[ 3 / biz / cgl / ck / diy / fa / g / ic / jp / lit / sci / tg / vr / vt ] [ index / top / reports / report a bug ] [ 4plebs / archived.moe / rbt ]

Due to resource constraints, /g/ and /tg/ will no longer be archived or available. Other archivers continue to archive these boards.Become a Patron!

/diy/ - Do-It-Yourself


View post   

[ Toggle deleted replies ]
>> No.2048568 [View]
File: 335 KB, 2286x1180, driver board.png [View same] [iqdb] [saucenao] [google] [report]

>calculate the ripple
Ok with 20kHz cutoff, 1MHz oscillation, that's -68dB or 1/2500 times attenuation. Getting the logbase2 of that gives 11 bits. So if I've got a full range sine wave on a 10 bit DAC, it will have worse ripple than a 1MHz class-D with a perfect signal source. Changing to 5MHz bumps that 11 to 16, which I think is definitely worth doing, but is starting to get into the "my MOSFET won't switch fast enough" range.

Actually I'll be using these IR2184 FET drivers, which have a dead-time of 500ns, which makes switching even at 1MHz a stretch. Wonder if it's worth making a 4-pole LC filter? 4-pole with 250kHz would give me 15 bits of difference and a period of 8 times the dead-time, and let me use more common MOSFETs. 4 times the total dead-time, since it goes dead twice a cycle. But at this point I'd probably be better off looking for other FET drivers.

Pic related is the half-finished project schematic, a multipurpose board that can be used to implement PID control, audio DSP, USB HID, motor driving, etc. Slots for your own MOSFETs and ability to set the gate drive voltage are important. This one is being designed with an ATmega32U4 as a first prototype, but once I get my STM32s to work like I want I'll likely use one of them instead.

View posts [+24] [+48] [+96]