YM2413 - HAAG - noisy

Page 3/4
1 | 2 | | 4

By Pencioner

Scribe (1476)

Pencioner's picture

13-02-2021, 19:58

Nope. Instead i followed the way of adding external output to my cartridges if they have none, and buying an audio mixer... So i can mix whatever i want the way i want, avoiding internal mixers of MSX computers

By sdsnatcher73

Prophet (2316)

sdsnatcher73's picture

13-02-2021, 20:24

Pac wrote:

Panasonic FS-A1WX schematics. In addition the Panasoft FM-PAC schematics if helps.

The original FM-PAC also sounds very clean!

By Daemos

Paragon (1954)

Daemos's picture

13-02-2021, 20:41

Scope images show white noise combined with 50hz sinal hum and 100hz filter residue. Is this from the PSU?? check and compare. Remember that noise does not equal hum. There is noise, high frequency pollution and hum and they all have different sources and sound very distinct.

If the PSU is your trouble maker then use some extra capacitors and if all fails use a capacitor multiplier and remember: voltage regulators are very bad AC suppressors and could make matters even worse introducing high frequency pollution.

High frequency pollution is extremely hard to suppress and silver mica caps are your best bet combined with some inductance. White noise can easily be killed using low pass filters and hum doesn't like lc filters and capacitpr multipliers.

How does it sound? The ear can really help defining the source of your audio polution.

Then audio mixers using opamps are great but prone to white noise. Thats a standard problem with opamps. There are low noise opamps but usualy sound sharp because of extra 3th harmonics (if you can hear the differences)

MSX computers all have "interesting" audio mixers. Overly complex and sometimes even downright badly designed (see nms8250).

Ohw and then there are groundloops. If you have those you are on your own Wink

By dublevay

Supporter (14)

dublevay's picture

16-02-2021, 00:45

Thanks for the responses. I will study the schematics supplied some more - particularly the FM-PAC.

I like the idea of the capacitor multiplier. I currently have 2x 1000uF where the power enters the board, and another 2x 1000uF in close proximity to the YM2413. I've also got the Horowitz and Hill open on my desk, and am going to have a good review of LC filters again - given that the MSX machines seem to make so much use of inductors.

By st1mpy

Paladin (838)

st1mpy's picture

25-02-2021, 00:51

I've got a ym2413 setup on a breadboard and connected to a fpga msx (de1). initially I had an oscillator just feeding into XIN. It was acting strange, and it was affecting the de1's internal fm sound (it was randomly skipping notes), and it is not because of using the same port (the external is write only). I replaced the oscillator with a crystal and capacitors, and this problem is gone. Interesting as it was mentioned in this thread earlier.

Now it is just noisy, but hopefully its because its on a breadboard... I'll solder these to a protoboard and see...

By sdsnatcher73

Prophet (2316)

sdsnatcher73's picture

25-02-2021, 06:46

I don’t think the noise will magically go away when mounted on a PCB. If you look at an FM-PAC you will see how elaborate the circuit is that they had to design to, amongst other, fight the noisy OPLL. They had to use caps that sink through the PCB to keep the cartridge thin. Nowadays this can probably be achieved with smaller components and e.g. an FM-PAQ lite from EBsoft is not noisy at all, so it is possible in a smaller design.

By dublevay

Supporter (14)

dublevay's picture

22-10-2021, 00:13

Thought I'd just resurrect this thread in order to share some things I've learned today. Below is my current stash of YM2413s.

I've now launched the product that I was originally calling the Dragon MSX 2+ board - though I eventually changed to name to the 6x09 SuperSprite FM+ board to avoid confusion and also avoid any trademark issues etc.

I tested all of the YM2413s that I had some time ago, on a breadboarded circuit - thus I knew that the HAAGs and THINK chips were all functioning. During development I discovered that the YM2413 was picking up some horrible mains hum, and so I inserted a 12 ohm resistor in the Vcc line of the YM2413, as had been done on an MSX YM2413 cartridge. It worked a treat, and the hum went away.

What I didn't realise at that time was that the "9950 THINK" chips would not work with this new 12 ohm resistor in the circuit. I don't know if they maybe already have something internal to fix the mains hum. As soon as I bridged the 12 ohm resistor the 9950 THINKs work again. I also have an "0728 THINK" chip which does work with the 12 ohm resistor. It's clear to me that these 'THINK' chips are re-marked (look at the different markings) - but what always seems strange is that their case type is so different from the HAAG chips.

Something else I've also learnt/confirmed tonight... The 9918 HAAG chips take about 30 seconds to become responsive. I don't know exactly why, but it's almost like they need to charge fully before pumping out their sounds. These 9918 HAAG chips are actually verified new-old stock (a guy in the UK has had them in his loft for 20 years or so - having bought them for a commercial project which stalled). This 30-second 'charge' wouldn't have been an issue back in the day, as disk drives would have taken a fair while to get code loaded to use the YM2413. But today with SD solutions, we obviously get to the executables much quicker.

With the exception of the "1025 HAAG" at the bottom of the third row, all other chips, including the YM2413B CMOS versions, are ready to play as quickly as I can get to my executable code! I suspect, therefore, that the "1025 HAAG" chips are also re-marked - the "10" probably gives it away really.

By Grauw

Ascended (10148)

Grauw's picture

22-10-2021, 13:51

A 30 second “charge time” is not normal, at all.

These chips are digital and they should function the instant they get power, there’s nothing in their design that would cause an initialisation time more than a few microseconds.

In my experience “fake” chips are usually not truly fake, they may be relabeled but if they’re pin compatible and produce sound then they’re still YM2413’s. Nobody’s made modern clone replacements of these Yamaha sound chips other than the few different variations Yamaha has made themselves.

So if you’re getting inconsistent results with these chips, maybe there is an issue with the board around it that makes them operate on the edge of what they’re capable of, e.g. too low voltage would affect NMOS sooner than CMOS.

By ro

Scribe (4530)

ro's picture

22-10-2021, 11:52

I love how this forum brings people together and collaborate on solutions. Keep it going guys!

By Pentarou

Champion (400)

Pentarou's picture

22-10-2021, 18:16

Grauw wrote:

Nobody’s made modern clone replacements of these Yamaha sound chips other than the few different variations Yamaha has made themselves.

~12 Years ago, a batch of ICs used in Viletim's FM master system sound board was found to be not 100% compatible. Read the SMS power forum post post if interested
I doubt Yamaha would produce partially compatible ICs, so these were probably fake/clones.
I still have 2 of these chips, they aren't the usual repainted +tinned + laser etched, but look almost the same as an original, the datecode/batch is 0108 HABG.

Besides, I find it strange that the OP has working ICs with the indent, as originals should have the dimple.

Page 3/4
1 | 2 | | 4