carlostex wrote:

@Skyscraper

There's another solution to your JUKO Board. Not a cheap solution but it can get you upper memory blocks as well. It's the lo tech 1MB RAM Board. I have one in my XT and if i set the memory to none with the dip switches the Board takes over and provides 1MB RAM in 64kb blocks. If you Build it yourself its probably cheaper depending how much farnell/rs/mouser will charge you for parts shipping.

In time I will probably buy/build (more likely buy) such a card(s) anyhow for the working XT Clone(s) (they have 640KB memory) and then I can test if it will revive the Juko Nest board.

Perhaps I can try the Juko nest BIOS in one of the working XT clone systems just to see if it's working and to be able to dump the BIOS ROM.

After I have played some Paku Paku I will test the V20 CPU from the Juko Nest board as a first step. 😀

I dug through my computer gear, sorting out cables and listing my "spares":

12 motherboards
38 graphics card
22 sound cards
6 NICs

And that's after I culled what I didn't "need".

I may just start to chuck them out, it's such as hassle selling what with testing, photographing, listing, getting postage quotes, packaging, posting etc etc.

ratfink wrote:

I may just start to chuck them out, it's such as hassle selling what with testing, photographing, listing, getting postage quotes, packaging, posting etc etc.

Er... not before you've catalogued them in the Vogons Wiki you don't! 😉

The V20 CPU from the Juko Nest baby XT motherboard did at least work! 😀

Removing a 8088 CPU when half of it is placed under the PSU with little clearance and then mounting a V20 without as much as disconnecting a single cable requires some gynecological skills. 😁

The CPU performance gain in Landmark is only 10% but the video performance gain is 30%! 😀

Note that the turbo mode video result with the V20 can't be compared with the turbo mode result with the 8088 I posted on the last page of this thread as I did not rerun the video test in turbo mode with the 8088. I fully expect the V20 would be the same 30% faster when it comes to video performance in turbo mode as it is in non turbo mode though.

Non turbo.

Turbo.

brostenen wrote:

FaSMaN wrote:

booting memtest+ it just instantly fails ALL tests, red lines everywhere.... Bad Co-processor , or bad Cache? no idea, was just interesting.

It's bad mem modules. Memtest+ only tests memory.

Not this time, when I swap the CPU out with one of my other ones it works fines, when I swap the ram out with the bad CPU it still reds out

Benchmarked a 3dfx Velocity 100 and compared it against the V3. Interesting card, it has 1 TMU disabled for Glide and OpenGL.

The latest V3 driver doesn't disable it though, so I manually disabled in through a registry entry to see what the performance is like.

In Quake II and Quake III the card takes a hit. In D3D the card is actually a tiny faster than a V3 3500 Compaq clocked at the same 143 MHz. The V3 3500 Compaq uses SDRAM, the Velocity 100 I have SGRAM. It might have righter memory timings also.

It also has only 8 MB of memory. Not a big deal, out of the things I tried only 3DMark2001 SE complained about that.

Investigated parts for an electronics project. A pretty small, cheap (primarily because I have stuff lying around already) and simple project that will come to light in the future if it works out.

Otherwise, I've been looking at old maps of my city on the internet as far back as the 1800s. More interesting are the maps from the 1920s to the 1940s where you can see my current neighborhood growing. Somewhere in the 1930s my house was built (I actually know for a fact that this particular one was built between 1936/7. The oldest on the estate are from the 1920s towards the western edge though the "Boot" houses were demolished in the early 2000s) and the next block along, the one they just knocked down, was the last block before the countryside started. There would literally have been nothing but fields there until some time in the early 50s when (I seem to think 1953 though the maps are too far apart to confirm this) Greatfield Estate started to appear. I knew some of this already, but it is interesting to see it happen over time and also fascinating to see what was there before, primarily farms and industry - again, I knew this, but had never seen it on a map as local museums like to only put up grainy pictures of medieval maps or similar.

Also learned my father's old house is significantly newer than I had always been told - though that fits, as I always felt the construction was wrong for the 1940s and it is indeed a 50s construction - also the house I am going to view soon is older than I thought. Whilst the designs appear the same, it is on the right side of the street so that it would be at least ten years older than the houses I have been in on that road. In fact, it would have been one of the last houses before another field for a number of years. This makes me wonder if I will find anything different hiding inside, perhaps the layout differs a little or the materials used were slightly different. The Amethyst house I handed back some time ago checks out with what I had already learned in that it was put up some time between 1953 and 1974, so most likely the 1960s as suggested. This also ties in with my father's old house as they were designed by the same person but the Amethyst house used lesser materials such that it definitely implies a gap of around ten years between those two. It also seems Murder Avenue did not exist yet which explains the inconsistencies in the post code around there and also explains why the block of shops there do not match the rest of the estate.

Also nice to see my suspicions confirmed that street names, for example, "Swanfield Road" were inherited from the farms which used to be there. In the case of that one the older maps simply show "Swanfield Farm" where today there is a road. Also nice to see how much railway was once there, I'd heard of it, read about it and even seen it in photos, but seeing the complex mass of lines on a map puts it into perspective. Those generally connected to the docks (this was a city built on fishing) and the city was for a long time mostly comprised of docks. You can still see obvious signs of some of the ones which no longer exist in the city today, I had always known Queens Gardens was formerly Queens Dock and the layout is consistent with the placement of the current roads and walkways.

And to think, back then most of my family were still living in little wooden caravans and on one side, weren't even living in the city, instead living very far to the North.

Finally after.. what 4 days? I can't even remember, lost track now. Got my AMD K7 up and running in XP side again and able to run games and everything.

I decided instead of running single-hard drive.. I would just switch it over to IDE Raid-0 since the onboard motherboard supports it and all.

Sitting here watching it run Final Reality Benchmark now, about to try and load up Diablo2 on it, and a very old pre-release leaked beta of Half-Life 2 I found last night and see how it does with that.

Up next tonight is trying to figure out how to make the Voodoo2 work in WindowsXP on this machine.

Threw my 8800GTX into a system to test it today. Amazingly it works with very little artifacts (at least at idle). But when i fire up Unreal Tournament 3 on it it hits (according to openhardwaremonitor, and it seems fairly correct) 125c without throttling. No wonder these things die so much.

So after yesterday's failed 233 P2 I wanted to do something that works so ended up grabbing some of my old parts and a case I was going to bin , and built a P1 166mhz rig for Dos and windows 95.

About 2 years ago I found a old AT P1-133mhz at the bottom of a freezer soaked in water , I purchased it for parts (SB16 , Ram, CPU) and was going to bin the board but decided to give it a second lease of life , so I soaked it in PCB cleaner and contact cleaner , and left it to dry (till I forgot about it and dug it out today) 2 years should be enough time to dry a Motherboard...

Anyhow hats off to gigabyte this thing just works, and wants to keep on working no matter what 🤣

OS: Dos 6.22+Win 3.11 and Windows 95 (plop)
Motherboard: GA-568ATV
CPU: Intel pentium 1 166mhz MMMX
Ram: 96MB mixed
Video: S3 ViRGE DX (need to find some more video ram for it)
Network: Realtech 8029
Sound: Sound Blaster 16Pnp
HDD: Seagate 15gig

To-do: Find and add a Voodoo1

Anyhow I am surprised that this Gigabyte board not only survived but also gigabyte kept updating the BIOS right up to 2003 🤣 , the only thing that doesn't want to work is the onboard Comports , bot the end of the world tho I threw in a XT comcard.

Benchmarking a GeForce256 and comparing performance with and without TnL on a faster machine.

Took some pictures of my Pentium 3 build, wrote a thread about it. Making some adjustments to my Pentium 233 MMX build.

I'm continuing to investigate what VGA cards my XT clone likes.

OTI-037C is another VGA chipset that seems to work in an XT, at least my 16bit ISA card worked at once. The XT do have a V20 CPU at the moment but the Trident 9000C set to 8bit BIOS did still not work so I'm not really sure if it matters.

h-a-l-9000 wrote:

Skyscraper wrote:

If anyone can think of any other reason for a card hindering a PSU that is powered on with a f*****g live mains switch from powering on other than beeing one big short please let me know. 😀

A tantalum cap may have shorted. Did you check the solder side for shorting pins? You can also remove many chips to exclude from being the culprit. Myself I would force some current from a regulated power supply in there and check which part gets hot.

I totally missed your post yesterday.

I did check for shorting pins on the back of the card and from what I can see there are none but I will investigate the card starting with removing all removable chips and also replacing the transistor that perhaps was shorted. The transistor only costs 0.2 euro or so but I need to order it at the same time as other stuff otherwise the shipping cost isn't reasonable. I have a half dead card with the same model of transistor and if i can't get it working properly I might use the transistor from that card. I would need to extend the middle leg though as I think it's too short to reach where it needs to go.

I have at least decided that I won't send the card back as it's little money and the card is too rare.

Not much time for retro activity yesterday--but I had bit of time in the morning to contribute to the Vogons wiki. Today will be another busy day, might have some time later this afternoon to try and finish up System Shock 1.

To sum up my findings when it comes to XT class systems and 16bit ISA VGA cards

16bit ISA cards that works in my XT clone.

All cards tested with Landmark to see their relative performance.

Trident TVGA 8900C <---- Sharded second fastest. Only monochrome picture in my system with a NEC V20, see below for fix
Trident TVGA 9000B <---- Middle of the pack. Only monochrome picture in my system with a NEC V20, see below for fix
Realtek RTG3106 <---- Shared second fastest.
OAK OTI-037C (I have tried three cards, all works) <---- These are rather slow.
Genoa Super VGA model 6400 GN007001 B <---- Also slow.
Ahead v5000 AVGA/Deluxe <---- Middle of the pack.
TSVGA ET4000AX <---- The fastest card, at least when it comes to Landmark.
Avance Logic ALVGA GUI1 003 C ALG2101 <---- Middle of the pack.

Member h-a-l-9000 has made a program that fixes the issue with no colour with Trident TVGA cards and other cards by switching the video BIOS to colour mode.

h-a-l-9000s program "coloron"

16bit ISA cards not working in my XT clone.

Trident TVGA 9000C, (I tried changing all jumpers using TH99 as reference but no luck, the card could be broken as I do not know if I have tested it before)
Orchid Fahrenheit 1280 (I did not try to mess with jumpers as I couldn't find a jumper guide)
Hercules Dynamite Tseng ET4000/w32i (I tried two cards, none of them worked but I did not try to mess with jumpers as I couldn't find a jumper guide)

This was all ISA video cards I could find, the rest are hidden away in systems. I might add a few more cards later. 😀

Please note that just because these cards worked in my XT clone there is no guarantee they will work in your XT class system, that said it's probably a good indication.

> but it only outputs monochrome
Investigating old ISA VGA cards

h-a-l-9000 wrote:

> but it only outputs monochrome
Investigating old ISA VGA cards

None of these two Trident cards have the LM339 chip and I'm using a "modern" 17" Viewsonic TFT.

I'm a bit slow tonight but I guess that indicates something else is the issue?

I could still test with an old 14" VGA screen later just to make sure it isnt some monitor detection issue. The Trident cards have a jumper for "standard" and "non standard" colour but I have tried with that jumper in both positions.

This is not really a critical issue for me as most other ISA VGA cards works perfectly fine in my XT system with colour and all but others might benefit from knowing. 😀

From the other thread.

h-a-l-9000
//

There are two different schemes for monitor detection, the more common and more complicated one of detecting the presence of a connection on the analog lines, and the other with the ID0/1 pins.

The first one is implemented on IBM VGA and many VGAs of other manufacturers, the second originated from IBM XGA and I've found it on ET3000/4000 boards and some 8-bit VGA I don't remeber the brand right now.

You can identify old cards by looking for a LM339 chip. If one is present, the card uses the first method. Newer ones, such as S3 Trio, have it integrated. You can check these by using a monitor and cable with separate BNC connectors: unplug red and blue, switch on the computer and let it POST, then plug them back in. If its in monochrome mode now, it used the analog line detection scheme.

\\

elianda
//

Early VGA standard defined a monitor detection with some pins on the vga connector, specifically Color / Monochrome.
Later the same pins were reused as DDC communication pins.

So if you have a DDC capable monitor and an older VGA card, the card detects 'something'. You can correct the behaviour if you cut it and configure the pin connections on the cards end to color monitor encoding.
(and leave the monitor ends DDC pins unconnected)

\\

The Trident chips could already have the LM339 integrated. But even if not the tool may be able to get back the color. It switches the video BIOS to color monitor.

If it does help, one can add it to autoexec.bat.

h-a-l-9000 wrote:

> but it only outputs monochrome
Investigating old ISA VGA cards

OK I will test the tool! 😀

I'm a bit tierd so I only read the later posts in that thread, I missed both the tool and that it worked for another Trident TVGA8900 card.