips on the SEPC go well with the yellow solder mask on the motherboard .

You mean for those tiny caps? How important do you think ESR is there?

Not at all. Actually, it's even better if the ESR is *not* too low, because they filter RAM Vtt, which is almost always generated from a linear regulator (on your motherboard, that would likely be that 6-pin TO-252 IC between the ATX connector and the RAM slots)

momaka, I think you will be very pleased to know that this PSU will be powering the machine (and it's also been powering the current machine for 11 months 24/7)

Macron!!!!!!!
Those things just won't quit no matter what. Looks like that one is a single-transistor forward design (a.k.a. forward converter). Nice!
*edit*
Oh, I see. It's actually a 2-transistor forward. Even better!
And with that recap job, it will keep going long enough even for your kids to remember.

[Pentium4]

Re: Recapping Asus PTGD1-LA

So my OCD got the best of me.... I ended up replacing every single non japanese cap. Every GSC cap tested perfect...So did every OST. The KZG were flawless as well. The four 1000uF KZG on the vrm high read 990uF, 1000uF, 1003uF, 1010uF. And all were 0.02Ω ESR. However, if you can spot those four 100uF 16V 6.3x7mm Ltec ST caps by the audio/realtek chip, one of those failed! It read 83uF and 8.82Ω ESR. These caps were old stock even in 2004 though. They were all manufactured on the 43rd week of 2001. I doubt they would have caused any major issues but it's nice to know that they are gone now They were all replaced with 100uF 16V KMG.

Was that after you recapped the board or before? Because those bulged caps on the RAM Vdd buck regulator could sure have caused that.

Before. That explains it They didn't even get that hot while running a memtest after the recap!

And I love that PSU heatsink wedged in there . Probably helps dissipate an extra Watt or two. With a 92 mm fan blowing on it, however, I doubt it will run hot at all.

The card used to run extremely hot. While in my core 2 system, when running HD netflix it would easily get up to 77C, the highest is ever hit is 84C. I know that's well below the thermal limit for the GPU but I don't like it running that hot. With new thermal paste and that heatsink wedged in there, it dropped the temperature a little bit. After that it never went higher than 73C

I guess I better take that suggestion back, lol.
Maybe it's because I run most of my PCs with the case open that the caps never bulged?

Maybe! And I'm sure they're fine most of the time. I can just see myself getting pissed off in the future when I see that those already 11 year old OST caps decided to bulge because I was too lazy to replace a few extra caps

Me three. In fact, I replaced those short UCC TMZ caps on the ASUS P5GC-MX you gave me exactly with Sanyo SEPC. I had other polymer cap choices too (Nichicon LF, UCC PSC, and Lelon OCRZ), but I settled on the SEPC, because I though the purple strips on the SEPC go well with the yellow solder mask on the motherboard

Nicely done!! I'm sure the CPU likes that. I have seen those short TMZ caps on those boards fail before, especially when the CPU was a Pentium D.

Not at all. Actually, it's even better if the ESR is *not* too low, because they filter RAM Vtt, which is almost always generated from a linear regulator (on your motherboard, that would likely be that 6-pin TO-252 IC between the ATX connector and the RAM slots)

Perfect! I replaced all of those 330uF 6.3V GSC (No series) caps with 330uF 10V Chemi-con KMG.

Macron!!!!!!!
Those things just won't quit no matter what. Looks like that one is a single-transistor forward design (a.k.a. forward converter). Nice!
*edit*
Oh, I see. It's actually a 2-transistor forward. Even better!
And with that recap job, it will keep going long enough even for your kids to remember.

Haha, let's hope so! I do love this PSU. It's so silent and runs very cool. I remember back then I didn't have a 0.47uF 50V cap to replace that GL cap on the primary, so I hope it doesn't have any issues because of that.

So it fired up no problem and ran a memtest for over an hour with no issues. So now it's time for an OS!

My DDR RAM supply wasn't as good as I thought it was. What do you think would perform better under Windows 7 32 bit. 2.375GB @ 400MHz, or 3.5GB @ 333MHz? It will just be playing movies, but I'm thinking Windows would benefit from the extra memory size over bandwidth.

The CPU it will be using is a Pentium 4 651. Which is 3.40GHz, Cedar Mill so it's 65nm, and the CPU Sspec is SL9KE revision D0, so it's the 65W TDP one It will do great with the big heatsink that Asus provided, originally to cool a Pentium 4 Prescott 530

Edit: I found an old board in a packed away box with DDR 400 in it. So it's going to have 3GB @ 400MHz

[Wester547]

Re: Recapping Asus PTGD1-LA

So my OCD got the best of me.... I ended up replacing every single non japanese cap. Every GSC cap tested perfect...So did every OST. The KZG were flawless as well. The four 1000uF KZG on the vrm high read 990uF, 1000uF, 1003uF, 1010uF. And all were 0.02Ω ESR. However, if you can spot those four 100uF 16V 6.3x7mm Ltec ST caps by the audio/realtek chip, one of those failed! It read 83uF and 8.82Ω ESR. These caps were old stock even in 2004 though. They were all manufactured on the 43rd week of 2001. I doubt they would have caused any major issues but it's nice to know that they are gone now They were all replaced with 100uF 16V KMG.

Sure it was the 43rd week of 2001? Usually, LTEC datecodes either look like this:

70308D (the 3rd week of October, 2007, factory "D", batch "8"?)

Or this:

3301T (March of 2003 or 2013, factory "T", batch "8"?)

Or this:

14B (April 2001 or 2011, factory B).

Most of those capacitors appear to have 2004 datecodes so I think the STs have 2004 datecodes (but please correct me if I'm wrong). That said, it's pretty sad when most tiny LTECs and Teapos are not capable of remaining in spec even under very little stress if any. I would guess that those Teapo SKs aren't in good condition either. I think the KZGs lasted because they were on the VRM input. They have a significantly higher failure rate on the VRM output.

Maybe! And I'm sure they're fine most of the time. I can just see myself getting pissed off in the future when I see that those already 11 year old OST caps decided to bulge because I was too lazy to replace a few extra caps

OST capacitors, besides the RLA series, rarely bulge when they fail. The hydrogen gas escapes too slowly so they just dry up. I think they'd be more likely to silently fail.

Nicely done!! I'm sure the CPU likes that. I have seen those short TMZ caps on those boards fail before, especially when the CPU was a Pentium D.

Sure it wasn't a heat-related failure? Curious to know how many TMZ capacitors you've seen go bad without merit - IMHO there isn't actually evidence that they're bad like KZG, more like KZH or KZE maybe (possibly a custom order of KZE in 8x11.5mm/1000uF 6.3V). Any wet electrolytic can dry up under high thermal duress. I'm not saying that it's impossible for TMZ to be a dud series, but I've only ever seen them dry up under very high heat is all. TMV and TMJ are definitely crap though, often found in the CPU VCORE on older ASUS boards.

All that to say, I don't mean to hijack your thread. Very nice work, as always!

[Pentium4]

Re: Recapping Asus PTGD1-LA

Sure it was the 43rd week of 2001? Usually, LTEC datecodes either look like this:

70308D (the 3rd week of October, 2007, factory "D", batch "8"?)

Or this:

3301T (March of 2003 or 2013, factory "T", batch "8"?)

Or this:

14B (April 2001 or 2011, factory B).

Oh that's right they do it different. You're probably right They all say 4301D. So I assumed they were 43rd week of 01. My bad!

Most of those capacitors appear to have 2004 datecodes so I think the STs have 2004 datecodes (but please correct me if I'm wrong). That said, it's pretty sad when most tiny LTECs and Teapos are not capable of remaining in spec even under very little stress if any. I think the KZGs lasted because they were on the VRM input. They have a significantly higher failure rate on the VRM output.

Yeah the other ones tested fine. Around ~96uF and about 0.8Ω ESR. I know the KZG are on the high side, but this is still KZG we're talking about KZG that are ~11 years old, and probably saw 35,000 running hours of mostly Precott CPU usage. They held up pretty well but there was absolutely no way I would have left them!

Sure it wasn't a heat-related failure? Curious to know how many TMZ capacitors you've seen go bad without merit - IMHO there isn't actually evidence that they're bad like KZG, more like KZH or KZE maybe (possibly a custom order of KZE in 8x11.5mm/1000uF 6.3V). Any wet electrolytic can dry up under high thermal duress. I'm not saying that it's impossible for TMZ to be a dud series, but I've only ever seen them dry up under very high heat is all. TMV and TMJ are definitely crap though, often found in the CPU VCORE on older ASUS boards.

I've seen them fail on a few occasions. One board had a Pentium D 3.40GHz, with a stock heatsink, and all 6 of the TMZ caps blew out their bottom bungs. I had pictures, unfortunately when my phone failed I hadn't backed up just 2 months of pictures (January and February 2015) so I lost those pictures. I have also seen TMZ caps fail under the stock heatsink of a Prescott P4. Just one out of the 6. Saw another fail powering a Core 2 system, after 43,000 hours in a very cramped and dusty case. These were all on Asus boards with a 945 chipset.

All that to say, I don't mean to hijack your thread. Very nice work, as always!

No worries, since it's still motherboard cap discussion. And thanks! Hope to see many more years of use out of this board.

[Pentium4]

Re: Recapping Asus PTGD1-LA

Also, I can't blame those Ltec caps that bulged too much. They were clearly the most stressed on the board besides the VRM out. After running that memtest, I physically touched all the caps in what I thought would be hot spots. The spots where the Ltec caps originally failed, were the hottest, by quite a bit. That's why I upped the specs of the caps in those spots. The upper one to the left of the RAM slots, I replaced with a 1000uF 6.3V nichicon HZ. The lower one by the southbridge chipset was replaced with a 1500uF 6.3V nichicon HZ.

[Wester547]

Re: Recapping Asus PTGD1-LA

25,000 hours or 35,000 hours? Your original post over three years ago says 25,000 hours. Or is it 35,000 hours?

Were those Teapo SKs *gasp* in spec?

Also, I can't blame those Ltec caps that bulged too much. They were clearly the most stressed on the board besides the VRM out. After running that memtest, I physically touched all the caps in what I thought would be hot spots. The spots where the Ltec caps originally failed, were the hottest, by quite a bit. That's why I upped the specs of the caps in those spots. The upper one to the left of the RAM slots, I replaced with a 1000uF 6.3V nichicon HZ. The lower one by the southbridge chipset was replaced with a 1500uF 6.3V nichicon HZ.

Well, the fact that one of them bulged in storage does suggest serious chemistry issues. IMO, LMB (I think LMB is worse than its equivalent KZG), LXY (equal to HD), TK, TH, and ST are all LTEC's worst series. The others are hit or miss, leaning on the "miss" side in most cases except Delta power supplies. Also, those Asus boards more likely had 680uF 4V TMJs or TMVs around the CPU, unless the values were 1000uF 6.3V (in which case they'd be TM or TMZ).

[Pentium4]

Re: Recapping Asus PTGD1-LA

25,000 hours or 35,000 hours? Your original post over three years ago says 25,000 hours. Or is it 35,000 hours?

25. You're right. My memory is having lots of ECC errors today

Were those Teapo SKs *gasp* in spec?

I know they're the same thing but these ones said Yageo on them. But yeah, perfectly! Even in those hot spots. Replaced, anyways I don't like 85C caps on motherboards, especially a P4 board!

Well, the fact that one of them bulged in storage does suggest serious chemistry issues. IMO, LMB (I think LMB is worse than its equivalent KZG), LXY (equal to HD), TK, TH, and ST are all LTEC's worst series. The others are hit or miss, leaning on the "miss" side in most cases except Delta power supplies. Also, those Asus boards more likely had 680uF 4V TMJs or TMVs around the CPU, unless the values were 1000uF 6.3V (in which case they'd be TM or TMZ).

Yeah, so the stress just sped up the degradation process. Pretty sure they were TMZ. I sent one of the (same model, not one of the mentioned failed) boards to momaka, and his had TMZ on it.

[ChaosLegionnaire]

Re: Recapping Asus PTGD1-LA

Macron!!!!!!!
Those things just won't quit no matter what.

isnt that also the same name of the endboss in quake 4? the same named joker just wouldnt die/quit either no matter what u threw at him and kept healing himself...

And with that recap job, it will keep going long enough even for your kids to remember.

yes its perfect for your family heirloom private pc museum to hand down to your kids:

well kids, long before your time back in grandpa's time during the 90s, we built things and pcs to last a long time. not the buy n' throwaway sheet u see today in 2050. everything was built with both IQ and EQ. today, things are built with neither IQ nor EQ. stupid designs, cost cutting decisions, totally brainless lack of IQ design there. EQ i dont need to say. if it spoils, they just say thats your own problem! and u're on your own.

ahh, they dont build things like they used to... it can be a family heirloom to be passed down to keep the spirit and flame of quality building alive. the grand kids can learn from the heirloom how to repair or refurb things to make them last and save things from the landfill...

I can just see myself getting pissed off in the future when I see that those already 11 year old OST caps decided to bulge because I was too lazy to replace a few extra caps

i've seen my fair share of failed osts on my p4 boards and none of em were bulging lol...

What do you think would perform better under Windows 7 32 bit. 2.375GB @ 400MHz, or 3.5GB @ 333MHz? It will just be playing movies, but I'm thinking Windows would benefit from the extra memory size over bandwidth.

hmm now that is a tricky question. playing movies doesnt actually require 3.5gb of ram tho esp. not if u use an efficient playback software like mpc. the large ram pool can help with caching the movie files but some playback software actually bypass the windows cache anyway and caching the movie files doesnt do much for performance anyway.

benchmarks have shown that p4s like high memory bandwidth rather than tight timings. the faster mem speed should also help with video decoding if video hardware acceleration (dxva) is not working/available for certain videos.

Edit: I found an old board in a packed away box with DDR 400 in it. So it's going to have 3GB @ 400MHz

is it 2x1gb modules and 2x512mb modules to get 3gb? that combo is important to get dual channel mode working for better mem bandwidth.

The CPU it will be using is a Pentium 4 651. Which is 3.40GHz, Cedar Mill so it's 65nm, and the CPU Sspec is SL9KE revision D0, so it's the 65W TDP one

i briefly had a p4 631 d0 stepping 65w tdp cedar mill too for a year as an interim stopgap while i waited for the e8000 series wolfdales and those cedar mill cpus are cool cucumbers. they were what prescott should have been from the get go. they were also excellent overclockers. i could overclock the 3ghz p4 631 to 4.5ghz with the stock intel cooler. cant beat a 50% overclock on stock cooling. best overclocking chip i ever had so far.

[Pentium4]

Re: Recapping Asus PTGD1-LA

Originally Posted by Grandpa Pentium4
well kids, long before your time back in grandpa's time during the 90s, we built things and pcs to last a long time. not the buy n' throwaway sheet u see today in 2050. everything was built with both IQ and EQ. today, things are built with neither IQ nor EQ. stupid designs, cost cutting decisions, totally brainless lack of IQ design there. EQ i dont need to say. if it spoils, they just say thats your own problem! and u're on your own.

Hahahaha. This cracked me up

hmm now that is a tricky question. playing movies doesnt actually require 3.5gb of ram tho esp. not if u use an efficient playback software like mpc. the large ram pool can help with caching the movie files but some playback software actually bypass the windows cache anyway and caching the movie files doesnt do much for performance anyway.

benchmarks have shown that p4s like high memory bandwidth rather than tight timings. the faster mem speed should also help with video decoding if video hardware acceleration (dxva) is not working/available for certain videos.

They do like the bandwidth don't they. Because of their long pipeline? Going with 3GB @ 400MHz seemed like the way to go. It's 1GBx2 in Channel A1/B1 and 512MBx2 in A2/B2. Scores a 4.6 in Windows 7 pro

So I just threw a hard drive in here that already had an install on it from a computer with the same chipset. It was 64 bit. Doesn't the fact that the OS is 64 bit "negate" half of the cache of the old P4? Would I see any benefit from installing a 32 bit version since it only has 3GB?

i briefly had a p4 631 d0 stepping 65w tdp cedar mill too for a year as an interim stopgap while i waited for the e8000 series wolfdales and those cedar mill cpus are cool cucumbers. they were what prescott should have been from the get go. they were also excellent overclockers. i could overclock the 3ghz p4 631 to 4.5ghz with the stock intel cooler. cant beat a 50% overclock on stock cooling. best overclocking chip i ever had so far.

They do run cool, but WOW, on the stock cooler?! That's really impressive. Did you see much of an increase in performance? Hope the mobo had a beefy VRM

[Pentium4]

Re: Recapping Asus PTGD1-LA

Here are some pictures. From the first post you'd think the VRM looked really wimpy...but 8 of the 12 FETs are on the bottom side.

Here's a picture of it all assembled. I had to remove the ghetto heatsink on the video card to make room for the PCI wireless card and the Optical audio card. I almost took things too far and even recapped the PCI wireless card



The hard drive was later cloned to an 80GB Seagate ST380811AS with a whopping 53,701 hours on it, with no errors. It makes a very loud high pitched whining sound though, so I'm probably going to swap it out soon for something else.

It was streaming full HD 1080P with GPU acceleration enabled, and the CPU was averaging around 15% usage.

Edit: momaka, that beautiful Sony sub in the left corner is rated for 300W RMS. It sounds AWESOME. It's 11 years old and sounds like it did the day we got it.

[momaka]

Re: Recapping Asus PTGD1-LA

So my OCD got the best of me.... I ended up replacing every single non japanese cap.

LIKE. A. BOSS.

However, if you can spot those four 100uF 16V 6.3x7mm Ltec ST caps by the audio/realtek chip, one of those failed! It read 83uF and 8.82Ω ESR.

If it's for the audio decoupling, it won't hurt anything. However, if you were to plug in low-impedance headphones (like <16 Ohms) directly to the PC, you might notice that one side is slightly off.

The card used to run extremely hot. While in my core 2 system, when running HD netflix it would easily get up to 77C, the highest is ever hit is 84C. I know that's well below the thermal limit for the GPU but I don't like it running that hot. With new thermal paste and that heatsink wedged in there, it dropped the temperature a little bit. After that it never went higher than 73C

73°C?!?
I think you should to put a fan under the video card or blowing to the side of it... or something. I've seen many modern passively-cooled GPUs die, even these low-end one. A 60-80 mm fan running at 5V should be both quiet and provide enough airflow to keep it cool.

All 12 of my Radeon HD 2400 video cards (yes, twelve ) have active coolers on them, and even under full load gaming for hours, I've never seen the GPU temps go above 58°C (with 30°C room temperature and semi-closed case with no exhaust fans). With 20C room temperature, closed case, and no case exhaust fan, I get 56°C max. In both cases, the GPU fan speed never goes above 20%, which is the same for both idle and under load. So I think even a little bit of air going over your video card will make a world of a difference.

Then again, I am probably just a teeny TINY bit paranoid when it comes to cooling BGA stuff (more like, A LOT). So feel free to ignore what I said above.

What do you think would perform better under Windows 7 32 bit. 2.375GB @ 400MHz, or 3.5GB @ 333MHz?

I don't know , I never found the memory speed/bandwidth to make enough significant difference in daily use, so I always go for more RAM. But that's just my *general* experience.

Pretty sure they were TMZ. I sent one of the (same model, not one of the mentioned failed) boards to momaka, and his had TMZ on it.

Actually, I just checked up on that. I had to, because let's face it: Wester547 knows his caps . When he says something about caps, he's probably right. And in this case he was! The original caps on the Asus board you gave me were indeed Chemicon TMV.

Heh... if your memory is having "lots of ECC erros", then I guess mine has whole chips blown off.
... Because, after all, I was the one who (wrongly) brought TMZ into the discussion.

Also, those Asus boards more likely had 680uF 4V TMJs or TMVs around the CPU

And that's exactly what they were (TMVs): 680 uF and 4 V.

Here are some pictures. From the first post you'd think the VRM looked really wimpy...but 8 of the 12 FETs are on the bottom side.

Yeah, I don't know why ASUS does that, but it does work. If there is one thing they DO right, it is CPU VRMs. The ASUS board you gave me has a 3-phase VRM with 6 MOSFETs total. It handles that 150 Watt Pentium D perfectly fine. Moreover, I check the motherboard manual, and what's interesting is that this board was actually built to handle primarily Pentium 4 and Pentium D CPUs - even the extreme edition ones. So that makes me believe ASUS really does put some good thought into their CPU VRMs.

Edit: momaka, that beautiful Sony sub in the left corner is rated for 300W RMS. It sounds AWESOME. It's 11 years old and sounds like it did the day we got it.

Nice!
I find most Sony speaker systems and headphones to have very good sound and pretty decent build quality.

Here, try this on it (preferably, turned up):
https://www.youtube.com/watch?v=NIUlgMF7aCo

[Pentium4]

Re: Recapping Asus PTGD1-LA

LIKE. A. BOSS.

If it's for the audio decoupling, it won't hurt anything. However, if you were to plug in low-impedance headphones (like <16 Ohms) directly to the PC, you might notice that one side is slightly off.

Good to know for future reference, but it doesn't matter on this board anymore since they're gone

73°C?!?
I think you should to put a fan under the video card or blowing to the side of it... or something. I've seen many modern passively-cooled GPUs die, even these low-end one. A 60-80 mm fan running at 5V should be both quiet and provide enough airflow to keep it cool.

That was during the summer in a hot room watching HD netflix, with the inefficient silverlight...Earlier I watched a 1080p mkv file and it hit 63C, and it was doing the decoding. You can definitely tell this is a low wattage card, since it idles at 59C. Before with the ghetto heatsink, it idled at 56C so it did make a difference

I'm on this machine right now. Both those scenarios were with a lower ambient temperature though. Now that the gas fireplace has been running all day, the room is very toasty and it's idling at 63C right now.

The P4 runs nice and cool @ 35C idle and 44C under heavy load. But the 92mm ball bearing fan bugs me. It only ranges between 1690-1710RPM no matter the load. The noise bugs me a bit so I'm going to splice a molex connector inline with the 12V and ground wires of the CPU fan to run the fan at a fixed 5V. This way It should be quieter but I can still use the 4 pin header later on if I want to. I figured since it's running so cool, it can handle the slower RPMs. I don't really care if the CPU averages 10C higher. Pentium 4s are some of the toughest CPUs out there. Other than that, ooooobviously that sleeve bearing Power Logic in the Macron is absolutely dead silent, and so is the 92mm Sunon MagLev sleeve bearing exhaust fan running at a nice ~900RPM

All 12 of my Radeon HD 2400 video cards (yes, twelve ) have active coolers on them, and even under full load gaming for hours, I've never seen the GPU temps go above 58°C (with 30°C room temperature and semi-closed case with no exhaust fans). With 20C room temperature, closed case, and no case exhaust fan, I get 56°C max. In both cases, the GPU fan speed never goes above 20%, which is the same for both idle and under load. So I think even a little bit of air going over your video card will make a world of a difference.

Ha! classic momaka Are you talking about those low profile ones? I love those cards but hate how they have that dumb DMS-59 connector or whatever it's called. They must be fairly silent at that fan speed? Amazing what just a little bit of air movement can do!

Actually, I just checked up on that. I had to, because let's face it: Wester547 knows his caps

Hahaha, this is true! He has (appropriately so) called me out on some statements, and he's undefeated in his knowledge We should call him captain capacitance

And that's exactly what they were (TMVs): 680 uF and 4 V.

Wester, that's what all those failed ones I mentioned were. Sorry for the misunderstanding.

Yeah, I don't know why ASUS does that, but it does work. If there is one thing they DO right, it is CPU VRMs. The ASUS board you gave me has a 3-phase VRM with 6 MOSFETs total. It handles that 150 Watt Pentium D perfectly fine. Moreover, I check the motherboard manual, and what's interesting is that this board was actually built to handle primarily Pentium 4 and Pentium D CPUs - even the extreme edition ones. So that makes me believe ASUS really does put some good thought into their CPU VRMs.

Yeah, I've noticed this too. The only ones who seem to do better on VRM at least from this era is MSI. That's interesting though, does the Intel 945 chipset support their extreme edition processors?? I thought you had to have the 955X or 975X chipset. i could be wrong though. I hope you're getting good use out of that Pentium D and board If my memory serves me (and is done erroring out ) it was a Pentium D 830?

Nice!
I find most Sony speaker systems and headphones to have very good sound and pretty decent build quality.

Here, try this on it (preferably, turned up):
https://www.youtube.com/watch?v=NIUlgMF7aCo

It seems so! I usually associate Sony with quality. I will definitely listen to that song when it isn't night time The cool thing is, the perfect bass zone is the center of the couch. I've listened to just 20 minutes of bass boosted SpewTube videos and it makes your ears/head hurt! But it hurts so good!

[ChaosLegionnaire]

Re: Recapping Asus PTGD1-LA

It was 64 bit. Doesn't the fact that the OS is 64 bit "negate" half of the cache of the old P4? Would I see any benefit from installing a 32 bit version since it only has 3GB?

no. rather there is actually a slight performance penalty of around 10% if u try to run a 32-bit program in a 64-bit os. so the "benefit" will be from seeing if u run mostly 32-bit media playback programs and codecs. if so, u may wanna install the 32-bit os instead for better performance in case there is a video that is dxva incompatible.

Did you see much of an increase in performance? Hope the mobo had a beefy VRM

im not very sure if there was a performance increase as i didnt run any benchmarks to check! lol! cuz it was an interim stopgap, i never bothered to benchmark it! also, i dont have the cpu anymore as i sold it off but the mobo had a 6 phase vrm. it was a gigabyte ep35-ds3 which has sadly become partially faulty due to a fuh-joo-yoo capped thermaltake psu.

Then again, I am probably just a teeny TINY bit paranoid when it comes to cooling BGA stuff (more like, A LOT). So feel free to ignore what I said above.

Ha! classic momaka Are you talking about those low profile ones? They must be fairly silent at that fan speed? Amazing what just a little bit of air movement can do!

nah momaka is just speaking from experience from the tons of bga failures he's seen on xboxes and playstations. bga chips just hate heat. leaded solder or not. im surprised the bga failure scandal didnt get blown up out of proportion like the bad caps scandal did. i wish it did and we could have a badbga.net spin off of this site. im quite disappointed there isnt a better or more reliable method of attaching high heat dissipation, high pin count interconnect chips to pcbs.

so im just gonna say here that the chinese have a saying: "ignore the wise man's counsel and your demise will happen in front of your own eyes." what it simply means is that if u ignore the advice of a wise man, u will bring about your own downfall/demise. prevention is better than cure. "cure" also does not guarantee that the problem can be fixed. sometimes the problem just cant be fixed... overkill cooling on bga chips is the way to go if u want them to last a long time...

i always have a 140mm fan above the expansion slot area blowing air down on it. i found a 140mm fan covers the entire width of the expansion slot area in a full size atx board. for a micro atx board, two 80mm fans side by side should cover the entire expansion slot area. i suggest ghetto modding the side panel case with mounting holes and fan intake ports for two 80mm fans.

[Pentium4]

Re: Recapping Asus PTGD1-LA

nah momaka is just speaking from experience from the tons of bga failures he's seen on xboxes and playstations. bga chips just hate heat. leaded solder or not. im surprised the bga failure scandal didnt get blown up out of proportion like the bad caps scandal did. i wish it did and we could have a badbga.net spin off of this site. im quite disappointed there isnt a better or more reliable method of attaching high heat dissipation, high pin count interconnect chips to pcbs.

so im just gonna say here that the chinese have a saying: "ignore the wise man's counsel and your demise will happen in front of your own eyes." what it simply means is that if u ignore the advice of a wise man, u will bring about your own downfall/demise. prevention is better than cure. "cure" also does not guarantee that the problem can be fixed. sometimes the problem just cant be fixed... overkill cooling on bga chips is the way to go if u want them to last a long time...

i always have a 140mm fan above the expansion slot area blowing air down on it. i found a 140mm fan covers the entire width of the expansion slot area in a full size atx board. for a micro atx board, two 80mm fans side by side should cover the entire expansion slot area. i suggest ghetto modding the side panel case with mounting holes and fan intake ports for two 80mm fans.

Haha, sorry. When I was saying "classic momaka" I was referring to him having 12 of those video cards

I know BGA hates heat. I may try and get another fan in there. But, if it fails, it fails. I'll try and up the cooling a little. But, silence is my main goal here... You know what, I have a Cooler Master Cool Viva Pro SE. Maybe I could remove the fan and just trow that heatsink on there. This is only a 25W TDP card.

no. rather there is actually a slight performance penalty of around 10% if u try to run a 32-bit program in a 64-bit os. so the "benefit" will be from seeing if u run mostly 32-bit media playback programs and codecs. if so, u may wanna install the 32-bit os instead for better performance in case there is a video that is dxva incompatible.

Hmm. I may just do a fresh install of 32 bit on a different drive since I was already planning on swapping this one out.

the mobo had a 6 phase vrm. it was a gigabyte ep35-ds3 which has sadly become partially faulty due to a fuh-joo-yoo capped thermaltake psu.

Looks like a good board! Unless the voltage skyrocketed from the Fuhjyyu caps, I doubt it was damaged. All those polymers would have been able to handle all the extra ripple. Not that you want bad caps powering your system though..

[Pentium4]

Re: Recapping Asus PTGD1-LA

I listened to that video 3 times with the volume CRANKED up Cool song for bass. The sub even hit those really low notes pretty well. Lots of things rattling in the house...I wonder what my neighbors thought

It even woke up my deaf cat! She can't hear anything but she came downstairs to "see what was up" She must have felt it