• Welcome! The TrekBBS is the number one place to chat about Star Trek with like-minded fans.
    If you are not already a member then please register an account and join in the discussion!

General Computer Thread

All kinds of media crap that needs to be decoded and put into the memory of your machines/graphics card which means it has to work hard so the fans will go to ludicrous speed.

So, found a casing, InWin A500, and drives for the P-II 333, graphics card, ATI Rage Pro II (I think) Realtek based NIC 10/100 a Adaptec SCSI adaptor for the CD Writer, and a Soundblaster AWE 32 ISA soundcard, found a Seagate drive, not sure how large it is, a 250MB tape drive, a 32 speed Philips CD-ROM (IDE) and a Ricoh CD Writer, as for memory, I have SDRAM, think I'll go with 256MB or so, more won't be needed also a the early P-II's had an artificial caching limit, they don't cache memory above 512MB as for an OS, Win2K. :mallory:

*edit*
Found memory, one stick of 128MB and two 64MB sticks so yeah indeed 256MB RAM, found a fan for the CPU cooler, it had been missing, so now there's a 60x60MM fan fixed to it with tie wraps, sounds a little dodgy but the thing isn't going anywhere.:mallory:
I also found a suitable PSU, one made by FSP so it's a high quality unit, also have another one lying around, no idea where it came from, it's a King Year, can't find much about that one, cleaned and tested it anyway, jump started it and it does run without exploding.:biggrin:
Just need to do connect power and cables, the latter will be quite a challenge with the layout of the A500, and then see if it will work.:biggrin:
 
Last edited:
So, the TLC board is not working, it boots up but none of the IDE controllers work, also the floppy controller is out, otherwise it does work, going to test that thing later on again, so I needed a plan B which I found, an old Compaq EP board with a Pentium II 400Mhz chip, so a little faster and a 100Mhz frontside bus speed, problem one, the chip has a BIG passive heatsink, the A500's PSU mounting place is above the CPU slot to extract the heat from the slot area, there's not enough clearance for that big heatsink, found an old standard sized heatsink, fitted it to the chip and AARGHH?!! there's no way to mount any fans, the thing must have had some plastic frame of somekind to fit a fan.. so there are mounting holes in the casing near the CPU slot, they were used for mounting a harddrive there which I am not going to mount there so.. found some L shaped metal things you usually use in the garden, mounted another piece to that and a 60x60mm fan, after some adjusting it is now blowing the air onto the right spot of the heatsink, a bit industrial but hey. :D

Removed the SCSI drive, found an old (and beige) CD writer and an old DVD drive(also beige), less cable mess because I found a few rounded IDE cables HA!
The Seagate drive is a 40GB drive, I've got another one from the same era and probably the same series it's a 10GB unit though.
As for why all this? It's much better than to add to the mountain of ewaste, also they'll get rare in time, on ebay they seem to go for 99 bucks minimum and 200+ for the better systems..
Ah well, in any case it was fun messing around.. :mallory:
 
All kinds of media crap that needs to be decoded and put into the memory of your machines/graphics card which means it has to work hard so the fans will go to ludicrous speed.

But wouldn't gaming be a bit more demanding then visiting the IMDB or searching for things via google or other search engines? The fans never really seem to go nuts when gaming, but anything web related just yeah things get loud.
 
First users report NVIDIA RTX 4090 GPUs with melted 16-pin power connectors

Be very careful about how you handle the RTX 4090 16-pin power connectors.

It could be a fire hazard.


To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.

To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.
 
Last edited:
To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.

To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.
I already watched the Jayz2Cents video, this could get the Consumer Protection Side of the FTC involved and force a product recall for fire hazard.

NZXT had this problem over a bad PCB for the Fan controller.
 
I always watch Jayz videos. But GN does good stuff too.
I primarily watch GN, only occaisionally watch Jayz2Cents videos if it's important info like today's fire incident with the 12VHPWR.

I swear, nobody in the process of approving the standard bothered testing the cables installation in a PC.

They would've easily figured out that they needed the 35 mm of strain relief that CableMods Cable recommends.
jQmVRMh.png

RltjwPa.jpg

Somebody didn't read the CableMod's Cable Safety Pamplet, you'll notice that the plug is bent sideways.
 
Nvidia's approach swap out burned plug, replace, keep testing.
I think it was Jayz video he mentioned something along the lines of 25 amps going through that connector..
 
Nvidia's approach swap out burned plug, replace, keep testing.
I think it was Jayz video he mentioned something along the lines of 25 amps going through that connector..
Yup it was, and if the cable has enough pressure that the receptacle contacts aren't properly seated on the plug pins, then you create a lot of resistance on that attached wire and potentially start a fire.

CableMods Cables tested and noted that you needed 35mm of strain relief right behind the plug to prevent the pins from getting moved out of spec and place by bending the cables in funny ways.

Somehow, either Intel, nVIDIA, PCI-SIG or the Adapter manufacturer didn't catch this basic fact and end users have ways of screwing that up easily.

Ergo it becomes a HUGE fire hazard and HUGE liability for nVIDIA, Intel, PCI-SIG, or the Adapter manufacturer.

This could force the recall and open up those companies to law-suits if anybody's houses burns down or dies because of a fire.
 
At the moment not really interested in buying a way overpriced stupidity that can draw 600 watts and certainly not in combination with such a moronic design of a connector, in a lot of cases (punny!) there's not enough room to give 35mm of strain relief, you know, because the card is way too large.

Back to that P-II machine, the Seagate is rather troublesome, doesn't work on the TZA board, gave some problems on the Compaq board as well, might be on its way out, found a 20GB WD Caviarm smaller, noisier but 7200RPM instead of 5400 so pretty spiffy and it works fine, installed Win2K on the machine updated all the weay up to the Unofficial SP5.1 which holds every update there was for 2K, now the fun part begins, finding drivers, software and all the old stuff I used in the past. :D
 
Saw this pop up on my feed and it made me smile. The thumbnail made me instantly curious, and I immediately clicked on it. :D

To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.

in a lot of cases (punny!) there's not enough room to give 35mm of strain relief, you know, because the card is way too large.

Yeah, in a trend towards smaller, there's the simple fact that many cases are simply not big enough to accommodate the monstrosity. This might actually end up requiring a whole new category of sizing.
 
Yeah, in a trend towards smaller, there's the simple fact that many cases are simply not big enough to accommodate the monstrosity. This might actually end up requiring a whole new category of sizing.
Design Flaw #1)
Igor of Igor's Lab has done a great analysis of this issue:
The horror has a face – NVIDIA’s hot 12VHPWR adapter for the GeForce RTX 4090 with a built-in breaking point
Go to website for pics of the shoddy craftsmanship that nVIDIA solely outsourced to "Astron", to design & manufacture.
However, the “safe” is only valid if e.g. the used supply lines from the power supply with “native” 12VHPWR connector have a good quality and 16AWG lines or at least the used 12VHPWR to 4x 6+2 pin adapter also offers what it promises. Which brings us directly to the real cause of the cases that occurred: It’s the adapter solution exclusively provided by NVIDIA to all board partners, which has fire-dangerous flaws in its inner construction!
...
A total of four thick 14AWG wires are distributed over a total of six contacts, with the two outer leads soldered to one pin each and the two middle leads soldered to two pins each. The solder base is a mere 0.2 mm thin copper base with a width of 2 mm per incoming wire, which then results in 4 mm per pair for the middle connections. Soldering one or even two 14AWG wires to it is very sporty.
For some reason, Astron thought it was a good idea to deviate from the standard of using 6x 16AWG copper wires to 4x 14AWG copper wires. Then do a shoddy welding job to a very thin copper bus bar to share power.
Why they thought this was a good idea, is beyond me.

CableMods Cables discovered issues with using nVIDIA's provided adapters and created this users safety guide
jQmVRMh.png
One where many of the users who reported melting power adapter cables clearly didn't get the memo or followed.

Design Flaw #2)
To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.
Buildzoid of AHOC YT channel also found another design flaw.
The Female contact terminals inside the 12VHPWR plug use the cheaper "Double Split" design over the more expensive "Single Split" design that CableMods and other PSU manufacturers use. It's probably a savings of a few cents, but somebody must have approved it or didn't know the 3rd Party Manufacturer was going to do that to save a few cents on their end. Either way, that design is more prone to splitting open when end users rock the cable side by side, ergo opening up the contact terminals and increasing resistance.

I think that between the two issues of bad soldering joint + bad soldering terminal design + bad contact terminal choices = fundamentally flawed design.

I could've probably thought of a more reasonable / safer design that only costs a bit more on BOM costs.

Instead of combining 4x ~4"-6" 14AWG cables into one like a squid, have one longer cable using 6x 15AWG wire pairs to the 12VHPWR cable + the 4x sense lines that connects to a simple Hexagonal or Pentagonal PCB with 4x 8-pin PCIe power plug terminals spread on each flat side. Then wrap the PCB inside a simple Aluminium box and use non-electrically conductive thermal pads to help cool down the internals by connecting them to the Aluminium box. One side of the box will have properly strain relived wire direct attached to the PCB. The other side at the plug will also have strain relief.

Yes my design would cost more, but it would be nicer looking, and a fundamentally safer design IMO.

Having the ugly "Squid like tentacles" come out of your $1600 Video Card is unsightly.

I'd rather just have a longer single adapter dongle that can be routed to somewhere in the back and hidden behind the MoBo and obscured by the PC case panel.

nVIDIA, stop being effing cheap skates. I'm sure I'm not the only person who could design a better power adapter for the masses. This is basic stuff that many people who are technologically minded can figure out.
 
Last edited:
^^ Cool clip. :mallory: I've got two C64's somewhere..

But at the moment I am going to diagnose and possibly fix the two not so well working Slot 1 boards.. I am gong to try a few other IDE drives and types of cables, it wasn't unusual in the old days that certain boards, cables and HDD's didn't like eachother.

As for the 4090 and cases, every other graphics card out there can fit, everyone knows what the flipping wiggle room is when it comes to cases, they should have known better, they could have designed a PCB with the connecter somewhere else..
 
Yeah, to paraphrase a famous quote, just because they could, doesn't mean they should.' I think Nvidia took the entire industry aback by going in that direction, when theoretically, other manufacturers, such as Intel, have shown that it's still possible to have a powerful card that's not a huge brick. Intel cards have driver problems, but given time, I think they could be a powerful contender. We need more competition in this industry. And frankly, Nvidia seems to have priced them so high as to feel like they've abandoned the consumer market.
 
Intel's drivers will get better, there's no doubt about that, it is very much NOT easy to get up to AMD's or Nvidia's level so out of the box.
I actually have two Intel cards, they're both around 20+ years old though :D
https://www.anandtech.com/show/199
I actually have the Diamond Stealth II G460 and a Biostar i740 card, that link contains a review of the Diamond card by Anandtech.. it's from 1998 :D
 
In the interests of environmental needs why go to higher power demands?

Why can't designers design graphics cards / chips which can still give good performance and also lower power usage, why does it even need 600 watts of power?
 
I think part of it is complacency. Consumers are generally used to having cards become more powerful with each generation, along with better cooling, and the increased power consumption. But no company seems to have realized that we may have hit the theoretical ceiling where all of this starts to become impractical. Instead of trying to find powerful yet energy efficient cards, manufacturers like Nvidia have gone full-on crazy with their power demands.
 
^^ Also, game and other software developers have become sloppy and lazy producing bloat code that's inefficient and relying on hardware getting ever more powerful instead of streamlining and making their code run efficiently.

Tested three old slot 1 mainboards, the Biostar M6TLC, first time I tried it the poor old thing didn't want to detect harddrives, today I found an old Maxtor 20GB, no idea where it came from but it had a sticker indicating that it came from a local shop, so made a test setup, mobo, ATI Rage Pro gfx card, SDRAM the HDD and a PSU, fired up the thing and it actually found the harddrive configured it correctly and it was good to go, so that's an indication that the Seagate drive I wanted to use was iffy.
Second board was the Compaq board, it also detected and configured it correctly, HDD LED is still not working though, guess something is up with that, otherwise the board is fine.
Last board I tested was an Intel SE440BX probably one of the most expensive Slot 1 boards ever made, it does show it, built quality is about IBM level, it also worked, that last one has been a bit of a troublesome board, really strange, it has been playing dead for at least four years and now it runs again.. ah well, so I have three complete Slot 1 machines, one with a QDI EX board running the almighty Celeron 333A, two machines with a Biostar M6TZA, one P-II 400 and the other one with a P-III 500.
Besides Slot 1 I have one Pentium III 1Ghz machines and a spare board, I also used to have a Celeron 633 but that mainboard has fallen victim by the capacitor plague, I will have it repaired though, it's just that hunting down the right capacitors is a bit of a problem mainly because of the physical dimensions of the caps.
 
^^ Also, game and other software developers have become sloppy and lazy producing bloat code that's inefficient and relying on hardware getting ever more powerful instead of streamlining and making their code run efficiently.

I think that's most of the answer right there. In the 80s an 90s storage was very limited and memory also limited so code had to do all the things required yet also be compact and tight. There was no sloppiness or work arounds.
 
If you are not already a member then please register an account and join in the discussion!

Sign up / Register


Back
Top