root42 wrote:

and can't be generated via DCDC converters on the board itself.

why not?

Well, maybe it could. I am not an electrical engineer. But I thought that would require quite large components for a current of several hundred milli Ampere. But maybe there are small DCDC converters that can do this? I dunno...

they do come in rather small packages, but off the shelf DC/DC converters aren't exactly cheap.

Yeah, my colleague always salvages them from scrap devices. Now I know why.

This is such a cool project.

I do have a random question though, not necessarily for this board but as a theoretical question as far as DUAL OPL2 support goes.

Would it be possible to make a DUAL OPL3 board and have it be compatible with DUAL OPL2 games?

You can also find DC/DC Converters replacing 7805 for example.

I also removed the 7805 from my SB Card and added a DC-DC Module for example.
If you have several cards and each one has one of this buggers you can save a little bit of energy, and also reduce heat inside the computer case.

i also generated the -12V supply with a ICM7660 for a DAC, some time ago.
also worked well. But only for small currents.

7805SR-C 500ma: ~2€ - there also bigger one`s

root42 wrote:

WAs the -12V is the biggest load on the board, I think it is fair to assume this will have to come via the ISA bus, and can't be generated via DCDC converters on the board itself.

No, +12V rail has the biggest load (but still within specs for ISA). The problem is in current paths if molex will be implemented. +12V and -12V will go physically different ways and that can cause problems.

imi wrote:

they do come in rather small packages, but off the shelf DC/DC converters aren't exactly cheap.

matze79 wrote:

You can also find DC/DC Converters replacing 7805 for example.

Good DC-DC converters will take additional space on the board:

Bad ones will put out a lot of noise, which I try to eliminate with linear regulators in the first place.

As long as measurement of real hardware will stay within ISA current specs - I will not consider adding DC-DC converters.

Fagear wrote:

Bad ones will put out a lot of noise, which I try to eliminate with linear regulators in the first place.

As long as measurement of real hardware will stay within ISA current specs - I will not consider adding DC-DC converters.

For sure. Linear regulators are the way to go in analog and mixed audio circuits. Buck converters are noisy, but good ones are too bulky. I use LM2931 in BlasterBoard with -90 dB PSRR for powering DACs and 317L & 337L for opamps. These give very good results in terms of noise and temperature if used correctly, with proper grounding and placement.
Small tip: The biggest noise source for ISA is a monitor connection with VGA cable (60Hz peak), so maybe you should consider a dedicated filter on your power rails. And also locate your power regulation circuit as close as possible to ISA power pins. The most interference comes from ISA slot itself due to physical connection characteristics.

matze79 wrote:

You can also find DC/DC Converters replacing 7805 for example. […]
Show full quote

You can also find DC/DC Converters replacing 7805 for example.

I also removed the 7805 from my SB Card and added a DC-DC Module for example.
If you have several cards and each one has one of this buggers you can save a little bit of energy, and also reduce heat inside the computer case.

i also generated the -12V supply with a ICM7660 for a DAC, some time ago.
also worked well. But only for small currents.

7805SR-C 500ma: ~2€ - there also bigger one`s

7805 dropin.jpg

When some people tried to replace the 7805 in their SNES units these things caused problems. Apparently you're trading heat and efficiency for a bit of noise. Not sure how much it would intrude on audio but it certainly caused problems for the video signal on those old consoles.

Nothing particularly interesting to add, but i just wanna say i'm very impressed by Fagear's work. The optimization of the modules os nothing shirt of amazing, and i've wondered about the possibility of a lite version of this card including only CSM, 1*SID (instead of duosid), GameBlaster and Tandy + MPU.

carlostex wrote:

Nothing particularly interesting to add, but i just wanna say i'm very impressed by Fagear's work. The optimization of the modules os nothing shirt of amazing, and i've wondered about the possibility of a lite version of this card including only CSM, 1*SID (instead of duosid), GameBlaster and Tandy + MPU.

Once full-length FMonster is done there will be nothing stopping from producing many more shorter variants just excluding blocks. Each block is more or less independent from others.

That's awesome i really like the modularity of the blocks. Full card prototype next year?

On a Mini Version it would be also interesting to get some kind of mass storage, this would be very nice for ultra compact XT systems with only one ISA Slot 😁

Look at what I've got on the left side of the board... 😈

AY block required 4 different clock sources, so I've had not only add one more generator (4th one) into clock-gen block, but also to heavily optimize and re-route bottom part of 2xSID block and ISA bus part of FMonster. 😵
But I've made it fit. 😎

Fagear wrote:

Look at what I've got on the left side of the board... 😈

What? No headers for the CSM joystick headers? 😈

Kidding aside, the multiple clock settings for the CSM is a nice touch. Will most DOS games sound different but comprehensible at those speeds, or are those settings better suited to tracker playback of some sort?

MrSmiley381 wrote:

Will most DOS games sound different but comprehensible at those speeds, or are those settings better suited to tracker playback of some sort?

For DOS games, expecting Covox Sound Master user should select 1.7898 MHz as is on original CSM.
Other frequencies are listed for listening music from according platforms with AY synth. Software is not yet existing for it, I presume.

Now working on some minor optimizations of "big board":
- tidying up silkscreen
- tidying up shared logic (Reset inverter, /IOR and /IOW repeaters, high-address decoder)
- adding debug LEDs (reset/power rails/MIDI TX-RX)

Board now decodes 15 of 16 bits of address bus (instead of usual 10 bits).

Also I've made an "ISA extender" for ease of debugging:

And of one live stream there was "incidental" try out of old (mono) OPL2 block. And it works pretty well right off the bat! 😎

After CSM (AY) block will be done, I think I'll rearrange blocks on the board again. Probably new order will be "SN - 2xSAA - 2xOPL2 - AY - 2xSID - clockgen - OPL3".

For the moment I'm thinking about DAC block.
Does anyone here has any expirience or knowledge about how players and trackers work in "LPT1+LPT2" mode with Covox Speech Thing?
I know about "strobed" variants of stereo CST on one LPT port, but that will require many more ICs to implement.

So I've stick to "LPT1+LPT2" solution for stereo.
All I want to know is how data is sent to LPT1 and LPT2 during playback? In strict interleaving order? Or not?
That knowledge is critical for "auto-stereo" detection logic.

There is also a variant of auto detecting Stereo ability.. maybe i find the schematic for this. let me look around in my file mess..

If you implent this too maybe you can also add DSS Support then.

Here is shock_`s latest schematic, without warranty.

i also have a unpopulated prototype pcb left. which has 2-3 minor design flaws 😀 which i can spare.

Those are descripted here:
https://dosreloaded.de/forum/index.php/Thread … ded-DE-Edition/
Sorry German.

matze79 wrote:

There is also a variant of auto detecting Stereo ability.. ... Here is shock_`s latest schematic, without warranty. se are desc […]
Show full quote

There is also a variant of auto detecting Stereo ability..
...
Here is shock_`s latest schematic, without warranty.
se are descripted here:
https://dosreloaded.de/forum/index.php/Thread … ded-DE-Edition/

Sorry, but that is a type of circuit I've been talking above - it is a "strobed" stereo Covox Speech Thing. It uses only one LPT port and differentiates "left/right" via additional pin.
Use of those additional pins requires much more complex circuitry for implementing my own "LPT port" (since I'm not using connection to an existing ones).
"Normal" LPT port has three ports/registers. Data register on a base address and two additional control/status registers, which support both write and read operations. That will require many more ICs to implement, thus it's not viable in the limits of FMonster project.
So, I'm implementing only first, data register. And only write operation. No control register, thus no strobe pins. So, for stereo I have to use two LPT ports (0x378/0x278). It is an available configuration in a software that can work with Covox Speech Thing.
And on the DAC itself in that circuit above - that's a AD7528, stereo DAC. All great but it can not write to both left AND right channels at once. It has to be written separately. Also not sutable for my application, where my logic should auto detect mono data and write the same data to both DACs at once. So my implementation will use two AD7524s.

Thinking of wider usable range of the DAC. I will not implement DSS support (clock gen, FIFO buffer).
But I'm thinking about adding some more addresses where the same DAC can be available.
For example, Covox Sound Master has the same type of DAC. It can be accessed via DMA (not happening in my case) and via direct port write at base+0x002 AND base+0x00F. Available base addresses: 0x220/0x240/0x280/0x2C0. That's 8 additional port to have that DAC on.
As far as I know, Covox Voice Master also has this DAC on 0x22F.
Has anybody more suggestions on other cards with similar DACs and what port those use?

There will be complications in address decoder and some compromises, but I'm willing to cover as many options as I can.

I would gladly apply for a card, whenever this is project reaches final form. 😀 Please put me on the list, if it exists of some form. This looks like an awesome card to have!