VOGONS


Reply 340 of 895, by NachtRave

User metadata
Rank Member
Rank
Member
rasteri wrote on 2022-02-24, 22:28:

Cool! I bet you can repair the torn-up one too - only a some of the pads have traces that run right through them so a few wire bodges should fix it

That's what it looks like. There are a few that I can spot that connect other circuitry. Not many, but a few. I would need a lot more powerful magnifying glass equipment. Maybe this could be something for Necroware or NorthridgeFix at best, alas.

Reply 341 of 895, by Mu0n

User metadata
Rank Member
Rank
Member
NachtRave wrote on 2022-02-24, 22:39:
rasteri wrote on 2022-02-24, 22:28:

Cool! I bet you can repair the torn-up one too - only a some of the pads have traces that run right through them so a few wire bodges should fix it

That's what it looks like. There are a few that I can spot that connect other circuitry. Not many, but a few. I would need a lot more powerful magnifying glass equipment. Maybe this could be something for Necroware or NorthridgeFix at best, alas.

I'd be able to repair it I bet. I have a cheap 40 USD usb microscope which would be enough. Just gotta add up solder mask and a uv light to cure up a replacement isolating layer after you've added tiny pieces of wires in the appropriate places. I see similar work done on the regular in the vintage Mac community, they are constantly repairing varta bombed logic boards from the early 90s to mid 90s.

1Bit Fever Dreams: https://www.youtube.com/channel/UC9YYXWX1SxBhh1YB-feIPPw
DOS Fever Dreams: https://www.youtube.com/channel/UCIUn0Dp6PM8DBTF-5g0nvcw

Reply 342 of 895, by NachtRave

User metadata
Rank Member
Rank
Member
Mu0n wrote on 2022-02-24, 23:00:

I'd be able to repair it I bet. I have a cheap 40 USD usb microscope which would be enough. Just gotta add up solder mask and a uv light to cure up a replacement isolating layer after you've added tiny pieces of wires in the appropriate places. I see similar work done on the regular in the vintage Mac community, they are constantly repairing varta bombed logic boards from the early 90s to mid 90s.

Wow, I've gotten both you and rasteri offering to repair it. Lemme send you both a DM and we can figure out who should do so. I can wrap it up in bubblewrap and get it on its way. =) Would be nice to save 200$ worth of otherwise good hardware.

@Rasteri, could you actually provide me with the weeCee Win98 desktop background image you're using, as was on LGR's video? Also, how did you get that weeCee logo in the system properties general tab, under the "Manufactured and supported by:" label? I'd love to use those on my setup just for a touch of authenticity. =)

Also, @Muon/@Rasteri, how did you get the IDE controllers, under Device Manager, to not be in compatibility 16-bit mode? Was there something special you did there, or? I tried to follow your earlier discussion but I can't seem to find the answer there. I noticed in LGR's videos that there isn't a caution! warning, so you must have figured something out?

Reply 343 of 895, by Mu0n

User metadata
Rank Member
Rank
Member

Wait, I did not offer to repair it, at least not on a short notice. Try it with rasteri if you want. I'm tied up in the next week at work and I bet his repair skills are above mine atm. I know the gist of trace repair having seen lots of footage of it, but I haven't attempted any of it yet.

As for 16-bit compatibility mode for IDE, that remains sadly an unsolved problem for me. I've accepted it. This means you'll get massive slowdowns when:

1) you transfer large files through the network while in win98se
2) you have to install big programs while a cd image is mounted virtually
3) you download big files from the internet

None of these problems occur if you use the mtcp suite inside dos 6.22, or mount .iso images with SHSUCD. Very weird! GO DOS POWER!

1Bit Fever Dreams: https://www.youtube.com/channel/UC9YYXWX1SxBhh1YB-feIPPw
DOS Fever Dreams: https://www.youtube.com/channel/UCIUn0Dp6PM8DBTF-5g0nvcw

Reply 344 of 895, by rasteri

User metadata
Rank Member
Rank
Member
NachtRave wrote on 2022-02-25, 15:19:

@Rasteri, could you actually provide me with the weeCee Win98 desktop background image you're using, as was on LGR's video? Also, how did you get that weeCee logo in the system properties general tab, under the "Manufactured and supported by:" label? I'd love to use those on my setup just for a touch of authenticity. =)

Also, @Muon/@Rasteri, how did you get the IDE controllers, under Device Manager, to not be in compatibility 16-bit mode? Was there something special you did there, or? I tried to follow your earlier discussion but I can't seem to find the answer there. I noticed in LGR's videos that there isn't a caution! warning, so you must have figured something out?

Haha that was all LGR's work. I dunno how he got 32bit mode working either TBH...

Reply 345 of 895, by NachtRave

User metadata
Rank Member
Rank
Member
rasteri wrote on 2022-02-25, 17:53:

Haha that was all LGR's work. I dunno how he got 32bit mode working either TBH...

Maybe he just closed the group to act like there wasn't an issue there? Or maybe just ignored it somehow? OR maybe he actually googled for it like I should probably do right now but won't because I am too busy knee deep in DOS game folder renaming, which for some reason just sounds way more fun than doing actual real work like fixing drivers.

Reply 346 of 895, by rasteri

User metadata
Rank Member
Rank
Member

Yeah I'm not massively interested in perfecting windows support, in my mind it's a DOS gaming machine that uses Windows 98 as a launcher.

A true Windows 98 gaming machine would need 3D acceleration and that's never gonna happen.

Reply 347 of 895, by NachtRave

User metadata
Rank Member
Rank
Member
rasteri wrote on 2022-02-26, 11:58:

Yeah I'm not massively interested in perfecting windows support, in my mind it's a DOS gaming machine that uses Windows 98 as a launcher.

Honestly, which makes it even better at running DOS games than DOS. Because editing your main autoexec.bat and config.sys to get a thing to work inevitable breaks something before that that did work. Plus, you just need a simple GUI to access the stuff anyways. Doesn't even need to be Windows, it's just super convenient on top of it.

Another thing I noticed, actually, was there is an option in the BIOS (edit: Yeah, it's the Protect/Unprotect option of the SST via SPITOOL.EXE, along with the read-only FDD boot modes) about disabling writing to the HD in use. I found this particular feature really interesting because it basically makes it so that the contents of the drive stay the same upon each bootup, meaning that while you cannot necessarily "save a game" you avoid, entirely, corruptions of the FAT32 file system due to sudden system reboots and such. It wasn't uncommon in the past to have to "start over" a computer from a fresh install of Windows 98 - I remember it being almost a defacto standard practice of the time. A lot of that of course eventually went away, largely in part to the journaled file systems of EXT3/NTFS, also just better software support/less buggy software overall, but the ability to "time freeze" a drive like that is not just something that industrial applications can use. This helps with overall system stability, in general.

rasteri wrote on 2022-02-26, 11:58:

A true Windows 98 gaming machine would need 3D acceleration and that's never gonna happen.

You know, there is a guy who reversed engineered the Voodoo 5 6000 ( https://www.tomshardware.com/news/3dfx-voodoo … etter-than-ever ), and I bet he would have a lot of intimate knowledge of the VSA-100 chips that were the last ones 3dfx came out with. I bet you you may just be able to get one of those put together, too. I agree that it would be a pipedream to attempt, but hey, the Orpheus sound card exists because somebody said "Challenge accepted."

I'm not saying you should but if you did, you would be a God among us mere mortals.

Reply 348 of 895, by digger

User metadata
Rank Oldbie
Rank
Oldbie
rasteri wrote on 2022-02-26, 11:58:

A true Windows 98 gaming machine would need 3D acceleration and that's never gonna happen.

Well, the Vortex86DX SoC does happen to have a PCI bus, so never say never. 😉

NachtRave wrote on 2022-02-26, 20:05:

You know, there is a guy who reversed engineered the Voodoo 5 6000 ( https://www.tomshardware.com/news/3dfx-voodoo … etter-than-ever ), and I bet he would have a lot of intimate knowledge of the VSA-100 chips that were the last ones 3dfx came out with. I bet you you may just be able to get one of those put together, too. I agree that it would be a pipedream to attempt, but hey, the Orpheus sound card exists because somebody said "Challenge accepted."

I'm not saying you should but if you did, you would be a God among us mere mortals.

A VSA-100 chip integrated into a future version of the weeCee would be amazing indeed.

Two challenges, though:

  • Since the VSA-100 chip isn't being manufactured anymore, it would become increasingly hard to source. (It's hard already.)
  • I think the VSA-100 chip would blow up the power requirements of the weeCee considerably, possibly by an order of magnitude.

It does make one wonder, though: are there any GPU chips still being manufactured that can interface with a "classic" PCI bus and offer some level of hardware-accelerated 3D graphics? The chances seem slim, since such functionality is usually integrated in SoCs these days. It wouldn't make sense from a cost or complexity perspective to manufacture such chips separately.

So for a very niche retro application like this, the FPGA route seems like the way to go. 3dfx Voodoo technology is more than a quarter century old now. Would it really be so infeasible to recreate it in an FPGA design at this point?

A future version of the weeCee with an affordable-yet-sufficiently-powerful FPGA hooked up to the PCI bus of the Vortex86 SoC (and perhaps also sitting between the SoC and the VGA port, passing the regular VGA signal through by default) might be something worth designing.

Reply 349 of 895, by rasteri

User metadata
Rank Member
Rank
Member

I didn't mean a tiny windows gaming PC isn't't technically possible, it's just that using an vortex86 module would not be the way to go about it. There are already small motherboards with PCI slots that would make a much better base for a 3D accelerated mini PC, some even have onboard acceleration.

Reply 350 of 895, by NachtRave

User metadata
Rank Member
Rank
Member
rasteri wrote on 2022-02-27, 12:31:

I didn't mean a tiny windows gaming PC isn't't technically possible, it's just that using an vortex86 module would not be the way to go about it. There are already small motherboards with PCI slots that would make a much better base for a 3D accelerated mini PC, some even have onboard acceleration.

You know one of the libraries I manage is kind of in the same spot. There are better alternatives to what I have, and what I have was only ever meant to work on the first generation of this one IR sensor. It quickly became all way too involved, and the hardware just really wasn't there to support it. It isn't to say that it isn't possible with today's hardware, it's just to say that I didn't think it was really worth it because of how much extra time and effort I was putting into it.

Still, it would be really cool to pull off, I just never really gave my time to it after I got fed up with it the first time. I think it's okay to step away from something if you feel like there is just a deeper hole there than you felt comfortable with to begin with. At the same time, it doesn't mean it wouldn't be any less cool if it were to be pulled off.

You don't even need something highly complex, but I bet there is still enough space on the back of that board to maybe put in a 3d accelerator and a piece of RAM for it. Connect it up on the PCI bus. Something using some chip that is easy to come by and has a well documented and clean interface to make your job of hooking it up way easier. You never know what you can find until you go looking around.

I mean, you could name it the Super weeCee or something clever like that.

Reply 351 of 895, by Mu0n

User metadata
Rank Member
Rank
Member
NachtRave wrote on 2022-02-25, 15:19:

Also, @Muon/@Rasteri, how did you get the IDE controllers, under Device Manager, to not be in compatibility 16-bit mode? Was there something special you did there, or? I tried to follow your earlier discussion but I can't seem to find the answer there. I noticed in LGR's videos that there isn't a caution! warning, so you must have figured something out?

I went back and looked at his video. He never expands the node for the IDE controllers, so he never shows that there's no question mark. iirc. My thinking is that he also has the warning and he's also suffering from the same problem.

1Bit Fever Dreams: https://www.youtube.com/channel/UC9YYXWX1SxBhh1YB-feIPPw
DOS Fever Dreams: https://www.youtube.com/channel/UCIUn0Dp6PM8DBTF-5g0nvcw

Reply 352 of 895, by NachtRave

User metadata
Rank Member
Rank
Member
Mu0n wrote on 2022-03-08, 14:49:

I went back and looked at his video. He never expands the node for the IDE controllers, so he never shows that there's no question mark. iirc. My thinking is that he also has the warning and he's also suffering from the same problem.

Yeah, I figured he may of tried to force select a different driver and maybe had some success with such. I think this is probably easily the biggest Windows 98 issue (outside of a lack of 3D accelerator) that is probably holding the weeCee back. I really wish ICOP would either have an explanation as to how to make these IDE drivers work correctly in Win98 (even if it means force selecting different but compatible hardware) or write some compatible drivers. Alas, I understand rasteri's intention about it primarily a DOS machine, but maybe I am just being too optimistic. Either way, after exhausting every possible BIOS configuration option and Win98 file system properties I can think of, I have not been able to resolve the 16-bit mode drivers either. The only thing I haven't tried doing is starting to go through the entire base list of Win98 drivers and force selecting ones to see if any alternatives may work. Alas.

If anyone wants to crack that problem open and figure out a solution, I think that would be pretty incredible a find. Perhaps we could start looking at what kind of IDE controller it has and what, if any, architecture or make/model it's based off of? Perhaps this is a 640k!isenough level question?

Reply 353 of 895, by 640K!enough

User metadata
Rank Oldbie
Rank
Oldbie
NachtRave wrote on 2022-03-10, 10:30:

Perhaps we could start looking at what kind of IDE controller it has and what, if any, architecture or make/model it's based off of? Perhaps this is a 640k!isenough level question?

As I've said before, the cost of entry for this project is far too high to be worth considering for me, especially since there is no guarantee that I would be able to do anything useful to solve this problem.

My familiarity with the CS4237 came from another project that had much lower costs associated with it. Just the import fees on this one are out of the question.

Reply 354 of 895, by NachtRave

User metadata
Rank Member
Rank
Member
640K!enough wrote on 2022-03-16, 20:31:

As I've said before, the cost of entry for this project is far too high to be worth considering for me, especially since there is no guarantee that I would be able to do anything useful to solve this problem.

Not even a worry, it is more likely an ICOP thing anyways. I have even grabbed a few books on VXD coding for Win9x, and it's some of the most contorted C++ code I've seen in a good while. Even just wrapping one's head around an entire IDE controller, let alone one meant for talking to a microSD card, let alone one that is not very well documented, just seems like a pretty crazy amount of effort. Definitely would need someone who has worked on such things before and is able to familiarize themselves with the SD card controller of the SOM (which I have found very little published for). That's easily a big time amount of effort, and who knows, maybe ICOP didn't put the effort into writing one because the gains wouldn't be enough for it to matter. There are a lot of unknowns, so.

I agree it's definitely a high cost barrier. On the flip side though, if you ever wanted a weeCee to test with.... :p I actually just finally got the non-SST NAND Flash version bought from DSL/ICOP in the UK, and to get just that one SOM shipped to here in the US, wow! 279 pounds they ask for (that's including 155 pounds for the SOM, and 124 pounds for the "Export Carriage, Insurance & Packing"), but after you add in wire transfer fees and everything else, a 180 USD SOM turns into 458.68 USD expense. The ones you can buy here in the US, from WDL, they only offer the SOM in the form that includes the SST NAND Flash module. So, hence where I am coming in from.

Reply 355 of 895, by 640K!enough

User metadata
Rank Oldbie
Rank
Oldbie
NachtRave wrote on 2022-03-17, 17:47:

I agree it's definitely a high cost barrier. On the flip side though, if you ever wanted a weeCee to test with.... :p I actually just finally got the non-SST NAND Flash version bought from DSL/ICOP in the UK, and to get just that one SOM shipped to here in the US, wow! 279 pounds they ask for (that's including 155 pounds for the SOM, and 124 pounds for the "Export Carriage, Insurance & Packing"), but after you add in wire transfer fees and everything else, a 180 USD SOM turns into 458.68 USD expense.

Ouch! No thanks. I'll tell you what: when I win the lottery, I'll make sure to build one of these and see if I can't be useful somehow. 😀

For the time being, I did a little digging that might lead one of you to to find a workable driver. Apparently, there are different revisions of the SOM. The later revisions are apparently compatible with the standard Intel PIIX ATA drivers. Anyone with the older revision isn't so lucky. They didn't design that part from scratch, however; apparently it uses a D1011 or D1012 from RDC Semiconductor (language selection drop-down in the top right). I don't know if a driver for that specific chip ever existed for Windows 9x, though I suppose you could write to them and ask.

Apparently, that controller is also compatible with the ITE Tech IT801x series, specifically the IT8011 or IT8012. Some of my search results seem to suggest that a Win9x driver did exist, but I am not in the mood to try and weed out the shady sites from those that might actually provide a malware-free download. If you can locate those drivers and get them working (you may need to force installation, as the RDC device IDs may not be present), it seems it may not be a good idea to enable UDMA support, so try at your own risk; it would still be better having a native driver than resorting to compatibility mode.

Reply 356 of 895, by rasteri

User metadata
Rank Member
Rank
Member

Damn it, I know I said I wasn't interested in fixing this but....

640K!enough wrote on 2022-03-18, 07:45:

Apparently, that controller is also compatible with the ITE Tech IT801x series, specifically the IT8011 or IT8012.

Yeah I found the docs that say the IDE controller is an IT8011 - I can't find anything about that chip. Could it possibly mean the IT8211?

I found the IT8211 drivers on archive.org :
http://web.archive.org/web/20121205120458if_/ … ad.ashx?file=13

I can force them to install by adding the hardware info to the inf file :

%ITE.DeviceDesc0% = iteatapi, PCI\VEN_17F3&DEV_1011&SUBSYS_101117F3

But it doesn't like it, gives the standard "device is not present, not working properly, does not have all drivers installed" (etc) (Code 10) error.

Also weirdly it shows up under "SCSI controllers"

Also tried a couple other versions of the 8211 drivers and even the 8212 driver (although I think it's more of a RAID controller)

Reply 357 of 895, by 640K!enough

User metadata
Rank Oldbie
Rank
Oldbie
rasteri wrote on 2022-03-18, 13:18:

But it doesn't like it, gives the standard "device is not present, not working properly, does not have all drivers installed" (etc) (Code 10) error.

Also weirdly it shows up under "SCSI controllers"

Also tried a couple other versions of the 8211 drivers and even the 8212 driver (although I think it's more of a RAID controller)

It could be that this applies only to Linux, as the driver for the ITE chips used to be shared with the RDC chip upon which the SOM's functionality is based. However, this was split off when the work-around for a chip bug (DMA failures/corruption) in the RDC version was improved. In the Windows world, extra initialisation logic may be required to get the RDC version into a usable state, which the ITE driver would likely lack.

The other possibility is that you haven't found the right driver yet. There were suggestions that the ITE chip can be used either as a standard ATA controller or as a RAID controller, possibly with separate chip revisions and/or drivers. I didn't really look into that. It could also be that no ITE driver will work with the RDC version.

Sadly, with the RDC chip being a more recent design (compared to Win9x), I am not sure a driver was ever developed or released. Unless someone wants to contact the company to ask, or write one themselves, you may be out of luck.

Reply 359 of 895, by Mu0n

User metadata
Rank Member
Rank
Member

I thought I was knowledgeable about computers but looks like a hard NOPE after all.

1Bit Fever Dreams: https://www.youtube.com/channel/UC9YYXWX1SxBhh1YB-feIPPw
DOS Fever Dreams: https://www.youtube.com/channel/UCIUn0Dp6PM8DBTF-5g0nvcw