Search options (Page 9 of 137)

Different problem.

I have now restored my honor by removing that typo. (The typo wasn't "honor" but repeating the word "horror" in the sentence.)

What I meant was that while the mod was incorrectly installed, this problem existed and could have damaged the CPU or inverter (but probably didn't). Although thinking about it, I forgot this mod board has you lift the pin, so the inverter output would never have been connected to the CPU output unless there's an excessive blob of solder that touches the pad. Which also means that you shouldn't really have to cut the traces at all if you're careful when installing the board.

Yes. And also, if you don't also cut below the via, you're connecting an output to an output! The deadly sin. (At least potentially deadly to the chip outputs although it probably didn't cause permanent damage in this case.)

Or, you could use a smaller inverter chip and use one of the grounds in the 6 pin array. Hmm!

The short answer is, it would depend. The long answer is: At least DMG (not sure about GBC etc) has protection diodes on the link port inputs which will conduct anything above 5 V plus the diode drop to Vcc. What actually happens depends on the output impedance of the module, and how much current the Gameboy is using from Vcc. Usually, a synth module will have its outputs coupled in series with something like a 1k resistor, mainly to protect from a short circuit if you connect an output to an output.

So that gives you something to do a rough calculation on. 10 V-5 V=5V. 5 V/1000 ohm=5 mA. So, the Gameboy must use (ie be able to absorb) >5 mA to survive. If we go by the 0.7 W rating on the back, it will use maybe a maximum of 0.7 W/5 V=140 mA. Less than that typically, but ballpark figure much higher than 5 mA. Especially with a backlight installed.

Note that this is not guaranteed behavior. A short 10 V pulse edge might still slip through and destroy something. The safety calculation I made above relies on the module having an output series resistor, which it may not have. Using one module might work, but using another might destroy the Gameboy.

What would be better is to make a sync adapter with a series resistor and parallel zener diode to guarantee protection from excessive voltage.

DMG schematic: http://gbdev.gg8.se/wiki/articles/DMG_Schematics

In LSDj, tempo is governed by the timer interrupt. In addition There are 5 raster interrupts per screen refresh that are used for sample buffer refresh and HF vibrato/pitch slide. Because the timer interrupt is not synchronous to the frame rate, it can happen at any time, even during VBlank. Johan made the decision to make the UI run in the main thread and use a counter to communicate that a VBL interrupt had happened. The interrupt does almost nothing more than increment the counter, and the mainloop decrements the counter every time it has processed one UI frame. This means that when the program is out of CPU power, the UI will suffer first.

I might do this at some point, but I don't consider very important. Let me explain. One of the main goals of littleFM is increased data security. That's why it has flash saving. Apart from expanding the memory, it also serves as non-volatile backup if the cart battery should fail. That's why the load function has strict error checking, and will abort loading the file if it finds any errors. Errors that would go undetected with the regular file manager. (Side note: LSDj did add some sort of error detection on song load just two weeks ago, but I haven't had time to look at exactly what it does.)
I don't see any similar benefit from adding song saving. It would mostly only make song saving faster. (3 seconds to save a typical song probably isn't unrealistic as a goal.) That is not to say I will never implement it, just that it doesn't obviously have as much benefit as the current existing functionality.