I was one of the designers at Intel that worked on the 430TX, that was back in 1994. Just retired from there last year, and now Granite Rapids is shipping, which is the last project I worked on. As the old saying goes, those who respect the law and enjoy sausage have never seen either being made.
Ha, I know what you mean. I'm myself in the software engineering since decades and there it's absolutely the same. I had a colleague, he worked for a big airplane manufacturer and since then he never took a plane again. Anyhow, thank you for your comment. Doesn't matter how good or bad it went internally, you still did a great job and it's an honor to mention it here. Thank you very much.
@gregebert5544 Ah gotcha! I am indeed up in Oregon in the R&D fab. It's all small batch stuff and test chips right now to get new nodes up and running.
Thank you for demonstrating that process of opening up the RTC module without desoldering it from the board, I just used that to fix one of these boards up myself! I am still bad at soldering (shaky hands) so I don't trust myself with desoldering the module from the board and replacing it with a socket, but I will get better with more practice and install the socket eventually. Until then, this is working perfectly! If I can make a recommendation, I used a dremel to cut a line on the top of the module on the right half, opposite from the pin 1 side. I then used a screw driver and hot air to chip away at the resin using the cut on the top as my entry point with the screw driver. It felt safer starting the chipping away from the top rather than the side where you could slip and accidentally hit some of the pins on the RTC and damage them. I left the left side of the module (the pin 1 side) intact so that it could hold the crystal oscillator in place and removed just enough so that I could remove the dead coin cell battery and solder the wires for a coin cell battery holder onto the positive and negative pins on the IC. I used some velcro stickers to keep the battery holder securely on the bottom of the case for easy removal. Also there is a Mr. BIOS available for this motherboard! The files can be found here: www.vogons.org/viewtopic.php?p=1162370#p1162370
I have so much memories of these boards... Used one in nineties with P200MMX and overclocked it to 250MHz. Later got Voodoo1 and played Unreal through. Need for Speed 3 and 4 was in game too.
And that was really a thing, I did the same with the overclocking. Can you imagine a Pentium 4 with 2GHz being overclocked to 2.5GHz and just working? Some CPUs today can do that but back in the day the hardware was unbelievable reliable.
25 years ago I had the ATX version of this board, and I recently bough this board in AT version to build myself a nice retro machine. Some of the parts I kept for 25 years and I use it again now. It's just a fun little project which brings back great memories.
The only issue. Was the cache slot was mechanically glitchy. Back in the day the headaches these caused as they got a little age on them, was fairly massive. The only mechanical retention for those was the electrical contacts of the slot. Particularly in a tower were these things would sit on its side with gravity pulling on the module. It would start to develop bad connections, and a crashy computer. There is a reason these started to disappear midway through the HX chipset run to never appear again.
@Alcochaser good thing chip 🍟 manufacturing technology advanced to the point you could have larger caches built directly into the CPU die so you didn't need an external cache module after a while anymore. I believe in the 90s one of the motorola 68k chips 🍟 had built in cache in chip but it was a tiny 256 bytes 🤏🏾 unlike the KBs you were getting on those cache modules
I just love your videos! I did not have much time over the last couple of months to tinker with retro hardware but just watching your videos makes me want to go down in my basement and just put something together!
I recently got an IBM PC 300 GL from 1998 or late 1997 with its original CR2032 battery still reading over 3.0V! I even tried powering a red LED briefly from it, and it worked beautifully. They really knew how to make things back then :) And yes, I was the first one to open that computer in over 25 years :)
I always joked that I overclocked my P200 to 208 .. But it actually was a nice boost because it was 83mhz FSB! Later I used a Tomshardware guide to unlock some voltages and multipliers to use a K6-2 500, which worked but the whole thing was limited by the chipset and memory I think. Good memories though!
We see good NVRAM/RTC modules every now and then from the late 80s in old Sun equipment...but as you said, it's because they were used in an always-on application (machine controllers). Still always impressive to see!
That was my first board. The PC stood in the basement during a flood many years ago when the PC was no longer in use, the board is still running. Back then 32MB RAM and a Pentium 200 MMX. Now the Bios Chip has been socketed and maybe I will upgrade to a clock module with 2032 battery. 256MB RAM and an AMD K6-3 400MHz and the system always runs rock-solid more than 25 years after I first got it. One hell of a board!
I had one of these.. upgrading from a genuine Intel advanced ZP 'Zappa' board with pipeline burst cache running the amazing 3rd party Microid Research BIOS, at that time a lot of other boards had fake writeback cache, at that time one of our suppliers were stuck with a lot of boards with fake cache and terrible performance, but thats another story, after seeing this board reviewed on Toms Hardware back in the day, after looking I run it 83mhz bus no problems with an SB16 and a diamond stealth 64 video vram - never had a problem with it.
He probably has about 20 of them. They just have a habit of appearing in the household nobody quite knows how and from where. The problem with them is that they need 7-10 minutes to get up to temperature, which is why i don't use them when i just need one blob of hot melt glue either.
@@nalinux That's not the problem. The problem is that they take too long to heat up. Apropos pink. I broke down and bought a cheap pink cordless drill. I love it. I want more pink tools.
I STILL have a T2P4 and it probably still works! It is a 256KB cache version with a cache slot - properly known as COAST or "cache on a stick" module. and on my board the Dallas module is socketed. It carries a 400mhz K6-2 which probably can be taken to 450mhz by upping the bus frequency. There are also a few BIOS mods done so it supports larger hard drives, plus I added a full screen logo of Butters in a mechanic coverall as seen in S4E16.
It's interesting seeing computer components that were very high tech and very expensive back then having the price stickers on them like you'd see at a dollar store or ma and pa grocery store. :p
The Asus P/I-P55T2P4 supports not only the 75MHZ mode for Cyrix CPUs, but also a 83MHZ mode! This makes it fastest socket 7 motherboard by far untill Super7. With the 166MHZ Pentium MMX, changing the busspeed to 83MHZ runs it at 207.5MHZ, but also increases the speed of the cache and memory!
I have hacked a few Dallas modules the ugly way, but the way it can be disassembled quite neatly was new for me. Although I usually do not care about the look, dremeling creates quite much black dust, so I will try this method next time.
i think by the time the battery can be desoldered it should be done right away before continuing with heating the resin. These batteries can explode pretty violently if heated for too long.
I am a fan of the VX chipset. It has USB and IR support, at least my DFI board from back in the day does. I had it hooked up to my Scanjet 4C back in the day over USB!
IR support is just built on one of the of the UART ports of the SuperIO chip, replacing one of the COM ports, and USB is likely a separate PCI device, such as 83c572.
I love this asus pentium 1 boards, I have a few different variants of it. Just surprised why modified rtc instead use yours necroware rtc on pcb. I've ordered some recently on black and green pcb, using then on my retro builds. Again thank you for great work, this pentium 1 boards are great for Quake and voodoo 1
I made a video about it in the past. Since I open sourced my RTC project, people started to buy the required chips like crazy. Meanwhile it's hard to find a reliable distributor who would sell them for a sane price. I already got a lot of fakes, so you never know what you get. Of course I have some of my modules in stock, but now I use them only if I can't restore the old module. That also means less waste, since the chips inside of those Dallas modules are still alright as we can see, so why not using those, right?
I had a 430vx board which had cache slot. 256kb on the board and another 256kb could be added to make 512kb total. It had jumpers on the board similar to the board in this video to configure the cache.
yes, it was very common, that vx boards had 512kb cache too. But main difference is, that they could not cache memory above 64MB limit. So anything stored above 64MB memory area, was not cached, so access to that data was slower. 430hx on other side, could cache up to 512MB memory.
Everything looks good but running the heat gun with the chip in the motherboard seems a bit risky. You worked it with no problems but ideally you would take the chip out before you tear into it. Great videos.
@@necro_ware sorry wasn’t trying to be the ackshully guy. I was just thinking about people who don’t know what they are doing. I should have qualified it. You were good.
I have exactly the same board like the one in your first scene. And I also already prepared it with pin headers for a VRM module. Do you have a spare one for me? That board is somewhat special as it can be equipped with a bigger tag ram to cache the max. of 192mb ram, whereas later P5 bords can only tag the first 64mb.
You mean the PA-2000? That is an interesting case. It's not quite simple and that project is ongoing. Hopefully it will be finished in the near future and then I can share my results.
I had one of those, running a Pentium 133 overclocked at 83x2=166 with no issues for years. I don't remember the cache card so maybe I had the version that wasn't expandable.
I need to research the cache on Triton II once again, then. I've read that due to a chipset bug, no more than 256k can be used, for some reason, but maybe it's untrue. I have an Intel Marl motherboard, with 256k cache, that I wanted to upgrade to 512k using chips I got from broken Slot 1 Pentium II, but got discouraged from doing that - and also add USB ports to the onboard controller, that were unpopulated. I kept looking, but I havent' seen a 512k or USB-equipped version anywhere.
Okay now you have me second-guessing which chipset my Socket 7 rig had back in the 90's. I could have sworn it was a 430HX, but then I'm fairly sure it had SD RAM. Now I have to go dig it out of storage...
Those dallas clock modules are designed to be very hard to replace for those without technical skills. Why can't it be a simple replacement for the battery like on some video game consoles?
Ich denke, die Benchmarq-Chips, die du für deinen Dallas RTC-Ersatz benutzt hast, sind wohl schwer zu finden. Da macht die Mod des alten Dallas RTC mit der CR2025-Batterie mehr Sinn. Und dadurch entsteht weniger Elektronik-Abfall, auch ein Plus.
I always suspected the RTC battery would be under less load on a powered system, great to confirm. Curiously, is the RTC powered by 5VSB on modern systems to further reduce the load?
Yes, standby power holds up the clock and NVRAM when available. I would like to suggest though that RTC battery depletion is not dominated by current consumption but by chemical aging; so the crucial difference could be manufacturing variations and storage temperature rather than how the computer was used in its heyday.
Stand up power on AT board? no, AT board is always turn off (from the AC). Only ATX has stand up power. It HAD to be running 24/7 in order to keep battery charged. Which is unlikely. I would suspect, someone had already soldered new Dallas before Necroware got it.
@@warrax111 Of course on AT there is no standby line. But that Dallas has a manufacturing mark year 1996 and there are no signs of rework on the board at all there, it wasn't replaced. The design life on that Dallas is 10 years, but they last somewhere about 20 years on average, so some last short of 15, others last 25, it's just random, there's just natural variation to them. Even if the Dallas was replaced, well it's running the moment it's manufactured and put into the package, the NVRAM is being held up, the clock is ticking some nonsensical date and time. And like it's obvious it was made in 96. So like it would have experienced the same aging whether it's in the board or not.
@@SianaGearz This is already 28 years. Battery discharge after some years, if board is turn off, cannot last more than 6-7 years. So only explanation would be, that board was running for like 22 years out of that 28 years. Those capacitors wouldn't survive it , I very doubt it. So in case of constant running of board, at least they would needed to be changed. If there are no signs of changing the capacitors (board was running contantly 28 years), and also battery is not changed, it is a miracle then. And these kind of miracles in physics just doesn't happen. :) Not single capacitor is buldged. After 28 years of constant run? That's could be possible on sold state capacitors, but eletrolytes?
dad bought the machine as a base model 33mhz \ 60mhz it when i read an article about some company's selling the same thing with a jumper moved it peaked my young mind to investigate this thing called Front Side bus .
I'm too lazy. I have a strap, but I usually just touch the case of the grounded PSU regularly. Just don't do what I do, never take advice in electronics from a software engineer ;)
Yes, that's a good point! I always wear protective glasses when doing this. Luckily, none exploded yet and I made this mod now on many boards. Probably 80°C is not hot enough, but better prepared than sorry.
It's not connected to the board. There is no need for a short test in a flying setup. That CPUs didn't get as hot and only needs a fan in a closed case under longer load.
You have one tag on board for the first 256K, then you have another on the COAST module for the second 256K. That gives you 8-bit tag for 512K which is enough to cache 64M of RAM. If you want to cache up to 512M of RAM, you need 11-bit tag and then you have to populate that socket on the board.
4E is a very late error code and means "Show all error messages on screen". That happens after all critical parts have been initialized. Are you sure that your graphics card is alright? May be just incompatibility and you could try another one?
512 MEGA bytes of cache? Running without the CPU fan is detrimental to the poor thing; those old CPUs didn't even have the thermal failure shutdown feature.
I guess, that if "wetern" economy collapses that far, then all economies will collapse. They are much more bound to each other, than most people without proper economical education think. In simple words, should it collapse, there will be nobody to trade old hardware, we would then probably fall back to agriculture and barter exchange. So, just enjoy that old hardware now and don't make plans about it beyond the end of the world as we know it.
@@necro_ware this. Even factories in taiwan would not be able to make new chips, they will need to shut down. So after some time, any computer will be good at some tasks (like running simple databases, doing simple computation). Even good old Pentium 1 ones. Maybe 386 and lower, would be too unusable. But I'm sure, if anything new could be made, even 386 will have some use. To scrap Pentium 1 CPU for gold, is sin. Big one. This civilization can learn it hard way. Also chipsets could be unsoldered from motherboards before scrapping (there could be law for it) New PCB could be made even after collapse, but new chips not.
@@necro_wareподскажи пожалуйста, изначально в модуле RTC (под компаундом) установлена Батайка или аккумулятор? Если был именно по заводу аккумулятор, то можно ли так просто ставить вместо него Батарейку?
@@PowerGood. Там батарейка изолированная от внешнего питания. У оригинальных чипов Hitachi/Motorola которые использовал IBM, 286 и ранние 386 клоны питание было на 24ом пине, который был специфицирован от 2 до 5В, куда подключался внешний аккумулятор через пару диодов вместе с основным питанием. Когда комп был включен, аккумулятор заряжался, иначе питал RTC. На интегрированных модулях типа Dallas на этом пине исключительно питание от 5В, а батарея подключается к ранее не нужному пину 20, который изолирован внутри чипа от основного питания. Соответственно батарея не может заряжаться.
I was one of the designers at Intel that worked on the 430TX, that was back in 1994. Just retired from there last year, and now Granite Rapids is shipping, which is the last project I worked on.
As the old saying goes, those who respect the law and enjoy sausage have never seen either being made.
Ha, I know what you mean. I'm myself in the software engineering since decades and there it's absolutely the same. I had a colleague, he worked for a big airplane manufacturer and since then he never took a plane again.
Anyhow, thank you for your comment. Doesn't matter how good or bad it went internally, you still did a great job and it's an honor to mention it here. Thank you very much.
@@necro_waresame here, I am in software as well and since I worked in logistics I am grateful every day I can buy food.
Hello from RP1! Kinda funny finding somebody who was very possibly a coworker in the comments of a random video lol.
@@DigitalJedi At that time, I was in FM-2. RP1 sounds like it's in Ronler Acres; I retired last year and spent the last 17 years in RA-2.
@gregebert5544 Ah gotcha! I am indeed up in Oregon in the R&D fab. It's all small batch stuff and test chips right now to get new nodes up and running.
Thank you for demonstrating that process of opening up the RTC module without desoldering it from the board, I just used that to fix one of these boards up myself! I am still bad at soldering (shaky hands) so I don't trust myself with desoldering the module from the board and replacing it with a socket, but I will get better with more practice and install the socket eventually. Until then, this is working perfectly!
If I can make a recommendation, I used a dremel to cut a line on the top of the module on the right half, opposite from the pin 1 side. I then used a screw driver and hot air to chip away at the resin using the cut on the top as my entry point with the screw driver. It felt safer starting the chipping away from the top rather than the side where you could slip and accidentally hit some of the pins on the RTC and damage them. I left the left side of the module (the pin 1 side) intact so that it could hold the crystal oscillator in place and removed just enough so that I could remove the dead coin cell battery and solder the wires for a coin cell battery holder onto the positive and negative pins on the IC. I used some velcro stickers to keep the battery holder securely on the bottom of the case for easy removal.
Also there is a Mr. BIOS available for this motherboard! The files can be found here: www.vogons.org/viewtopic.php?p=1162370#p1162370
you're my favorite vintage pc hardware channel, hands down
Nice work, always enjoy seeing how quickly you get stuff like this done.
I have so much memories of these boards... Used one in nineties with P200MMX and overclocked it to 250MHz. Later got Voodoo1 and played Unreal through. Need for Speed 3 and 4 was in game too.
And that was really a thing, I did the same with the overclocking. Can you imagine a Pentium 4 with 2GHz being overclocked to 2.5GHz and just working?
Some CPUs today can do that but back in the day the hardware was unbelievable reliable.
25 years ago I had the ATX version of this board, and I recently bough this board in AT version to build myself a nice retro machine. Some of the parts I kept for 25 years and I use it again now. It's just a fun little project which brings back great memories.
The only issue. Was the cache slot was mechanically glitchy. Back in the day the headaches these caused as they got a little age on them, was fairly massive. The only mechanical retention for those was the electrical contacts of the slot. Particularly in a tower were these things would sit on its side with gravity pulling on the module. It would start to develop bad connections, and a crashy computer. There is a reason these started to disappear midway through the HX chipset run to never appear again.
@Alcochaser good thing chip 🍟 manufacturing technology advanced to the point you could have larger caches built directly into the CPU die so you didn't need an external cache module after a while anymore.
I believe in the 90s one of the motorola 68k chips 🍟 had built in cache in chip but it was a tiny 256 bytes 🤏🏾 unlike the KBs you were getting on those cache modules
Always a pleasure to watch you work, man. Good stuff!
I just love your videos! I did not have much time over the last couple of months to tinker with retro hardware but just watching your videos makes me want to go down in my basement and just put something together!
I recently got an IBM PC 300 GL from 1998 or late 1997 with its original CR2032 battery still reading over 3.0V! I even tried powering a red LED briefly from it, and it worked beautifully. They really knew how to make things back then :) And yes, I was the first one to open that computer in over 25 years :)
I would be less nervous to desolder the RTC vs tearing it apart while it's still attached to the board. Great video!
It's really fun to see those old boards again.
Die besten Freitage sind die mit einem Nercoware Video.
The best Fridays are the ones with a Nercoware video.
33.9K subs, congratulation, you deserve it, excellent content ( 🤫/Nostalgia ).
Thx for the Video of this legendary Board ❤
I always joked that I overclocked my P200 to 208 .. But it actually was a nice boost because it was 83mhz FSB! Later I used a Tomshardware guide to unlock some voltages and multipliers to use a K6-2 500, which worked but the whole thing was limited by the chipset and memory I think. Good memories though!
my first Pentium mainboard, although I had the ATX version, XP55T2P4. An awesome board!
We see good NVRAM/RTC modules every now and then from the late 80s in old Sun equipment...but as you said, it's because they were used in an always-on application (machine controllers). Still always impressive to see!
That was my first board.
The PC stood in the basement during a flood many years ago when the PC was no longer in use, the board is still running.
Back then 32MB RAM and a Pentium 200 MMX.
Now the Bios Chip has been socketed and maybe I will upgrade to a clock module with 2032 battery.
256MB RAM and an AMD K6-3 400MHz and the system always runs rock-solid more than 25 years after I first got it. One hell of a board!
Happy that the Dallas modules can be repaired this way..they will all need it sometime!
Amazing repair -- Very impressive. Keep up the good work!
I had one of these.. upgrading from a genuine Intel advanced ZP 'Zappa' board with pipeline burst cache running the amazing 3rd party Microid Research BIOS, at that time a lot of other boards had fake writeback cache, at that time one of our suppliers were stuck with a lot of boards with fake cache and terrible performance, but thats another story, after seeing this board reviewed on Toms Hardware back in the day, after looking I run it 83mhz bus no problems with an SB16 and a diamond stealth 64 video vram - never had a problem with it.
Someone spend this poor guy a hot glue gun!
Wut?
@@ruxandy 15:25
He probably has about 20 of them. They just have a habit of appearing in the household nobody quite knows how and from where. The problem with them is that they need 7-10 minutes to get up to temperature, which is why i don't use them when i just need one blob of hot melt glue either.
@@SianaGearz That's why mine is an awful pink color.
you can't miss it :)
@@nalinux That's not the problem.
The problem is that they take too long to heat up.
Apropos pink. I broke down and bought a cheap pink cordless drill. I love it. I want more pink tools.
I remember I had this Asus motherboard before switching to celeron 300A overclocked at 450 mhz .Asus boards of those times were really rock solid
Отличная тема как можно разобрать старый Даллас, мне понравилось 🤔👍
Gosto tanto desse canal que clico no curtir logo na vinheta. E sempre certeza de um excelente vídeo.
btw thanks for this trip down memory lane!
I believe this boards supports PCI 0 Waitstate. That feature i missed so hard from my 440BX and later Chipsets
Good music at 15:40
Thanks, seems not everybody likes it, but obviously tastes are different :)
@mentalplayground kinda reminds me of something from streets of rage
I STILL have a T2P4 and it probably still works! It is a 256KB cache version with a cache slot - properly known as COAST or "cache on a stick" module. and on my board the Dallas module is socketed. It carries a 400mhz K6-2 which probably can be taken to 450mhz by upping the bus frequency.
There are also a few BIOS mods done so it supports larger hard drives, plus I added a full screen logo of Butters in a mechanic coverall as seen in S4E16.
It's interesting seeing computer components that were very high tech and very expensive back then having the price stickers on them like you'd see at a dollar store or ma and pa grocery store. :p
Ohh, the T2P4 board. MEMORIES! I had one back then, let's watch!
I had the Gigabyte version. With TAG & 512MB win2k was cool 😎
The Asus P/I-P55T2P4 supports not only the 75MHZ mode for Cyrix CPUs, but also a 83MHZ mode! This makes it fastest socket 7 motherboard by far untill Super7.
With the 166MHZ Pentium MMX, changing the busspeed to 83MHZ runs it at 207.5MHZ, but also increases the speed of the cache and memory!
QDI Speedeasy Titanium I / IB series, had also 83 MHz with last BIOS.
I have hacked a few Dallas modules the ugly way, but the way it can be disassembled quite neatly was new for me. Although I usually do not care about the look, dremeling creates quite much black dust, so I will try this method next time.
i think by the time the battery can be desoldered it should be done right away before continuing with heating the resin. These batteries can explode pretty violently if heated for too long.
I am a fan of the VX chipset. It has USB and IR support, at least my DFI board from back in the day does. I had it hooked up to my Scanjet 4C back in the day over USB!
IR support is just built on one of the of the UART ports of the SuperIO chip, replacing one of the COM ports, and USB is likely a separate PCI device, such as 83c572.
2 USBs are built into the south bridge and are USB 1.0. Very slow. Better to disable onboard USB and add a USB 2.0 PCI card.
@@simontay4851 at the time it was awesome there was no USB 2.0
I love this asus pentium 1 boards, I have a few different variants of it. Just surprised why modified rtc instead use yours necroware rtc on pcb. I've ordered some recently on black and green pcb, using then on my retro builds. Again thank you for great work, this pentium 1 boards are great for Quake and voodoo 1
I made a video about it in the past. Since I open sourced my RTC project, people started to buy the required chips like crazy. Meanwhile it's hard to find a reliable distributor who would sell them for a sane price. I already got a lot of fakes, so you never know what you get. Of course I have some of my modules in stock, but now I use them only if I can't restore the old module. That also means less waste, since the chips inside of those Dallas modules are still alright as we can see, so why not using those, right?
I have Triton II on some boards too. It's good one and CPU SL27K / MMX 166 is too one of my favorites.
I had a 430vx board which had cache slot. 256kb on the board and another 256kb could be added to make 512kb total. It had jumpers on the board similar to the board in this video to configure the cache.
yes, it was very common, that vx boards had 512kb cache too.
But main difference is, that they could not cache memory above 64MB limit.
So anything stored above 64MB memory area, was not cached, so access to that data was slower. 430hx on other side, could cache up to 512MB memory.
I prefer the drilling method.
Everything looks good but running the heat gun with the chip in the motherboard seems a bit risky. You worked it with no problems but ideally you would take the chip out before you tear into it. Great videos.
As I said, that was just an experiment and the heatgun was set to only 100°C. That should be safe.
@@necro_ware sorry wasn’t trying to be the ackshully guy. I was just thinking about people who don’t know what they are doing. I should have qualified it. You were good.
Super video nice board :)
I wonder if speedsys see any difference between the onboard cache and the coast module ?
Nope, it's just L2 cache
Man, what do you do with all of these boards you test and fix?
13:55 I remember that smell of the heated epoxy.
Oohhhh yyaaaa I loved that bord, first with a Pentium 133 MHZ later a K6-2 400.
I have exactly the same board like the one in your first scene. And I also already prepared it with pin headers for a VRM module.
Do you have a spare one for me?
That board is somewhat special as it can be equipped with a bigger tag ram to cache the max. of 192mb ram, whereas later P5 bords can only tag the first 64mb.
You mean the PA-2000? That is an interesting case. It's not quite simple and that project is ongoing. Hopefully it will be finished in the near future and then I can share my results.
Thanks!
Bravo!!!
I had a 233 MMX running at 290 MHz, I think with a TX chips.
There's a not documented 83 MHz bus setting.
Yes. This topic should pop up in one of my next videos.
I had one of those, running a Pentium 133 overclocked at 83x2=166 with no issues for years. I don't remember the cache card so maybe I had the version that wasn't expandable.
The 430HX could also support SMP(dual cpu's) so could the VX(smp) it could cache up to 512mb ram the FX,VX & TX could not.
True, except VX couldn't SMP.
I have one of these motherboards, but the capacitor failed and I'm looking for time to fix it ,thanks for the video !
And here I was struggling with the hot glue gun :)
It's a lot faster to use it that way 😀
I need to research the cache on Triton II once again, then. I've read that due to a chipset bug, no more than 256k can be used, for some reason, but maybe it's untrue. I have an Intel Marl motherboard, with 256k cache, that I wanted to upgrade to 512k using chips I got from broken Slot 1 Pentium II, but got discouraged from doing that - and also add USB ports to the onboard controller, that were unpopulated. I kept looking, but I havent' seen a 512k or USB-equipped version anywhere.
Okay now you have me second-guessing which chipset my Socket 7 rig had back in the 90's.
I could have sworn it was a 430HX, but then I'm fairly sure it had SD RAM.
Now I have to go dig it out of storage...
Those dallas clock modules are designed to be very hard to replace for those without technical skills. Why can't it be a simple replacement for the battery like on some video game consoles?
what is pci expansion brown part for?
Apparently it's for: "Asus Media Bus" for Asus Media Bus Revision 2.0 cards only, shared with PCI slot 4
Some ASUS videocards had a builtin SB16 on board. Better to stick with separate card IMO.
Была у меня такая четверть века тому назад (25 лет) , правда ревизия платы немного отличалась.
Ich denke, die Benchmarq-Chips, die du für deinen Dallas RTC-Ersatz benutzt hast, sind wohl schwer zu finden. Da macht die Mod des alten Dallas RTC mit der CR2025-Batterie mehr Sinn. Und dadurch entsteht weniger Elektronik-Abfall, auch ein Plus.
Can you check the height of the socketed Dallas modul? Do oversized sound cards still fit in the last slot? I once had problems with a socketed modul.
It is still lower, than an ISA slot.
mate, nice you are back! just yesterday i wondered when u will do you next video, greaT!!
I always suspected the RTC battery would be under less load on a powered system, great to confirm.
Curiously, is the RTC powered by 5VSB on modern systems to further reduce the load?
Yes, standby power holds up the clock and NVRAM when available.
I would like to suggest though that RTC battery depletion is not dominated by current consumption but by chemical aging; so the crucial difference could be manufacturing variations and storage temperature rather than how the computer was used in its heyday.
Stand up power on AT board?
no, AT board is always turn off (from the AC). Only ATX has stand up power.
It HAD to be running 24/7 in order to keep battery charged.
Which is unlikely.
I would suspect, someone had already soldered new Dallas before Necroware got it.
@@warrax111 that's why I was asking about modern systems (ATX).
You are correct, AT was completely dead when powered off (except for the RTC battery).
@@warrax111 Of course on AT there is no standby line. But that Dallas has a manufacturing mark year 1996 and there are no signs of rework on the board at all there, it wasn't replaced. The design life on that Dallas is 10 years, but they last somewhere about 20 years on average, so some last short of 15, others last 25, it's just random, there's just natural variation to them.
Even if the Dallas was replaced, well it's running the moment it's manufactured and put into the package, the NVRAM is being held up, the clock is ticking some nonsensical date and time. And like it's obvious it was made in 96. So like it would have experienced the same aging whether it's in the board or not.
@@SianaGearz This is already 28 years.
Battery discharge after some years, if board is turn off, cannot last more than 6-7 years.
So only explanation would be, that board was running for like 22 years out of that 28 years.
Those capacitors wouldn't survive it , I very doubt it. So in case of constant running of board, at least they would needed to be changed.
If there are no signs of changing the capacitors (board was running contantly 28 years), and also battery is not changed, it is a miracle then. And these kind of miracles in physics just doesn't happen. :)
Not single capacitor is buldged. After 28 years of constant run? That's could be possible on sold state capacitors, but eletrolytes?
The original opaque looks of the case would have totally covered up that butchered dallas :D
the first mainboard i ever moved a jumper on. i remember moving the fsb jumper to get moe mhz! lol
dad bought the machine as a base model 33mhz \ 60mhz it when i read an article about some company's selling the same thing with a jumper moved it peaked my young mind to investigate this thing called Front Side bus .
If you add the COAST slot to the board that is missing it, will you be able to have 768KB of L2 cache?
430HX supports only up to 512K of cache.
I have one of these mobos in my DOS/Win 95 machine. It's a beast. I think I have a 500MHz K6-2 in it right now.
@necroware when working with electronics, do you ground yourself or do you consider this unneccesary?
I'm too lazy. I have a strap, but I usually just touch the case of the grounded PSU regularly. Just don't do what I do, never take advice in electronics from a software engineer ;)
These chipset run with either a standard 8MB or if you had cash to spare the 16MB of ram.
8MB was normal on the first generation pentium systems. As HX and TX hit the market (MMX) it was already more likely to have 32MB
We want to see SUPER Socket 7
I should have something in my todo box as well.
I had a asus p2b but now i stil have asus cuv4x-e
Warning... Be careful! Protect the lithium cell from direct hot air! It can end as a bullet, so protect your eyes!!!
Yes, that's a good point! I always wear protective glasses when doing this. Luckily, none exploded yet and I made this mod now on many boards. Probably 80°C is not hot enough, but better prepared than sorry.
Is there a reason why your heatsink fan is not spinning?
It's not connected to the board. There is no need for a short test in a flying setup. That CPUs didn't get as hot and only needs a fan in a closed case under longer load.
@@necro_ware All right makes sense. Thanks for answering.
If you have 512k cache how come you don't need the tag ram
You have one tag on board for the first 256K, then you have another on the COAST module for the second 256K. That gives you 8-bit tag for 512K which is enough to cache 64M of RAM. If you want to cache up to 512M of RAM, you need 11-bit tag and then you have to populate that socket on the board.
Have the same board with post code 4E4A.Who knows what is that?
4E is a very late error code and means "Show all error messages on screen". That happens after all critical parts have been initialized. Are you sure that your graphics card is alright? May be just incompatibility and you could try another one?
@@necro_ware Thank you! It works with is a card. I hope make a video with media bus card.
512 MEGA bytes of cache? Running without the CPU fan is detrimental to the poor thing; those old CPUs didn't even have the thermal failure shutdown feature.
A good heatsink is sufficient. TDP Is barely 10 W
Oh did I say MB instead of KB somewhere? The CPU can be used briefly without a cooler. It doesn't get hot as fast as later models.
@@necro_ware yeah, 1:29 i came to the comments to see if i was misunderstanding you there :)
Y asi es como Necroware sube otro video de hardware, que quedará registrado para la posteridad
Save those old computer parts. When the western world economy collapses, we'll be trading what we can't make anymore for necessities of life.
Don't worry then! All of those components are made in Taiwan, even if the entire western economy collapses, you just need to move to Asia
I guess, that if "wetern" economy collapses that far, then all economies will collapse. They are much more bound to each other, than most people without proper economical education think. In simple words, should it collapse, there will be nobody to trade old hardware, we would then probably fall back to agriculture and barter exchange. So, just enjoy that old hardware now and don't make plans about it beyond the end of the world as we know it.
@@necro_ware this.
Even factories in taiwan would not be able to make new chips, they will need to shut down.
So after some time, any computer will be good at some tasks (like running simple databases, doing simple computation). Even good old Pentium 1 ones.
Maybe 386 and lower, would be too unusable. But I'm sure, if anything new could be made, even 386 will have some use.
To scrap Pentium 1 CPU for gold, is sin. Big one. This civilization can learn it hard way. Also chipsets could be unsoldered from motherboards before scrapping (there could be law for it)
New PCB could be made even after collapse, but new chips not.
привет и добро пожаловать ....
Аналогично ;)
@@necro_wareподскажи пожалуйста, изначально в модуле RTC (под компаундом) установлена Батайка или аккумулятор? Если был именно по заводу аккумулятор, то можно ли так просто ставить вместо него Батарейку?
@@PowerGood. Там батарейка изолированная от внешнего питания. У оригинальных чипов Hitachi/Motorola которые использовал IBM, 286 и ранние 386 клоны питание было на 24ом пине, который был специфицирован от 2 до 5В, куда подключался внешний аккумулятор через пару диодов вместе с основным питанием. Когда комп был включен, аккумулятор заряжался, иначе питал RTC. На интегрированных модулях типа Dallas на этом пине исключительно питание от 5В, а батарея подключается к ранее не нужному пину 20, который изолирован внутри чипа от основного питания. Соответственно батарея не может заряжаться.
@@necro_ware спасибо за подробный ответ.
the background music sounds like being taken from a German 1970s porn movie. Isn't the commentary in itself sufficient?
I watched tons of German 1970s porn to find the best tracks. I'm sorry you didn't like it, it was a lot of work 😞
@@necro_wareafter watching tons of 1970s German porn you want to become a monk! Very hairy (indeed).
🤣🤣🤣🤣💀
i ruined up my Shuttle HOT-555 during operation of trying to add battery holder to this goddamn dallas RTC :(