Show off your old and retro computer parts

[Bitter]

@da na

I wonder if an oven Hail Mary might fix some of those cards, they're of the cracking solder era aren't they?

[da na]

1 hour ago, Bitter said:

@da na

I wonder if an oven Hail Mary might fix some of those cards, they're of the cracking solder era aren't they?

Yep, I now have around 20 cards and motherboards dead/dying due to bad solder. The issue is finding an oven in which to bake them.

[Average Nerd]

1 minute ago, da na said:

The issue is finding an oven in which to bake them.

Maybe a used kitchen oven? A lot of them can reach temps of over 250°C with no issues, which should cause the solder to melt (unless there's an issue with that that I'm not aware of)

[Bitter]

10 minutes ago, da na said:

Yep, I now have around 20 cards and motherboards dead/dying due to bad solder. The issue is finding an oven in which to bake them.

My toaster oven with a fan in it will hit 450F/230C. You can just slightly insulate the temp control to trick it higher and just use an external thermistor to monitor temps or get your own controller rigged up via a relay between the oven and wall. A little work but probably fairly cheap way to get a reflow oven?

[da na]

6 minutes ago, Average Nerd said:

Maybe a used kitchen oven? A lot of them can reach temps of over 250°C with no issues, which should cause the solder to melt (unless there's an issue with that that I'm not aware of)

Well, I would not want to eat out of said oven afterwards.

(And that's coming from the person who eats while soldering...)

 

I could buy another just for repair, but that does not seem practical for my luxurious 160 square foot apartment.

[da na]

3 minutes ago, Bitter said:

My toaster oven with a fan in it will hit 450F/230C. You can just slightly insulate the temp control to trick it higher and just use an external thermistor to monitor temps or get your own controller rigged up via a relay between the oven and wall. A little work but probably fairly cheap way to get a reflow oven?

That's a fair point, could look for a small toaster oven.

Now I have the mental image of slotting two graphics cards into a toaster and waiting for them to pop up once the reflow is complete.

[da na]

It is so funny how old computers sometimes refuse to work for no explicable reason whatsoever.

Picked up a lovely rare Artist Edition Pavilion DV6 a few weeks ago - 2.4ghz Turion Ultra, ATi HD 3200, 4GB DDR2. 

Instant WHEA error in any Windows version - whether booting from an installer CD or HDD with OS on it.

I had tried a good deal of solutions, was on the verge of buying a new motherboard.

As a last resort, I swapped the DV6's Turion CPU with the 2.1ghz Turion Ultra from my DV4 Special Edition. 

 

Not only did the CPU swap fix the WHEA error, but the 2.4ghz CPU which caused the error in the first place works flawlessly in the other laptop.

The two boards use the exact same socket, northbridge, and southbridge. The CPUs have the same FSB speed and cache. There is not a single reason under the sun that a CPU would POST but be unstable in one, yet a near identical CPU would resolve this issue, while the original CPU works fine in a different machine. I would understand if I'd put this CPU in after-the-fact, but the 2.4ghz chip has been in there since 2008!  

[SimplyChunk]

8 hours ago, da na said:

It is so funny how old computers sometimes refuse to work for no explicable reason whatsoever.

new computers do this sometimes too XD

[flibberdipper]

4 hours ago, SimplyChunk said:

new computers do this sometimes too XD

When my friend and I (mainly me lol) built his new PC, we could not get it to post to save our lives so we were both panicking a little bit, him because he thought his 4080, 7800x, or whatever Gigashite board was broke, and myself because I didn't want him to think I broke it.

 

Turns out it just will not post with one specific flash drive plugged in.

[SimplyChunk]

25 minutes ago, flibberdipper said:

When my friend and I (mainly me lol) built his new PC, we could not get it to post to save our lives so we were both panicking a little bit, him because he thought his 4080, 7800x, or whatever Gigashite board was broke, and myself because I didn't want him to think I broke it.

 

Turns out it just will not post with one specific flash drive plugged in.

:rolleyes:

[Bitter]

3 hours ago, flibberdipper said:

When my friend and I (mainly me lol) built his new PC, we could not get it to post to save our lives so we were both panicking a little bit, him because he thought his 4080, 7800x, or whatever Gigashite board was broke, and myself because I didn't want him to think I broke it.

 

Turns out it just will not post with one specific flash drive plugged in.

I had a problem with an AMD Ryzen build one time semi recently where it was 90% stable but would start throwing memory errors under high memory loads, weird. Swapped in different RAM, ran RAM at different speeds/timings, swapped CPU, nopw nope nope all problems. The RAM and CPU worked fine in another much lower end board, no memory errors. Must be a bad board? Nope, defective assembly...I just had the RAM in the wrong 2 slots. The cheaper board only has 2 slots. It wasn't the first Ryzen build I'd done, I don't know how my brain slipped so hard, I caught it after reading through the board manual because I was getting nowhere with it. Super facepalm.

[da na]

3 minutes ago, Bitter said:

I had a problem with an AMD Ryzen build one time semi recently where it was 90% stable but would start throwing memory errors under high memory loads, weird. Swapped in different RAM, ran RAM at different speeds/timings, swapped CPU, nopw nope nope all problems. The RAM and CPU worked fine in another much lower end board, no memory errors. Must be a bad board? Nope, defective assembly...I just had the RAM in the wrong 2 slots. The cheaper board only has 2 slots. It wasn't the first Ryzen build I'd done, I don't know how my brain slipped so hard, I caught it after reading through the board manual because I was getting nowhere with it. Super facepalm.

I had the memory in my Ryzen board wrong for almost 7 months until @schnoz told me I was a stupid idiot.

I thought the channels went 1-1-2-2 like in... well, every other motherboard I own.

Nope, they're 1-2-1-2. So I skipped a slot thinking I did it right; I in fact put two on Channel 1.

Seemed logical at the time because it's what I was used to. And the motherboard slots were not colored to indicate which should be filled first.

That whole Ryzen board is a fiasco

[da na]

Well here's another fun thing. Core 2 Quad in a laptop, the closest we'll ever get to dual CPUs on a mobile platform.

Only like 5 mobile C2Qs were released and they are compatible with just a couple chipsets so actually finding one in the wild is very rare.

That's the Precision M4400, pretty remarkable to have essentially dual processors as well as Nvidia graphics and 1920x1200 display in a 15" laptop.

[da na]

On 4/3/2024 at 8:00 PM, da na said:

It is so funny how old computers sometimes refuse to work for no explicable reason whatsoever.

Picked up a lovely rare Artist Edition Pavilion DV6 a few weeks ago - 2.4ghz Turion Ultra, ATi HD 3200, 4GB DDR2. 

Instant WHEA error in any Windows version - whether booting from an installer CD or HDD with OS on it.

I had tried a good deal of solutions, was on the verge of buying a new motherboard.

As a last resort, I swapped the DV6's Turion CPU with the 2.1ghz Turion Ultra from my DV4 Special Edition. 

 

Not only did the CPU swap fix the WHEA error, but the 2.4ghz CPU which caused the error in the first place works flawlessly in the other laptop.

The two boards use the exact same socket, northbridge, and southbridge. The CPUs have the same FSB speed and cache. There is not a single reason under the sun that a CPU would POST but be unstable in one, yet a near identical CPU would resolve this issue, while the original CPU works fine in a different machine. I would understand if I'd put this CPU in after-the-fact, but the 2.4ghz chip has been in there since 2008!  

Well to add onto this, Turions have very bizarre compatibility problems.

It seems that support for older chips is removed from newer BIOSes.

With a Core 2 Duo, there are older chipsets that don't support newer chips, but I've never had an older chip not be supported on a newer chipset as long as it fits in the socket.

Turion 64 is not this way at all. A non-Ultra chip won't work in a machine designed for the Ultra chips, and the only difference is the clockspeed. A machine built for the later 2ghz non-Ultra models also will not accept the early 1.6ghz models built on a larger process node. 

I still find it really funny that AMD branded their Turion chips above the blazing fast speed of...... two whole gigahertz!!!! as "ULTRA" chips, and considering that a 2.1ghz "ULTRA!!!!!" performs like a 1.6ghz Core 2 Duo while consuming almost triple the power... 

 

EDIT: They ALSO do not modulate their clocks whatsoever, it seems. Unlike a Core 2 Duo, where a 2400mhz chip can run at any speed between 200 and 2400mhz, a comparable Turion chip has two speeds: 2400mhz, and off.

 

[WhitetailAni]

On 4/3/2024 at 8:26 AM, WhitetailAni said:

a 500 in brown..??, and a 554 in tan. The 554 I got for a single dollar because the antique store was closing...

Forgot to mention this, but the system has a pulse dialing to DTMF converter inbuilt! When I dial on one of my rotary phones I can hear it play the tone after the pulses finished. I wasn't expecting it to do that, it's handy!

[Bitter]

On 4/4/2024 at 12:25 PM, da na said:

I had the memory in my Ryzen board wrong for almost 7 months until @schnoz told me I was a stupid idiot.

I thought the channels went 1-1-2-2 like in... well, every other motherboard I own.

Nope, they're 1-2-1-2. So I skipped a slot thinking I did it right; I in fact put two on Channel 1.

Seemed logical at the time because it's what I was used to. And the motherboard slots were not colored to indicate which should be filled first.

That whole Ryzen board is a fiasco

Yeah I was supposed to have both sticks in the furthest channel. It was 1,2,1,2 I think and I had them in 1,1 but it was 2,2 when only using 2 of 4 slots. I think. I just always consult the manual now, it's a good time saver.

[da na]

After the nice Core 2 and fPGA surprise I found in the satellite receiver, I purchased another - this one is 1U instead of 2U and has a release date of early 2004 instead of ~2008.

I think its design is a little more attractive.

Being a smaller unit, there is not quite as much inside. Just like the newer unit it features an ITX motherboard, standard hard drive, and ATX SFF power supply, as well as custom cards.

The board is based on a VIA chipset (Oh joy.) 

I had guessed the CPU of this machine would be a mobile Athlon XP; I figured a Pentium 4 Williamette chip would be way too hot for this small enclosure. 

I was right about the second part - International Datcasting didn't use a Pentium 4. However, they stuck with Intel, instead using a Socket 370 Celeron. 

I had never actually seen a Pentium III before now. Got a good laugh out of the die-size-to-PCB ratio on this big guy.

The chip has alright specs, more than enough for a radio receiver for certain. 1GHZ clock, 100mhz FSB, and 128KB cache. 

The system packs a reasonable 256-mb of PC133 RAM. This RAM gives us an approximate date code of about when this particular unit might have been manufactured - late 2004.

The satellite interface card, just like the newer unit, runs on an Altera FPGA with 8MB of dedicated memory. It is connected to the board's disk controller over IDE. 

An interesting thing to note is, since the main IDE cable was used for this FPGA card, the HDD is actually connected to the system's secondary IDE channel, which is the thinner 44-pin connector mostly seen on laptops. Makes a ton of sense why this card would not be run in master-slave mode with the HDD and instead the HDD would get its own odd adapter, I presume it would be very difficult to use the FPGA card's video capture function were there only 1 IDE channel since I am not sure there would be enough bandwidth to read from the card and write to the drive simultaneously.

I'll get Windows XP running on this soon. 
If it's like the last unit, it will have shipped with Fedora Linux. I'll keep the original hard drive around and mess with it in the future.

[da na]

She lives!

Wouldn't boot with the original RAM, which had corrosion along the slot. Replaced it with another 133mhz stick while I left the original to soak in an alcohol bath.

Also unplugged (probably forever) the system's two incredibly bothersome fans. On the Core 2 Duo unit I replaced the small axial CPU fan with an 80mm radial set on top of the heatsink, will probably do the same here. Cools the CPU just as well but almost silently, and provides enough intake of its own so as not to need another noisy system intake fan.

HDD still works! The system software runs on Red Hat Linux. It boots for a little while then performs a soft reboot after around 30 seconds. The Core 2 Duo unit did this one too; I am not sure why. Probably why both of them ended up for sale in thrift stores. 

These units are awesome but they reinforce my belief that computers should not be forced into places they do not belong, because look what happens! Your $2,000 piece of rack gear is now useless because someone pulled the wrong plug and Linux got corrupted. 

 

EDIT: Got the OS booting properly!

It has a pretty ballin' GUI, seems to run a real desktop environment. I thought it would just be CLI. 

After the GUI loads, an ancient Firefox version opens and loads into a cached website for the Public Radio Satellite System's FTP server management site. It seems at some point this unit was run as a satellite fileserver. Rad.

[da na]

Well I hate to spam this thread but this seems like the best place to ask if anyone has experience with something similar.

It seems I have very little clue how to actually use a computer that doesn't hold my hand through the entire process.

 

I've been trying to get this board working for a while. It boots flawlessly into the copy of Red Hat Linux on the original HDD. 

Windows XP installer runs fine until when it reboots for the first time, when the chipset/video driver causes a BSOD.

Windows 7 installer loads its files but the system locks up due to an IRQL error when the GUI is loading.

Putting in a HDD with Windows 7 already loaded, the system boots to "Starting Windows" then hard reboots. Entering the recovery environment and doing a system restore throws an "Application xxxx tried to reference memory at xxxx" error.

So, memory seems to be the problem.

I've tried clocking the memory down, disabling caching/shadowing, disabling all of the optional accessories I can that might be reserving memory range. Only thing I cannot do is disable the onboard graphics since I do not have a PCI video card, but I have a feeling that is what is causing the memory issues. (A Matrox Millennium 4MB is en route.) My theory is that the memory range that the integrated graphics reserves is the same memory range that the operating system wants to use, but there does not seem to be an option to change the AGP card's memory allocation, just the parallel ports'. I might be completely wrong there. In fact, I likely am.

 

I would assume it was a hardware problem with the memory or chipset, but the fact it boots and then runs Red Hat Linux flawlessly really makes me think that something has gone awry on the BIOS side. 

 

To be fair I've owned three VIA boards and, although none have been quite this problematic... well, the other two weren't too much better.

[da na]

Well, here's something incredibly boring with a bit of an interesting history.

This is a prototype Chromebook from 2010. Codename "Mario Fish" or the Google 2330. 

This was the first hardware prototype for the next generation of e-waste; Google showed this off to prospective Chromebook manufacturers to give a general idea of what a Chromebook should be like.

And... well, I actually have nothing to say about it. 

This is genuinely the most boring laptop of all time.

It's like if you searched "Laptop Clipart" and 3-D printed it. Made entirely out of solid black plastic, has no defining features whatsoever, and it has three ports and one of them is the charger. So yeah, it's a Chromebook alright!

Under the battery lies what makes this laptop the slightest bit worth having around, though - the sticker signifying its existence as a prototype unit.

I am not supposed to have this. No-one is supposed to have this. These were never released. 

A few were given to review outlets, who pretty much all said, "So, it's a web browser and nothing else? Who would buy a computer like that?". 

 

I do not have the charger; it uses a different one than I would expect. Despite this being remarkably boring in design and specifications (1.6ghz Atom, 2GB RAM, 32GB storage), I cannot wait to boot it up. The prototype here is not only the hardware, but also in the ChromeOS software.

Of course, I will try to hack it. Maybe see if it can boot from USB/SD or if Google left a drive interface header on the board somewhere. 

 

[Kilrah]

4 hours ago, da na said:

So, memory seems to be the problem.

Would suspect so, do you have other RAM to swap in?

[da na]

5 hours ago, Kilrah said:

Would suspect so, do you have other RAM to swap in?

Forgot to include in my post that I tried a few other RAM sticks.

Tried a 256MB Crucial ECC stick, Infineon and Centon 128MB stick. The Crucial produced the same condition, the 128MB ones weren't large enough to load the RAMdisk for Windows 7 recovery so I am not sure how well they worked.

I cleaned out the memory slot with alcohol, since the original RAM seemed to have corrosion along the edge. 

[Bitter]

So with as functional as these are....can you hook some dishes or antenna to them and catch signals?

Also IDE makes sense, high bandwidth low latency, up to 133MB/S both ways. Much faster than anything else that board has on it.

[da na]

1 hour ago, Bitter said:

So with as functional as these are....can you hook some dishes or antenna to them and catch signals?

Also IDE makes sense, high bandwidth low latency, up to 133MB/S both ways. Much faster than anything else that board has on it.

Don't see why that would not work. I lack dishes or antennas to plug into them, but if the DVB protocols haven't changed since 2004 there's no reason for it not to still be able to do its job.

 

 

EDIT: The next post in this thread will be the 10,000th!

[SimplyChunk]

8 hours ago, da na said:

EDIT: The next post in this thread will be the 10,000th!

Well done everybody!!