The low volume order will kill you, concerning such low density parallel srams.

Pricing will have a proportional relationship with newness and reliability, such that newly made srams from a reputable vendor, like mouser or digikey, will be quite good quality and more expensive, where 'likely rebadged, salvaged, or fake chips' from alibaba or ebay, will be 'much cheaper, but much more likely to not work or behave strangely.'

I mean I don't know where you guys generally source old HW, but my go to is eBay. I've found a seller in the UK, with a set of chips identical to what is on the Trident. I'm just trying to figure out if it's work the £20 to upgrade the card OR wait and source another ISA card with more ram for the same money.

I just checked both those sites and none of them stock HY51C4256S10

An electrically identical parallel sram of the same speed will work fine.

Look up the spec sheet for the existing srams, then use mouser or digikey's search tools to filter for that kind of package.

In this case, it's a 20 pin dip package. I cant tell if that is 10ns or 70ns for the speed. Image is too blurry. If that number is correct, it's 10ns, and this is the datasheet.

https://www.datasheets360.com/pdf/-2391777341851843209

Source a compatible modern 256x4bit dip 20 sram.

Recently I searched for FP and EDO RAM chips on aliexpress and the cheapest I found was about $2.50 for one piece. I think these chips was salvaged from electronic scrap.

wierd_w wrote on 2023-08-13, 16:59:

An electrically identical parallel sram of the same speed will work fine. […]
Show full quote

An electrically identical parallel sram of the same speed will work fine.

Look up the spec sheet for the existing srams, then use mouser or digikey's search tools to filter for that kind of package.

In this case, it's a 20 pin dip package. I cant tell if that is 10ns or 70ns for the speed. Image is too blurry. If that number is correct, it's 10ns, and this is the datasheet.

https://www.datasheets360.com/pdf/-2391777341851843209

Source a compatible modern 256x4bit dip 20 sram.

FYI SRAM <> DRAM
I've noticed you twice mentioning sram while you should've been using dram.
Using the wrong term adds unnecessary confusion.

and dram tends not to be 10ns especially those old ones. it's more likely to be 70ns

Derp. Yes, 100% correct.

Protip, dont shop for parts past your bedtime.

It's a dip20 256k x 4bit FP mode DRAM.

I have one 512k version of TVGA8900 (JAX-8216A) and it has two -10 soldered and two -8 socketed...KM44C256's. My 1Mb TVGA8900 JAX JA-8232E/V3 has eight TMS-44C256 -7's...
Both have jumpers to allow working in a 8bit ISA slot, the one pictured here does not appear to have them.....just for reference.....

analog_programmer wrote on 2023-08-13, 11:29:

Actually the Orchid Voodoo 1 with these mechanical contacts in the relays is more prone to failure, so it's not superior to Diamond Monster 3D card.

I remember reading that Orchid cards have better IQ compared to most other Voodoo cards and those relay switches are the main reason they have less noise. I have two V1s, one is Diamond and other is Skywell and both have equally shitty IQ, so it may be that Orchid is indeed better in this regard.

Don't know if this is true, however, because I don't have Orchid.

analog_programmer wrote on 2023-08-13, 11:29:

Here comes some "voodoo hype"...

mkarcher wrote on 2023-08-13, 01:22:

The pass-through cable is a standard male-to-male VGA cable, and as I understand it, you can re-use the usual detachable LCD VGA cables as pass-through cables.

Wrong. It's male-to-female VGA cable and you can't use standard male-to-male VGA cable as pass-through.

So much for assuming stuff. So a LCD monitor VGA cable is the wrong cable, but a VGA extension cable should do.

analog_programmer wrote on 2023-08-13, 11:29:

mkarcher wrote on 2023-08-13, 01:22:

Contrary to the Diamond Monster 3D, the Orchid version uses mechanical relays to forward the 2D VGA signal to the monitor, whereas the monster uses FET-based switches. Except for the audible clicking in the switching process, which is a matter of taste, the relay-based version is usually considered superior.

Actually the Orchid Voodoo 1 with these mechanical contacts in the relays is more prone to failure, so it's not superior to Diamond Monster 3D card.

So, it might be superior as long as the contacts work fine, but in the long run, the FET-based cards outlive the relay-based cards.

wierd_w wrote on 2023-08-13, 15:51:

Trident cards DO have primitive 2D accelleration inside them, and scitech display doctor DOES know how to use it. When using SDD, the software univbe handler will thus provide hardware accellerated 2d modes, at higher resolutions.

The Trident TVGA series (TVGA8800, TVGA8900, TVGA9000) do not have 2D acceleration inside them. The later TGUI chips do. "GUI" in their name refers to the capability of acceleration graphical user interfaces. Furthermore, it is true that some versions of SDD provide accelerator features via the VBE/AF software interface, but software using this interface to make use of hardware accelerated 2D graphics is very rare.

Hiddenevil wrote on 2023-08-13, 09:41:

From what I've read now, the Trident 8900 might not be the ideal card to install in a 486 system, it might not be able to handle Doom, Duke, e.g games from the latter half of the 90's. Would that be a fair assessment?

A 486 DX2/66 is only complete on a mainboard with VLB slot paired with a VLB card 😀
Any ISA card is a bottleneck for such a system

analog_programmer wrote on 2023-08-13, 11:29:

mkarcher wrote on 2023-08-13, 01:22:

The pass-through cable is a standard male-to-male VGA cable, and as I understand it, you can re-use the usual detachable LCD VGA cables as pass-through cables.

Wrong. It's male-to-female VGA cable and you can't use standard male-to-male VGA cable as pass-through.

Just add a gender changer

HanSolo wrote on 2023-08-13, 22:12:

analog_programmer wrote on 2023-08-13, 11:29:

Wrong. It's male-to-female VGA cable and you can't use standard male-to-male VGA cable as pass-through.

Just add a gender changer

Yeah Or use a short VGA extension cable. The blue DIN is standard female, black DIN is a male according to some pictures.

mkarcher wrote on 2023-08-13, 21:52:

wierd_w wrote on 2023-08-13, 15:51:

Trident cards DO have primitive 2D accelleration inside them, and scitech display doctor DOES know how to use it. When using SDD, the software univbe handler will thus provide hardware accellerated 2d modes, at higher resolutions.

The Trident TVGA series (TVGA8800, TVGA8900, TVGA9000) do not have 2D acceleration inside them. The later TGUI chips do. "GUI" in their name refers to the capability of acceleration graphical user interfaces. Furthermore, it is true that some versions of SDD provide accelerator features via the VBE/AF software interface, but software using this interface to make use of hardware accelerated 2D graphics is very rare.

Come to think of it, I was using a 9440, if memory serves, and that is indeed a tgui card.

Still, the latter isn't quite true. While doom, doom2, and heretic all just fiddle with the bog standard vga interface, descent, duke3d, rise of the triad, strife(?), warcraft 2, and quite a few others all use VBE, and will cheerily use SDD. I would not call it rare.

mkarcher wrote on 2023-08-13, 21:49:

analog_programmer wrote on 2023-08-13, 11:29:

Here comes some "voodoo hype"...

mkarcher wrote on 2023-08-13, 01:22:

The pass-through cable is a standard male-to-male VGA cable, and as I understand it, you can re-use the usual detachable LCD VGA cables as pass-through cables.

Wrong. It's male-to-female VGA cable and you can't use standard male-to-male VGA cable as pass-through.

So much for assuming stuff. So a LCD monitor VGA cable is the wrong cable, but a VGA extension cable should do.

Again: the pass-through cable is male-to-female and you call it as you want - standard, extension... etc.

mkarcher wrote on 2023-08-13, 21:49:

analog_programmer wrote on 2023-08-13, 11:29:

mkarcher wrote on 2023-08-13, 01:22:

Contrary to the Diamond Monster 3D, the Orchid version uses mechanical relays to forward the 2D VGA signal to the monitor, whereas the monster uses FET-based switches. Except for the audible clicking in the switching process, which is a matter of taste, the relay-based version is usually considered superior.

Actually the Orchid Voodoo 1 with these mechanical contacts in the relays is more prone to failure, so it's not superior to Diamond Monster 3D card.

So, it might be superior as long as the contacts work fine, but in the long run, the FET-based cards outlive the relay-based cards.

Again for those intoxicated by the clicking of the relays: In the long term the contacts and coils of mechanical relays are much less reliable, but it is obviously unnecessary to explain in detail why this is so to prejudiced theorists. Orchid's engineers are ingenious, but no one else uses their "ingenious solution" even in Voodoo 2...

I didn't expect you to admit your mistakes as most people don't.

analog_programmer wrote on 2023-08-13, 11:29:

Actually the Orchid Voodoo 1 with these mechanical contacts in the relays is more prone to failure, so it's not superior to Diamond Monster 3D card.

While the relays may wear out, it should be a pretty straight forward replacement for anyone inclined to do a repair. And replacement relays are readily available.

Shponglefan wrote on 2023-08-14, 03:43:

While the relays may wear out, it should be a pretty straight forward replacement for anyone inclined to do a repair. And replacement relays are readily available.

No one who doesn't understand at least a little bit of diagnostics in electronics will anyway go to change even a relay soldered on a PCB. And no easier detection of a 2D-3D switching problem i.e. potential ease of repair was the main idea in using mechanical relays in this case.

I am kinda shocked that they used mechanical relays at all. There will be so much electrical bounce from that!! The collateral this would have elsewhere in the circuit design just... .Why would you even do this?

The FET is just a better device for this, in every way! It's not like this is a high current application where a mechanical relay would make sense or anything... The FET switches faster, and does so SO MUCH CLEANER-- why? Why would they do this?

wierd_w wrote on 2023-08-14, 04:42:

I am kinda shocked that they used mechanical relays at all. There will be so much electrical bounce from that!! The collateral this would have elsewhere in the circuit design just... .Why would you even do this?

The FET is just a better device for this, in every way! It's not like this is a high current application where a mechanical relay would make sense or anything... The FET switches faster, and does so SO MUCH CLEANER-- why? Why would they do this?

I assume that Orchid's engineers decided to use a true electrically isolated bypass (mechanical relay instead of MOSFET circuit) assuming that this way they will provide cleaner signal in the 2D pass-through mode without any idea that these videocards would still be in use after more than a quarter century. But as I have already noted, over time the mechanical relays deteriorate their contact resistance and it is more likely to get purely mechanical problem or in the coil itself, so the MOSFET solution is the better option in terms of long term operation in this particular case.

P.S. From a purely "idiocratic" point of view: There is always a possibility that some clever engineer has literally understood the term "bypass relay switch" 😁

If you want to suppress noise in the circuit, wouldn't a choke or a resistor-capacitor pair on ground for the signal line, be a valid and much longer lived option?

The bounce on the video signal on hardware switching would still occur, even if you are avoiding other sources of circuit hum by using a literal isolated switching circuit. (Granted, analog monitors would probably handle that fine, but still. Depending on the monitor in question, this might be enough to cause strange sync or signal identification issues with multisync monitors.)

wierd_w wrote on 2023-08-14, 06:50:

If you want to suppress noise in the circuit, wouldn't a choke or a resistor-capacitor pair on ground for the signal line, be a valid and much longer lived option?

The bounce on the video signal on hardware switching would still occur, even if you are avoiding other sources of circuit hum by using a literal isolated switching circuit. (Granted, analog monitors would probably handle that fine, but still. Depending on the monitor in question, this might be enough to cause strange sync or signal identification issues with multisync monitors.)

I'm not disputing your knowledge of analog and digital electronics, because you're absolutely right. I was just commenting on what can cause such irrational decision (using mechanical relays), which some nerds admire 😀