mkarcher wrote on 2026-01-24, 20:01:

butjer1010 wrote on 2026-01-24, 19:27:

Those 0.33uf i bought for Amiga RAMs aren't ok?

They are a perfect electronic fit (0.33 or 0.1µF doesn't matter), and it seems they are also a perfect mechanical fit. As long as you don't intend to modify the card to use 32k*4 memory chips you can not find anyway, there is no need to populate U18..U25 and the respective caps, though.

Thanks a lot, will change those caps for new one, just in case they are the problem. In Monday i will order those 2 LS's i don't have (32 and 153), so i can change them also.

I have tried to change RAM chips for "quicker" ones (see picture), still freezes in POP, but work in CK - of course, with screen tearing. No change of behavior with both RAM chips.

butjer1010 wrote on 2026-01-24, 20:04:

Thanks a lot, will change those caps for new one, just in case they are the problem. In Monday i will order those 2 LS's i don't have (32 and 153), so i can change them also.

Just to make sure we didn't misunderstand each other: Your Amiga caps are a perfect replacement for the small capacitors in the glass tube ("like a diode" as you said), even though they are 0.33µF instead of 0.1µF. They are not a perfect replacement for that orange 4.7µF/16V tantalum capacitor next to the ISA connector and the RCA socket.

mkarcher wrote on 2026-01-24, 20:25:

butjer1010 wrote on 2026-01-24, 20:04:

Thanks a lot, will change those caps for new one, just in case they are the problem. In Monday i will order those 2 LS's i don't have (32 and 153), so i can change them also.

Just to make sure we didn't misunderstand each other: Your Amiga caps are a perfect replacement for the small capacitors in the glass tube ("like a diode" as you said), even though they are 0.33µF instead of 0.1µF. They are not a perfect replacement for that orange 4.7µF/16V tantalum capacitor next to the ISA connector and the RCA socket.

Hehe, yep, i figured this out. Need to order this one too (4.7).

The caps this card has remind me of military grade ones, basically like SMD ceramics but hermetically sealed in glass. Resistant to some corrosive factors and less prone to mechanical damage and stress/cracking during soldering. I doubt these are bad but if replacing the LS chips doesn't help then the caps are not that much more work.

If another set of RAM chips didn't cure the problem then I would start suspecting the PCB and ASICs. Is any of these getting hot with the game running? Sorry if I missed it, did you pull them out of the sockets at least once? These sockets very rarely fail but if there was any oxidation then pulling and re-inserting the chips will cure that by scraping the pins and contacts a little bit.

Deunan wrote on 2026-01-24, 23:47:

The caps this card has remind me of military grade ones, basically like SMD ceramics but hermetically sealed in glass. Resistant to some corrosive factors and less prone to mechanical damage and stress/cracking during soldering. I doubt these are bad but if replacing the LS chips doesn't help then the caps are not that much more work.

If another set of RAM chips didn't cure the problem then I would start suspecting the PCB and ASICs. Is any of these getting hot with the game running? Sorry if I missed it, did you pull them out of the sockets at least once? These sockets very rarely fail but if there was any oxidation then pulling and re-inserting the chips will cure that by scraping the pins and contacts a little bit.

No hot chips, even if i play for 15-20 minutes. Yes, i did pull them out ( You can see the marks on bigger one which i did 😀 ), clean them and sockets, and insert them back to sockets.

I don't know what is the problem with Prince of Persia? I played Monkey Island for at least half an hour, no problem at all.... Will try to play some other games to see if the problem is only on POP.

Given the tearing in Keen, I would suspect an issue with the status register or something that PoP chokes on but not Keen. There’s a fallback check in Keen if it waits too long for a vblank that will bail and keep the game running instead of halting.

keenmaster486 wrote on 2026-01-25, 19:48:

Given the tearing in Keen, I would suspect an issue with the status register or something that PoP chokes on but not Keen. There’s a fallback check in Keen if it waits too long for a vblank that will bail and keep the game running instead of halting.

I have played Monkey Island today, again with no issues.
One thing is maybe important - After i replaced U7 and U8, i still didn't get this "broken" screen on startup. I need to change U9 and U2 when they arrive, and than i will try POP again.

mkarcher wrote on 2026-01-24, 16:25:

I like the way you are thinking, but except for half-broken chips as used for example for the 32KB "extra memory" in the ZX spectrum, the number of bits in DRAM chips at that time always were a power of four, so there were no half-dense chips, just quarter-dense chips. IBM used 16K*4 chips on the original EGA (in 1984), while the P82C435 only supports 64K*4 chips according to the data sheet.

I started thinking that I should look for a very similar VGA chipset, perhaps derived from the P82C435. I found the P82C441 on vgamuseum, and the example there of "Rainbow EGA", which despite being a VGA capable chip, was sold labeled as EGA with a supposed EGA BIOS. Further, it is clear it was stripped of Super EGA function, and there is a PCB section that would support 8(!) ICs, possibly of the 64K*4 chips, but it goes a step further of having a 16K*4 section as well. Thanks for confirming that I was not missing a real 32K*4 in my search, I didn't really want to look that far for one.

Yes, I was also thinking of half-broken chips. I think they were creating options on how to use whatever precious stock of RAM they could find.

I'm still curious about the BIOS, especially since it is starting to look the problem is specific to POP. They do share a bit between EGA and VGA implementations and thinking what a broken BIOS or wrong BIOS could do. Maybe there are clues in the programming of the BIOS or POP.

the3dfxdude wrote on 2026-01-26, 01:35:

mkarcher wrote on 2026-01-24, 16:25:

I like the way you are thinking, but except for half-broken chips as used for example for the 32KB "extra memory" in the ZX spectrum, the number of bits in DRAM chips at that time always were a power of four, so there were no half-dense chips, just quarter-dense chips. IBM used 16K*4 chips on the original EGA (in 1984), while the P82C435 only supports 64K*4 chips according to the data sheet.

I started thinking that I should look for a very similar VGA chipset, perhaps derived from the P82C435. I found the P82C441 on vgamuseum, and the example there of "Rainbow EGA", which despite being a VGA capable chip, was sold labeled as EGA with a supposed EGA BIOS. Further, it is clear it was stripped of Super EGA function, and there is a PCB section that would support 8(!) ICs, possibly of the 64K*4 chips, but it goes a step further of having a 16K*4 section as well. Thanks for confirming that I was not missing a real 32K*4 in my search, I didn't really want to look that far for one.

That's a good way of researching it, but your result is still slightly wrong: it's not a 16K*4 section, but a 64K*1 section. As the section is filled, you can easily look up the chip type M5K4164, and another clue is the DIP16 package. Both 16K*4 and 64K*4 use DIP18. I don't think any C&T chipset later than the 8240 (82c431..83c434) supports 16K*32 banks. The STB AutoEGA uses discrete logic to fake a single 64K*32 bank from two 32K*32 banks, so this is not a feature related to the chipset at all.

the3dfxdude wrote on 2026-01-26, 01:35:

Yes, I was also thinking of half-broken chips. I think they were creating options on how to use whatever precious stock of RAM they could find.

The theory they intended to use half-broken chips is possible, but weak. The card has no way to choose which half to use, it still passes the most significant column bit to all the RAM chips, so you can use chips with the lower half bad in one bank, and chips with the upper half bad in the other bank. Needing an equal amount of both types of bad chips poses a logistic challenge that might kill the ecenomic viability of that approach.

the3dfxdude wrote on 2026-01-26, 01:35:

I'm still curious about the BIOS, especially since it is starting to look the problem is specific to POP. They do share a bit between EGA and VGA implementations and thinking what a broken BIOS or wrong BIOS could do. Maybe there are clues in the programming of the BIOS or POP.

The theory by keenmaster486 that the "vertical retrace" status bit might stick sometimes is a possible explanation. A sticky retrace bit can also be caused if the card inadvertently switches from CGA-like (status at 3DA) to MDA-like (status at 3BA). A bad status read would most likely point to the primary ASIC, the P82C435 being bad. Inadvertant switching of the base address would point to the bus interface ASIC being bad. I don't see how the BIOS (which is most likely not used at all while the game is running) can cause a freeze after some indeterminate time.

BIOS dump if needed...

Today remaining LS chips arrived, changed them (and 4.7uf 16Vtantalum also), but still the same. It works in any games i have played, but it freeze in POP. I still have caps near RAM to change, and if that doesn't work, i quit! 😀

I did changed caps near RAM, but neither that didn't changed Prince behavior. All the other games are working normally, there is no "broken startup" anymore, but Prince of Persia (latest 1.3 version) freezes randomly. All in all, I'm satisfied with this card, because all other games and apps works normally.
Thanks everyone for help!!!!!

butjer1010 wrote on 2026-02-02, 12:31:

I did changed caps near RAM, but neither that didn't changed Prince behavior. All the other games are working normally, there is no "broken startup" anymore, but Prince of Persia (latest 1.3 version) freezes randomly. All in all, I'm satisfied with this card, because all other games and apps works normally.
Thanks everyone for help!!!!!

So actually, there were two issues with your card. The "broken startup", which you fixed by replacing U7-U9 (I suppose U8 is the most likely culprit), and the freezing in Prince of Persia, which you couldn't fix (and we don't know whether it's some kind of incompatibility, a bug in Prince of Persia or a defect in one of the ASICs). Have fun with that card!

mkarcher wrote on 2026-02-02, 20:41:

butjer1010 wrote on 2026-02-02, 12:31:

I did changed caps near RAM, but neither that didn't changed Prince behavior. All the other games are working normally, there is no "broken startup" anymore, but Prince of Persia (latest 1.3 version) freezes randomly. All in all, I'm satisfied with this card, because all other games and apps works normally.
Thanks everyone for help!!!!!

So actually, there were two issues with your card. The "broken startup", which you fixed by replacing U7-U9 (I suppose U8 is the most likely culprit), and the freezing in Prince of Persia, which you couldn't fix (and we don't know whether it's some kind of incompatibility, a bug in Prince of Persia or a defect in one of the ASICs). Have fun with that card!

He he, thanks man, will do 😉

Great you managed to fix the broken startup screen! That's the only issue I remember having with the card.

I suppose you don't have another EGA card at hand to try Prince of Persia with?

MMaximus wrote on 2026-02-03, 08:44:

Great you managed to fix the broken startup screen! That's the only issue I remember having with the card.

I suppose you don't have another EGA card at hand to try Prince of Persia with?

I have another EGA card, everything worked ok before. Need to try again, but i tried also with CGA card i have, and it worked normally.

How does Keen work now? Stuttering gone?

keenmaster486 wrote on 2026-02-03, 17:17:

How does Keen work now? Stuttering gone?

I didn't play Keen after changing all the chips and caps, but last time i played, before caps and other 2 LS, it had. Rest of the games worked normally.