VOGONS


Reply 60 of 98, by Nexxen

User metadata
Rank l33t
Rank
l33t

After desoldering the Vram and cleaning the holes, soldering the sockets and testing with other Vram chips:

The attachment IMG_20261029_004030105_HDR.jpg is no longer available

I still have to fine tune the Mclock and Videoclock but the outcome of a full recovery is quite likely.
Hitting F1 or Del still yields out of range.

Well, looks like helping goes a long way.

PC#1 Pentium 233 MMX - 98SE
PC#2 PIII-1Ghz - 98SE/W2K

- "One hates the specialty unobtainium parts, the other laughs in greed listing them under a ridiculous price" - kotel studios
- Bare metal ist krieg.

Reply 61 of 98, by mogwaay

User metadata
Rank Newbie
Rank
Newbie

Sorry not read the whole thread here, so sorry if this isn't relevant, but I made my own ICS2494 clock generator chip for my ET4000AX card as it wasn't working. Here link to the board files and code if it's helpful:

https://github.com/moogway82/ics2494-replacement

It worked a treat for me. You might need a different system clock speed and reprogram the clock generator for your specific chip, the ICS2494 datasheet has the clocks for specific chip sets.

Reply 62 of 98, by Nexxen

User metadata
Rank l33t
Rank
l33t
mogwaay wrote on 2026-10-28, 23:57:

Sorry not read the whole thread here, so sorry if this isn't relevant, but I made my own ICS2494 clock generator chip for my ET4000AX card as it wasn't working. Here link to the board files and code if it's helpful:

https://github.com/moogway82/ics2494-replacement

It worked a treat for me. You might need a different system clock speed and reprogram the clock generator for your specific chip, the ICS2494 datasheet has the clocks for specific chip sets.

Yes, your project was linked in another message some pages ago.
It is in fact something I'd like to get 😀

Btw, awesome project.
I still have to understand what I need to download to print the PCB and the parts to populate it.

Last edited by Nexxen on 2026-10-29, 00:10. Edited 1 time in total.

PC#1 Pentium 233 MMX - 98SE
PC#2 PIII-1Ghz - 98SE/W2K

- "One hates the specialty unobtainium parts, the other laughs in greed listing them under a ridiculous price" - kotel studios
- Bare metal ist krieg.

Reply 63 of 98, by Nexxen

User metadata
Rank l33t
Rank
l33t

Last pictures

PC#1 Pentium 233 MMX - 98SE
PC#2 PIII-1Ghz - 98SE/W2K

- "One hates the specialty unobtainium parts, the other laughs in greed listing them under a ridiculous price" - kotel studios
- Bare metal ist krieg.

Reply 64 of 98, by Nexxen

User metadata
Rank l33t
Rank
l33t

Nope, with my main screen I get:

The attachment IMG_20261029_012523407_HDR.jpg is no longer available
The attachment IMG_20261029_012352215_HDR.jpg is no longer available

The problem is my 4:3 screen with this card.
I tried to modify a setting, Clock on the menu and won't display again.
All other cards do though... F@@@@@@@@@@@@@@ck... 🤣

PC#1 Pentium 233 MMX - 98SE
PC#2 PIII-1Ghz - 98SE/W2K

- "One hates the specialty unobtainium parts, the other laughs in greed listing them under a ridiculous price" - kotel studios
- Bare metal ist krieg.

Reply 65 of 98, by Nexxen

User metadata
Rank l33t
Rank
l33t

H. freq (kHz) Resolution V. freq (Hz)
31.5 640*350 70
31.5 720*400 70
31.5 640*480 60
35.0 640*480 67
37.5 640*480 75
35.2 800*600 56
37.9 800*600 60
46.9 800*600 75
49.7 832*624 75
48.4 1024*768 60
60.0 1024*768 75
69.0 1152*870 75
71.8 1152*900 76
63.9 1280*1024 60
80.0 1280*1024 75

SCANNING
Vertical refresh rate 56 Hz-76 Hz
Horizontal Frequency 30k Hz-83 kHz
VIDEO
• Video dot rate 140 MHz
• Input impedance
- Video
75 ohm
- Sync
2.2K ohm
• Input signal levels 700m Vpp
• Sync input signal
Separate sync

Composite sync
Sync on green

• Sync polarities Positive and negative
• Video interface D-Sub (analog)

Last edited by Nexxen on 2026-10-29, 01:23. Edited 1 time in total.

PC#1 Pentium 233 MMX - 98SE
PC#2 PIII-1Ghz - 98SE/W2K

- "One hates the specialty unobtainium parts, the other laughs in greed listing them under a ridiculous price" - kotel studios
- Bare metal ist krieg.

Reply 67 of 98, by wierd_w

User metadata
Rank l33t
Rank
l33t
Nexxen wrote on 2026-10-28, 23:44:
wierd_w wrote on 2026-10-28, 20:28:

The trimming cap is the expensive bit.

Otherwise its approx 100nH inductor + 247pF ceramic cap, for ~32mhz.

What would I need beside the above to create this?

some transistors, and or an opamp chip.

The theory and principle can be found here:
https://www.eleccircuit.com/simple-transistor … tts-oscillator/

The basic components of the lc resonator (the inductor and the cap), will "bounce" a stored charge on the capacitor, converting it back and forth between charge and EM field in the inductor, at a frequency paired with the values of these components. Because these have resistance, the signal quickly dies out.

However, you can use a set of suitable transistors and an op-amp to supply more input voltage, and collect the sinusoid output of the tank so that it does not die out. Some example circuits are provided in the link. Just substitute suitable components for your 32mhz signal, where appropriate.

Logic level devices will want square waves, not sinusoids though. So, using 1 more transistor on the output sinusoid across the gate, with a voltage source, will turn it into a square wave. (the peaks of the sinusoid are rectified, then fed to the gate of a transistor. When the peak is present, the transistor conducts. When it is not, it does not. This will give you an amplified square wave, derived from the tuned sinusoid. The transistor needs to have a switching frequency greater than the desired output frequency.)

You can create a circuit that you can manually adjust the output signal from this way, from some discrete parts.
It will draw more power though. The idea here is to find out what the frequencies you actually need are, and then get a suitable crystal, not to substitute the part. (but, I suppose if you dont care about it being bulky, ugly, and power hungry, you could.)

More or less, "Here is a way to make a tunable clock source, for diagnostic purposes."

Reply 68 of 98, by Nexxen

User metadata
Rank l33t
Rank
l33t

Tomorrow I'll open my 19" LCD to check for dead components, I wouldn't be surprised if some capacitor were dead.
It works with all the other cards, even PCIe... go figure...

The card is really basic.

PC#1 Pentium 233 MMX - 98SE
PC#2 PIII-1Ghz - 98SE/W2K

- "One hates the specialty unobtainium parts, the other laughs in greed listing them under a ridiculous price" - kotel studios
- Bare metal ist krieg.

Reply 69 of 98, by Nexxen

User metadata
Rank l33t
Rank
l33t

Some benchmarks.

Are aaa1m304p-07 chips good in this card?

Image has those ugly columns.

Thanks to everybody so far.

PC#1 Pentium 233 MMX - 98SE
PC#2 PIII-1Ghz - 98SE/W2K

- "One hates the specialty unobtainium parts, the other laughs in greed listing them under a ridiculous price" - kotel studios
- Bare metal ist krieg.

Reply 70 of 98, by rasz_pl

User metadata
Rank l33t
Rank
l33t

imo columns are DAC iref/vref pumping due to insufficient number of capacitors on board, basically every time vga chip fetches data (again afaik it fetches in batches) voltage drops slightly
that last test - 974 KB/s 😮 try fitting/removing JP6, should be waitstates

https://github.com/raszpl/sigrok-disk FM/MFM/RLL decoder
https://github.com/raszpl/FIC-486-GAC-2-Cache-Module (AT&T Globalyst)
https://github.com/raszpl/386RC-16 ram board
https://github.com/raszpl/Zenith_ZBIOS Zenith Z-386 MFM-300 ZBIOS disassembly

Reply 71 of 98, by analog_programmer

User metadata
Rank Oldbie
Rank
Oldbie

Good job, Nexxen! But where is your UM9502 chip? Is it burnt?

rasz_pl wrote on 2026-10-28, 18:33:
analog_programmer wrote on 2026-10-28, 08:37:

I have a question for you, since you're much more into electronic spare components/parts, than I am. Are there any suitable 512 x8 memory chips (SOJ or DIP), that could be routed to these four 4-bit data lines (with some kind of DIY PCB adapter), so we could get 1 MB of RAM in total for this card?

I have no more clue than what mkarcher already wrote in Video RAM chips for a UMC TK-85C408VGA ISA Video Card?

Sorry, I know this thread, but we discussed only 4-bit RAM chips (and defective 4-bit chips used as 3-bit). I'm not sure if it will be possible the two pairs of 4-bit data lines to be combined on this particular card, so two 8-bit RAM chips with bigger capacity to be used (2 x 512x8) with some kind of DIY PCB adapter plugged into four of the card's VRAM sockets, but basically this is my idea for the 1 MB VRAM upgrade. Just found, that there are FPM 512x8 chips like Samsung KM48C512D.

The word Idiot refers to a person with many ideas, especially stupid and harmful ideas.
This world goes south since everything's run by financiers and economists.
This isn't voice chat, yet some people overusing online communications talk and hear voices.

Reply 72 of 98, by Nexxen

User metadata
Rank l33t
Rank
l33t

I tried the 9504 and the card works. That is not preventing anything to start.
What I believe is that the 14.318MHz crystal is dead, even the one I replaced it with. Also, usually you have 2 x pF caps connected to ground, here they are missing. Circuit isn't that good.
Makes me wonder if they are inside the 9504... or simply not because it's a cheap card and they cheaped the hell out of it.

I also have to add an external working 14.318 to this 9504 to see if it wasn't because of it.
I doubt it but I'd check it too.

I made a mistake, I could have tested the oringinal Vram chips to find the dead one. It skipped my mind until I woke up. 🤣
First thing after "coffeeeeeee" thought arose.

PC#1 Pentium 233 MMX - 98SE
PC#2 PIII-1Ghz - 98SE/W2K

- "One hates the specialty unobtainium parts, the other laughs in greed listing them under a ridiculous price" - kotel studios
- Bare metal ist krieg.

Reply 73 of 98, by Nexxen

User metadata
Rank l33t
Rank
l33t
rasz_pl wrote on Yesterday, 03:44:

imo columns are DAC iref/vref pumping due to insufficient number of capacitors on board, basically every time vga chip fetches data (again afaik it fetches in batches) voltage drops slightly
that last test - 974 KB/s 😮 try fitting/removing JP6, should be waitstates

Design is cheap.
JP6 is IRQ2 setting. To me it is now a test card for dead boards. An expendable 😀

PC#1 Pentium 233 MMX - 98SE
PC#2 PIII-1Ghz - 98SE/W2K

- "One hates the specialty unobtainium parts, the other laughs in greed listing them under a ridiculous price" - kotel studios
- Bare metal ist krieg.

Reply 74 of 98, by analog_programmer

User metadata
Rank Oldbie
Rank
Oldbie
Nexxen wrote on Yesterday, 14:03:
I tried the 9504 and the card works. That is not preventing anything to start. What I believe is that the 14.318MHz crystal is d […]
Show full quote

I tried the 9504 and the card works. That is not preventing anything to start.
What I believe is that the 14.318MHz crystal is dead, even the one I replaced it with. Also, usually you have 2 x pF caps connected to ground, here they are missing. Circuit isn't that good.
Makes me wonder if they are inside the 9504... or simply not because it's a cheap card and they cheaped the hell out of it.

I also have to add an external working 14.318 to this 9504 to see if it wasn't because of it.
I doubt it but I'd check it too.

I made a mistake, I could have tested the oringinal Vram chips to find the dead one. It skipped my mind until I woke up. 🤣
First thing after "coffeeeeeee" thought arose.

There are no holes or pads for capacitors near crystal resonator on my card either. I doubt those are included inside the UM9502 circuit. I often see old motherboards with no decoupling(?) capacitors connected to crystal resonators, so probably they are not so important.

If your UM9502 chip is still in good working condition, just try the card with brand new or known good 14.31818 MHz crystal.

This card could be not so bad, if it had no VESA BIOS extensions implemented in its BIOS. Unfortunately it doesn't and I'm not skilled enough to inject such things into videocard BIOSes.

The word Idiot refers to a person with many ideas, especially stupid and harmful ideas.
This world goes south since everything's run by financiers and economists.
This isn't voice chat, yet some people overusing online communications talk and hear voices.

Reply 75 of 98, by wierd_w

User metadata
Rank l33t
Rank
l33t

How large is the UMC TSR?

There's that empty padding, and the mystery optionrom in the existing dump.

It might be possible to 'massage' the tsr program to fit in the padding, with checksum and header?

Reply 76 of 98, by analog_programmer

User metadata
Rank Oldbie
Rank
Oldbie
wierd_w wrote on Yesterday, 14:44:

How large is the UMC TSR?

There's that empty padding, and the mystery optionrom in the existing dump.

It might be possible to 'massage' the tsr program to fit in the padding, with checksum and header?

A couple of kilobytes - see the UTIL directory in the attached archive. If I'm not mistaken, the VESA driver/TSR was named "VESA.COM":

The attachment UMC-VGA.zip is no longer available

P.S. But there are also this lines in the READ.ME file:

RAMBIOS device driver: […]
Show full quote

RAMBIOS device driver:

- RAMBIOS SYS ...a program to speed up video read and write operation.

Put DEVICE=RAMBIOS.SYS as the first device driver command in CONFIG.SYS
file. It should be installed before any other device driver in order to be
compatible with most application programs.

- RAMBIOS EXE ...a program for AUTOEXEC.BAT or under dos prompt.

The word Idiot refers to a person with many ideas, especially stupid and harmful ideas.
This world goes south since everything's run by financiers and economists.
This isn't voice chat, yet some people overusing online communications talk and hear voices.

Reply 77 of 98, by wierd_w

User metadata
Rank l33t
Rank
l33t

COM files are "more or less" capable of being homed in a rom, with little effort.

It might be massagable with something like exe2rom
https://exe2rom.sourceforge.net/

And, failing that, Com2Rom, with some UMB available, might also be a possible thing to wrap it with.
Re: CGA game on option ROM

Then the TSR would get triggered by the BIOS optionrom scan, hook the needed vectors, and that would be that.

Reply 78 of 98, by analog_programmer

User metadata
Rank Oldbie
Rank
Oldbie

I want to extract only the usable VBE code from the .com-file, but I'm not good enough to read and analyze the entire disassembled code. And I don't know where/when these VBE functions should be called inside the BIOS.

The word Idiot refers to a person with many ideas, especially stupid and harmful ideas.
This world goes south since everything's run by financiers and economists.
This isn't voice chat, yet some people overusing online communications talk and hear voices.

Reply 79 of 98, by wierd_w

User metadata
Rank l33t
Rank
l33t

VBE is just another hook in the vector table, I thought? A program wanting to use VBE, checks the vector table to see if there is an entry, then uses the normal bios extension calling convention to pass instructions to that handler.

From perusing the spec document ( https://pdos.csail.mit.edu/6.828/2014/reading … rdware/vbe3.pdf ), this appears to be an expanded version of int10h.

VBE Function Reference This chapter describes in detail each of the functions defined by the VBE standard. VBE functions are cal […]
Show full quote

VBE Function Reference
This chapter describes in detail each of the functions defined by the VBE standard. VBE
functions are called from real mode using the INT 10h software interrupt vector, passing
arguments in the 80X86 registers. The INT 10h interrupt handler first determines if a VBE
function has been requested, and if so, processes that request. Otherwise control is passed to the
standard VGA BIOS for completion. Starting with VBE 3.0, VBE functions can also be called
directly from protected mode via the protected mode entry point. When called via the protected
mode entry point, the VBE functions execute as 16-bit protected mode code which can be called
directly from any 16-bit or 32-bit protected mode operating system or application.
All VBE functions are called with the AH register set to 4Fh to distinguish them from the
standard VGA BIOS functions. The AL register is used to indicate which VBE function is to be
performed. For supplemental or extended functionality the BL register is used when appropriate
to indicate a specific subfunction.
Functions 00h-0Fh have been reserved for Standard VBE function numbers; Functions 10h-FFh
are reserved for VBE Supplemental Specifications.

So, letting the normal vga bios run, followed by the opt-rom-ified TSR, would have the TSR code take over int10h, and provide VBE services.

Just enapsulating the tsr, and getting it to run as an injected optrom should be all that's needed?