Linoleum wrote on 2025-03-20, 01:29:
I finished my refurbishing job of this filthy and rusty AT power supply. Happy with the results. Good for another 35 years!
Wow, nice work!
How did you get the metal looking nice and shiny again like that? Rust-converter paint? Or just regular paint + sanding?
I'm curious, as I've encountered a lot of rusty cases over the years... even have one in the basement waiting to decide what to do with it (it's seriously rusty!) Would be curious to know how you got everything nice-looking like that.
JustJulião wrote on 2025-03-15, 21:24:
Traveled 80kms to pick several CRT monitors for free.
Unfortunately, the bottom one, a 21 or 22", turns on but doesn't display anything.
The attachment IMG20250315221920.jpg is no longer available
Looks like a Sony Trinitron build (and FWIW, Viewsonic did use Sony as the OEM for some of their monitors, among others.)
First thing to check is for bad solder joints on the power board - particularly for the the heater, B+, and around HV-drive areas. Speaking of which, does the monitor make that "static electricity" noise during power up? If not, then that might suggest the high-voltage drive is not working. On late Trinitrons, HV drive is separate from HOT from what I remember.
On the other hand if you do get the static electricity noise to indicate the HV is coming up, then a missing heater voltage could be the source of no picture.
RetroPCCupboard wrote on 2025-03-14, 08:48:
MS Office peaked in 2003. After that it was just useless features and a more confusing interface. Haha. Give me back the menus over the ribbon interface any day. 😀
+1
Menus are more intuitive if you don't use Office too frequently and forget where to find stuff. Ribbons are quicker if you use Office daily *and* on the same monitor / PC.
FWIW, I remember when our university switched some of the PCs to Office 2007 (from 2003). Going between different classrooms with different size of monitors meant all of the ribbon buttons moved all the time. Moreover, different PCs had different ribbon button hidden and shown... not that this was any different in Office 2003. But at least in Office 2003, I could quickly find something within the menu if I couldn't on a ribbon button that was not in plain view. In Off.2007, good luck with that. It's like menus were a forbidden item or something.
I don't know what set off that trend, by I remember it as early as the first Chrome browser and some of the other 3rd part browsers in the late XP era. It's a fad I really wish to never have seen the light of day... but it did.
Menu systems are functional and most importantly, consistent.
Ribbon buttons are quicker *only* once you learn where everything is and *if* their layout doesn't change (which it often does.)
momaka wrote on 2025-03-11, 06:57:
The only thing I haven't tried yet is do both at the same time (OC and UV the V_core), but might do that later when I finish testing the board.
And now, to follow up on that...
I can indeed OC the CPU and under-volt it at the same time. I did reduce the OC a little, though - went with 115 MHz for the bus, mostly because the NB heatsink seems to run really hot with anything over 500 FSB and I didn't feel like modding it with a proper cooling. With 115 MHz bus, the resulting CPU frequency was 2.65 GHz - just a tiny OC over the stock 2.3 GHz (15%). Meanwhile, I lowered the CPU V_core even further, down to 1.325V. Had no problems running any benchmark or game... so, yay!
While at it, I decided to do test a good deal of the AGP cards I have acquired in the last 3-4 months. I tested the following:
* MSI GeForce 4 MX 400 SDR with 128-bit bus and 64 MB (I think?) of 5 or 6 ns RAM
* no-name GeForce FX 5200 with 64-bit bus and 128 MB 6 ns RAM (eww!! 😜 )
* no-name GeForce FX 5200 with 128-bit bus and 128 MB 5 ns RAM
* no-name GeForce 4 MX 440 SDR with 128-bit bus and 64 MB 4 ns RAM
* Sapphire Radeon 9600 Pro Advantage! 128-bit bus 128 MB "rare" 3.6 ns TSSOP RAM (nice!)
Compared to the Radeon 7000 64-bit 32 MB 6ns RAM video card I was using initially to test the system, all of the above were quite a bit faster - even the MX 400 and 64-bit FX 5200. That said, I didn't really test that many games... actually only two:
* Need For Speed High Stakes (full-grid race, 2 laps on Celtic Ruins, daytime weather, all graphics options set to max with D3D as the render, and tested under resolutions of 1024x768 and 800x600, both in 16-bit and 32-bit color modes.)
* Counter-Strike 1.6 (De_Dust 2, with 5 bots, both OpenGL and D3D modes under 1024x768.)
Honestly, though, that was more than enough to show the strengths and weaknesses of the above cards.
Well, the Radeon 9600 Pro clearly had no problems to achieve 60 FPS with any of the above settings and clearly outclassed all of the other cards. The 128-bit FX 5200 also did quite well, but not nearly as much as the MX 440. Now, I know that people here have mentioned it many times that the MX 420, 440, and 460 are all faster cards than the FX 5200. But for some reason, I was always left with the impression from back in the days that the MX cards were slow turds, so I had to check this for myself (perhaps because I probably only witnessed some gimped 64-bit MX with slow RAM??). To my surprise, though, even the MX 400 seemed to do better than the 64-bit FX 5200... well in more of the cases anyways (particularly under 16-bit color mode). With 32-bit color mode, the performance reduced quite a bit and the 64-bit FX 5200 came out faster. But between the MX 440 and the 128-bit FX 5200, I think the MX 440 came out as much faster in just about every scenario. In particular, the 128-bit FX 5200 performed quite well even in 32-bit color mode in NFS HS... but there were a few areas that did make the FPS dip just ever so slightly under 60... whereas with the MX 440, this didn't happen.
So for DX7 titles, I think the MX 440 brings quite a bang for the buck... and even the MX 400 is quite acceptable. That said, I wouldn't quite fully dismiss the FX 5200 cards - even the 64-bit one. After all, the above tests were done on a system with really fast (for the games I tested) CPU. For systems with much slower CPUs (i.e. Pentium II or very early P3 500 MHz or less), I think the difference in FPS could drop quite a bit.. so in such case, even the 64-bit FX 5200 may not be a terrible choice.
FWIW, I used the same (XP) drivers for the MX and FX cards, so overhead should be similar.
Anyways, I did the GPU testing over the course of several days. Also tested some IDE optical drives and floppy drives while at it. This also allows me to say that both IDE ports and floppy port on the motherboard were tested. Oh, and I tested most of the mice I've collected from the dumpster (and cleaned, of course) - a few needed their buttons opened and contacts cleaned, but no biggie. Finally getting through the piles of hardware I've been collecting over the last half a year... which is absolutely necessary, as I have even more stuff waiting to be diagnosed, fixed, recapped, and etc... then tested.
At this rate, I might need another lifetime to do it all. But at least everything tested really has been checked out thoroughly.
