pentiumspeed wrote on 2023-02-20, 17:35:

Again that is non-sense. There is no new allocations or something. To make the compatibility to work the VLB i/o board had to be exactly same design as ISA in hardware, irq and addresses as ISA so DOS and windows to work otherwise would require a IDE VLB driver for the VLB IDE chip.
If the MPU-IPC-T worked while the hardMPU failed then something is not quite right with HardMPU. This needs to be debugged properly and revise the HardMPU's hardware or firmware. There was same issue on a clone soundcard and the person who designed this found the issue in firmware for the sound card clone and fixed it.

Cheers,

All I can say is what I am observing. Hence the comment of 'I think'. It does appear to be something with the way the HardMPU gets the resources though. On one of the Socket 7 boards I have I couldn't get the HardMPU to work until I changed the IRQ to 3. Then it worked okay. Thinking this was strange I went into the BIOS and I saw that the first PCI IRQ allocation was IRQ9 and then incremented by one for each PCI slot through PCI Slot 4. On a curiosity I increased the IRQ for each PCI slot by one increment, so instead of starting at IRQ9, it started at IRQ10. After doing this the HardMPU worked at IRQ2/9. I don't know why this was. PCI slot 1 was not populated with a card, so why would the HardMPU would fail to use IRQ9? When I changed the PCI Slot IRQ it started working. I don't know enough about system resource allocation and resource conflict to know why it didn't work. Only that it didn't. I would have guessed that since there was nothing using the PCI slot that IRQ9 would be free to use for whatever. At least that was how I thought resource allocation worked on PCs outside of hard fixed addresses. But for the HardMPU card this didn't work. If you know what could be causing this I would be really interested in knowing. It seems very strange to me.

nathanieltolbert wrote on 2023-02-21, 19:14:

pentiumspeed wrote on 2023-02-20, 17:35:

Again that is non-sense. There is no new allocations or something. To make the compatibility to work the VLB i/o board had to be exactly same design as ISA in hardware, irq and addresses as ISA so DOS and windows to work otherwise would require a IDE VLB driver for the VLB IDE chip.
If the MPU-IPC-T worked while the hardMPU failed then something is not quite right with HardMPU. This needs to be debugged properly and revise the HardMPU's hardware or firmware. There was same issue on a clone soundcard and the person who designed this found the issue in firmware for the sound card clone and fixed it.

Cheers,

All I can say is what I am observing. Hence the comment of 'I think'. It does appear to be something with the way the HardMPU gets the resources though. On one of the Socket 7 boards I have I couldn't get the HardMPU to work until I changed the IRQ to 3. Then it worked okay. Thinking this was strange I went into the BIOS and I saw that the first PCI IRQ allocation was IRQ9 and then incremented by one for each PCI slot through PCI Slot 4. On a curiosity I increased the IRQ for each PCI slot by one increment, so instead of starting at IRQ9, it started at IRQ10. After doing this the HardMPU worked at IRQ2/9. I don't know why this was. PCI slot 1 was not populated with a card, so why would the HardMPU would fail to use IRQ9? When I changed the PCI Slot IRQ it started working. I don't know enough about system resource allocation and resource conflict to know why it didn't work. Only that it didn't. I would have guessed that since there was nothing using the PCI slot that IRQ9 would be free to use for whatever. At least that was how I thought resource allocation worked on PCs outside of hard fixed addresses. But for the HardMPU card this didn't work. If you know what could be causing this I would be really interested in knowing. It seems very strange to me.

Documentation of your issues is important to state everything to make sense. I will ask, what is the motherboard and the chipset it has?

Thanks for replying, cheers.

I will go through all of the boards I have and note their chipsets. Of the two PCI 486 boards I have that I have tested with they are both SIS496/497 I believe? Socket 7 boards I have tested with are 4 separate PA-2010+ boards. They have the VIA VT82C585VPX and VT82C586A. And they all work. But I don't know if that is because I use Windows 95/98 on them or something else
3 VLB boards use a chipset of UMC UM898F - These are the Markvision Flex boards rebadges from Techmedia - https://theretroweb.com/motherboards/s/biosta … 14-umc498f-2vlb
1 VLB board uses a chipset of OPTI 82C495 SX which has stopped functioning. - This is a Shuttle HOT-409 it no longer identifies RAM. The ram just doesn't seem to lock into the simm sockets now - https://theretroweb.com/motherboards/s/shuttl … ot-409-ver.-2.2
1 VLB board uses a chipset of SiS 85C471 - This is an ASUS 486SV2G that is one of the later 486SV2GX4 boards that doesn't have the 3.3V circuitry populated So the board looks like this without the VRM and resistors and parts - https://theretroweb.com/motherboards/s/asus-vl-i-486sv2gx4

Those are the numbers on the largest chips. If this is not the correct information, please let me know which other chips I need to pull the information from so I can get it for you.

In regards to the PCI board with the strange PCI allocation that was one of the 2 that uses 496/497 chipset. The motherboard is a LS-486E motherboard. It's a Rev C2. Here's the Retroweb link to the board - https://theretroweb.com/motherboards/s/lucky- … e-rev:c1-rev:c2

The other PCI board is an A-Trend ATC1425A motherboard. Here's the board info on Retroweb - https://theretroweb.com/motherboards/s/a-trend-atc-1425a

-Edit - It took me a while to get to the storage unit and check all of the boards. I don't know if any of the information I have provided will help, but this is what I have tested with. I may be backward on the pci boards on which does the weird IRQ allocation issue. I unfortunately don't have any of the HardMPU boards that I bought and built. I gave them to other retro streamers that are my friends so they could use midi hardware. I only have my MPU-401IPC-T card that someone here on Vogons was gracious enough to sell me 10 years ago or so now. Thank you still to this day.

Does the HardMPU-WT support MIDI in? I'm looking for a host for a Turtle Beach Rio that uses that pin for feedback.

I'm not 100% certain, but I don't think the HardMPU-WT supports midi in. But I of course could be wrong. I think it was mentioned earlier in the thread whether it does or doesn't though.

Ok so to be clear: HardMPU (WT) as-is doesn't support MIDI In. The project page states "playback only" and in another talk I had with somebody knowledgeable on the subject I was also made aware that it's playback only.

nathanieltolbert wrote on 2023-01-17, 00:28:

mbarszcz wrote on 2022-12-22, 04:50:

I can give you a rundown of how I had mine made and which options I used if you like.

I have 3 of them in different PCs and they're all working great.

I would love to get a rundown on this. I have never made my own run of PCBs before, and so any tips and suggestions is super appreciated.

Sorry it's been so long, but I figured I'd still reply anyway in case anyone is looking for the information.

I ordered my PCBs from JLCPCB, and since these PCBs are ISA cards, I selected a few extra options. Unforunately some of them makes the PCBs cost a bit more than the default cheap options, but it also makes it more durable and better suited to being an ISA card.

Via Covering: Tented (Standard), this covers the vias with solder mask so they are not exposed
Deburring/Edge rounding: No (Not necessary here)
Gold Fingers: Yes、30° (This provides a thicker, more durable, and corrosion resistant coating on the pads that go into the ISA slot and chamfers the edges at 30° for easier insertion. If you don't get gold fingers, the lead plated contacts won't hold up very well being inserted/removed from an ISA slot).
Surface Finish: ENIG Gold Thickness: 1 U" (Standard is HASL/Lead). This provides a very smooth, flat, and durable gold finish on the pads (and is requires because of the gold fingers)
PCB Thickness: 1.6 (Standard)

The board cost for 5 of these was $26.04.

To order them, download all the files in the gerber directory from the github repo: https://github.com/PickledDog/hardmpu-wt/tree/main/gerbers, put them into a zip file and upload them to JLCPCB (or another manufactuer of you choice).

Select the options you want, and check out.

I got together with several friends and did a run on making PCBs and the components needed to build the HardMPU-WT boards as well. I spoke directly with PickledDog, and they were super nice and really helpful in getting me the information I needed to be able to submit the files to get the boards made, as I had never done one before. Really great person. Afterward, they also updated their the repository so I had the information I needed to be able to program the ATMega chips with my TL866 programmer, which is just fantastic. The only thing I didn't have was I didn't realize that it uses SB-Midi cables. So I now need to make some SB-Midi Cables. PickledDog also has a PCB with some minimal surface mount soldering to make the cable. I need to get an order of those done as well so I can make the cables. I just haven't had the money or the time lately. I have a bunch of boards sitting here waiting for my left hand to recover from surgery, and I am dominant Left Handed.

Hello,

I am having a weird issue with HardMPU v1.1 and MT32 (Old version) in Wing Commander 2. When using the HardMPU, some instrument are wrong. It's not a clipping noise, just the instrument playing sounds wrong. If I use the SoftMPU with my sound card or the PCMIDI from my Orpheus I card, I don't get the problem.

Is there a Firware upgrade that could fix the issue?

Thanks.