Well if you had the place, you simple could use 2 sound chips on the bord, 1 for windows and 1 for dos.
Maybe if the cpu got a sound onboard, or to use sound chip on the vodoo/vga card, if you don't have the space for it. Under windows it would not care if it's on a other bus than a ISA bus.

Eivind wrote on 2024-10-19, 20:01:

Been kinda busy these last days, but here's a quick update of my findings: - The TL3.3 board looks alright, it does work (like t […]
Show full quote

Been kinda busy these last days, but here's a quick update of my findings:
- The TL3.3 board looks alright, it does work (like the previous prototypes) and I can play games on it! 😀
- The MXM connector works with sdz's Voodoo4 card, I'm getting VGA and HDMI through MXM and out from the respective connectors on the TL board!
- My Mini-PCIe-to-MXM adapter works fine as well, tested with my SM750-based card.
- I think there's some kind of bug with my power circuit - It takes two pushes of the power button to turn on the board, one to power it down. I tried two different USB-C PD power bricks, identical result. Will investigate this further.
- My CM4 with eMMC works just fine even though the board also has an external SD card slot for the Pi. I guess that means you can use either the eMMC-equipped CM4 or the Lite version - I need to get one of those for testing.
- Haven't gotten to experimenting with the I2C IO expander and diagnostic LEDs yet.
- The SATA connector might not have been the best idea. I kinda forgot that SATA devices also require "SATA power", and there's obviously no ATX power supply with these power connectors at play here. What might be a much smoother solution is using an mSATA or M.2 connector. Problem is, there's no room currently on the board for such a connector. A good quality microSD card works very well though, so not sure it's worth pursuing really.
- I don't like the position of the power- and reset buttons. Will have to think long and hard about those, I do want them to be 90-degree ones, but they all have through-hole pins or stabilizer studs which would interfere with the MXM connector on the bottom side.
- My biggest concern right now, however, is PicoGUS's apparent lack of Win98 support. It works amazingly marvellous in DOS, but I'm really struggling getting either the GUS or SB mode working in Windows. Wish it had a WSS mode... 😁 Any and all suggestions are welcome!

Edit: a little bird told me that a WSS mode might be in the cards...

centurion2024 wrote on 2024-10-24, 22:20:

Well if you had the place, you simple could use 2 sound chips on the bord, 1 for windows and 1 for dos.
Maybe if the cpu got a sound onboard, or to use sound chip on the vodoo/vga card, if you don't have the space for it. Under windows it would not care if it's on a other bus than a ISA bus.

The voodoo4 card could (in theory at least) have a PCI sound chip. However, the mini-PCIe cards use PCIe directly without a PCIe-to-PCI converter IC. PCIe is point-to-point, so this solution is not something I'd want to consider unless all other options are exhaused. I'll wait and see what might be possible with the rp2040.

Brain-dump update:
- I think I've figured out the cause of the double-press to power on issue, I was asserting/deasserting the PD controller's EN pin, apparently confusing it, instead of just enabling/disabling the main power mosfet.
- To solve both issues with power/reset button positions and SATA, I'll try to make a small riser board to sit above portions of the front area of the main board (above the sd card slots). This riser board will have buttons, probably some of the LEDs and an M.2 SATA connector for a 2230 or 2242-sized card (tbd).
- In order to more easily be able to play with different ISA sound solutions, I'm contemplating moving all the audio logic to a separate slim board, connected using the same Hirose board-to-board connectors the CM4 uses. There should be enough room underneath the MXM card. Mixer, opamp, etc can stay on the main board.

Eivind wrote on 2024-10-25, 23:10:

Brain-dump update: - I think I've figured out the cause of the double-press to power on issue, I was asserting/deasserting the P […]
Show full quote

Brain-dump update:
- I think I've figured out the cause of the double-press to power on issue, I was asserting/deasserting the PD controller's EN pin, apparently confusing it, instead of just enabling/disabling the main power mosfet.
- To solve both issues with power/reset button positions and SATA, I'll try to make a small riser board to sit above portions of the front area of the main board (above the sd card slots). This riser board will have buttons, probably some of the LEDs and an M.2 SATA connector for a 2230 or 2242-sized card (tbd).
- In order to more easily be able to play with different ISA sound solutions, I'm contemplating moving all the audio logic to a separate slim board, connected using the same Hirose board-to-board connectors the CM4 uses. There should be enough room underneath the MXM card. Mixer, opamp, etc can stay on the main board.

I think that would be super awesome to have the sound card to become a card. That way you can update the sound card independently with all kinds of options.
If you choose that route, please consider all kinds of options (with wave table and without) pins so we can come up with all kinds of cool options.
We might be able to even have versions based on transferring ASICS from existing sound cards. This will also be a nice upgrade/evolution for the Llama Blaster project.
Man, this project is evolving... I love it

Just came from dusting off my TinyLlama 3 components laying on the corner of my bench... 😀

I have day job running IT projects and Scope Creep is clear risk.
Have you considered forking base version into some sort of public V1 Limited run stable version? To give vogoners with higher risk appetite possibility to start playing around with wide range of legacy software few dedicated beta testers will not be able to cover.

ahmadexp wrote on 2024-10-26, 19:46:

If you choose that route, please consider all kinds of options (with wave table and without) pins so we can come up with all kinds of cool options.

What did you have in mind? Whatever MIDI signal is coming out of the sound card is gonna be routed to the CM4 to be synthesized there. Were you talking about also exposing a full-sized wavetable header? Setting aside the issue of finding space on the board for this, I kind of feel it wouldn't be needed seeing how capable and versatile the CM4 is. Do you disagree?

vutt wrote on 2024-10-26, 20:18:

Just came from dusting off my TinyLlama 3 components laying on the corner of my bench... 😀

How could you have gathered the components needed for the TL3 before I've released any BOMs? 😁

vutt wrote on 2024-10-26, 20:18:

I have day job running IT projects and Scope Creep is clear risk.

Yes, I won't argue with that! However, seeing as this project only has a single contributor (me), no bosses and nobody I'm beholden to, I'd rather take my time and get it right... 😉

vutt wrote on 2024-10-26, 20:18:

Have you considered forking base version into some sort of public V1 Limited run stable version? To give vogoners with higher risk appetite possibility to start playing around with wide range of legacy software few dedicated beta testers will not be able to cover.

That sounds like way too much hassle, sorry!

Eivind wrote on 2024-10-29, 19:27:

ahmadexp wrote on 2024-10-26, 19:46:

If you choose that route, please consider all kinds of options (with wave table and without) pins so we can come up with all kinds of cool options.

What did you have in mind? Whatever MIDI signal is coming out of the sound card is gonna be routed to the CM4 to be synthesized there. Were you talking about also exposing a full-sized wavetable header? Setting aside the issue of finding space on the board for this, I kind of feel it wouldn't be needed seeing how capable and versatile the CM4 is. Do you disagree?

As usual you are spot one. However, since the board to board connector will have more pins that perhaps going to be unused, we can have the all the wavetable UART, line in/out, ISA etc... pins routed and provisioned. What I have in mind is to have sound cards that we can reuse in other projects that CM4 might not be present. It is always easy to neglect an existing pin and not use it, but its not easy to have the need for a pin that is not routed. I hope it makes sense.

ahmadexp wrote on 2024-10-30, 00:47:

As usual you are spot one. However, since the board to board connector will have more pins that perhaps going to be unused, we can have the all the wavetable UART, line in/out, ISA etc... pins routed and provisioned. What I have in mind is to have sound cards that we can reuse in other projects that CM4 might not be present. It is always easy to neglect an existing pin and not use it, but its not easy to have the need for a pin that is not routed. I hope it makes sense.

Ah, yeah I can definitely set aside pins for all the common sound card signals, and just not use them on the TL3 (or the initial PicoGUS card), for future use. 👍

any chance you would be interested in releasing the pre-MXM schematics? I'd love to build one of those as I don't really have a way to build the Voodoo card yet and I'd really like to play with the compact one!

@Vridek

I just saw this -
https://www.reddit.com/r/retrobattlestations/ … _our_new_retro/

Nice!
Shouldn't you credit Eivind and rasteri for inspiring the board designs though?

Duffman wrote on 2024-11-03, 21:01:

@Vridek […]
Show full quote

@Vridek

I just saw this -
https://www.reddit.com/r/retrobattlestations/ … _our_new_retro/

Nice!
Shouldn't you credit Eivind and rasteri for inspiring the board designs though?

The Pixel project has been in the works since the fall of 2022, with a chief engineer hired in February 2023. Over 1000 hours have been spent on the project and it has cost $21,790 to date.
That's not counting the countless emails I've had to write to get to certain datasheets and solve technical problems. In total we have about 5 prototype boards where DX1, DX2, DX3, VEX, VEX2 were used.
We use our own designs, our own solutions and we have come to what we have now the hard way. Can I know where the hell the problem is? Do you have any idea how much trouble it was to get something like SocCfgTool the official way?
I’ve never been annoyed with someone like you for a long time!!

Vridek wrote on 2024-11-03, 22:49:

The Pixel project has been in the works since the fall of 2022, with a chief engineer hired in February 2023. Over 1000 hours ha […]
Show full quote

Duffman wrote on 2024-11-03, 21:01:

@Vridek […]
Show full quote

@Vridek

I just saw this -
https://www.reddit.com/r/retrobattlestations/ … _our_new_retro/

Nice!
Shouldn't you credit Eivind and rasteri for inspiring the board designs though?

The Pixel project has been in the works since the fall of 2022, with a chief engineer hired in February 2023. Over 1000 hours have been spent on the project and it has cost $21,790 to date.
That's not counting the countless emails I've had to write to get to certain datasheets and solve technical problems. In total we have about 5 prototype boards where DX1, DX2, DX3, VEX, VEX2 were used.
We use our own designs, our own solutions and we have come to what we have now the hard way. Can I know where the hell the problem is? Do you have any idea how much trouble it was to get something like SocCfgTool the official way?
I’ve never been annoyed with someone like you for a long time!!

pretty cool looking projects. I'm curious, on the smaller board with the DM&P SOM, what BIOS are you using or is it completely custom?

Vridek wrote on 2024-11-03, 22:49:

The Pixel project has been in the works since the fall of 2022, with a chief engineer hired in February 2023. Over 1000 hours ha […]
Show full quote

Duffman wrote on 2024-11-03, 21:01:

@Vridek […]
Show full quote

@Vridek

I just saw this -
https://www.reddit.com/r/retrobattlestations/ … _our_new_retro/

Nice!
Shouldn't you credit Eivind and rasteri for inspiring the board designs though?

The Pixel project has been in the works since the fall of 2022, with a chief engineer hired in February 2023. Over 1000 hours have been spent on the project and it has cost $21,790 to date.
That's not counting the countless emails I've had to write to get to certain datasheets and solve technical problems. In total we have about 5 prototype boards where DX1, DX2, DX3, VEX, VEX2 were used.
We use our own designs, our own solutions and we have come to what we have now the hard way. Can I know where the hell the problem is? Do you have any idea how much trouble it was to get something like SocCfgTool the official way?
I’ve never been annoyed with someone like you for a long time!!

I urge you to respect the community. Duffman only asked a question. Also, it does not hurt to acknowledge people even if you just got the idea from them.
Your Pixel x86 really looks like a modern version of Rastri and the ITX one looks like the Llama-ITX with PCIe. Your order board looks like a flat version of the TL2.
So, if you came up with both ideas on your own, good for you. Do not get mad at people when they ask why your product looks like such as such. This will just make you look bad.
It is similar to my Time Card project (https://github.com/opencomputeproject/Time-Ap … aster/Time-Card) which I put it up open source then comes this person (https://www.timebeat.app/) and copies the entire thing and renames it and start selling it for an arm and a leg under his own original product. When people as him, why did not you acknowledge the Time Card authors, he responds something similar to what you responded.
I personally do not give a damn, neither Rastri nor Eivind does since we all put our project out there for the community as open source.
It is up to the others how they take things and how they reuse them and if they are decent enough to acknowledge the source.
The similarity of your projects to Rastri and Eivind are definitely more than a coincidence, but hey, there is a long list some famous inventions that were developed by two different people around the same time like the telephone (Alexander Graham Bell and Elisha Gray), the theory of evolution (Charles Darwin and Alfred Russel Wallace), calculus (Isaac Newton and Gottfried Leibniz), and the lightbulb (Thomas Edison and Joseph Swan) and perhaps we can add the Llama-ITX (Eivind and Vridek) and Pixel x86 DX (Rastri and Vridek).

Hope you feel better now.

ahmadexp wrote on 2024-11-05, 03:32:

I urge you to respect the community. Duffman only asked a question. Also, it does not hurt to acknowledge people even if you jus […]
Show full quote

Vridek wrote on 2024-11-03, 22:49:

The Pixel project has been in the works since the fall of 2022, with a chief engineer hired in February 2023. Over 1000 hours ha […]
Show full quote

Duffman wrote on 2024-11-03, 21:01:

@Vridek […]
Show full quote

@Vridek

I just saw this -
https://www.reddit.com/r/retrobattlestations/ … _our_new_retro/

Nice!
Shouldn't you credit Eivind and rasteri for inspiring the board designs though?

The Pixel project has been in the works since the fall of 2022, with a chief engineer hired in February 2023. Over 1000 hours have been spent on the project and it has cost $21,790 to date.
That's not counting the countless emails I've had to write to get to certain datasheets and solve technical problems. In total we have about 5 prototype boards where DX1, DX2, DX3, VEX, VEX2 were used.
We use our own designs, our own solutions and we have come to what we have now the hard way. Can I know where the hell the problem is? Do you have any idea how much trouble it was to get something like SocCfgTool the official way?
I’ve never been annoyed with someone like you for a long time!!

I urge you to respect the community. Duffman only asked a question. Also, it does not hurt to acknowledge people even if you just got the idea from them.
Your Pixel x86 really looks like a modern version of Rastri and the ITX one looks like the Llama-ITX with PCIe. Your order board looks like a flat version of the TL2.
So, if you came up with both ideas on your own, good for you. Do not get mad at people when they ask why your product looks like such as such. This will just make you look bad.
It is similar to my Time Card project (https://github.com/opencomputeproject/Time-Ap … aster/Time-Card) which I put it up open source then comes this person (https://www.timebeat.app/) and copies the entire thing and renames it. When people as him, why did not you acknowledge the Time Card authors, he responds something similar to what you responded.
I personally do not give a damn, neither Rastri nor Eivind does since we all put our project out there for the community as open source.
It is up to the others how they take things and how they reuse them and if they are decent enough to acknowledge the source.
The similarity of your projects to Rastri and Eivind are definitely more than a coincidence, but hey, there is a long list some famous inventions that were developed by two different people around the same time like the telephone (Alexander Graham Bell and Elisha Gray), the theory of evolution (Charles Darwin and Alfred Russel Wallace), calculus (Isaac Newton and Gottfried Leibniz), and the lightbulb (Thomas Edison and Joseph Swan) and perhaps we can add the Llama-ITX (Eivind and Vridek) and Pixel x86 DX (Rastri and Vridek).

Hope you feel better now.

PS: 1000 hours of engineering work for those products seems to be an over kill, especially when Eivind and Rastri's GitHub projects are available to the public. Anyways, let's be nice to each other.

Vridek wrote on 2024-11-03, 22:49:

The Pixel project has been in the works since the fall of 2022, with a chief engineer hired in February 2023. Over 1000 hours ha […]
Show full quote

Duffman wrote on 2024-11-03, 21:01:

@Vridek […]
Show full quote

@Vridek

I just saw this -
https://www.reddit.com/r/retrobattlestations/ … _our_new_retro/

Nice!
Shouldn't you credit Eivind and rasteri for inspiring the board designs though?

The Pixel project has been in the works since the fall of 2022, with a chief engineer hired in February 2023. Over 1000 hours have been spent on the project and it has cost $21,790 to date.
That's not counting the countless emails I've had to write to get to certain datasheets and solve technical problems. In total we have about 5 prototype boards where DX1, DX2, DX3, VEX, VEX2 were used.
We use our own designs, our own solutions and we have come to what we have now the hard way. Can I know where the hell the problem is? Do you have any idea how much trouble it was to get something like SocCfgTool the official way?
I’ve never been annoyed with someone like you for a long time!!

With regards to your Pixel x86 project, maybe you can consider the wishlist that I gave Rastri and he dismissed.

Here is a short version of it:

Adding HIDman to support USB Keyboard and Mouse
Adding a CM4 to do the modem as well as munt and perhaps ethernet connection and Floppy/CD emulation.
Adding PicoGUS or even PicoGUS pro.

Please let us know.

@Eivind, just remembered to ask, since we have an HDMI connector now, shouldn't we have a way to send the audio to HDMI via the digital formal as well. We can have the option enabled when the 3.5mm audio jack is not connected. Also this can be somehow related to the LPC sound blaster that @Rastri was asking @sdz on the M4800 thread.

It is similar to my Time Card project (https://github.com/opencomputeproject/Time-Ap … aster/Time-Card)
[/quote]

That's your project? Wow! puts my designs to shame. you've got some EDA skills!
Also, referencing this topic. I made a device based off Eivind's TinyLlama and ITX llama smashed together but A. I have no intention to monetize or sell it and B. I spoke directly with Eivind about it and credited him on the board.
It's ok to use other people's designs since that's the beauty of open source but I agree that the 3 of these together look WAY too similar to the WeeCee, TinyLlama and ITX llama.
I'm all for mass producing these things so that less tech savvy people can get a hold of them and play with them but only with proper credit given to the original designed.

ahmadexp wrote on 2024-11-05, 03:47:

With regards to your Pixel x86 project, maybe you can consider the wishlist that I gave Rastri and he dismissed. […]
Show full quote

Vridek wrote on 2024-11-03, 22:49:

The Pixel project has been in the works since the fall of 2022, with a chief engineer hired in February 2023. Over 1000 hours ha […]
Show full quote

Duffman wrote on 2024-11-03, 21:01:

@Vridek […]
Show full quote

@Vridek

I just saw this -
https://www.reddit.com/r/retrobattlestations/ … _our_new_retro/

Nice!
Shouldn't you credit Eivind and rasteri for inspiring the board designs though?

The Pixel project has been in the works since the fall of 2022, with a chief engineer hired in February 2023. Over 1000 hours have been spent on the project and it has cost $21,790 to date.
That's not counting the countless emails I've had to write to get to certain datasheets and solve technical problems. In total we have about 5 prototype boards where DX1, DX2, DX3, VEX, VEX2 were used.
We use our own designs, our own solutions and we have come to what we have now the hard way. Can I know where the hell the problem is? Do you have any idea how much trouble it was to get something like SocCfgTool the official way?
I’ve never been annoyed with someone like you for a long time!!

With regards to your Pixel x86 project, maybe you can consider the wishlist that I gave Rastri and he dismissed.

Here is a short version of it:

Adding HIDman to support USB Keyboard and Mouse
Adding a CM4 to do the modem as well as munt and perhaps ethernet connection and Floppy/CD emulation.
Adding PicoGUS or even PicoGUS pro.

Please let us know.

I have a KiCad version of the WeeCee project that I made and I removed the PS/2 in favor of the HIDman. I think it would be cool to try to get a PicoGus and CM4 on there! it would be tricky but I'm sure it's feasable.

snipe3687 wrote on 2024-11-05, 20:46:

ahmadexp wrote on 2024-11-05, 03:47:

With regards to your Pixel x86 project, maybe you can consider the wishlist that I gave Rastri and he dismissed. […]
Show full quote

Vridek wrote on 2024-11-03, 22:49:

The Pixel project has been in the works since the fall of 2022, with a chief engineer hired in February 2023. Over 1000 hours ha […]
Show full quote

The Pixel project has been in the works since the fall of 2022, with a chief engineer hired in February 2023. Over 1000 hours have been spent on the project and it has cost $21,790 to date.
That's not counting the countless emails I've had to write to get to certain datasheets and solve technical problems. In total we have about 5 prototype boards where DX1, DX2, DX3, VEX, VEX2 were used.
We use our own designs, our own solutions and we have come to what we have now the hard way. Can I know where the hell the problem is? Do you have any idea how much trouble it was to get something like SocCfgTool the official way?
I’ve never been annoyed with someone like you for a long time!!

With regards to your Pixel x86 project, maybe you can consider the wishlist that I gave Rastri and he dismissed.

Here is a short version of it:

Adding HIDman to support USB Keyboard and Mouse
Adding a CM4 to do the modem as well as munt and perhaps ethernet connection and Floppy/CD emulation.
Adding PicoGUS or even PicoGUS pro.

Please let us know.

I have a KiCad version of the WeeCee project that I made and I removed the PS/2 in favor of the HIDman. I think it would be cool to try to get a PicoGus and CM4 on there! it would be tricky but I'm sure it's feasable.

With regards to the CM4, you can always place it on the bottom side.