Could you get a front lit picture in higher resolution?

Also, where's the RAM?

There are two scratches, but I checked and the connections are ok. The DRAM chips are fast enough. I also have it on another ET4000. 70ns chips make no difference.

There may also be another problem but I don't think the RAM chips are fast enough.

The ET4000 manual says "the SCLK's [System Clock] cycle time should be equal to the CAS low pulse width and less than 25ns." - Page 24, http://dosdays.co.uk/media/tsenglabs/Tseng_La … roller_1990.pdf

The 44.9MHz crystal gives 22.2ns(?), so the RAM CAS needs to be less than that?

"tc514256p-100" have a CAS access time of 30ns. https://html.alldatasheet.com/html-pdf/146037 … /TC514256P.html

Also, one of those jumpers (JP1?) is IRQ I think and will not make the card work if switched wrong.

Reinserting the chips will help connection as well (RAMDAC, BIOS, etc)

You are right. But the card is even dead with 70ns DRAM chips. The card must do anything? The 100ns is a minimum speed and can definitely be faster. There are the same chips in another card and the card works perfectly at 1024x768 and 256 colors without any problems ...

If you have an mda card & monitor you can add it to the computer with the et4k card and verify if the vga bios is available at the usual location.

You'll likely have to set the bios for mono .

I have only a 486 PCI/ISA Mainboard for testing with DOS 6.22 and Windows.

And a Atari ST TOS 68k Computer with a ET4000 Adapter Card. With the Atari all ET4000 cards work without any BIOS Rom on the graphiccard the card.

The card is on both dead !

Thanks for sharing the additional pictures, that helps a lot 😀

You say it remains dead, does that mean no VGA output or the BIOS gives the "no VGA" beep code error?

It seems to be very tough to find any information on your particular card, the closest I've been able to find is this one:

Not the same card revision but clearly the same PCB layout. Looks like there are some differences between the TC6100AF and TC6058AF chips so your card having both a PLL and a regular crystal should be fine as-is.

#1 - please verify the ICS PLL chip - Is that definitely installed the right way around? More significantly, is it even the right chip?
ICS2494 pinout, should cover the ICS2494N as well:

Looking at the traces to the crystal on the back of your card, I'm not sure the ICS2494 is even the right clock generator - see the traces to the XTAL are in the middle of the pins on your card but are on the edge of the ICS2494 pinout shown in the datasheet. Unless the ICS2494N has a different pinout that's not documented anywhere, this may well be the main problem on your card.
Looking at other cards, this Diamond card has an ICS1394 clock generator - that's got the XTAL traces in the right location:
http://www.amoretro.de/2013/03/diamond-speeds … ax-1mb-isa.html

This looks better:

Can you pull the clock generator and verify which pins on the socket connect to the 5V rail?
It may even be an ICS1494 looking at the big data book - check out the application circuits to see which one matches your board layout (PDF page 20): http://www.bitsavers.org/pdf/integratedCircui … s_Data_Book.pdf
Or maybe even a CH9204...

#2 - Check the BIOS / VESA chip, it should be a 27c256 EPROM and you can compare with other BIOS files here, where there are a few to compare with: Tseng 4000AX Weird Characters
(I'm not totally sure those will work since there may be differences between TC6100AF vs TC6058AF BIOS files, idk)
The one labelled "SYNC" is the translation ROM as far as I can tell and it's not essential for the card to run.

#3 - resolder the QFP pins 😀

#4 - The transistor at Q1 looks kinda twisted, make sure that's not shorting

Thermalwrong wrote on 2023-11-22, 16:44:

You say it remains dead, does that mean no VGA output or the BIOS gives the "no VGA" beep code error?

... no Output the Monitor is black, no Sync. No beep

Thermalwrong wrote on 2023-11-22, 16:44:

#1 - please verify the ICS PLL chip - Is that definitely installed the right way around? More significantly, is it even the righ […]
Show full quote

#1 - please verify the ICS PLL chip - Is that definitely installed the right way around? More significantly, is it even the right chip?
ICS2494 pinout, should cover the ICS2494N as well:
Screenshot 2023-11-22 at 16-18-55 The DatasheetArchive - Datasheet Search Engine - ics2494_ics.pdf.png

Looking at the traces to the crystal on the back of your card, I'm not sure the ICS2494 is even the right clock generator - see the traces to the XTAL are in the middle of the pins on your card but are on the edge of the ICS2494 pinout shown in the datasheet. Unless the ICS2494N has a different pinout that's not documented anywhere, this may well be the main problem on your card.
Looking at other cards, this Diamond card has an ICS1394 clock generator - that's got the XTAL traces in the right location:
http://www.amoretro.de/2013/03/diamond-speeds … ax-1mb-isa.html

This looks better:
Screenshot 2023-11-22 at 16-41-27 ICS1394 datasheet.png

Can you pull the clock generator and verify which pins on the socket connect to the 5V rail?
It may even be an ICS1494 looking at the big data book - check out the application circuits to see which one matches your board layout (PDF page 20): http://www.bitsavers.org/pdf/integratedCircui … s_Data_Book.pdf
Or maybe even a CH9204...

I´m not sure the ICS2494 is the right chip? See my picture ...

Thermalwrong wrote on 2023-11-22, 16:44:

#2 - Check the BIOS / VESA chip, it should be a 27c256 EPROM and you can compare with other BIOS files here, where there are a few to compare with: Tseng 4000AX Weird Characters
(I'm not totally sure those will work since there may be differences between TC6100AF vs TC6058AF BIOS files, idk)
The one labelled "SYNC" is the translation ROM as far as I can tell and it's not essential for the card to run.

i will check when I have more time ...

Thermalwrong wrote on 2023-11-22, 16:44:

#3 - resolder the QFP pins 😀

#4 - The transistor at Q1 looks kinda twisted, make sure that's not shorting

The Transistor is ok

That's good, that means the VGA BIOS section is probably working and you don't need to worry about the LS244 / LS245 chips that connect up the ROMs 😀

Ah, I thought that might be the VCC trace. That means we can rule out the ICS1394 / ICS 1494 then.
I wonder what clock generator could have the VCC at pin 10??

Can you pull the clock generator and look inside the socket to see if there are any markings?

Thermalwrong wrote on 2023-11-22, 18:20:

That's good, that means the VGA BIOS section is probably working and you don't need to worry about the LS244 / LS245 chips that […]
Show full quote

That's good, that means the VGA BIOS section is probably working and you don't need to worry about the LS244 / LS245 chips that connect up the ROMs 😀

Ah, I thought that might be the VCC trace. That means we can rule out the ICS1394 / ICS 1494 then.
I wonder what clock generator could have the VCC at pin 10??

Can you pull the clock generator and look inside the socket to see if there are any markings?

Just a thought, Does the computer boot with the card installed?
edit: also it seems like your idea of the ICS1394 still seems to fit best, since the datasheet mentions that CPSEL pin 10 is be pulled up to VDD if left open. so seems like it's fair to just hook it up directly to +5V . and the rest of the pins also seem to match up.

are there any other +5v connections on the board to the ics chip ?

Pin1 is marked on the soldering side of the PCB with a square via.

Pin1 = GND
Pin2
Pin3 = over 10R (R38) to VCC
Pin4
Pin5 = Quartz 14,xxx Mhz
Pin6 = Quartz 14,xxx Mhz
Pin7 = Quartz oscillator 44,xxx Mhz
Pin8 = GND
Pin9
Pin10 = VCC

Pin11
Pin12
Pin13
Pin14
Pin15
Pin16
Pin17 = GND
Pin18
Pin19
Pin20

VCC = +5 Volt

I buy a ICS1394 from Ebay and test it ...

Yep, that's worth a go 😀
To be sure, you could measure the resistance between pin 10 and pin 3, if that's 0 ohms then the two pins are probably connected underneath the socket. Hopefully that's the case.

Frank, I guess you mean R30, since R38 is in the unpopulated section of the board. 10ohm is low enough not to pose much of a voltage supply problem.
Thermalwrong, it's more likely that he'll measure 10 ohm between pin 10 & pin 3 😀

Here's to hoping the ICS1394 does the trick.

The card now works with the ICS1394...

Windows 95 works with the standard Windows driver at 640x480 at 16 colors. If the Windows 95 TSENG ET4000 driver is installed, nothing works anymore, i.e. the screen goes black and stays black.

The problem is that the card doesn't have a name and I can't search for the right drivers for Windows?

DOS SYSCHK reports the memory access is 16bit and the I/O access is 8bit. On the Atari, the graphics card adapters are 16bit except for one that can also use 8bit for I/O. The card works with 1024x768 and 256 colors.

Another question would be what kind of RamDAC is originally on the card? An 8bit, 15/16bit or even 24bit RamDAC ...

Frank wrote on 2023-11-25, 16:57:

Another question would be what kind of RamDAC is originally on the card? An 8bit, 15/16bit or even 24bit RamDAC ...

16 bit, better said, 5 R, 6G, 5B, according to the datasheet for the SC11487. HiColor 65k color support.

The card works better with a MUSIC TR9C1710 DAC but on the Atari only at 1024x768 and 640x480 under 2/16/256 Colors. 800x600 in any colors give a black screen.

I check it not with Windows ...

Frank wrote on 2023-11-25, 16:57:

The problem is that the card doesn't have a name and I can't search for the right drivers for Windows?

You usually don't need a card specific driver for Windows for an ET4000AX card. The different cards differ only in mode initialization (programming a high-color DAC to the correct color mode, selecting the correct clock from a clock chip) and whether a translation ROM is present for hardware-assisted CGA/MDA/EGA emulation. The Windows driver delegates mode setup to the BIOS (and there is one set of standard mode numbers for ET4000 BIOSes), and the Windows driver expects the card to not be in any kind of backwards compatibility mode (I don't know for sure, but it's likely that the ET4000 BIOS cancels any backward compatibility mode as soon as an SVGA mode is initialized). In essence, this means: Not knowing the card's name might make it difficult to find a suitable BIOS, but is likely irrelevant for selecting a Windows driver.

While I wouldn't be surprised if a big OEM (like Compaq, HP, Dell) had custom BIOSes shipped with their systems integrating an ET4000AX assigning "conventional OEM mode numbers" instead of ET4000 mode numbers to these modes, this will most likely apply to on-board ET4000 chips. Non-standard ROM interfaces are extremely unlikely to be present on No-Name off-the-shelf ISA cards.

Congratulations on getting the card running again 😀
The RAMDAC issue you're describing with Win95 sounds like when I try to run 16-bit modes on an 8-bit DAC that doesn't have any such mode. But surely 256 colours should still work, maybe the DAC to use depends on the BIOS setup?

See the picture I posted above of the similar card and that uses a regular 8-bit RAMDAC the KDA0476CN-80. Perhaps try a basic DAC to see if it likes working with that.