BitWrangler wrote on 2021-06-01, 15:59:

I am skeptical of the "doesn't work with EDO" I think it's like "Doesn't work with PC-100" ... it's just that higher density DRAM more common on the newer standard and it's those the board doesn't support.

EDO is more than just something that can work faster (PC66/100/133). EDO outputs data long after the bus cycle is over, so that a new cycle can start while data is still being read, offering sort of a pipeline operation that helps to push performance.

The consequence here is that if the host is not designed to handle it, the data from previous cycle will cause a bus conflict with current one which leads to a swift crash. That is why EDO doesn't work on things not designed for it.

486 boards that work with EDO have data bus buffers between the CPU bus and DRAM bus (either external or built into the chipset), not unlike in all Pentium1 designs (there's usually two "data path drivers" sitting between memory and CPU bus).

It only does that if the chipset holds a line longer, if it doesn't support the full features of EDO, which is FPM with a bolt on, it doesn't do that and the EDO doesn't do that. The whole PC-100 thing was just to emphasise how newer tech can be higher density than older tech and is a much more likely reason for not functioning. Then there's a whole other thing where an EDO SIMM could have ECC which will stop a board working if it can use parity, but doesn't support ECC. If it just flat out ignores parity it's all good. If it just wants normal parity SIMM it's thrown for a loop by ECC. Again, it's not that it's EDO, it's some other parameter you've n00bed up the match on.

I have two 16MB sticks, with identical electrical layout, with same kind of 4x 4M chips with identical internal oganization and refresh requirements, and the only difference is one stick has EDO and other has FPM DRAMs on them. One works with a typical 486 board, other fails the POST, It is not a problem of having parity, voltage, layout or any other difference.

If you look at the timing diagrams in the datasheets of EDO and FPM chips you can clearly see that EDO holds its output during page mode cycles, which is the only difference and the reason for the name. It is the only reason why a board will not even POST when an EDO stick is inserted, while it has no problem with the stick with FPM DRAMs. FPM and EDO are not meant to be directly compatible and they are not.

I am not pulling stuff out my rear here, and most people who say EDO doesn't work on a typical 486 board do not either though they may not know the actual technical details.

The EDO doesn't hold it's output, the motherboard holds the EDO output. The only time that could happen in a non EDO board is if the manufacturer halfassed some EDO forward compatibility and never got around to updating the BIOS to handle it.

If it was a motherboard feature there wouldn't be a whole new sub-family of DRAMs. The whole point of EDO DRAMs is that they hold their output between !CAS cycles, saying that doesn't happen or has no effect in the situation seems to imply that one is not familiar at all with the details of operation of these memories and devices they are used with...

If Edo was a function of the chip set then FPM would work as EDO in those motherboards which it doesn't.

BitWrangler wrote on 2021-06-02, 14:16:

The EDO doesn't hold it's output, the motherboard holds the EDO output.

No, it's the EDO itself. The main difference between FPM and EDO RAM chips is the behaviour of the RAM chip when /CAS goes inactive (i.e. high) after a read cycle, while /RAS stays active. On FPM chips, the data lines get tristated in this situation, whereas on EDO, the output is held on the data lines (hence the name "extended data out"). You could get EDO chips to release the bus while still keeping the page open if you deactivate the /OE pin ("output enable"), but this pin is not on the PS/2 simm connector. On PS/2 EDO SIMMs, /OE is permanently active (i.e. low).

I used to think that this is a flaw of the PS/2 SIMM standard that makes page-mode operation of EDO SIMMs nearly impossible, but I finally got to re-read the SiS 496/497 data sheet. While it provides a RAMOE pin that could be used to connect to the RAM /OE line, this signal is actually meant to be connected to a buffer chip between RAM and the chipset (and is intended to be ignored when there is no buffer chip, as on most consumer boards). Instead, the SiS 496 memory controller pulls a clever trick in EDO mode: After a read burst, it pulses /WE active while keeping /CAS inactive. According to EDO datasheets, activating /WE should be enough to free up the data bus, so that the processor can apply the new data that is going to be written as soon as /CAS will go low (activating /WE before /CAS is what RAM datasheets call an "early write", as /WE is earlier than /CAS). As the SiS 496 never pulls /CAS low, there is no actual write cycle happening (you could call it an "aborted early write" or "incomplete early write", but I never heard those terms). This killed my plan of modding an EDO module to have /OE on an NC pin and adding a bodge wire to the mainboard, as this is most likely unneeded. Instead, the RAM detection routine in the BIOS is the prime suspect now. It should detect if EDO RAMs are installed, and switch the chipset into EDO mode, so it generates the dummy /WE pulses, and thus prevent the bus conflict. In practice, all SiS 496/497 boards I had access to yet (Soyo SY-4SAW2, Asus PVI-486SP3) plain refuse to boot with EDO chips. The Asus board is extremely misleading by dropping into "flash recovery" mode (or trying to do so) if EDOs are installed, although the flash chip itself contains a good BIOS image.

The worst board I've had is Gainward 5VPA, super socket 7 with VIA MVP3 chipset. I really can't understand it.

- CPU & memory clock jumpering is just confusing. Even though everything should be right, sometimes the board just don't want to boot. Next day it might be fine again. Like it had volatile fault memory or something.
- AGP slot powering is too weak. It can't handle period correct cards. ATI 7000 is the best that works so far. Voodoo 3 is too much, computer boots when loading Windows. There must be more than power issues.
- VIA IDE controller driver messes up something and Windows protection error occurs.

If something positive, I can run K6-2+/3 on it. Just is unavailing due to issues above. I still have the board because it's the only MVP3 one I have.

IDE controller is the one reason I had to retire my Luckytech P5MVP3. No matter what I'd do, Secondary channel would always show up with a exclamation sign, making the drive work in compatibility mode. I ended up replacing it (for now) with a Soyo 6BA+IV /P3 650MHz build but I'm hoping to switch it to a more simple ASUS P2B Rev1.02.

PCCHIPS M919 that I have laying around doing nothing. Not sure what's wrong with it. I bought it at RE-PC in 2018 and it's been bricked the whole time I've had it. Starting to wonder if the Flash BIOS is borked because I tried flashing it in my FIC 486 PVT and it said the chip was not a flashable chip (weird). Motherboard POST card stops before POST. It's a future weekend project.

As nostalgic as I am as it was my first 486, the ZEOS GMB-4500-95B 386/486 Upgradable Systems mainboard is another one. Blown HDD Controller, Serial Ports as flaky as a Pillsbury Biscuit, constant battles with the Game Port between it and the one on my SoundBlaster....I got run through the ringer on my first "retro PC" in 2001 (Creeping Net 1). The blown HDD controller though was a bit of a blessing in disguise because I doubled the IDE Channels using a Future Domain Power-IDE card in there with it's own BIOS.

AST Advantage Pentium 60 - this was early adopter syndrome defined.

creepingnet wrote on 2021-06-03, 23:10:

PCCHIPS M919 that I have laying around doing nothing. Not sure what's wrong with it. I bought it at RE-PC in 2018 and it's been bricked the whole time I've had it. Starting to wonder if the Flash BIOS is borked because I tried flashing it in my FIC 486 PVT and it said the chip was not a flashable chip (weird). Motherboard POST card stops before POST. It's a future weekend project.

Not that weird. A non-flashable chip (sometimes still erasable by UV light) was cheaper. Cheap as (PC)chips so to speak.
I suggest you find a proper replacement EEPROM chip and try again with that.

evasive wrote on 2021-06-04, 08:09:

creepingnet wrote on 2021-06-03, 23:10:

Starting to wonder if the Flash BIOS is borked because I tried flashing it in my FIC 486 PVT and it said the chip was not a flashable chip (weird).

Not that weird. A non-flashable chip (sometimes still erasable by UV light) was cheaper. Cheap as (PC)chips so to speak.

Most likely that's the issue.

The other possiblility is that the chip is indeed flashable, but the tool used on the FIC board doesn't know how to flash that chip. It's easy to find out if you read the model number of the BIOS chip (you might need to pull the sticker for that, and you might want to consider whether this damages the "original retro look", decide for yourself). If the model number of the flash chip starts with 27, it needs to be erased using a UV light. If it starts with 27, and doesn't have a window, there is no way for UV light to get in, so erasing that chip isn't sensibly possible (although I heard stories about erasing such chips using a dentist's X-Ray device, or milling away the top cover). On the other hand, if the model number starts wit 29, it should be flashable using suitable software (like UNIFLASH) in the FIC board. If it starts with 28, you have to check datasheets for the required programming voltage (which might be 5 or 12V) and set the jumpers accordingly (assuming the FIC board has a 5V/12V flash jumper - most boards of that time do).

I've seen motherboards use OTP chips rather than EEPROMS.

OTP chips are probably EPROMS with no erase window (so you can't erase them). Would be cheaper to make than a proper EPROM with erase window.

I don't remember exactly which models had seriously design issues like PCI cards too close to other mainboard components or with the Socket 7 cpu heatsink in the middle of their space.. some were not exactly stable too.

What I can't stand it anymore are most mainboards with capacitors gone and I've seen too many Socket A, Socket 478, Socket AM2 board having those problems and changin capacitors you never know if it will be easy or not depending on many variable and often I've lost patience changing them (capacitor that could not be desoldered or reinserted in the holes etc..).
For the older boards I don't like all the Dallas battery boards and obviously the Ni-Cd battery leaked boards that might be easy to repair or a nightmare until you really analyze the area and discover how damaged the board is.
For the more modern boards I don't like the opposite basically you can't even analyze them, they integrate too many BGA ICs and components that are simply not possible to repair for the average soldering experienced user. Also I think it's become too difficult to debug modern boards... can be broken anything from a mosfet to some ics or whatever without even showing signs of the damage.. I don't know if there're comparison of the rate of dead mainboards nowdays compared to the past (beside the capacitor period problems).

canthearu wrote on 2021-06-04, 08:42:

OTP chips are probably EPROMS with no erase window (so you can't erase them). Would be cheaper to make than a proper EPROM with erase window.

The OTP chips you can find as BIOS chips on mainboards (and on some peripherals as BIOS extensions) are exactly that. They usually even have the same chip type number as the corresponding EPROM. A major contribution to cost reduction is due to the fact that you can't easily mount the quartz glass window for the erasing light into a plastic case, so proper EPROMs need to use a ceramic case.

There were older "PROMs" in the 80s that worked in an entirely different way, but you wouldn't find them in variants that are big enough to contain a 486 or Pentium BIOS.

Strangely enough, the worst mobo I've ever worked with is also kinda the best one.

It is QDI Synactix-1, and, comparing it to the previous i815 microATX board that I've used with my only retro rig - the wonderful DFI CS32-TC, - it has quite a lot of unpleasant quirks:

- HDD size is limited to 137 GB by the BIOS even with the latest firmware update;
- CD/DVD-ROMs connected via an 80-pin ribbon cables do not work for some reason (a 40-pin cable solves this);
- the ATX connector is located in a weird place, but that's a problem with just my picoPSU-like Chinese power supply solution;
- the AGP is one slot below its standard position; where it should normally be, there is a now-useless CNR connector instead;
- there is one PCI slot less thanks to the presence of the damned CNR;
- initializing of a new FSB frequency (for example, switching from 100 to 133 with jumpers) requires disconnecting the power cord!
- accepting a new FSB frequency requires going into the BIOS and saving the changes, otherwise there'll be a warning each time;
- BIOS chip is soldered on and cannot be hot-swapped or replaced easily;
- IRQs cannot be assigned directly onto specific PCI slots; you can only reserve IRQs in succession until the right one is assigned;
- no mention of supporting Tualatin CPUs in the manual other than "future CPUs", but they *should* work as 1.5 voltage is there.

Literally none of that is the case with DFI CS32-TC, but, despite all that, I still love the Synactix-1, because it has a real and working PCPCI /SB-Link connector 😀 The slot locations are also not that bad since I've still managed to cram all the stuff I needed into a tiny all-aluminum PC case (after applying slight alterations involving power tools).

There's definitely something in accepting a motherboard for what it is, rather than what you wish it to be. Some people might think for instance that a 430HX motherboard is the highest embodiment of classic Pentium P54 class architecture... or others might think instead that it's a crappy motherboard because it doesn't even support SDRAM or ATA-100.

PcBytes wrote on 2021-06-03, 20:10:

IDE controller is the one reason I had to retire my Luckytech P5MVP3. No matter what I'd do, Secondary channel would always show up with a exclamation sign, making the drive work in compatibility mode. I ended up replacing it (for now) with a Soyo 6BA+IV /P3 650MHz build but I'm hoping to switch it to a more simple ASUS P2B Rev1.02.

I forgot to add that to my post. I cannot use UDMA in the soyo 5EHM (via MPV3), windows simply cant install or boot in this mode. I tried EVERYTHING with no success, so i have to use it in PIO4 which eats a lot of resources from the poor k6-2 500Mhz

Trigem Cognac, for some bizarre reason installing anything newer than Windows 2000 causes one hell of a headache until you change the Computer driver to Standard PC rather than ACPI (or whatever the default is called, I forget). If you don't do that then any optical drive access just results in a wall of errors in the event viewer and nothing being read.

Before anyone asks, yes I have upgraded the BIOS to the newest version available but no change. And swapping the optical drive makes no difference either. However it did allow me to up the memory to 512MB instead of 256MB, before the BIOS update it wouldn't even allow me to do that.