fosterwj03 wrote on 2025-02-22, 22:20:

I was thinking XP and 7 as well, but sure, it could do 11, I guess.

I use 11 on my daily driver, so I don't have much nastalgia for it yet. Whereas, I ran Windows 2000 on my daily driver 20 years ago.

Use Win10 instead, Win11 23H2 is the last decent version of 11 without too much AI infection or forced bitlocker. MS is currently on a AI/Bitlocker push.

Trashbytes wrote on 2025-02-22, 23:50:

fosterwj03 wrote on 2025-02-22, 22:20:

I was thinking XP and 7 as well, but sure, it could do 11, I guess.

I use 11 on my daily driver, so I don't have much nastalgia for it yet. Whereas, I ran Windows 2000 on my daily driver 20 years ago.

Use Win10 instead, Win11 23H2 is the last decent version of 11 without too much AI infection or forced bitlocker. MS is currently on a AI/Bitlocker push.

I was thinking about creating a Windows 10 retro rocket, and it might be close to time. I'm mostly worried about license activation with Windows 10 long-term. I also don't want annoying "Upgrade to Windows 11" notifications when I'm trying to preserve a Windows 10 install.

One other issue I have with multi booting my retro rockets is constantly switching out the graphics cards. I have a PCI Matrox G200 for NT 3.51, GeForce 6800 GS for NT 4, Radeon x800 XL for Windows 98/Me, and a GTX 480 for Windows 2000. Adding XP and 7 to the boot rotation on the same system would add my GTX 980 Ti to the mix (I'm waiting for prices to drop to get an RTX 3080 for Windows 7). That makes for a lot of component swapping (to say nothing of sound card swaps).

fosterwj03 wrote on 2025-02-23, 01:42:

Trashbytes wrote on 2025-02-22, 23:50:

fosterwj03 wrote on 2025-02-22, 22:20:

I was thinking XP and 7 as well, but sure, it could do 11, I guess.

I use 11 on my daily driver, so I don't have much nastalgia for it yet. Whereas, I ran Windows 2000 on my daily driver 20 years ago.

Use Win10 instead, Win11 23H2 is the last decent version of 11 without too much AI infection or forced bitlocker. MS is currently on a AI/Bitlocker push.

I was thinking about creating a Windows 10 retro rocket, and it might be close to time. I'm mostly worried about license activation with Windows 10 long-term. I also don't want annoying "Upgrade to Windows 11" notifications when I'm trying to preserve a Windows 10 install.

One other issue I have with multi booting my retro rockets is constantly switching out the graphics cards. I have a PCI Matrox G200 for NT 3.51, GeForce 6800 GS for NT 4, Radeon x800 XL for Windows 98/Me, and a GTX 480 for Windows 2000. Adding XP and 7 to the boot rotation on the same system would add my GTX 980 Ti to the mix (I'm waiting for prices to drop to get an RTX 3080 for Windows 7). That makes for a lot of component swapping (to say nothing of sound card swaps).

Install Win10 21H2 and use Gpedit to lock it to updates for that version only. It will never ask you to upgrade to any other version of windows and you can also turn of automatic restarts ...I have a machine setup this way and its rock solid with a 6 month uptime.

There are always ways to stop MS from controlling your machine, I use a well known little program called incontrol https://www.grc.com/incontrol.htm, its the one that lets you lock down windows updates and prevent any random unwanted upgrades.

I also use this to lock my Windows 11 machines to 23H2 which is the last version before they forced Recall AI and decided that Bitlocker is enabled by default. (Enabled but not active, just one less step they have to take later)

Socket3 wrote on 2025-02-21, 18:38:

PcBytes wrote on 2025-02-19, 18:18:

My guess is only "MMR" models might have had 1GHz+ cores - I have a 700MHz "MPR" T-Bird and the core is also rated for 700.

(the reason I say 1GHz + is because I just recently saw a 900MHz model with a 1100 core.)

I wonder why AMD released them at lower speeds, as this is a pretty common occurrence. My 750MHz Thunderbird core CPU is marked "850" on the die, and will happily run at that speed @ stock voltage.

I believe it has to do with binning, when being tested for operation at the rated speed it didn't quite meet the QC threshold at that speed, so they released it at a lower speed. Or it could simply be a matter of supply if they had an abundance of higher speed dies on hand and needed to fill market demand for lower speed CPUs.

Brawndo wrote on 2025-02-23, 04:18:

Socket3 wrote on 2025-02-21, 18:38:

PcBytes wrote on 2025-02-19, 18:18:

My guess is only "MMR" models might have had 1GHz+ cores - I have a 700MHz "MPR" T-Bird and the core is also rated for 700.

(the reason I say 1GHz + is because I just recently saw a 900MHz model with a 1100 core.)

I wonder why AMD released them at lower speeds, as this is a pretty common occurrence. My 750MHz Thunderbird core CPU is marked "850" on the die, and will happily run at that speed @ stock voltage.

I believe it has to do with binning, when being tested for operation at the rated speed it didn't quite meet the QC threshold at that speed, so they released it at a lower speed. Or it could simply be a matter of supply if they had an abundance of higher speed dies on hand and needed to fill market demand for lower speed CPUs.

From experience, it had to do with an overabundance of higher speed dies.

Almost every single SLOT-A CPU I have pulled apart has had a higher rated due than what it was sold as.

For the ones before Thunderbird, it was better to buy specific speed ones because they had faster L2 cache chips than certain ones that were sold as higher speed CPUs.

This is because they used a divider for the L2 cache so certain ones running 1/2 speed cache would end up having the higher speed L2 cache than the higher ones that were sold with the cache running at 1/3 or 2/5.

It has been a while so I looked it up. The best to have as far as getting higher speed cache would have been the 700 or the 850:

https://www.vogons.org/viewtopic.php?t=91885#p1160688

Well, I had a rather expensive day, but it's not a bad kind. No, instead I finally got something I've wanted for a while.

But first I will need to call some random guy...

Yes, I am now slowly constructing a tower of MIDI, and have added an MT-32 to the mix alongside my AWE64 gold and 1.0 ROM SC-55 for the "unique sound" side of things.

The best part though?

It's a 1.0.7 from the factory. Certainly makes it a lot easier to use without having to think too much. I've got it working through a dumb mixer now, and am using my YMF card and SoftMPU to drive it.

cyclone3d wrote on 2025-02-23, 05:27:

From experience, it had to do with an overabundance of higher speed dies. […]
Show full quote

Brawndo wrote on 2025-02-23, 04:18:

Socket3 wrote on 2025-02-21, 18:38:

I wonder why AMD released them at lower speeds, as this is a pretty common occurrence. My 750MHz Thunderbird core CPU is marked "850" on the die, and will happily run at that speed @ stock voltage.

I believe it has to do with binning, when being tested for operation at the rated speed it didn't quite meet the QC threshold at that speed, so they released it at a lower speed. Or it could simply be a matter of supply if they had an abundance of higher speed dies on hand and needed to fill market demand for lower speed CPUs.

From experience, it had to do with an overabundance of higher speed dies.

Almost every single SLOT-A CPU I have pulled apart has had a higher rated due than what it was sold as.

For the ones before Thunderbird, it was better to buy specific speed ones because they had faster L2 cache chips than certain ones that were sold as higher speed CPUs.

This is because they used a divider for the L2 cache so certain ones running 1/2 speed cache would end up having the higher speed L2 cache than the higher ones that were sold with the cache running at 1/3 or 2/5.

It has been a while so I looked it up. The best to have as far as getting higher speed cache would have been the 700 or the 850:

https://www.vogons.org/viewtopic.php?t=91885#p1160688

that is very informative, thank you!

cyclone3d wrote on 2025-02-23, 05:27:

Almost every single SLOT-A CPU I have pulled apart has had a higher rated due than what it was sold as.

Key word - almost. Because two of my other Athlons were marked accordingly on both die and case - Thunderbird 700 has a 700 marked core, Pluto 700 has a 700 marked core.
The only one I have with a higher rated die is a 600 Pluto which I resistor modded to 800 @1.7v and am currently searching a new board for, as my recent MS-6191 purchase died.
I think it might do 900 if I bump the voltage a bit.

Picked up some goodies at the recycler/thrift store this weekend.
A beautiful packed full Dell XPS T700R time capsule. With a TNT2, A3D Vortex 8830 & a 700Mhz Slot 1 !
+ 1.25GB's Of PC 133 a Working Riva 128ZX 8MB, a random HT omega sound card and some cables.

Feels 🥰.

AGP4LIfe? wrote on 2025-02-23, 21:25:

Picked up some goodies at the recycler/thrift store this weekend. A beautiful packed full Dell XPS T700R time capsule. With a T […]
Show full quote

Picked up some goodies at the recycler/thrift store this weekend.
A beautiful packed full Dell XPS T700R time capsule. With a TNT2, A3D Vortex 8830 & a 700Mhz Slot 1 !
+ 1.25GB's Of PC 133 a Working Riva 128ZX 8MB, a random HT omega sound card and some cables.

Feels 🥰.

The attachment PXL_20250223_032343370~2.jpg is no longer available

The attachment PXL_20250223_032916225.jpg is no longer available

Was the 1.25 gigs installed and working? On what chipset?

H3nrik V! wrote on 2025-02-23, 22:32:

AGP4LIfe? wrote on 2025-02-23, 21:25:

Picked up some goodies at the recycler/thrift store this weekend. A beautiful packed full Dell XPS T700R time capsule. With a T […]
Show full quote

Picked up some goodies at the recycler/thrift store this weekend.
A beautiful packed full Dell XPS T700R time capsule. With a TNT2, A3D Vortex 8830 & a 700Mhz Slot 1 !
+ 1.25GB's Of PC 133 a Working Riva 128ZX 8MB, a random HT omega sound card and some cables.

Feels 🥰.

The attachment PXL_20250223_032343370~2.jpg is no longer available

The attachment PXL_20250223_032916225.jpg is no longer available

Was the 1.25 gigs installed and working? On what chipset?

Only 256MB in the XPS. The other 1.25GB's (3 Sticks) is loose for other machines.

Ozzuneoj wrote on 2025-02-18, 15:15:

Alexraptor wrote on 2025-02-18, 14:35:

UPS just dropped off this lovely "upgrade" today. […]
Show full quote

UPS just dropped off this lovely "upgrade" today.

The attachment GkE4ugKXAAAc1rY.jpg is no longer available

I was just looking for a little more juice in the system, going from 145W to 200W. But the fan on this unit turned out to be whisper quiet, which honestly made me tear up a little.

My ears are in heaven now. 🤤

Oh man, the yellow label made me twitch a little from PTSD. That isn't a relabeled Bestec is it?

It's a Delta, I am 100% sure.
Any time you see "DPS"-something in the beginning of the model number, rest assured it's a Delta PSU.
And if you see "NPS"-something, that's a Newton PSU... which I believe is either a daughter or subsidiary company of Delta, so you can expect the same solid quality.

That said, I still wouldn't plug this in any of my PCs right away without a proper check of the 5VSB rail first, to make sure it's at 5V and/or +- 0.15V. Anything more could be indicative of trouble with the caps or a run-away 5VSB circuit. FWIW, Delta's early PSUs often used 2-transistor self-oscillating circuits for the 5VSB, and these are quite prone to going over-voltage over time and taking out the SIO/LPC and/or Southbridge and other TTL circuitry. So always check old ATX PSUs for that... and if in double, open the PSU, take pictures, post here, and let others see what is going on.

Ozzuneoj wrote on 2025-02-18, 19:02:

Funny thing is... they really haven't changed at all. […]
Show full quote

cyclone3d wrote on 2025-02-18, 18:28:

Yeah, Dell did some weird stuff back in the day. A PSU that slide-locked into the bottom of the case was one of them.

Funny thing is... they really haven't changed at all.

Their midsize desktops are still a disastrous mix of wasted engineering time\energy and planned obsolescence.

I get the impression that this is some kind of deeply ingrained aspect of the way the company operates. Like, multiple levels of people creating nonsense busy-work to justify the existence of their job and to make use of engineering educations that would probably be better used in other fields.

Glad to see I'm not the only one that fees this way.
Several years back, I had two cases from old i7 Studio XPS 8700 towers. Both needed a deep cleaning, so I decided to not only take off the front bezels, but also take apart the bezels completely. It took me nearly 1 hour to take out all of the little pieces and plastics on these, as there were THREE layers of different plastic parts clipped into each other. Half-way through the process, I remember stopping and admiring the creativity and CAD skills someone (or the team) that came up with the design of all parts - it was so damn complex!
But for what?! Just to have some of the trim pieces made of different plastics as well as the vents a different color plastic than the shroud around it?
Absolute waste of engineering talent, like you said... or someone just trying to justify all of their OT at work maybe?
Anyways, I cleaned it all up by way of washing with a brush (water + dish detergent) as there was no other way to intricately clean this bezel... and a good clean up really was needed, because I suspect the shop I got these PCs from had roaches. Thus, I had to check out every nuk and cranny, just in case.

AGP4LIfe? wrote on 2025-02-15, 01:16:

These two absolute beauties came in the mail today! 😍
A Terratec GeForce FX 5800 & a Winfast A250 Ultra (GF4 Ti 4600).

Just a heads up - that Leadtek GF4 TI 4600 will need a recap in the long term. I have the same card, but with TI 4400 core. Caps are the same though - a mix of polymers and very crappy GSC electrolytics. The GSC's have to go ASAP. The polymers should be fine.

smtkr wrote on 2025-02-15, 04:08:

I always loved the cooler on those leadteks

It's actually an adequately-sized cooler for these cards, for once. So I like that part about it too. What I don't like are the dual tiny screamer fans attached to it. I haven't replaced mine yet, but probably gut them for a single 80 mm or 92 mm fan for better cooling and less noise.

New old stock 200W AT power supply came up for the price of a used low end ATX. I think these are good quality, has Nichicon caps. Just have to get the musty smell out of it!

Just above someone was mentioning having PTSD from Bestec PSUs 😁 I personally never used or even seen that brand before, so can't comment though.

Almost every brand has good and bad models, some brands being known more for bad models.

The brand I try to use in almost everything is Seasonic.

Even AT systems get a nice, shiney Seasonic and an ATX to AT converter. There used to be an eBay seller that sold adapters with the -5v built in, but I will have to start either making my own adapters, mod the motherboards or use something like a voltage blaster.

momaka wrote on 2025-02-24, 00:44:

AGP4LIfe? wrote on 2025-02-15, 01:16:

These two absolute beauties came in the mail today! 😍
A Terratec GeForce FX 5800 & a Winfast A250 Ultra (GF4 Ti 4600).

Just a heads up - that Leadtek GF4 TI 4600 will need a recap in the long term. I have the same card, but with TI 4400 core. Caps are the same though - a mix of polymers and very crappy GSC electrolytics. The GSC's have to go ASAP. The polymers should be fine.

Yea I noted two or three of the caps were very very slightly bulged. The card tested just fine through the stress test, but I'm sure it would be good to do so eventually.

Thanks for the extra motivation!

Bought these babies. I'll eventually upgrade the RAM to 1024KiB.

These three S.7 motherboards, I like them quite a lot, both because they are S.7, and because I didn't have these cards among those in my collection, let's see what 😨 they are:

A PCChips TXPRO (aka ALI chipset), in the period when these cards were on sale, I kept away from them, I always preferred 😻 those with Intel chipsets, however today re-evaluating, it could be interesting 🧐, even if I doubt it can get excellent results in the bench.

An ASUS VX97, here is a card that I have never had, I have a fair number of i430VX cards, but none of the ASUS, brand that I have been fond of since the first 386DX that I bought, reading on the retroweb there are some power supply problems, exactly I do not know if it is solvable, we will see when it will be between my paws 🐾, the only visible defect, is the lack of SDRAM slots, if a large quantity is not needed, the four 72 PIN slots will be enough (?).

Finally, a Chaintech 5TDM2 M101, also this never had, so it's already only for this, quite interesting 🧐, if I'm not mistaken, it has automatic recognition of the CPU, and any modification must be made by BIOS, it should be quite similar to the system used by QDI.

All three cards are untested, and therefore not working (?), but it is possible that one or even all, can work, I don't think that normally anyone dedicates time to solving problems, except of course 🙄 those who like me, dedicate their free time only to make old motherboards work, in short a technician usually limits himself to suggesting an upgrade, with something more recent, like a P3 or a P4, because it would cost less than a few hours of work on an S.7.

Bought another 3d Prophet 4500.
Guess it's gonna be my first time with a Kyro card, given the other I got decided to stop working out of the blue. It did have some shady repairs done to it.

Got this with a box of PC parts:

Compatible with DOS/9X/XP. It might be a useful card for my test bench. Not sure I'd use it in a build though. But will be interesting to test it out.

I see Philscomputerlab did a video on this chipset

https://youtu.be/oHKa-TE-U4Y?si=V3PajWwpResqrDuk