CharlieFoxtrot wrote on 2024-08-21, 09:47:

Mondodimotori wrote on 2024-08-21, 06:01:

I'll probably do that if the new cooler and new psu don't get the job done. Unfortunately old case with good ventilation are quite hard and/or expensive to find. Guess I'll go with a new one.

I wouldn’t worry about case cooling beforehand too much. I had a Thuderbird 1400 system back in the day in very basic Enlight ATX case without fancy airflow and it worked just fine. There is a huge difference on temp load with modern systems where people have 200W+ GPUs in so even if your CPU is a hot one, there is significantly less heat generated.

I currently have “ultimate late 2001 win98se full AMD build (well, GPU is still ATi in 2001, but you get the gist)” with TB1400, KG-7 RAID using AMD 760 chipset, 512MB DDR 2-2-2-5, ATi Radeon 8500 + Creative Voodoo2 12MB for older Glide stuff and SSD in a Chieftec Dragon clone case without any extra case cooling and it is completely fine. Granted, I recapped a vintage 365W Enermax PSU for the system with dual fans, but I wouldn’t say that my system has high air flow by any means.

Yeah, that was my plan. Just get the new cooler and a better PSU (either a new one with decent 3.3 and 5v rail or an used enermax) and see if the cooling gets better. As you can see, my case is a really old style, of those that were OK for an MMX machine, maybe not so much for a 462. After all, even the Duron that was in there ran pretty hot, at 55° during gaming. I used it in those conditions between 2002 and 2009 and, today, it still works. But I don't want to push my luck more with the Athlon 1400.

Of course, if with both the new cooler and PSU I won't get decent temps, I'll probably get a cheaper case that gives better airflow: After all, currently I have basically zero airflow. So anything will be better than zero. But I still hope that I don't have to do that and I can keep using the original case.

About recapping: Currently it's not a skill I have. Yet. I plan to find a course for electrical repairs and soldering, because may be a useful skill to have when being into old hardware .

Mondodimotori wrote on 2024-08-21, 10:06:

About recapping: Currently it's not a skill I have. Yet. I plan to find a course for electrical repairs and soldering, because may be a useful skill to have when being into old hardware .

These PSUs aren’t that difficult to recap. Sure, there are lot’s of capacitors in the secondary side of the PSU, but everything is through hole stuff and thus pretty simple. Some rails are pretty beefy, so good quality and powerful iron is needed and desoldering gun makes of course things easier, but isn’t necessary.

I have now recapped three vintage 5V heavy Enermax PSUs from early 2000s. They all had PCE tur and tul series caps. They aren’t high quality, but on the otherhand not disasters like many cheap capacitors from that era were, especially ultra-low ESR ones which were used in MB VRM. Every cap I tested after removing them were still in spec, but it just gives some peace of mind to get them replaced. These PSUs are from the worst time of capacitor plague when cheap chinese and Taiwanese companies outbid pretty much all japanese manufacturers. And it is impossible to know what kind of hours these used PSUs have been used.

In general, though, these old Enermax PSUs are probably one of the most reliable from that era.

CharlieFoxtrot wrote on 2024-08-21, 10:43:

These PSUs aren’t that difficult to recap. Sure, there are lot’s of capacitors in the secondary side of the PSU, but everything […]
Show full quote

Mondodimotori wrote on 2024-08-21, 10:06:

About recapping: Currently it's not a skill I have. Yet. I plan to find a course for electrical repairs and soldering, because may be a useful skill to have when being into old hardware .

These PSUs aren’t that difficult to recap. Sure, there are lot’s of capacitors in the secondary side of the PSU, but everything is through hole stuff and thus pretty simple. Some rails are pretty beefy, so good quality and powerful iron is needed and desoldering gun makes of course things easier, but isn’t necessary.

I have now recapped three vintage 5V heavy Enermax PSUs from early 2000s. They all had PCE tur and tul series caps. They aren’t high quality, but on the otherhand not disasters like many cheap capacitors from that era were, especially ultra-low ESR ones which were used in MB VRM. Every cap I tested after removing them were still in spec, but it just gives some peace of mind to get them replaced. These PSUs are from the worst time of capacitor plague when cheap chinese and Taiwanese companies outbid pretty much all japanese manufacturers. And it is impossible to know what kind of hours these used PSUs have been used.

In general, though, these old Enermax PSUs are probably one of the most reliable from that era.

Yeah, I know it's not hard, but first you need to know how to solder. Unfortunately I don't know how (YET).
I plan to learn to do it in the near future because I notice more and more that is a useful skill to have when you have this passion with retro hardware, not just PCs and capacitors.

Mondodimotori wrote on 2024-08-21, 11:09:

Yeah, I know it's not hard, but first you need to know how to solder. Unfortunately I don't know how (YET).
I plan to learn to do it in the near future because I notice more and more that is a useful skill to have when you have this passion with retro hardware, not just PCs and capacitors.

It sure is a useful and fun skill to have. Fixing old things is fun and satisfying, but there are also countless of fun DIY projects you can do. For example, I recently built Kiselev’s Micro 8088 project with backplane and FDD/serial controller ISA board. I need to finish the VGA card at some point too, but strictly speaking I don’t even need it.

IMO best way to learn soldering is just getting the gear for the job and practice. You can first use some scrap circuit boards and learn the basics there. Or then just jump straight in with some easy through hole job like I did back in the day. If you haven’t before practically held a soldering iron in your hand, you can get the theory basics from countless of YT repair channels, although you can’t learn soldering in any other way than doing it and improving your skills gradually tackling more challenging projects such as SMD stuff and multi-layered PCBs.

CharlieFoxtrot wrote on 2024-08-21, 11:35:

Mondodimotori wrote on 2024-08-21, 11:09:

Yeah, I know it's not hard, but first you need to know how to solder. Unfortunately I don't know how (YET).
I plan to learn to do it in the near future because I notice more and more that is a useful skill to have when you have this passion with retro hardware, not just PCs and capacitors.

It sure is a useful and fun skill to have. Fixing old things is fun and satisfying, but there are also countless of fun DIY projects you can do. For example, I recently built Kiselev’s Micro 8088 project with backplane and FDD/serial controller ISA board. I need to finish the VGA card at some point too, but strictly speaking I don’t even need it.

IMO best way to learn soldering is just getting the gear for the job and practice. You can first use some scrap circuit boards and learn the basics there. Or then just jump straight in with some easy through hole job like I did back in the day. If you haven’t before practically held a soldering iron in your hand, you can get the theory basics from countless of YT repair channels, although you can’t learn soldering in any other way than doing it and improving your skills gradually tackling more challenging projects such as SMD stuff and multi-layered PCBs.

Held a soldering iron only once, during orientation for choosing high school. Unfortunately I went to the more scientific one than the technical one. I still learned lots about PCs and how to use them/how they work, but never got my head around learning the hardware side of it. Kinda a youth regret I have.
A regret I plan to fix it. At this point I'll look around on the internet for online classes and videos on how to do it, it will surely save me the money of an actual course.

Here's a free intro... https://learn.adafruit.com/adafruit-guide-exc … llent-soldering

BitWrangler wrote on 2024-08-23, 13:29:

Here's a free intro... https://learn.adafruit.com/adafruit-guide-exc … llent-soldering

Free infos are always welcome, thank you!

momaka wrote on 2024-08-19, 21:48:

80C is wayy too much for these socket A CPUs, as is 70C. Heck, even 60C is no good. Remember, socket A CPUs don't have built-in […]
Show full quote

80C is wayy too much for these socket A CPUs, as is 70C. Heck, even 60C is no good.
Remember, socket A CPUs don't have built-in temperature probes.
The way most socket A motherboards measure CPU temperature is through an NTC thermistor or thermal diode in the CPU socket... and typically not touching the CPU package at all... meaning, those temperature readings are a little "passive" and quite slow. This is why a socket A CPU can die instantly without a heatsink, even if the motherboard has an over-temperature shutdown mechanism. By the time the NTC thermistor or diode under the CPU gets hot enough for the motherboard to register an overheating condition, the CPU die would be long gone and dead.

So for this reason, socket A CPUs should be kept at a lot lower temperature. IMO, 55C is about as high as I'd like to see one... and usually isn't that hard to achieve in a decently-large case with OK cooling and a good CPU heatsink.

This 15-year old video is showing very well, when two Athlons rapidly die...

What Happens when the CPU cooler is removed? (Tom's Hardware Guide)

Palomino and later cores had an on-die thermal diode, but there was no internal thermal protection. They relied on the hardware monitor on the motherboard, which obviously doesn't always work. I have a laptop that can shut down quickly enough to protect the CPU with no heatsink (which I accidentally tested), but that seems to be the exception. It used a completely separate temperature measurement chip instead of the VIA southbridge hardware monitor.

I remember a bunch of generic copper coolers with different fans (either the usual thin 60mm or one of the ultra-high-speed ones). They might have been made by Dynatron (OEM for a lot of Antec stuff back then), or they were just clones. Aside from that, I don't remember anything that other people haven't already posted.

melbar wrote on 2024-08-24, 04:18:

This 15-year old video is showing very well, when two Athlons rapidly die...

What Happens when the CPU cooler is removed? (Tom's Hardware Guide)

🤣🤣🤣 370 degrees celsius!! 🤣🤣🤣

I'm glad I shut down my sistem when the sensor reported 60°, it was propably higher on the die itself. Now I'm even more surprised that the previously installed Duron survived for more than a decade at usual reported temps of around 55° (wich means actually higher temps on the die itself).
I've alredy picked up (or I'm in the process of) picking up better coolers and a new PSU. If push comes to shove, I'll even tinker with the case itself to make it more airflowy.

lti wrote on 2024-08-24, 05:17:

Palomino and later cores had an on-die thermal diode, but there was no internal thermal protection. They relied on the hardware monitor on the motherboard, which obviously doesn't always work. I have a laptop that can shut down quickly enough to protect the CPU with no heatsink (which I accidentally tested), but that seems to be the exception. It used a completely separate temperature measurement chip instead of the VIA southbridge hardware monitor.

I remember a bunch of generic copper coolers with different fans (either the usual thin 60mm or one of the ultra-high-speed ones). They might have been made by Dynatron (OEM for a lot of Antec stuff back then), or they were just clones. Aside from that, I don't remember anything that other people haven't already posted.

Yesterday arrived one of the two dynatron I've ordered. It's the smaller one, I plan to put it into the 370 machine and pair it with a nice 1.0 GHZ pentium III. Still have to decide if I'll get the one with the IHS or not.
The other one, the dinatron c26, should be arriving in the next week or so. I also plan to pick up a Titan next week from a local seller.

Mondodimotori wrote on 2024-08-24, 12:25:

🤣🤣🤣 370 degrees celsius!! 🤣🤣🤣 […]
Show full quote

melbar wrote on 2024-08-24, 04:18:

This 15-year old video is showing very well, when two Athlons rapidly die...

What Happens when the CPU cooler is removed? (Tom's Hardware Guide)

🤣🤣🤣 370 degrees celsius!! 🤣🤣🤣

I'm glad I shut down my sistem when the sensor reported 60°, it was propably higher on the die itself. Now I'm even more surprised that the previously installed Duron survived for more than a decade at usual reported temps of around 55° (wich means actually higher temps on the die itself).
I've alredy picked up (or I'm in the process of) picking up better coolers and a new PSU. If push comes to shove, I'll even tinker with the case itself to make it more airflowy.

The socket thermistors generally had something like a 15-20C offset to guesstimate actual core temp, it wasn't raw temp they reported. They still had lag and averaging due to poor thermal coupling.

BitWrangler wrote on 2024-08-24, 12:35:

The socket thermistors generally had something like a 15-20C offset to guesstimate actual core temp, it wasn't raw temp they reported. They still had lag and averaging due to poor thermal coupling.

Oh yeah, I've noticed it in MBProbe, they give you space to insert manually the termal offset of the reading. You should take into account the discrepancy based on your MOBO and chipset.
Unfortunately the manual of my mobo doens't say anything of the like, so I just plan to get temps as lower as possible. It would be sad to cook up a nice and functioning athlon classic.

Talked about it before, and ended up getting one: A new power supply that claims to have 26A on the 3.3, 25 on the 5 and a maximum of 190W on both rails combined.
It also comes with a 20+4 connector, 2 molex connectors and one Floppy connector.
And a nice 120mm top fan.

I'm gonna test it this weekend and see how it perfom with the Athlon 1400C, I'll let you know how it goes, since brand new PSUs that can actually deliver those kind of performances and feautures are becoming a rarity.

OH YEAH! I also picked up a TITAN TTC D5B for cooling, so yeah, it's going to be a busy weekend.

Ok, got out of work early and couldn't wait, so I tested both elements:
The PSU seems to be able to cope with the Athlon 1400C. Played Max Payne at max settings for over an hour and it didn't miss a beat. Apparently those numbers on the box are true, and being a new PSU should be much more reliable than older used units. Probably gonna keep it with this build, just need to find a better way to organize all the extra cables that are used in a contemporary build, but are not useful in this one.

About the cooler now: It surely runs much cooler, even with the fan not at max speed (and luckly so, because at full speed it sounds like a jet engine and I could't cope for more than 15 minutes with that ruckus).
With the cooler fan mounted in the right direction and the 120mm fan of the PSU over it, it hovered around 54/55° under prolonged gaming stress, with fan at half speed, for over an hour (with ambient temps still closer to 30°).
That's not a bad result, both considering that the old cooler got the CPU to 60° in just 10 minutes, and the two fans, in that configuration, were fighting each other for air intake.
Unfortunately I didn't had a small enough screwdriver to revert the cooler's fan direction. Gotta go out tomorrow and get one from the hardware store. I'll then proceed to test the setup all closed up with the cooler's fan inverted, thus shooting all the hot air into the 120mm PSU's fan and out of the system.

I also had a slightly sick idea involving a the free slots on the front of the case (between the CD drive and the Floppy drive) and a couple of Noctua fans... But that's a test for another day. 😈

Well, time to shares some updates:
I've reverted the fan on the cooler, and it still gets to around 53/55 under heavy loads, but this time with the case fully closed.
Now, it's not that bad, but I think I can improve it: I plan to get one or two 80 mm noctua fans and install them in the front of my case (I have two 3pin fan connectors on the mobo, just begging to be used) and I have more than enough space in the slots between the CD drive and the Floppy drive. I've also managed to move cables out of the way, so air should reach the CPU cooler basically uninterrupted, and maybe be able to create a positive pressure setup inside the case. I'll also test one of those fans on the CPU cooler itself, since the original fan of the Titan is not that silent even at half speed, and at that speed it move much less air than a noctua fan at full speed.
I've also finally got my hand on an Enermax PSU, and I've decided to install this one in the Socket A build, putting the modern PSU in the 370 build. Why, you ask? Because, under heavy loads, the new PSU (on Everest, I don't have yet a multimeter) drops the 5V rail to 4.66V.
Reading around, the minimum allowed by the ATX specs is 4.75, and at around 4.6 you start getting system instability on an Athlon like mine and sistem shut downs. So I've decided to use the Enermax instead, wich keeps a more reliable 4.88V even under heavy load.

Looks brand new and works, but has a whiny issue I've reported in a separated thread... (Enermax 495W PSU high pitch whine when connected to power: some capacitors on their way to PC heaven?)
Maybe it's time to learn how to replace capacitors, don't ya think?

Of course, I could keep using the new PSU even if it drops at 4.66V under stress, after all I haven't actually experience any sistem instability even by running 3dmark99 for almost an our (CPU max temp 51/52°), but what's yall opinion on this? Should I worry about 4.66V under heavy load on the 5V rail or not? The only peripheral I plan to add to this build is an audio card, and that's it. Would I be pushing the system to the brink of collapse with that PSU?

My experience with Socket A and PSUs is that if the +5 begins to drop, it will grow worse and the system will eventually become unstable. I've had to cycle through several power supplies that had specs that ought to have been good enough, but then began to sag under load. This is with a KT600 and several different CPUs and video cards - high power draw video cards and overclocking do contribute. One PSU that couldn't take the load in that system has been working great in my Soyo Dragon Socket A system, I suppose because it has fewer fans, a low power draw (comparatively) video card, and its limited ability to overclock.

Repo Man11 wrote on 2024-09-07, 18:10:

My experience with Socket A and PSUs is that if the +5 begins to drop, it will grow worse and the system will eventually become unstable. I've had to cycle through several power supplies that had specs that ought to have been good enough, but then began to sag under load. This is with a KT600 and several different CPUs and video cards - high power draw video cards and overclocking do contribute. One PSU that couldn't take the load in that system has been working great in my Soyo Dragon Socket A system, I suppose because it has fewer fans, a low power draw (comparatively) video card, and its limited ability to overclock.

So, with use and wear, that 4.66V under load could drop even more... Finally reaching instability territory and shutting down my sistem?
That's not good news. That's not good news at all! (cit.)

I don't plan in adding a more powerfull GPU (I only have PCI slots, and the 9250 gets the job done alredy), but I wanted to add, at least, a sound card. And, I bet, those still used to runs on the 5V rail.

Guess I better solve the whiny issue on the Enermax and use that one, because apart from this HKC, I haven't been able to find ANY NEW PSU with more than 20 amp on 3.3V and 5V and more than 140W on those rails combined. Even top of the line 1000W+ PSU from reputable brands don't have enough amperage and power delivery on those rails. And I've even read that those too don't like to overuse the 5V rail and keep the 12V to rest, thus dropping voltage.

I have a feeling that old hardware won't die because the component themselves fails. But because we won't be able to feed decent power to them.

melbar wrote on 2024-08-24, 04:18:

This 15-year old video is showing very well, when two Athlons rapidly die...

What Happens when the CPU cooler is removed? (Tom's Hardware Guide)

This video is also bullshit, because there are no practical scenarios when the radiator could be removed while a PC is running. Not to mention that Socket A motherboard options were a bit... sketchy. Early implementations of shutdown protection worked fine for a failing fan scenario.

lti wrote on 2024-08-24, 05:17:

Palomino and later cores had an on-die thermal diode, but there was no internal thermal protection. They relied on the hardware monitor on the motherboard, which obviously doesn't always work. I have a laptop

Pentium 3 didn't have it either, even Coppermine. It relies on newer chipset i8xx thermal protection.

The Serpent Rider wrote on 2024-09-07, 19:25:

Early implementations of shutdown protection worked fine for a failing fan scenario.

About this... I'm pretty sure I've read something about shutdown protection in the manual of my Socket A mobo.

Of course, I don't intend to test it.