Show off your old and retro computer parts

[Dabombinable]

@da na
I had similar issues on my Gigabyte GA-7IXEH. With the added trouble of DMA being enabled on ODD causing lockups (HDD were fine). Slot 3 (which had salt all over it - sea air sucks) was the cause even though it wasn't being used as I thought it was dead. I cleaned the slot as best I could (used wood toothpick to gently scrape each pin clean), and then worked the slot using a dead stick of RAM to ensure that the contacts were free to move and completely unobstructed.

Solved every issue that I had been having with the board (until I damaged the PCB near the socket removing the Titan cooler I bought for it).

[da na]

17 minutes ago, Dabombinable said:

@da na
I had similar issues on my Gigabyte GA-7IXEH. With the added trouble of DMA being enabled on ODD causing lockups (HDD were fine). Slot 3 (which had salt all over it - sea air sucks) was the cause even though it wasn't being used as I thought it was dead. I cleaned the slot as best I could (used wood toothpick to gently scrape each pin clean), and then worked the slot using a dead stick of RAM to ensure that the contacts were free to move and completely unobstructed.

Solved every issue that I had been having with the board (until I damaged the PCB near the socket removing the Titan cooler I bought for it).

Oh great, good to know! 

So odd that it doesn't crash in Linux, though.

 

 

Update on the XPS M2010: 

All of the replacement graphics cards I bought to replace the artifacting one are even worse - no video output whatsoever. So unfortunate that literally only one video card was ever made for this system, and they are getting very rare given the M2010's low sales figures compared to other XPS models.

The PCIe slot on the machine is the same as the one used on several other Dell laptops of the era, but it has a different display connector (which plugs straight into the graphics card). I took the Quadro FX 1500M from my XPS M1710 and the Radeon X1300M from my Inspiron E1505 and installed them in the M2010, connecting each card to a display with a compatible LVDS connector. They both POSTed and output video to the Inspiron's display I had connected. So it is fully possible to run a non-X1800M graphics card in the machine, just not on its screen sadly.

 

I did manage to repair the original card so it no longer freezes and crashes every few minutes - a bad capacitor was the cause. Now it just introduces a bit of a ...unique challenge to the gaming experience.

Spoiler

 

 

[da na]

1.8" uSATA (Micro SATA) server drive.

The amount and size of discrete SMD components absolutely blew my mind. The top side of the drive has 54 conspicuously large bright orange capacitors.

200GB drive - because server drives have to be weird capacities because we can't have nice things. Case in point, the first SAS/SCSI server hard drives were 9.1GB and for some reason we stuck with that standard for the next 30 years, scaling to 18.2gb then 36.4 and 72.8 and 143 and etc. Why on earth.

 

This thermal interface material is something I've never seen before - this seems to be a "liquid thermal pad". It has the same consistency as thermal pads - wet, a little sticky, and a little firm. I would think this is quite literally thermal pad material that was squeezed from a tube onto the chips.

[Bitter]

I'm pretty sure that's thermal putty! Also wow someone must have had a lot of filtering caps laying around. Is that all SLC?

[da na]

4 hours ago, Bitter said:

I'm pretty sure that's thermal putty! Also wow someone must have had a lot of filtering caps laying around. Is that all SLC?

They look to be. 

And thank you for the term "Thermal Putty". 

 

 

 

And thank you @dabombinable for the help with the odd Pentium III board. The issue definitely seems to be on the driver/BIOS side; I thoroughly cleaned the motherboard's RAM slot and the RAM's connectors, it passes memtest86 just fine and works great on Linux, but always crashes at the same stage of Windows XP setup or Windows 7 boot.

I will try to get W98 running and see if that does any better - have to find my boot floppy first. 

 

[flibberdipper]

12 hours ago, da na said:

snip snap patty whack

I wonder if all those caps are actually for dumping data to the drive in the event of a power loss... That would be very server-esque.

[Bitter]

7 hours ago, flibberdipper said:

I wonder if all those caps are actually for dumping data to the drive in the event of a power loss... That would be very server-esque.

Probably not, the total charge in those would be very small, Pico farads worth probably not even one micro farad. What's the numbers on them?

[da na]

Those fPGA cards are neat, but they can't run Crysis, can they?

Quadro 2000 (GF106 die, single-slot/75-watt GTS 450) in the radio receiver. Its DVI port nicely screws into one of the many empty DB15 cutouts in the back of the case.

Bus interface is a pretty major limiting factor, seeing as this board does use a laptop chipset it is limited to PCIe 1.1 x1 - 1/32 the bandwidth this card is intended to run at.

I originally intended to install a 9800 GTX+ in this system, but the 180-watt power supply only gives 90 watts on the 12v rail. I didn't manage to get very far with the 150-watt geForce, so instead opted for this Quadro, which performs almost exactly like a 9800GT with a 75-watt power requirement. 

[da na]

4 minutes ago, Bitter said:

Probably not, the total charge in those would be very small, Pico farads worth probably not even one micro farad. What's the numbers on them?

They aren't labeled with any direct values and since they're in-circuit, each one tests slightly differently with a multimeter

[Bitter]

1 hour ago, da na said:

They aren't labeled with any direct values and since they're in-circuit, each one tests slightly differently with a multimeter

Part number at least? I'm curious now what kind they are. Sure looks cool! Plot twist, turns out they're all 0 value blanks and it's all just for looks to sell product.

[Kilrah]

SMD caps typically have no markings.

You can easily get 10uF per cap at the size I'm guessing it is though.

[Kilrah]

Also searching for the model you get results for one that while different specifically boasts using caps to buffer writes, so it's very likely this one has it too.

https://www.thessdreview.com/our-reviews/smart-storage-systems-xceediops-2-200gb-emlc-6gbps-enterprise-ssd-review/

 

Quote

The processor is accompanied by 32 individual OxiCap Niobium Oxide Capacitors. These are the yellow squares between the processor and the NAND dice. They provide power for the SSD to flush all data to NAND in the case of an unplanned power loss.

 

[da na]

Just now, Kilrah said:

Also searching for the model you get results for one that while different specifically boasts using caps to buffer writes, so it's very likely this one has it too.

https://www.thessdreview.com/our-reviews/smart-storage-systems-xceediops-2-200gb-emlc-6gbps-enterprise-ssd-review/

 

 

Wow, that's absolutely fascinating. 

[SimplyChunk]

13 minutes ago, Kilrah said:

Also searching for the model you get results for one that while different specifically boasts using caps to buffer writes, so it's very likely this one has it too.

https://www.thessdreview.com/our-reviews/smart-storage-systems-xceediops-2-200gb-emlc-6gbps-enterprise-ssd-review/

 

 

 

9 hours ago, flibberdipper said:

I wonder if all those caps are actually for dumping data to the drive in the event of a power loss... That would be very server-esque.

Nailed it lol

[Bitter]

13 minutes ago, Kilrah said:

Also searching for the model you get results for one that while different specifically boasts using caps to buffer writes, so it's very likely this one has it too.

https://www.thessdreview.com/our-reviews/smart-storage-systems-xceediops-2-200gb-emlc-6gbps-enterprise-ssd-review/

 

 

I'm shocked they have enough capacity for that. I guess I'm used to raid cards with nicd battery packs leaking all over.

[da na]

Just now, Bitter said:

I'm shocked they have enough capacity for that. I guess I'm used to raid cards with nicd battery packs leaking all over.

I suppose it takes a lot more time and power to dump 512MB of DDR SDRAM to an array of 16 HDDs than it would take to dump the what, 16mb? DRAM cache to NAND.

[luckybob77]

You can get a standard MLCC capacitor in a 0602 package up to 330uF

 

you can also get super small, super caps too.  https://www.mouser.com/ProductDetail/Seiko-Semiconductors/CPH3225A?qs=3etwrb1wR%2BhUOph6lAO7eg%3D%3D

 

They 100% have enough power to save the dram cache to the flash before the caps fully discharge.

[SimplyChunk]

Have acquired an Epson Stylus Photo 1200 large format printer.  Cartridges seem to be plentyful and reasonably priced for a .c 1999 printer.  I can't find one on fleabay though.  Anyone got an inkling on what it may be worth?  It's got support for both LPT and USB and the drivers are still available on the Epson website

[da na]

8 minutes ago, SimplyChunk said:

Have acquired an Epson Stylus Photo 1200 large format printer.  Cartridges seem to be plentyful and reasonably priced for a .c 1999 printer.  I can't find one on fleabay though.  Anyone got an inkling on what it may be worth?  It's got support for both LPT and USB and the drivers are still available on the Epson website

Nice printer. 
I'd put it in a similar price range as the venerated Pixma Pro9000, those go for like $80-240 depending on condition.

The big issue with those old printers is that they are quite expensive to maintain. For example, my Pixma Pro9000 MK.II takes eight ink cartridges, which Canon sells for $25 each, and it can burn through them in under 40 prints on high DPI mode. And aftermarket ink is a lot cheaper, but clogs the printheads and is incredibly annoying to clean. They aren't attractive to many buyers because, though the printer might be cheap, the continual cost of actually using it gets high quickly.

[SimplyChunk]

56 minutes ago, da na said:

They aren't attractive to many buyers

that's what I was thinking.  To be honest i've not even plugged it in yet to test it and the heads may be, like you say, glogged

[da na]

10 minutes ago, SimplyChunk said:

that's what I was thinking.  To be honest i've not even plugged it in yet to test it and the heads may be, like you say, glogged

They run great if used frequently, and there are some die-hards still hanging onto these old large format printers (me). But I'd think everyone who wants one already has one.

[luckybob77]

meh.  Inkjets have been clog machines since conception.  Now my HP 7475A plotter, never once clogged.  the pens did dry out pretty quick though...

 

but seriously.  unless you sit there and print something on a daily basis, they arent worth it.  At that point you just go color laser.

[Dabombinable]

Finally have a working SS7 board - can see if I killed my old K6-2 500 with the mosfet short, and test my 72pin SIMM.

Spoiler

Had to swap the cooler or the board wouldn't fit. And the Medalist finally died after years of bad sector warnings so that was also replaced, and the Fireball's performance is on point.

Spoiler

Board appears to have great performance as well despite the 8bit TAG (128MB cacheable RAM). Will be getting a K6-3+ of some clock speed (Jan's patched BIOS and 2V support means perfect compatibility).

[da na]

Little Coppermine ITX board is up and running in Windows 98 SE.

VIA VT8604 chipset, S3 ProSavage AGP integrated graphics, and of course, Celeron 1000.

I'd forgotten how massive of a difference a good hard drive makes on Windows 98. This 5-platter Hitachi Deskstar ATA is so much snappier than the CompactFlash card in my Compaq Win98 laptop, it's unreal. It almost consistently tops out the ATA interface with its sequential reads/writes.

[da na]

Unlikely hero came to the rescue tonight. Bought RimWorld on Steam, quite excited to play it but my workstation said "no, you can't have fun!" and just wouldn't install the game. But someone would...

Now I'm happily rimming. 

Quite enjoyable. Runs excellently at 1920x1200 on the Quadro FX4800, sadly does not utilize SLI:( 

As for sound, I've attached three tactile transducers to the back of the Apple monitor - one for left, one for right, and a considerably larger one for bass. The solid aluminum back panel of the screen sounds really nice when it vibrates, although the bass sounds... metallic at times. Had to tighten the stand because the bass actuator kept making the hinge slip and tilting the screen down at a certain frequency.

I tried the TouchSmart on the left first and the game was unplayable, I don't say that lightly. Seems to require DX10 and since the TouchSmart IQ770's GeForce 7600M had only DX9 support, all acceleration was handed over to the weak CPU. Of course it has no issues with a Core 2 Quad and fast GPU, though.