Mu0n wrote on 2022-12-21, 23:16:

No slowdowns at all under DOS, even when I transfer a mass amount of files through mtcp's ftpsrv. […]
Show full quote

Duffman wrote on 2022-12-21, 06:38:

where do you normally notice the slowdowns?

20221221_142836_weecee_ide.jpg

No slowdowns at all under DOS, even when I transfer a mass amount of files through mtcp's ftpsrv.

However, under win98se and transfering large files like CD ISOs or higher is when I notice the alternance between copying slowly for a few seconds, followed by a total pause of 15ish seconds, then another sluggish copy interval of a few seconds, and on and on.
Transfering small files doesn't seem to have that effect or it's not noticeable. It's so terrible for large files that I just do the copying under modern hardware.

Same exact situation I've seen, too. File transfers will pause, things will lockup, and then you just gotta give it time to go.

Even with the slow-down it doesn't make it unusable by any means, but it does run sub-optimally, for sure. The file system being in compatibility mode reeks havoc on things like virtual drive emulation, but who knows... These SOMs are fairly complex, so. I know there is some shared pins, for example, and however they do that sharing I haven't any idea.

Duffman wrote on 2022-12-21, 14:00:

uummm.... Direct3D on the WeeCee... maybe? […]
Show full quote

uummm.... Direct3D on the WeeCee... maybe?

20221221_213624_weecee_3d_maybe_1.jpg

20221221_213807_weecee_3d_maybe_2_resized.jpg

FPS is a little low - not sure if this is fully working or not.

driver: https://ufile.io/57fx8bp2

Makes my anti-virus squeal a bit -- is that normal?

@Nachtrave

Well, I have made modifications to the driver, nothing malicious though.

I did a test where I replaced the newer smaller xgigrm.vxd with the older larger xgigr.vxd that's used with the Volari V5/V8

Win98 gave a BSOD on next boot, bugger!
it might be possible to tweak this to work for the Volari Z9 but I'm out of ideas, so I'll stop for now.

Duffman wrote on 2022-12-22, 10:57:

I did a test where I replaced the newer smaller xgigrm.vxd with the older larger xgigr.vxd that's used with the Volari V5/V8

Win98 gave a BSOD on next boot, bugger!
it might be possible to tweak this to work for the Volari Z9 but I'm out of ideas, so I'll stop for now.

Yes! Punch the keys!

I put the zip file onto the NR google drive and announced link to it on NR twitter, so others can also experiment. We noted that it is experimental, but I think even just getting something to run is an impressive feat with that.

I'll start putting it on the SD cards we ship with if it turns out to work fairly well enough. Sucks that it makes the Win10 anti-virus bark, but it barks at all sorts of stuff.

I gave VBEMP9x a go just to see how it would go.

In UT99 it was slower than the official driver by a 3-4 FPS.

So even though the official driver is itself slow, it's still faster than VBEMP9x.

I think it'd be cool if we could get a SOM without the Volari Z9 and use the spare PCI lane to hook up a Geforce FX 5500 PCI that you see on ebay all the time to add some 3D acceleration plus support for Table fog and 8-bit palleted textures.
It's bigger than the WeeCee though and we'd probably need rasteri's help to hook it up.

Alas, just a dream.

It's absolute nostalgia 😀. I had the same problem with 3D when I was a kid, and sadly watched the demo of Redline Racer on the CD cover but Sam and Max and Day of The Tentacle made up for it 😁.

Question for folk: Is there a way to get a USB-based external CD-ROM to work under DOS? To register as a device and be accessible?

Seems like one of the issues that came up was that when you allow the BIOS to do the CD-ROM emulation of it it will come up as an A:\ drive and only show the first so many bytes of the disk.

I've always used SHSUCD and Daemon Tools and such so that I don't have to switch CDs, but I'm not sure where to go with this one.

@Nachtrave

There's a USBCD1.SYS and USBASPI.SYS driver included with this MS-DOS 7.1 installer

try them.

https://archive.org/details/ultimate-msdos-7.1

@Nachtrave @rasteri @Paul_V

I've got some experimental modded XGI Volari BIOS ROMs (32KB and 64KB variants) that I want to test, but I'm wary of risking bricking the 1 SOM I have.

Have any of you got a spare SOM that you'd be willing to try a modded XGI Volari BIOS ROM on?

Duffman wrote on 2022-12-24, 09:13:

@Nachtrave @rasteri @Paul_V

I've got some experimental modded XGI Volari BIOS ROMs (32KB and 64KB variants) that I want to test, but I'm wary of risking bricking the 1 SOM I have.

Have any of you got a spare SOM that you'd be willing to try a modded XGI Volari BIOS ROM on?

Hi,

I can try test it on my long-suffering ebox motherboard which I'm experimenting on now. It has the same DX CPU and Volari VGA.
You can PM me the files and brief how\what\with_what to test instructions. I'll see what I can do.

@Paul_V

Here is my modded XGI Volari BIOSs - Both A 32KB Z9 and 64KB V5/V8 version.

Along with Winflash that was bundled with these ROMs. I'm not sure if this is just standard winflash or if it's a special version for XGI.
Hopefully this flash tool works.

I'm assuming just open winflash.exe - I'm guessing just select the modded ROM and flash it?

My theory is that the extra BIOS code in the 64KB ROM is what allows 3D to work? needs testing to verify.

If the BIOS mod then works I'll send you the Volari V5/V8 driver.

Duffman wrote on 2022-12-24, 12:08:

@Paul_V I'm assuming just open winflash.exe - I'm guessing just select the modded ROM and flash it? […]
Show full quote

@Paul_V
I'm assuming just open winflash.exe - I'm guessing just select the modded ROM and flash it?

My theory is that the extra BIOS code in the 64KB ROM is what allows 3D to work? needs testing to verify.

If the BIOS mod then works I'll send you the Volari V5/V8 driver.

Vortex uses own flash tools, like BIOSMP, XFLASH etc. Didn't try others though.
What I meant is how I'm to test whether 3D is working (if at all) as intended or not. 3D accel games was not exactly my era of active gaming )

Right now, i have successfully booted with 32KB vga BIOS you provided. Now I get extra BIOS VGA message before main BIOS boot.

UPD:
64KB is a no go. And it halts the boot process, so I can't reflash through autoexec.
JTAG connector on EBOX is an ugly one, I may need some time to hook up to it and reflash.

@Paul_V

Ah, shame the 64KB ROM is a no go. Thanks for testing that.

When you've reflashed the working ROM -
First run setup.exe in this archive to install this XGI driver.

https://ufile.io/c4hix8ew

Then check with dxdiag to see if Direct3D says if D3D acceleration is available or not. or install UT99.

Duffman wrote on 2022-12-24, 09:13:

@Nachtrave @rasteri @Paul_V

I've got some experimental modded XGI Volari BIOS ROMs (32KB and 64KB variants) that I want to test, but I'm wary of risking bricking the 1 SOM I have.

Have any of you got a spare SOM that you'd be willing to try a modded XGI Volari BIOS ROM on?

I have a broken SOM that has its entire IDE circuit ripped to shreds and I don't think necessarily boots (it does post tho and you can go in and do stuff to the BIOS and what have you)... You could at least use it as a test-bed to wreck as you please. I was originally going to send it to rasteri to take a look at maybe fixing, although it's super tiny and delicate circuitry. You'd need a really good microscope and very steady hands to work on the kinds of trace repair you'd need to bring it back to functional with the microSD.

Would be great to be able to figure out a way to still boot off the USB or something (should still work?), then you could do anything you wanted to that SOM - I am unsure if you can just skip the IDE detection part on it (I suppose it can?), but I'd have to dig it out. The reason it's damaged is because it was my first attempt at removing the SST NAND Flash, and that went very, very badly. I could try plugging it in and seeing if I can get USB to boot at least, but I'd still need help with shipping.

Otherwise: US Distributor of SOM304RD-VI (where I buy mine from) -- you will need to go through the inquiry process to get the VINE1 (instead of VIBE1) version without the SST NAND Flash, otherwise you will need to desolder/tombstone a resistor to disable it (as I have done in the past in this thread, many pages ago). The total cost is 201$ USD for the VINE1 version, and 227$ USD for the VIBE1 version, along with 3-4 wk lead time. The only real initialization on the BIOS I do beside loading optimal defaults is modify the IDE Primary Pin Select to SD Card (under Advanced->IDE Configuration tab), and correctly set the time (which I do with an atomic clock sync program in Win98SE).

But yeah do be careful of the SOM. I honestly don't have any replacements besides this half broken one that I still need to send to rasteri to look at to see if he can't get it going. Last time I tried mailing it I guess the postage was hand-written all weirdly and the British post really didn't like it. Sent it back. Said undeliverable. So this next time I'm just going to do it all electronically - I have it ready to ship back out I am just lazy. It's also 70$ for shipping that I don't wanna spend right yet.

Merry Christmas! Everyone!

@Nachtrave

I'd say it'd need to not just POST but also be in a bootable state in order to test if the 3D acceleration is functional or not.
Anyways @Paul_V has taken the risk with his hardware and it looks like the 64KB ROM isn't in a workable state. Hopefully he can recover back to the 32KB ROM and report back on his findings.

Regarding the IDE issue - I notice that the BIOS misreports the size of the SD card as 24.9GB when it's 128GB.

Perhaps the BIOS might be part of the problem? has anyone tried 16GB SD cards?

Merry Xmas from us here as well!

Duffman wrote on 2022-12-25, 04:13:

Regarding the IDE issue - I notice that the BIOS misreports the size of the SD card as 24.9GB when it's 128GB.

Perhaps the BIOS might be part of the problem? has anyone tried 16GB SD cards?

I also noticed that - it doesn't seem to affect functionality, but who knows. I believe it has something to do with overflow due to LBA/non-LBA supported code, or related to the display of it (such as it is with patched fdisk.exe - again with digit/size overflow). You can certainly try with a smaller card, as I think it would be really great to see what other things we can see happen. Unfortunately given that we have seen these issues across very different builds it's likely to be a software limitation or some other issue with Win9x and its IDE implementation. The thing is, the MicroSD card talks to the SOM along a standard SPI connection -- it's a well documented standard, so it's unlikely to be the culprit here as the circuitry there is incredibly similar to existing MicroSD circuits that this SOM was built to work with. You see pretty much the same design used everywhere else.

Perhaps the solution may lay in looking at how this IDE thing might be handled in say WinMe or Win2k, and maybe find a way to back-port that missing/wrong/incorrect functionality to Win98SE to work with the weeCee. We could possibly ask the folk who do the Win98 unofficial patches project their input on this issue as well -- perhaps we're just not even looking at the right spot, at all. I dunno, I'm not a driver expert by any means, I'm just spitting out ideas.

Edit: Oh, and there is also a modified driver file edit out there for Win9x that is for larger >128GB drives (needed if say you're doing a 256GB or 512GB card). 128GB is distributed because it's a) cheap, b) doesn't require driver modifications due to being <=128GB, and c) seems to work just as well as the others do.

NachtRave wrote on 2022-12-25, 23:29:

Merry Xmas from us here as well!

The thing is, the MicroSD card talks to the SOM along a standard SPI connection -- it's a well documented standard, so it's unlikely to be the culprit here as the circuitry there is incredibly similar to existing MicroSD circuits that this SOM was built to work with. You see pretty much the same design used everywhere else.

It could be the culprit though, I think, If it's not a Standard IDE Connection so something else might be happening.

Is there an SPI driver for 98SE?
What does SPI stand for anyway? I'm unfamiliar with it.

You know those cheap IDE to SD card adapters on ebay? The ones that work with 98SE? They always have a chip on them, is there any documentation out there on what that does?

EDIT: According to this review - https://goughlui.com/2019/02/03/tested-generi … dapter-sd35vc0/
They translate microSD into CF - I guess that explains why they work reliably.

It's not, but look at this: https://www.wdlsystems.com/SD-1917-IDE-to-microSD

But there's no chip on that though, unlike the Sintechi ones on ebay that are known to work with 98SE.