FRS' MSX page WIP

Prophet (3675)

03-02-2017, 00:20

I'm opening this thread to post news about software and hardware projects in development while keeping it under a single thread instead of one for each project/release (what would be annoying).

And the 1st news is:

JoyTest v2.4 is out! It's the well known utility that lets you test a lot of Human Interface Devices that can be connected to the MSX, either directly or by using adapters.

It adds support to test the following devices:

So, if you would like to be sure if your light gun, light gun adapters or light pen are working properly, just give this software a run.

A special thanks go for Erik Luppes, of www.msxinfo.net, for donating a rare Sanyo MLP-001 Lightpen. The support for this device wouldn't have been possible without that.

بواسطة rderooy

Paladin (686)

03-02-2017, 09:32

I have added the MLP-001 to the Wiki, if you want to add some details like pictures :-)
https://www.msx.org/wiki/Sanyo_MLP-001

بواسطة sd_snatcher

Prophet (3675)

17-08-2017, 16:29

I released a new set of drivers for the MSX Turbo-R BIOS in my page, with many improvements over its original set of drivers, including way better performance.

1) Adds the missing drivers for these devices:

MSX-Paddles (or any PWM device, just like the original)
Light-Pen

2) Replace the existing drivers with much better ones for these devices:

Mouse, now with supported for the extended mouse wheel protocol and much faster than the original
Trackball, with angular to linear conversion for a much improved response
Touchpad: 10 times faster on Z80, 120 times faster with the R800
PAUSE Key: Will now work correctly with Konami games running on R800
Key click

3) Other enhanced BIOS routines:

BEEP: Doesn't reset the system timer anymore

The drivers are loaded by an MSX-DOS executable, and remain as part of the DRAM-mode BIOS until unloaded either by the same utility, a reset or activation of the ROM mode.

بواسطة sd_snatcher

Prophet (3675)

17-08-2017, 16:31

And here's another announcement some people have been demanding me.

The two Panasonic MSX Turbo-Rs have a very good audio circuit design that has everything it should have for its price range. But the result was poor because the circuit was terribly calibrated, resulting amongst other problems, in:

The slot sound was too loud, meaning you could barely hear the PSG drumkit on when an SCC cartridge was used.
OTOH, the OPLL sound was too quiet, so the soundtracks of Micro Cabin games, Valis-2, Starship Rendezvous and others that used the OPLL and FM simultaneously was severely affected. Most accords were derailed, and Micro Cabin's trick of using the PSG as a 3rd operator for the OPLL didn't match anymore. The PSG kept screaming all the time.
The OPLL low-pass filter was even stronger than the already incorrect Yamaha reference design, resulting in a terribly muffled FM sound that made the previous problem even worse.
The global audio-out low-pass filter was also too strong, to the point that the borders of the square wave output of the PSG were totally round, degrading many beautiful accords of famous Konami soundtracks for this chip
The global audio-out level was too weak, resulting in poor SNR and weak bass.

I published this fix recalibrates the audio circuit for maximum performance, and what astonishing difference that it does. It then blows any other MSX sound quality out of the water.

This recipe corrects all these problems, and normalizes the mixing volume between the slots (usually SCC or MSX-Audio), PSG and OPLL, to match the levels of the Panasonic MSX2+ and Sony MSX2+ machines.

بواسطة ricbit

Champion (438)

17-08-2017, 18:06

Do you plan to release source code for the new tr drivers? I could add my carcara basic improvements in the patch.

بواسطة wolf_

Ambassador_ (10135)

17-08-2017, 20:59

Sd_snatcher: can you elaborate on that third operator remark? Maybe it's a language/expression thingy, but are you saying that the psg influences the FM operators in the same kind of way a genuine 3-op FM synth could??

بواسطة Grauw

Ascended (10821)

17-08-2017, 21:22

Well, an operator doesn’t need to be a modulator…

بواسطة wolf_

Ambassador_ (10135)

17-08-2017, 21:28

Within an FM context, using the term 'operator' for something non-FM is a bit like saying 'that is clear as a whistle' when talking about trumpets.

بواسطة Grauw

Ascended (10821)

17-08-2017, 22:18

Well, operator is really just Yamaha terminology for oscillator…

I think the point of phrasing it like this is, it implies that the PSG is played in unison with the FM to build the timbre of the FM sound (synthesis), rather than considering them two separate layered instruments. As you know, many 4-op and 6-op algorithms contain lone carriers, which serve the same purpose.

(Btw, talking about 2- or 3-modulator chains, only two of the OPL3/4’s 4-op algorithms have these, the other two algorithms can be built from 2-op channels playing in unison of which the OPL4 has plenty. Why they didn’t put four proper 4-op algorithms in there…

The YM2151 OPM sure did a better job there, out of its 8 algorithms, five are properly 4-op, and that one doesn’t have a bounty of 2-op channels so those three simpler algorithms actually have a use.)

بواسطة Grauw

Ascended (10821)

17-08-2017, 22:50

Anyway, I think if I am going to apply this sound fix to my turboR, might as well also re-cap the thing. Maybe get rid of those jailbars. Need to take inventory of what’s in there and order parts.

It looks like Digikey has 1n5/50V polypropylene film caps in stock, but in case one uses C0G ceramic caps, does the voltage rating matter much? Those don’t seem to come in a 50V variant. From my knowledge for caps you can always use one with a higher voltage rating, but I’m not an electronics engineer… [Edit: ok, I clearly didn’t search well enough -_-;;]

Btw, it was rumoured that the GT sound circuitry was improved from the ST, and maybe that later revisions of the ST or GT also had improvements there. Does that match your findings?

بواسطة sd_snatcher

Prophet (3675)

18-08-2017, 00:19

Bingo! Grauw has explained wonderfully how the 3rd operator works in Micro Cabin games.

Quote:

Anyway, I think if I am going to apply this sound fix to my turboR, might as well also re-cap the thing. Maybe get rid of those jailbars.

Is your turboR an ST or GT? Why do you think it is necessary to recap it? It's rarely the case for MSX machines since they have excellent Japanese capacitors. I would only recommend to change them when it's really necessary. If you describe the problem, maybe we can find its specific solution.

Quote:

Btw, it was rumoured that the GT sound circuitry was improved from the ST, and maybe that later revisions of the ST or GT also had improvements there. Does that match your findings?

I didn't analyze the PCM part of the circuit, so please keep that in mind when reading my conclusions ahead.

IMHO, there isn't such a big improvement between the circuit of the two models. It's 99% identical, and change that seems to be responsible in the perception of an improvement was in fact just a tweak.

The changes are:

1) A more elaborate mute circuit. On the ST it's just a transistor that grounds the audio line when necessary. On the GT, they used one analog switch to cut the audio. This shouldn't affect the audio quality in normal operation, but instead just make a better mute.

2) They added a 4K7 pull-down to the PPI click sound line. This will affect only things that use the key click

3) And finally, they tweaked the global opamp to raise the output volume a bit. I think that this is the change that makes people associate the GT with a better sound output, as the audio-out is a bit louder the SNR also gets a bit better.

There can also be some changes related to the PCB layout that can have a significant impact specially in noise/interference. But I didn't analyze that. It's too difficult on the TR PCBs since they're tinted black to hide the tracks.