jamesfmackenzie wrote on 2025-04-08, 22:18:

What a fantastic result! 😎🏆
Do you think the adapters will be sold on eBay too? I have a Supermicro motherboard ready to go!

I encourage my friend to do so. Close to the summer I suppose.
I also may recommend you stick with this Z87M-Pro4 board because of overclocking, PCI availability and lack of IRQ conflicts: both 5 and 7 are available - this maybe PITA with other boards. And as a dialup fan with lot of modems, this board has a COM header.
With this board soldering point for LDRQ1 quite apparent. I have also Gigabyte and Asrock B85M's (D3H rev 1.x and Pro4) with PCI slots with LDRQ1 also soldered and proven works. They both support K CPUs overclocking.
Those boards are quite available and attractive for 30-50$ w/o and w/CPU, so I bought another two $10 w/CPU H61 MSIs (2xPCIe 1x, no PCI), both works with adapter and DOA Asus B86M-E .
Check boardview for LDRQ1 availability before you buy.
In most cases if the board you evaluating has SuperIO of some sort, there will be high chance LDRQ1 availability, if not - very low. This is not binary rule, but very close to. I spent almost a week evaluating boards and looking for/in boardviews and the conclusion was as stated.

rasteri wrote on 2025-04-08, 21:58:

Very nice work! I might grab a Asrock Z87M-Pro4, I haven't been able to find many motherboards that work and I kinda want one with builtin PCI

Thanks man. This is all yours efforts encourage us. I'm eyeing this project for past two years having no soldering skills and board printing and hoping the day will come.

You may maintain some MB compatability list. It would be nice to have it at one place (wiki/github/...) instead scattered over this huge thread. Just a simple table with MB type, location of LDRQ, free IRQ availability and test status.

I personally prefer Gigabyte MBs as some are rich of legacy stuffs like PCI, COM and LPT port. Many others may have COM but LPT not often. I also found classic award BIOS on gigabyte 6x series a bit better with more legacy options than UEFI ( I have one P67 MB that has available both award and beta AMI UEFI so I tried to flash it and back...

Great idea - perhaps something for Vogons Wiki?

For awareness: I have a Supermicro X10SAE (C226 chipset) and looks like it has LDRQ# on the TPM header already.

RayeR wrote on 2025-04-09, 12:43:

You may maintain some MB compatability list. It would be nice to have it at one place (wiki/github/...) instead scattered over this huge thread. Just a simple table with MB type, location of LDRQ, free IRQ availability and test status.

I personally prefer Gigabyte MBs as some are rich of legacy stuffs like PCI, COM and LPT port. Many others may have COM but LPT not often. I also found classic award BIOS on gigabyte 6x series a bit better with more legacy options than UEFI ( I have one P67 MB that has available both award and beta AMI UEFI so I tried to flash it and back...

jamesfmackenzie wrote on 2025-04-09, 14:44:

For awareness: I have a Supermicro X10SAE (C226 chipset) and looks like it has LDRQ# on the TPM header already.

I would take SM docs with grain of salt. They're copypasting LPC header pinout from revision to revision. I even saw Z170 overclocking gaming board with LDRQ LPC header.
As for WIki I'm in, I may post imagery with LDRQ1 solder points for mobos I have already proven works.
Just noted, that Necromancer's DOS navigator has IRQ vectors information and indeed IRQ 5 and 7 are free to use on Z87m-Pro4. Nice tool to check out what's available in UEFI times.

I didn't explore the Vogons wiki, it seems not much there, like 1% of content of forum threads. And it seems I need to get special permission/extra account other than standard account for forum login, am I right? But I think it's right place to be there, where it can be maintained/contributed by multiple users...

Hm, I'm not sure how this tools are reliable about IRQ free. HWinfo for DOS also displays IRQ usage. When I enable/disable LPT in ECP mode (with DMA) it reflects the change on IRQ5. BTW I still didn't find some relevant info if possible/how to override BIOS PnP enumerator who allocates IRQ resources (Windows sometimes can do this some way so it should be possible change it after boot). As there's no longer "reserve IRQx for legacy" option in SETUP it tends to allocate free IRQ5/7 to other PCI devices like USB *HCI, SATA, ethernet, etc. As I wrote, in case you have LPT that assigns IRQ5 for it and then you can reprogram the SuperIO to disable LPT completly and let IRQ5 be free. But for other MBs it may be problem...

RayeR wrote on 2025-04-09, 15:20:

I'm not sure how this tools are reliable about IRQ free.

ndn portrays vectors (F000:Fxxx for 5 and 7) while others pointed to somewhere else. I'll check them out latter in memeditor, what actually at that place, maybe some NOPs and RETF.

Okay, little chatgpt and ramview for Z87M-Pro4 CSM interrupt table:
IRQ 7 pointed to F312C which is IRET (CF)
IRQ 5 pointed to FFF54 which is JMP to EFA8A (E9 33 EA, offset -15CD) to zeroes nonsense
So they both free.
Later I realized IRQ 3 & 4 "partially" free (IRQ4 = COM1 onboard not used AND IRQ3 pointer points to the same address), coz there's no hardware callers connected and ready for action just in case.

Hm, maybe it's some default dummy vector. Also you may try if the HW is configured to actually generate the IRQ, it goes through some complicated routing and can be masked somewhere...

RayeR wrote on 2025-04-10, 16:21:

Hm, maybe it's some default dummy vector. Also you may try if the HW is configured to actually generate the IRQ, it goes through some complicated routing and can be masked somewhere...

Yeah, IRQ 3 & 4 are bound to the superIO. Duke3D setup says about IRQ conflict.

vsharun wrote on 2025-04-10, 16:34:

Yeah, IRQ 3 & 4 are bound to the superIO. Duke3D setup says about IRQ conflict.

I just checked the manual of Z87M Pro4. It has two serial (one normal, one infrared) and one parallel port headers. So normally all relevant IRQs (3, 4, 7) would be already taken.

Try disabling the port headers from BIOS you don't use nor connect. That may free the IRQs in question, which may then be grabbed by certain IRQ lines if it prefers those to the one it's currently using.

However, leaving the parallel port enabled with IRQ7 may kind of allow sound cards to work with that, if it behaves just like old motherboards where ISA sound cards were allowed to share IRQ7 with parallel port to some extent.

Still, you may want to use tools like PCICFG to inspect if any onboard PCI device is on its own IRQ line (perhaps onboard audio). Disabling such devices can also influence the system PIRQ assignments, though you won't have much control over the result.

EDIT: Just read your previous posts. If neither IRQ5 nor IRQ7 are being claimed by any PIRQ line (thus actually free) on this board, maybe its BIOS is indeed good enough to not make any PIRQ line prefer those IRQs.

LSS10999 wrote on 2025-04-11, 01:25:

I just checked the manual of Z87M Pro4. It has two serial (one normal, one infrared) and one parallel port headers. So normally […]
Show full quote

vsharun wrote on 2025-04-10, 16:34:

Yeah, IRQ 3 & 4 are bound to the superIO. Duke3D setup says about IRQ conflict.

I just checked the manual of Z87M Pro4. It has two serial (one normal, one infrared) and one parallel port headers. So normally all relevant IRQs (3, 4, 7) would be already taken.

Try disabling the port headers from BIOS you don't use nor connect. That may free the IRQs in question, which may then be grabbed by certain IRQ lines if it prefers those to the one it's currently using.

However, leaving the parallel port enabled with IRQ7 may kind of allow sound cards to work with that, if it behaves just like old motherboards where ISA sound cards were allowed to share IRQ7 with parallel port to some extent.

Still, you may want to use tools like PCICFG to inspect if any onboard PCI device is on its own IRQ line (perhaps onboard audio). Disabling such devices can also influence the system PIRQ assignments, though you won't have much control over the result.

EDIT: Just read your previous posts. If neither IRQ5 nor IRQ7 are being claimed by any PIRQ line (thus actually free) on this board, maybe its BIOS is indeed good enough to not make any PIRQ line prefer those IRQs.

Thanks.
Enabling or disabling in UEFI one/all COM's changes nothing. Those IRQ3/4 seems hardwired to the SuperIO. Ultimately, both IRQ5/7 are available while only one required. Also enabling/disabling LPT changes nothing.
Just in case - USB mouse (MS wireless mobile mouse 1000, small notebook sized mouse) works perfectly, no lag and identified as a PS/2 by Cute Mouse.
Still jemmex best in compatibility. Slowdowns, hangs and "not enough memory" with himem/himemx. Will try harder next days.

BTW is there any ways to unhardwire IRQ3/4, make them shareable like IRQ7 ? I assume is one IRQ can be shared = mean most can be in one way or another.

vsharun wrote on 2025-04-11, 06:19:

Thanks. Enabling or disabling in UEFI one/all COM's changes nothing. Those IRQ3/4 seems hardwired to the SuperIO. Ultimately, bo […]
Show full quote

Thanks.
Enabling or disabling in UEFI one/all COM's changes nothing. Those IRQ3/4 seems hardwired to the SuperIO. Ultimately, both IRQ5/7 are available while only one required. Also enabling/disabling LPT changes nothing.
Just in case - USB mouse (MS wireless mobile mouse 1000, small notebook sized mouse) works perfectly, no lag and identified as a PS/2 by Cute Mouse.
Still jemmex best in compatibility. Slowdowns, hangs and "not enough memory" with himem/himemx. Will try harder next days.

BTW is there any ways to unhardwire IRQ3/4, make them shareable like IRQ7 ? I assume is one IRQ can be shared = mean most can be in one way or another.

An IRQ claimed by a PCI IRQ line will be shared by all PCI devices on that same line (e.g. INTA#), and ISA devices will not be able to use it properly if that's the case.

By "changes nothing" you mean it still complains about IRQ conflict, or it no longer complains but things don't work correctly?

If it's the former, you may try checking with HWINFO and see what it reports. It's possible that IRQ 3/4 may have been taken by a PIRQ line after you disabled the serial or infrared port. If HWINFO can't provide reliable info you may use Ralf Brown's PCICFG utility which can dump info about all PCI devices (onboard and discrete) and you can check if the IRQ in question is indeed used by PCI devices.

If it's the latter, you probably need to manually configure the registers related to the particular IRQ (setting up interrupt vectors etc.), as modern BIOSes may not automatically do so, or set garbage values.

LSS10999 wrote on 2025-04-11, 08:12:

An IRQ claimed by a PCI IRQ line will be shared by all PCI devices on that same line (e.g. INTA#), and ISA devices will not be a […]
Show full quote

An IRQ claimed by a PCI IRQ line will be shared by all PCI devices on that same line (e.g. INTA#), and ISA devices will not be able to use it properly if that's the case.

By "changes nothing" you mean it still complains about IRQ conflict, or it no longer complains but things don't work correctly?

If it's the former, you may try checking with HWINFO and see what it reports. It's possible that IRQ 3/4 may have been taken by a PIRQ line after you disabled the serial or infrared port. If HWINFO can't provide reliable info you may use Ralf Brown's PCICFG utility which can dump info about all PCI devices (onboard and discrete) and you can check if the IRQ in question is indeed used by PCI devices.

If it's the latter, you probably need to manually configure the registers related to the particular IRQ (setting up interrupt vectors etc.), as modern BIOSes may not automatically do so, or set garbage values.

Thanks. Will check it out next days.

RayeR wrote on 2025-04-09, 12:43:

You may maintain some MB compatability list. It would be nice to have it at one place (wiki/github/...) instead scattered over this huge thread. Just a simple table with MB type, location of LDRQ, free IRQ availability and test status.

I'm compiling a list on Github here (fork from original repository): https://github.com/AranVink/dISAppointment
It's still a work in progress. Once I'm happy with the state I'll make a pull request towards rasteri.
Source is me going through some of the Vogons pages and making notes last few months, plus research on boards I encounter online. It's tedious and error-prone, so please make an edit if you see any mistakes or omissions. Also feel free to copy it over to the Vogons wiki if that makes more sense. (I started going down this rabbit-hole seeing the documentation on GitHub being only one line 😉 )

Here's LDRQ#1 location for Asrock B85M-Pro4, Gigabyte B85M-D3H rev 1.0 (I strongly recommend stick with rev 1.x because of better VRM), MSI MS 7788 1.0
Also MS7740 v1.2 LDRQ#0 hijack from superIO. Black wire and glued dots. Asrock Z87M-Pro4 LDRQ location in the post - Asrock's B85 & Z87 M's has the same place for LDRQ#1, looks like those boards are siblings, despite (slightly) different chipsets.

BTW Shuttle DS61 has LPC header with LDRQ.

Metalliferous wrote on 2025-04-12, 20:26:

I'm compiling a list on Github here (fork from original repository): https://github.com/AranVink/dISAppointment It's still a wor […]
Show full quote

RayeR wrote on 2025-04-09, 12:43:

You may maintain some MB compatability list. It would be nice to have it at one place (wiki/github/...) instead scattered over this huge thread. Just a simple table with MB type, location of LDRQ, free IRQ availability and test status.

I'm compiling a list on Github here (fork from original repository): https://github.com/AranVink/dISAppointment
It's still a work in progress. Once I'm happy with the state I'll make a pull request towards rasteri.
Source is me going through some of the Vogons pages and making notes last few months, plus research on boards I encounter online. It's tedious and error-prone, so please make an edit if you see any mistakes or omissions. Also feel free to copy it over to the Vogons wiki if that makes more sense. (I started going down this rabbit-hole seeing the documentation on GitHub being only one line 😉 )

Hi,
You may notice (X) in the mobo manual for X10 series, the same (X) i saw for their gaming Z170 overclocking mobo (imagine Supermicro and gaming, right ?)
I have X8SIL-F for years, and this mobo has terrible DOS support, no PCI slots, integrated VGA on PCIe 1x suffers from typical DOS workloads (low speed and compatibility). I would not recommend going Supermicro way in this (DOS retro) way.

Metalliferous wrote on 2025-04-12, 20:26:

I'm compiling a list on Github here (fork from original repository): https://github.com/AranVink/dISAppointment
It's still a work in progress. Once I'm happy with the state I'll make a pull request towards rasteri.
Source is me going through some of the Vogons pages and making notes last few months, plus research on boards I encounter online. It's tedious and error-prone, so please make an edit if you see any mistakes or omissions. Also feel free to copy it over to the Vogons wiki if that makes more sense. (I started going down this rabbit-hole seeing the documentation on GitHub being only one line 😉 )

Yeah this is a great idea, once I get some free time I'll contribute info from the boards I have

I'm currently waiting on a new response from my friend and Fintek about the previously talked about PCIe/ISA chip F85526. The design might be finalized as it's "possible" to buy F85526 and F85526D from china apparently. Friend has yet to hear back from Fintek or distributors from china though. (This chip has apparently been "around" since 2014 but the design has never been finalized and has never been for sale until recently (past few months))

I'm still not sold that it does what it says.

~

On a whim I checked out my wife's Gigabyte z390 Aorus motherboard which has a LGA1151 socket (intel coffee lake), PCH Z390 and ITE8688 (sio). Apparently the ITE8688 does have LDRQ# and connects to PCH z390. Though this is not exposed on the TPM header, the rest of the TPM header seems all there to support the dISAppointment. I'll have to build a dISAppointment to test it, but everything on the surface seems fine. I think this would be the most recent intel board that would have it? Unless i'm missing something here. https://www.slideshare.net/slideshow/z390-des … 10pdf/259237641

~

Since no one is selling boards (Unless im wrong, would love a link), It might be worth it to add a dISAppointment shared project on PCBWay/JLCPCB and/or electromaker for ease of purchase for anyone attempting who's not familiar with ordering pcbs from fabs. I doubt they would have the components for assembly, but would make it easier for pcb acquisition.