fluke

I can't see the ym2612 in the parts list for the DX7 II. The ym2612 is only 4 operator FM and the the DX7 is 6 operator FM anyways so it can't be the sound generator chip. Wouldn't it be easier to write a program to run on a megadrive to control the ym2612 than to try and source the chips and build custom hardware?

The Casio CZ series use Phase distortion synthesis which is a different concept to Frequency or Phase Modulation, though it did produce similar results. The main difference between FM and PM is that a DC offset in the modulator produces a pitch shift in the carrier with FM and only a phase shift with PM. This makes a PM synth more stable. See http://archive.cs.uu.nl/pub/MIDI/DOC/phase-mod for more information and some C source code.

The answer is - enough voltage to get 5mA flowing at the given resistance. ;) Have a look at http://www.midi.org/techspecs/electrispec.php for the MIDI electrical specification. MIDI in has an optocoupler (typical voltage drop of 1.3V for a 6N138) and a 220? resistor. So your MIDI out needs 2.4V to supply 5mA. If your supply is higher, add more resistance to drop the voltage. Eg an extra 120? if you're using a 3.3V part with a 3V minimum output high voltage or 380? if you've got a 5V part with a 4.3V minimum output high voltage. In the MIDI spec example (and the MIDIbox), they've been cautious and used 2 220? resistors (also 380? isn't a standard resistor value). My MIDISPORT 2x2 has a single 470? registor has its output. Speaking of which, the MIDISPORT 2x2 is a nice device to base a MIDI interface off. It uses the Cypress AN2131SC (same as the MBHP USB), an 8051 based microcontroller with hardware USB support. It downloads its firmware from the host computer so development cycles are fast. The stock firmware isn't class-compliant and doesn't work with Vista 64-bit. When i get some time i'll write new firmware for it, using the open source firmware for Linux (http://www.linux-usb.org/ezusb/) as a base.

I'm going to actually finish building my MIDIbox FM this year. I'm building it without a control surface, so i designed a stripboard layout for the Core, with banksticks and a 12pin connector to the OPL3 board. Then i had what i think is a really good idea. What if i move the OPL3 connection to other pins so that the hardware I2C interface on the PIC is freed up. Then i could connect the MIDIbox FM Core as an I2C slave to another Core that is being a MIDI Router without needing a PIC16F88 between them. And i could put my DB50XG in the same box for the complete 90s synth experience. :) Does this sound like a workable idea? I know i'd have to extend MIOS to support this, would it be a lot of work?

I find a small flat-blade screwdriver is the best tool for removing ICs. Just lever it in from from one end then the other 3 or 4 times, going up only a small amount each time so as not to bend the pins.

The SID is not a simple locked microcontroller. It is a full custom analog/digital IC, they can't even change feature size without needing a major redesign. Even if you do selectively etch away each metal layer to get a full layout of the SID, you'd still need them restart the 20 year old process that they originally used to make the SID to avoid needing a redesign. A better plan would be designing a SID clone either out of multiple chips (so hobbyist could make it and you wouldn't need analog and digital in the same chip) or a analog/digital FPGA.

512kbit = 64kbyte

The LM324 has lower bandwidth (1MHz vs 4MHz), is slower (0.5V/us vs 13V/us) and is probably noisier than the TL074. But they'll work, so have a listen and see how you like them. Just remember to socket them so you can change them later.

334 is the right code. I'm not sure of what exactly would have one, but something low tech with a design that hasn't changed in 10-20 years is more likely to have through-hole parts. Polyester or "greencap" capacitors are also suitable, you're more likely to find 330nF in those types than ceramic. Remember that it is also 0.33uF. If you really can't find that value, just substitute the closest value ceramic capacitor you can find. It's just a noise filter capacitor, so you could even get away with leaving it out. The construction is more important than the exact value here.

If i understand you correctly, you're referring to pins 2 (/IRQ) and 9 (TEST) of the YMF262 chip. If so, they can safely be left unsoldered as they have no connection. If you look at the PCB images, you'll see that the pads were only as large as the pins and didn't go anywhere.

The power supply will be noisy, leading to some noise on the audio signal. How much noise depends on the power supply filters, ripple rejection of the opamps, switching frequency (52kHz in this case), etc, etc.

That's a switch mode power supply, so it will be noisy. The one you posted is a +/-12V supply only, they also sell a +/-5V version which uses the same PCB, just different regulator chips, hence the confusion.

Without any rectification, the capacitors on the input side of the 7805 are simply acting as resistors to ground, with a resistance of -j/2?fC (ie 1/2?fC with 90° phase lag). So all you've got there is effectively 0.4? resistor to ground, which would explain the ripples.

MPLab does work under WINE for editing/assembling programs, but i couldn't get it to talk to a programmer. It did work under VMWare though.

I'm in NZ and pollin.de quoted me €22 + €3/kg for shipping. How did you get a rate of €7 to Australia? (Was it actually the rate to Austria instead? :) )

The slave select pin is for when the PIC is being an SPI slave. For master mode, just use any I/O pin to select the slaves, one per slave.

Which ADC are you referring to? The YMF262 is pure digital, the YAC512 is a DAC and the MIDIboxFM application doesn't support an AIN board.

The OPL3 (and other "frequency modulation" synths) actually use phase modulation instead of frequency modulation. The reason is that a DC offset (eg non-zero amplitude at 0Hz) will give a frequency shift in FM but only a phase shift in PM. See http://www.nic.funet.fi/pub/sci/audio/misc/pm-intro for more information.

Try to say OPL3 when you mean the FM synth chip. It will be less confusing. :) The sample rate is the same 49.716kHz. I'm not sure of the exact sample rate for the internal calculations, but it will be at least 12 bit mantissa as that's what the log-sin table in the ROM takes up. See this forum post on OPL Decapsulation for more information on the ROM contents. That sounds like an OPL3 chip. It's unlikely that Creative modified the chip as they'd need a custom production run of it and then it would more likely have a Creative logo on it instead. Emulation is free once you've written it once. Real chips cost real money, not just buying the chip, but the increased PCB area, soldering and testing it. The OPL3 has up to 18 voices. Channel means voice in this case. The OPL3 has a total of 36 operators (oscillator + enveloping), so either 18 2-op voices, or 6 4-op voices and 6 2-op voices, or 15 2-op voices and 5 percussion voices, or 6 4-op voices, 3 2-op voices and 5 percussion voices. MBFM uses the last mode (but doesn't give you access to the 2-op voices).

The OPL2 uses the YM3014 DAC, which has a dynamic range of 16 bits (its input is a floating point number with 10 bit mantissa and 3 bit exponent). As used in the AdLib and SoundBlaster cards it has a sampling frequency of 49.716 kHz. The OPL3 The OPL2 uses the YAC512 DAC, which has a dynamic range of 16 bits (its input is a floating point number with 10 bit mantissa and 7 bit exponent). As used in the SoundBlaster Pro and 16 cards it also has a sampling frequency of 49.716 kHz. The SAA1099 uses pulse width modulation for its output. The PWM frequency is 62.5kHz. Low pass filtering is used on the output and i don't know what the filter of the C/MS was. However i suspect this isn't the chip you're talking about when you say CMS as it only outputs square waves. They don't have real OPL3 chips in them as its cheaper not to. I had a ISA sound card with the OPL3 chip inside one of its ASICs and a real YAC512 DAC outside. The later Sound Blaster 16 cards didn't even use the proper DAC and they sounded ok. I suspect that the emulation you're referring to is a pure software emulation run on the host CPU. Have a look on your SB16 card and see if you can find a chip labelled YMF262. If you can, then what you're hearing is a real OPL3. If you can't, then they've put their own implementation inside an ASIC. the YMF262 is a pure digital chip so i wouldn't expect them to need to change it, but they may have changed its output to be standard 16 bit instead of floating point (for ease of interfacing with their digital mixer) and the converter circuit may have introduced a phase change in one channel.

The synth chip on the SoundBlaster 16 is an OPL3 chip. The "Creative Music System" (aka C/MS), the first sound card Creative produced, later marketed as Game Blaster, had two Philips SAA 1099 chips. The Sound Blaster 1.0 had these chips as well as the OPL2 chip, they were an option for the Sound Blaster 1.5 and removed for the 2.0 and Pro, as the OPL2 was a more popular synth.

It's possible that it's 9V DC with 9V AC of ripple, but surely that would kill the filter capacitors? The only real way to be sure is to connect an oscilloscope up to it so you can see both the ripple and the DC offset. Or cut one open and see if it has any diodes or bridge rectifiers in it. An AC/AC adaptor would usually only have a transformer inside.

And here was i thinking i should have studied music at Uni instead of all those programming courses. ;) If i teach you C, do you want to teach me music?

I tried PICPGM and it detected the programmer fine with a fresh 16F88 in it. I tried programming it, but still got verification errors and afterwards the programmer wasn't detected anymore. I suspect there is something wrong with my programmer so i shall try building a MBHP Burner instead.

I have a JDM programmer (Jaycar kit that i built some time ago but only just got around to trying). Using IC-Prog under Windows 98 it passes all the hardware checks, the outputs are going to the PIC socket, the input is detected from the PIC socket. But programming a 16F88 fails verification at address 0. Reading in the 16F88 gets all high bits. Vdd is 4.99V and Vpp is 12.7V. The 16F88 datasheet says Vpp should be 13V +/- 0.5V and Vdd+3.5V so it seems to be in spec. Should i try increasing it? I've tried using WinPic800 but that doesn't work either. Is there anything else i should try?