The new JLC build sounds cool, looking forward to seeing. A few dollars here or there doesn't make much difference, for what its worth, given that the entire project is a massive bargain at any conceivable price, compared with how much$ original products (or even modern remakes) cost for even half of this functionality.

vutt wrote on 2023-09-23, 09:32:

A few additional bucks will not change my mind. […]
Show full quote

A few additional bucks will not change my mind.

Yesterday I spend few hours running players and dos demos on my new 486 build. So HiNT CS8005 chipset can be added to PicoGUS compatible list now.
Following demos/players worked fine:

Copy code to clipboard

12nd reality and unreal by Future Crew, Crystal Dream I and II, DRIFT by WILD LIGHT, C N C D Inside, Legend and Project ANGEL by IMPACT Studios, Machines of Madness by Dubius, Contrast - OXYGENE
2also Cubic Player, Impulse Tracker, DMP v4 and XTC player

Only one fail - Electromotive Force - Verses (1994) . In 2 attempts music started well but stopped/was partially played in the middle of demo.

IIRC I have experienced the same inconsistency with Verses - sometimes reaching partway into the demo, and having the music abruptly stop.

I don't know what the cause is, though - whether it's the PicoGUS, or all GUSes, that demo's code, the machine specs/speed, or some vagary combination of these and other factors.

speaking of optimizing costs, you can replace PCM5102 I2S dac with 3x cheaper PCM5100 or even smaller PCM1754. I suspect even the cheapest clone of 30 year old TDA1311 design the $0.01 8211 will sound perfectly fine and comparable to original TDA1545 of Gus Classic. People still use 8211 https://mitxela.com/projects/rscope2

Shreddoc wrote on 2023-09-23, 20:47:

vutt wrote on 2023-09-23, 09:32:

Only one fail - Electromotive Force - Verses (1994) . In 2 attempts music started well but stopped/was partially played in the middle of demo.

IIRC I have experienced the same inconsistency with Verses - sometimes reaching partway into the demo, and having the music abruptly stop.

I don't know what the cause is, though - whether it's the PicoGUS, or all GUSes, that demo's code, the machine specs/speed, or some vagary combination of these and other factors.

Verses is one of the odd ones for me. It's perfect on my main development platform (Pentium 120, Intel 430FX chipset) almost all of the time, but when I start playing with how IRQs are handled it's probably the first thing I notice breaking. I need to do some more in depth bus captures when it fails to figure out what I can do to improve its stability. One hint - play with the audio buffer size (pgusinit's /a option) to see if it improves things.

rasz_pl wrote on 2023-09-24, 05:24:

speaking of optimizing costs, you can replace PCM5102 I2S dac with 3x cheaper PCM5100 or even smaller PCM1754. I suspect even the cheapest clone of 30 year old TDA1311 design the $0.01 8211 will sound perfectly fine and comparable to original TDA1545 of Gus Classic. People still use 8211 https://mitxela.com/projects/rscope2

I'm using the PCM5100A on the "Femto" PicoGUS and it sounds great. I will be using it on the new "mass produced" PicoGUS. The PCM1754 is a non-starter as it needs a system clock that is a multiple of the sampling rate. The major benefit to the PCM510xA series is that it will accept any arbitrary sample rate and synthesize its own clock internally based on the I2S LRCK signal. This is a major upside for GUS emulation where sample rates are oddball rates and can change continuously like in Impulse Tracker. The 8211 looks like a really neat option... looks like it's about $0.12 at JLCPCB compared to $0.68 for the PCM5100A. But I think I'll stick with the PCM5100A because I can continue using the I2S library in pico-extras as it's worked extremely well so far.

If I recall correctly, Verses was sensitive to ISA bus speed for me. Again, AFAICR, the music stopping symptom was easy easy to reproduce on my 486 DX2/80 machine with the ISA bus set higher than 8 or so MHz.

Please take that 30ish-year-old memory of mine with a grain of salt, as I could be misremembering.

I'm impressed - Zone66 is working without picogus init run or environment variables. Actually if I do init first my i486DX2 actually restarts PC. Very HC coding, devs are clearly with demoscene background..

polpo wrote on 2023-09-22, 14:21:

- How important is it that the wavetable volume be adjustable? On my roadmap is simultaneous MIDI with GUS or SB digital emulation.
- Would you pay $2-$3 extra to be able to adjust wavetable volume, on a card that will cost $30-$40 to buy?

Is there no possibility to make this volume option optional i.e. bypass missing/unpopulated thimbwheel with some jumper(s)?

A little lifehack/idea for e-waste conscious PicoGus owners - in case you don't know what to do with surplus PCBs.
It can be turned into nice ISA bus troubleshooting tool. Handy way to check BUS clock speed, data traffic or even voltages...

^^ that's a neat little idea.

I built and tested 3 of my spare PCBs (into PicoGUSes, obvs) then sold them locally, to recoup my costs - albeit at a cost of many hours of work on my part!

But I screwed another one up - lifted a track, by way of idiocy - so I still have that board spare, for such ideas.

Btw. we can add MOD4WIN 2.40 for Win 3.11 to working fine list.
I'm actually amazed that PicoGUS is working in Win 3.11 without any extra effort with v4.11 drivers linked in PicoGUS guide. Install it by adding new device in Windows CP. Point drivers location to ULTRASND\WINDOWS directory and then give io parameters when asked.
Haven´t messed around with midi part yet...

Adrian's Digital Basement has a new video about the PicoGUS! https://www.youtube.com/watch?v=bBYUTwKRyNk

When testing, Adrian ran into a few issues. For example we discovered that the PicoGUS (and real GUS!) do not work well with the XTIDE. If drive access happens simultaneously with GUS music playback, strange behavior can occur (freezing, VGA writes failing, other odd things). Also he had an issue where firmware writing wouldn't work on a 486 PCI board he has, and I hope to fix that one. The next release of pgusinit will include some extra fixes to make firmware writing more bulletproof.

polpo wrote on 2023-10-07, 16:40:

Adrian's Digital Basement has a new video about the PicoGUS! https://www.youtube.com/watch?v=bBYUTwKRyNk

When testing, Adrian ran into a few issues. For example we discovered that the PicoGUS (and real GUS!) do not work well with the XTIDE. If drive access happens simultaneously with GUS music playback, strange behavior can occur (freezing, VGA writes failing, other odd things). Also he had an issue where firmware writing wouldn't work on a 486 PCI board he has, and I hope to fix that one. The next release of pgusinit will include some extra fixes to make firmware writing more bulletproof.

Watched the video. Very cool to see this project getting attention.

Question about the XTIDE conflict... is that only when using the IDE header on the card, or does it still happen with drives connected to other controllers too?

Awesome video! Awesome product, great job on this!

I’m being totally selfish here by asking this but in the future do you think you might do a pure mp401 clone? I know a lot of us MIDI enthusiasts would love a pure mp401 card, and I’d feel kinda weird buying a whole picoGUS for that one feature.

On the other hand it’s such a cool product maybe I’ll do it anyway 😀 Great job again!

midicollector wrote on 2023-10-09, 19:17:

Awesome video! Awesome product, great job on this!

I’m being totally selfish here by asking this but in the future do you think you might do a pure mp401 clone? I know a lot of us MIDI enthusiasts would love a pure mp401 card, and I’d feel kinda weird buying a whole picoGUS for that one feature.

On the other hand it’s such a cool product maybe I’ll do it anyway 😀 Great job again!

That is the kind of thing I imagine could sell well (pre-assembled) in a store like Serdashop, for example.

danieljm wrote on 2023-10-09, 17:43:

Question about the XTIDE conflict... is that only when using the IDE header on the card, or does it still happen with drives connected to other controllers too?

It's only when using the IDE/CF interface on the XTIDE card, and only when drive activity on that connector is simultaneous with music playing or other GUS commands. This actually makes me want to design an "AT-IDE" card with built-in XTIDE Universal BIOS ROM socket, as 16-bit IDE interfaces are not affected by this issue, even when using the XTIDE Universal BIOS. And just about any computer that can effectively use a GUS has 16-bit ISA slots. A 16-bit IDE interface card capable of PIO would only take a few TTL logic chips.

midicollector wrote on 2023-10-09, 19:17:

I’m being totally selfish here by asking this but in the future do you think you might do a pure mp401 clone? I know a lot of us MIDI enthusiasts would love a pure mp401 card, and I’d feel kinda weird buying a whole picoGUS for that one feature.

The cost of the "sound card" portions of the PicoGUS bill of materials with integrated DAC are only about $3 (DAC is about $0.60, PSRAM is about $1.70, audio connector is $0.20, then add a few more cents for passives...) so to me it's not worth creating a version of the card only with MIDI output. Compared to the price of a PCMIDI card, a fully-populated PicoGUS is fraction of that so it's worth it even if you never use the audio portion of the card, and only use MIDI.

darry wrote on 2023-10-09, 19:49:

That is the kind of thing I imagine could sell well (pre-assembled) in a store like Serdashop, for example.

I'll be selling pre-assembled PicoGUS 2.0 boards in my Tindie shop as soon as I validate the prototype that I'll be getting in my hands today... I already have a lot of people on the waitlist on Tindie after Adrian's video was posted, so the demand is definitely there.

Wow, when you put it like that, that’s a great deal for all the sound features of the picoGUS plus the MIDI support. I think you just sold me on it 🤣

Thanks for the fast reply and for all your hard work on such a cool project!

Will the v1.1 boards still be supported by ongoing future firmware releases?

or, to ask another way, will the v2.0 boards get features/functionality that the v1.1's cannot do?

Shreddoc wrote on 2023-10-10, 01:35:

Will the v1.1 boards still be supported by ongoing future firmware releases?

or, to ask another way, will the v2.0 boards get features/functionality that the v1.1's cannot do?

All boards will run the same firmware! There are some hardware differences on the new board that have to be handled, so the firmware will detect that it's running on the board and adjust accordingly. Here's a breakdown of the differences:

PicoGUS v1.1/v1.1.1:
- Uses Pico or Pico W
- PCM5102 on module
- No wavetable header
- Micro USB-B port can be used for joysticks with powered USB OTG cable or a bodge, is internal to PC
- 3.5mm MIDI jack
I'm on the fence if I'll sell any of the v1.1.1 boards on Tindie - it's not much cheaper to make than 2.0, and people don't have to buy a Pico with 2.0. I'd say the only reason for me to sell these is to support the Pico W for future WiFi support (but it's inside the computer, so not the best for reception). I definitely want to keep this version 100% supported as it's the most DIY-friendly design, and that's important to me.

PicoGUS Femto w/ ISA adapter:
- Uses Pico or Pico W
- PCM5100 on board
- Wavetable header, analog volume thumbwheel for wavetable
- Micro USB-B port can be used for joysticks with powered USB OTG cable or a bodge (or jumper in next batch), is external to PC
- 2.5mm MIDI jack
I'll be putting some more of the Femto in my Tindie shop soon. I got the PCBs today but I still need to assemble the connectors on them.

PicoGUS v2.0:
- RP2040 direct on board
- PCM5100 on board
- Wavetable header, digital volume control for wavetable (set with pgusinit.exe)
- USB-A port so joysticks (or mice, or whatever else) can just plug in
- 3.5mm MIDI jack
The main reason for the digital volume control is that JLCPCB stopped carrying the analog thumbwheel I was using, and the digital one is a lot cheaper than a physical control. I asked a bit about this a few posts above. The goal of this version is to make something 100% able to be assembled by JLCPCB. I got the first prototype today and got it working in GUS mode tonight... I need to tweak a couple things but I think it's ready to be "mass produced" once the parts arrive at JLCPCB.

Congrats on the new revision! Looks good 😀

This is really nice, congratulations.

Just to clarify, is there code to allow USB connected mice or joysticks (HID) to be seen by the host computer as serial or gameport devices, respectively ? If so, that is a cool feature that should be mentioned in the feature summary, IMHO.

darry wrote on 2023-10-10, 12:43:

This is really nice, congratulations.

Just to clarify, is there code to allow USB connected mice or joysticks (HID) to be seen by the host computer as serial or gameport devices, respectively ? If so, that is a cool feature that should be mentioned in the feature summary, IMHO.

The USB support is still in progress and not finalized so I haven't put it on the official feature summary yet. Serial mouse is trivial so I haven't started on it yet, and gameport needs tweaking and maybe extra work in the PIO to support IOR without bringing IOCHRDY low to solve a problem with jittery movement.