$5 goodwill tower

[Wester547]

Re: $5 goodwill tower

I also abhor that they used plastic vents/grates at the back of those old Dell Dimension 8xxx towers. They're easy to bend permanently and that may potentially harm upon the passage of air. That, or it's very cheap metal. I also don't like how you have to turn the case over in order to open it.

[c_hegge]

Re: $5 goodwill tower

^
Agreed. I've also had to work on those clamshell cases, mostly on Hell Dementia 8300 and similar tower PCs. I've had several with hardware issues where I struggled to get them open.

[momaka]

Re: $5 goodwill tower

Not good- if cards shift and pins/lands bridge, you're gonna have fireworks!

Yeah, I personally despise the Dell clam shell cases. Had quite a few that I couldn't open because their mechanism got stuck due being handled just a little more roughly in an office environment.
The again, I never liked Dell cases, much less their SFF ovens. HP and other major OEMs were in a similar boat. Anyone remember the HP Pavilion and Dell cases from the Pentium 3 era? Those had so much unnecessarily plastic on them. Whoever "designed" them was an idiot.
The HP/Compaq Evo and other P4 towers were decent, though. Good amount of ventilation holes and generally well-built.

[Wester547]

Re: $5 goodwill tower

The cards are not "falling" into each other, so to speak, they're just becoming slightly loose even when firmly seated because Dell didn't use screws.

[kaboom]

Re: $5 goodwill tower

My only gripe with my clamshell Dell is that it doesn't use screws for the PCI slots (and they have no lock) so the cards fall out easily even when pushed in firmly. Then again, that whole design is plastic rather than steel so it's not hard to break off parts of the case.

Not good- if cards shift and pins/lands bridge, you're gonna have fireworks!

[Ami Sapphire]

Re: $5 goodwill tower

My only gripe with my clamshell Dell is that it doesn't use screws for the PCI slots (and they have no lock) so the cards fall out easily even when pushed in firmly. Then again, that whole design is plastic rather than steel so it's not hard to break off parts of the case.

Had a Dell Dimension 8200 as well. That was the one thing I hated about that case.

On the board, though, the green AGP lock piece eventually broke off. I had to remove it from the board.

[Wester547]

Re: $5 goodwill tower

My only gripe with my clamshell Dell is that it doesn't use screws for the PCI slots (and they have no lock) so the cards fall out easily even when pushed in firmly. Then again, that whole design is plastic rather than steel so it's not hard to break off parts of the case.

[momaka]

Re: $5 goodwill tower

Since my Dell Dimension 8200 has PC800 RDRAM, it (the RDRAM) runs rather hot if I don't keep the clamshell totally closed (it must wreck upon it airflow by opening it even in the slightest). If it's totally closed, it runs much cooler. Not an option for me, but I never noticed the fans combating for airflow. Though, they have become loud over time.

Ah yes, the clam shell Dells are a different story. I've worked on some GX260 and GX270 (non-SFF) clam shells, and they do seem to run cool and quiet.
The OptiPlex 170L do have the same fan setup (and thus same problems) as my Dimension 3000, though.

[Wester547]

Re: $5 goodwill tower

Most of the normal sized desktop P4 -based Dell Dimensions I have seen use a 92 mm case fan that is also used to cool the CPU (it draws warm air from the CPU heat sink through a duct and exhausts it out of the case). It's a pretty clever design, IMO. However, Dell still managed to fudge it. On my family's Dell Dimension 3000, the only air intakes are a small hole on the front of the case right above the front USB ports and another small vent on the back. With the case closed and if the ambient room temperature is cool (below 21C), all is good. However, when the ambient room temperature gets a little warmer (as in above 22C - which is quite cool IMO) and with the CPU under full load, the built-in thermistor in the 92 mm fan makes the fan speed up a notch or two. Since the air intake holes in the case are undersized, the result is low pressure inside the case. This makes the PSU fan ramp up as well so it can push out the same amount of air as it did before. This in turn causes even lower pressure in the case so the CPU gets warmer and the CPU/case fan ramps up again. In the end you get the CPU/case and PSU fans fighting for air from the air intakes. The hard drive stays really cool because there's a jet stream going right through it coming from that tiny hole on the front of the case. At the same time, the whole computer has just turned into a loud bitch. The only relief against that is to slightly open one side of the case about 2 to 3 cm. Once that is done, the PC goes all quiet again at the expense of a slightly raised HDD temps. But all in all, I find that well worth it.

Since my Dell Dimension 8200 has PC800 RDRAM, it (the RDRAM) runs rather hot if I don't keep the clamshell totally closed (it must wreck upon it airflow by opening it even in the slightest). If it's totally closed, it runs much cooler. Not an option for me, but I never noticed the fans combating for airflow. Though, they have become loud over time.

I agree, but my experience tells me that it just doesn't seem to matter that much. I have several other hard drives that run in the 40C+ range even in the winter with 5000+ hours and 1000+ power cycles, all still running fine.

I consider 5,000 hours and 1,000 power cycles still somewhat early in the HDD's life. Though, one cannot predict when a HDD will fail either way. And I meant many power cycles. ^^;

Me too, but I don't get a choice all the time (I'm using what I have - and that's mostly old HDDs that like to run hot).

Using active cooling helps. But I find the difference between idle and load temperatures in HDDs most staggering.

I don't know. I actually have several of these and they look like they have seen some use in their 12 year life. Even in a room with normal temperature around 23C, the CPU still reaches around 50 to 60 C (or at least that's what it feels like... when I get a thermocouple soon, I will know for sure) and it's always been running like that. Yes, I'm sure the leaded solder has definitely made a difference.

If there are no dreck packages like BGA, CSP, or QFN, that would be even better.

Reallocated Sector Count; 5
Reallocated Event Count: 5
They have been there since I started using the hard drive. Everything else is zeros.

That's not bad, though I think it's a tad early in its life to have reallocations. That said, that's not many and at least it's stable with respect to that. The only other thing I would look out for is errors reported in the S.M.A.R.T. log (you can view them with third-party apps known as GSmartControl and/or smartctl). It may be able to tell you if the drive is experiencing hard, physical errors even before any pending or reallocated sectors (if auto-offline data collection is enabled) surface.

[momaka]

Re: $5 goodwill tower

Well, in the case of Dell Dimensions, since the CPU fan is isolated away from the rest of the system, they may need a PSU fan running in order to keep the rest of the system well ventilated since they don't provide room for a case fan in such a design. As for hard drive temperatures, hard drives could only be kept cool without a PSU fan if it's mounted right near a vent and if the case is closed (preserving the passage of airflow). Then, though, with a PSU fan, it might be too cool, especially at load.

Most of the normal sized desktop P4 -based Dell Dimensions I have seen use a 92 mm case fan that is also used to cool the CPU (it draws warm air from the CPU heat sink through a duct and exhausts it out of the case). It's a pretty clever design, IMO. However, Dell still managed to fudge it. On my family's Dell Dimension 3000, the only air intakes are a small hole on the front of the case right above the front USB ports and another small vent on the back. With the case closed and if the ambient room temperature is cool (below 21C), all is good. However, when the ambient room temperature gets a little warmer (as in above 22C - which is quite cool IMO) and with the CPU under full load, the built-in thermistor in the 92 mm fan makes the fan speed up a notch or two. Since the air intake holes in the case are undersized, the result is low pressure inside the case. This makes the PSU fan ramp up as well so it can push out the same amount of air as it did before. This in turn causes even lower pressure in the case so the CPU gets warmer and the CPU/case fan ramps up again. In the end you get the CPU/case and PSU fans fighting for air from the air intakes. The hard drive stays really cool because there's a jet stream going right through it coming from that tiny hole on the front of the case. At the same time, the whole computer has just turned into a loud bitch. The only relief against that is to slightly open one side of the case about 2 to 3 cm. Once that is done, the PC goes all quiet again at the expense of a slightly raised HDD temps. But all in all, I find that well worth it.

I could never be comfortable with 45C, much less 55-60C. Hard drives are very fragile devices and they certainly don't like heat. I can understand reaching 40+C when running 24/7 but long power cycles and excessive heat are just too much for hard drives to take, I think.

I agree, but my experience tells me that it just doesn't seem to matter that much. I have several other hard drives that run in the 40C+ range even in the winter with 5000+ hours and 1000+ power cycles, all still running fine.

I personally prefer my HDD temperatures below 40C but above 30C

Me too, but I don't get a choice all the time (I'm using what I have - and that's mostly old HDDs that like to run hot).

Heat may eventually take its toll on that notebook, though perhaps I'm wrong, but I think temperatures that high aren't safe, though they are hardly the end of the world considering that it's leaded solder.

I don't know. I actually have several of these and they look like they have seen some use in their 12 year life. Even in a room with normal temperature around 23C, the CPU still reaches around 50 to 60 C (or at least that's what it feels like... when I get a thermocouple soon, I will know for sure) and it's always been running like that. Yes, I'm sure the leaded solder has definitely made a difference.

If you don't mind me asking, how many bad sectors does that Hitachi drive have, and are they reallocated, pending, or offline uncorrectable?

Reallocated Sector Count; 5
Reallocated Event Count: 5
They have been there since I started using the hard drive. Everything else is zeros.

[ratdude747]

Re: $5 goodwill tower

So the fact that my temps are around 70C at full load (I F@H on it) means that its a crap system?

If so, it looks I may need faster cooler fans (I swapped the stock ones for quieter ones with longer cables) and a set of ear plugs. The only other ways to cool it down would be to throw down $120 for a pair of full copper coolers or to toss the system and start over from scratch.

That number might be wrong. I just checked it and the highest temp I saw was 53C. One CPU always runs a bit hotter than the other... I think it may be because I admittedly used a rather skimpy amount of thermal paste... I know it doesn't need much but it was maybe a grain of rice amount. I may have fixed that though, although I am not sure (I have plenty more if I can find it).

[ratdude747]

Re: $5 goodwill tower

They're based on the Northwood core, in 130nm process. So, yes they do.

So the fact that my temps are around 70C at full load (I F@H on it) means that its a crap system?

If so, it looks I may need faster cooler fans (I swapped the stock ones for quieter ones with longer cables) and a set of ear plugs. The only other ways to cool it down would be to throw down $120 for a pair of full copper coolers or to toss the system and start over from scratch.

[Wester547]

Re: $5 goodwill tower

For what its worth, the Pentium 4 did have hyperthreading, though I have no idea how beneficial that was.

Off the top of my head, throttling temperature is 63C. There's a program called ThrottleWatch if you want to check.

Well, I just came to the cognizance that I think throttling temperature has to do with whether or not the CPU exceeds its TDP, which means the throttling temperature may be different for each Pentium 4, but, indeed, in the low to mid 60s (celsius). That's too low of a throttling temperature as far as I'm concerned. Those Pentium 4s ran hot as it is, you can't always help if they reach 60C at 100% load. That, however, must mean that those CPUs never came quite close to their maximum power dissipation (unless you overclock and/or overvolt them, maybe), and that they didn't exceed their TDP by much if at all. I have used ThrottleWatch before, though Dell don't have motherboard sensors.

Depends a lot on the motherboard. Most P3 and older era motherboards don't need a whole lot of cooling. The CPU fan usually provides enough airflow for the NB, SB, and caps. The only thing that may get just slightly warmer is the CPU VRM.

Well, in the case of Dell Dimensions, since the CPU fan is isolated away from the rest of the system, they may need a PSU fan running in order to keep the rest of the system well ventilated since they don't provide room for a case fan in such a design. As for hard drive temperatures, hard drives could only be kept cool without a PSU fan if it's mounted right near a vent and if the case is closed (preserving the passage of airflow). Then, though, with a PSU fan, it might be too cool, especially at load.

I could never be comfortable with 45C, much less 55-60C. Hard drives are very fragile devices and they certainly don't like heat. I can understand reaching 40+C when running 24/7 but long power cycles and excessive heat are just too much for hard drives to take, I think. I personally prefer my HDD temperatures below 40C but above 30C, though that stems of someone who does not run their PCs 24/7. Heat may eventually take its toll on that notebook, though perhaps I'm wrong, but I think temperatures that high aren't safe, though they are hardly the end of the world considering that it's leaded solder. If you don't mind me asking, how many bad sectors does that Hitachi drive have, and are they reallocated, pending, or offline uncorrectable?

[Th3_uN1Qu3]

Re: $5 goodwill tower

That explains why they run so cool. I imagine it was the same for the Pentium IIs. However, in a good cooling setup, there won't be a huge discrepancy between idle and load temperatures - more than a 15 degree (celsius) discrepancy alludes that the cooling is rather poor, or at least the thermal compound.

Not necessarily. I had an E2180 that would run at 25C idle, and over 70C load, under a big motha of a cooler (Scythe Mugen, still have it to this day). Needless to say the heatsink stayed cold... I tried everything. I lapped the IHS. I even took it off. Same thing. I eventually upgraded to an E6550 which is entering its 5th year IIRC... still going strong at 3.61GHz stable. Used to run 3.73 when i had only 2GB RAM, but with 8GB, the motherboard can't do the crazy FSBs no more.

The Pentium 4s were not a great successor. It wasn't until the Core 2 Duos that Intel had another winner in full. That said, I liked the Pentium 4s. I didn't think their throttling threshold was as low as 60C, though. Wouldn't the throttling threshold be 70C if a Pentium 4's maximum temperature is 74C (Northwood) or 76C (Willamette)?

I liked the Pentium 4 too... until i first got my hands on an AXP 2400+. That thing wiped the floor with my 3.0GHz HT, it was simply no comparison. At that time, the multicore era was still at its inception, so most applications weren't optimized for multithreading, and the P4 suffered a lot.

Off the top of my head, throttling temperature is 63C. There's a program called ThrottleWatch if you want to check.

OT: Do Gallatin core Xeons have throttling as well?

They're based on the Northwood core, in 130nm process. So, yes they do.

[momaka]

Re: $5 goodwill tower

If the PSU fan wasn't running, the RAM, the northbridge, the capacitors, and other components on the motherboard may get too hot for comfort.

Depends a lot on the motherboard. Most P3 and older era motherboards don't need a whole lot of cooling. The CPU fan usually provides enough airflow for the NB, SB, and caps. The only thing that may get just slightly warmer is the CPU VRM.
As for the RAM - desktop DDR and SDRAM runs quite cool.

I've always seen huge spikes in HDD temperatures alone when the PSU fan fails (granted, these were two 7200RPM drives in the same chassis, but still).

Yes, with no air movement the HDDs do get a little warmer than normal (low to mid 40s C in my experience, except maybe a few select brands/models that always run cool regardless). That still isn't anything too bad, though.
As long as it's under 45C, it doesn't matter too much. Heck, the HDD in my laptop regularly runs 55C in the summer. A few times it reached 60C and one time it reached 68C! Still lives, is being used, and doesn't have any more bad sectors that it did before. It's an old Hitachi Travelstar DK23EA-20, FYI. Almost 7300 hours of ON time and 1200 power cycles.

I don't know, ever since I got this Dell Latitude C600 laptop, everything else seems to run cool in comparison. In the summer, this laptop will burn your legs and normally it starts smelling like melted plastic when used for more than 1 hour. CPU heat sink is probably sitting in the 70 to 80C range and CPU fan turns on only when things get really toasty. Yet it never crashes and it's close to 12 years old now (granted I'm the 3rd or 4th owner, and I only started using it about 1.5 years ago). I still think it will work for quite a while regardless. But we will see.

I'll agree on the fans tho - up until the C2D era, Intel had a thing with small-ish, high RPM fans. The stock cooler on my Pentium 4 HT 3GHz (socket 478) sounded pretty close to a vacuum cleaner in the summer. The fan had a temperature sensor integrated in the hub, and when it felt things getting crispy, it ramped all the way up to 5200 rpm. Crazy.

Not all of them were that loud, though. I have 2 very similar in size stock Intel P4 heat sinks with the same size 70mm "upside-down" fans. One is fairly quiet and the other not so much. The louder one is rated for higher Amps though, so no surprise there. Both are regular, straight-up 12V fans. I forgot what brand, but I think one is NMB and the other is Nidec. Either way, they both aren't all that loud. Now the 60 mm Spire fan on my 933 MHz P3 PC - that thing is a jet engine. Motherboard reports close to 5300 RPM. I had to undervolt it to 7 V to make it quiet. At 7 V it doesn't move a lot of air, but it's quiet and enough to keep the CPU cool.

Do Gallatin core Xeons have throttling as well?

If it's Netburst -based - I would guess most likely yes.

[ratdude747]

Re: $5 goodwill tower

OT: Do Gallatin core Xeons have throttling as well?

[Wester547]

Re: $5 goodwill tower

Another interesting thing i've found about PIIIs, is that their load temp isn't much higher than their idle temp, they run nice and steady.

That explains why they run so cool. I imagine it was the same for the Pentium IIs. However, in a good cooling setup, there won't be a huge discrepancy between idle and load temperatures - more than a 15 degree (celsius) discrepancy alludes that the cooling is rather poor, or at least the thermal compound.

I'll agree on the fans tho - up until the C2D era, Intel had a thing with small-ish, high RPM fans. The stock cooler on my Pentium 4 HT 3GHz (socket 478) sounded pretty close to a vacuum cleaner in the summer. The fan had a temperature sensor integrated in the hub, and when it felt things getting crispy, it ramped all the way up to 5200 rpm. Crazy.

The fans in the Dell Dimension computers of that era were larger than the usual 70mm ones - they were 92mm and while their static pressure wasn't great, their CFM was lots - because their fans were not mounted on a heatsink but blowing air away from the heatsink, near a vent, with the plastic cover isolating the heat of the CPU away from the rest of the computer (and probably doing a better job of conducting heat away from the capacitors when the CPU gets hotter).

One thing i'll tell you tho - the PIIIs passed the test of time, the P4 didn't. The P4s actually had lower maximum operating temps than the PIIIs did, and they also ate a stupid amount of power. To prevent the chips from bombing out on you, intel added thermal throttling. When it got over 60-some degrees, your brand spanking new P4 (and like any other intel CPU, it cost a load of $$ when new), would turn into a very, very slow machine.

The Pentium 4s were not a great successor. It wasn't until the Core 2 Duos that Intel had another winner in full. That said, I liked the Pentium 4s. I didn't think their throttling threshold was as low as 60C, though. Wouldn't the throttling threshold be 70C if a Pentium 4's maximum temperature is 74C (Northwood) or 76C (Willamette)?

[Th3_uN1Qu3]

Re: $5 goodwill tower

That said, I agree with the rest. Even a Pentium III 1GHz (EB) Coppermine does not get much above 45C at 100% load (provided the fan isn't underpar), in hot weather. Its maximum power dissipation is only about ~36W so no surprise there at all.

My Tualies run at 50-55C pretty much all the time. Admittedly, this is a dual-CPU board with the processors pretty close to one another so there's no room for big fancy heatsinks, but the fan in the PSU works fine, and i also have a 120mm case fan. Okay, that runs at 5v, so maybe we can count that out. The CPUs are two of those. With Tcase spec'd to 69C, i don't think i have anything to worry about tho. Another interesting thing i've found about PIIIs, is that their load temp isn't much higher than their idle temp, they run nice and steady.

I'll agree on the fans tho - up until the C2D era, Intel had a thing with small-ish, high RPM fans. The stock cooler on my Pentium 4 HT 3GHz (socket 478) sounded pretty close to a vacuum cleaner in the summer. The fan had a temperature sensor integrated in the hub, and when it felt things getting crispy, it ramped all the way up to 5200 rpm. Crazy.

One thing i'll tell you tho - the PIIIs passed the test of time, the P4 didn't. The P4s actually had lower maximum operating temps than the PIIIs did, and they also ate a stupid amount of power. To prevent the chips from bombing out on you, intel added thermal throttling. When it got over 60-some degrees, your brand spanking new P4 (and like any other intel CPU, it cost a load of $$ when new), would turn into a very, very slow machine.

I believe most of the SFF PCs they were thrown into ran at throttling pretty much all the time. Had some Dells like that in the computer lab at high school, and they were so slow it wasn't funny. My dual Tualatin still runs circles around most P4s, and when it will be no longer useful as a desktop machine, i'm pretty sure it will last 5 more years as a server, if not more.

[Wester547]

Re: $5 goodwill tower

Old computers like that just don't generate a lot of heat, so even without the PSU fan running, they still wouldn't overheat.

If the PSU fan wasn't running, the RAM, the northbridge, the capacitors, and other components on the motherboard may get too hot for comfort. Old machines like that rely on the PSU fan to be a case fan as well, while the CPU has its own cooling assembly. I've always seen huge spikes in HDD temperatures alone when the PSU fan fails (granted, these were two 7200RPM drives in the same chassis, but still).

That said, I agree with the rest. Even a Pentium III 1GHz (EB) Coppermine does not get much above 45C at 100% load (provided the fan isn't underpar), in hot weather. Its maximum power dissipation is only about ~36W so no surprise there at all. Granted, I'm talking about the CPU I had in a Dell XPS B1000, and it had a Minebea (NMB) dual ball bearing fan spec'd at 3600RPM/61.8CFM/92x92x25mm. It did get loud quickly (and the bearings even sounded worn) so I'm guessing that though it was temperature controlled that it would often reach full speed or close. Compare that with the Pentium 4 Willamettes (a 1.7 GHz P4 of that generation having about ~80W in maximum power dissipation), many of which easily reached 55C at 100%, in warm weather, with a similar cooling setup (though larger heatsinks). As for the other temperatures, if it was an RDRAM board then it's likely the RDRAM ran the hottest in the system (in my experience, the ones with two heat spreaders easily get to 55-60+C). If it was SDRAM, though, there probably wasn't much heat at all.

[momaka]

Re: $5 goodwill tower

- Poor airflow in general with little if any bezel air intake and it depends on a generic fan in the generic PSU as the only exhaust. Eek! It's surprising this thing didn't fail yet, probably doesn't have a lot of hours on it.

Old computers like that just don't generate a lot of heat, so even without the PSU fan running, they still wouldn't overheat. I'm speaking from experience here. Have many Pentium 3 PCs like that. They can run at 30C room temperature and never overheat. In fact, this summer I under-volted the CPU fan on my 933 MHz P3 PC at my grandmother's house to 7V. With the room temperature at 28-29C, the thing never ran hotter than 42C under full load. Everything else in the PC wasn't that warm either. Second warmest were the hard drives - probably around 35-40C.