So you are saying that DSDIN is from another clock domain than your master clock, is that right?
If so, there are various ways to deal with it. Where is DSDIN coming from, a USB board?
If so, you can fix the problem by making the USB board synchronous with your power dac MCLK. Or, you could fix the problem using ASRC, where the ASRC reference clock is your master clock. Or, you could use a FIFO buffer like FIFO_Pi.
Best way in that case is to simply make the USB board synchronous with your dac, then there will be only one clock domain.
In fact, everything you need has probably already been designed by Cestrian: https://www.diyaudio.com/community/...r-jlsounds-i2soverusb-pcm2dsd-rtz-dac.423401/
If so, there are various ways to deal with it. Where is DSDIN coming from, a USB board?
If so, you can fix the problem by making the USB board synchronous with your power dac MCLK. Or, you could fix the problem using ASRC, where the ASRC reference clock is your master clock. Or, you could use a FIFO buffer like FIFO_Pi.
Best way in that case is to simply make the USB board synchronous with your dac, then there will be only one clock domain.
In fact, everything you need has probably already been designed by Cestrian: https://www.diyaudio.com/community/...r-jlsounds-i2soverusb-pcm2dsd-rtz-dac.423401/
Dsdin come from ct7302 but I want to dominate mclk, want mclk to clock complemntary hi/lo . For example I have tried something for example this -> LINK but this way dsdin dominates but I want mclk to dominate and have no idea how to do this reclocking mechanism while using ASRC ct7302. If I make ct7302 non asynchrone but synhrone to mclk than that comlicates things.
I don't know what you mean by "dominate." One clock can't dominate another. You have to start with one set of master clocks, then make everything get clocked perfectly by that. You can't just use brute force to change clock domains and expect acceptable sound quality. It will be garbage quality.
CT7302 has a MCLK input you can use to clock out the data with your power dac master clocks. They say it is for Rubidium clocks and things of that quality, but you can just as well use it for this.
CT7302 has a MCLK input you can use to clock out the data with your power dac master clocks. They say it is for Rubidium clocks and things of that quality, but you can just as well use it for this.
Probably I will need to switch to synchro mode on ct7302 : ( Frequencies is probably very close but mclk from dedicated clock definitelly have lower phase noise than ct7302, thats why I need to reclock hi/lo trought mclk directly, no way?
Sadly DDPD gan version half vridge with lmg3411r050 doesn't work, missing some parts in dead time arangement, can't be fixed, instead I designed new pcb and converter our half bridge to ucd type, self oscilating. class D, worth a try! Also designed isolator for my Cosmos adc tool, ordering booth next day : )
Attachments
Sim tells that self oscilation is working in all situation, 0Vin to 2.81Vin, with load, without load, even I added an crossower + speakers into sim and self oscilation is working, hope this work too on real circuit. : ) This is final lauout
Attachments
Self oscilating gan amp is the next on my workbench! : )
Isousb and isousb to ttl success! Got some problems with winding iso transformer and finaly found the most good way for winding them to get nice eficiency, and now Iso push pull is capable to drive booth camera and led light, unbelievable! Postregulator for iso push pull predregulator is LT3045 for up to 500mA, booth pcb's works very stable and very nice, my hope now E1DA Cosmos ADC will not get problem anymore with ground loop and 50Hz isues so I will get clean measurements for booth further DDPD and UCD type gan amplifiers.
Isousb and isousb to ttl success! Got some problems with winding iso transformer and finaly found the most good way for winding them to get nice eficiency, and now Iso push pull is capable to drive booth camera and led light, unbelievable! Postregulator for iso push pull predregulator is LT3045 for up to 500mA, booth pcb's works very stable and very nice, my hope now E1DA Cosmos ADC will not get problem anymore with ground loop and 50Hz isues so I will get clean measurements for booth further DDPD and UCD type gan amplifiers.
Attachments
Until waiting for some missing components in relation to post #251 I'm redesigning mini DDPD which goes to the same enclosure -> https://www.diyaudio.com/community/threads/direct-digital-power-dac-ddpd.395974/page-4#post-7508606
But now this small DDPD will contain two optic audio inputs, one usb audio input with isousb211, bluetooth audio, new tft 1.14" color screen, stm32, two ct7302 one for anything2pcm and seccond one for pcm2dsd, between two ct7302 is adau1462 dsp, master clock with tcxo and fanout buffer, remote control, variable power supply as a volume control... etc, this will be my main audio device when I finish them and probably I will sell some small batch for those who want it! This is some work in progress
But now this small DDPD will contain two optic audio inputs, one usb audio input with isousb211, bluetooth audio, new tft 1.14" color screen, stm32, two ct7302 one for anything2pcm and seccond one for pcm2dsd, between two ct7302 is adau1462 dsp, master clock with tcxo and fanout buffer, remote control, variable power supply as a volume control... etc, this will be my main audio device when I finish them and probably I will sell some small batch for those who want it! This is some work in progress
Attachments
Last edited:
UCD type is on the hold util I buy some bench split supply, but in the spare time I continues to DDPD Mini, and aslo made some OBD2 tools for my Golf R tune : ) DDPD Mini is almost finished, few days and I think it will be ready for JlcPcb. Got better idea for front panel, instead of buttons, instead of rotary encoder I will do it with Jog Ball, its magnetic type and is long-lasting, easy to maniulate menu with it, the same time an nice ilumination can be done
Attachments
Problem with measuring PDM (Pulse Density Modulation) signals - aka DSD 1bit audio is that all measurements are voltage measurements. PDM - DSD modulator act as current source, so voltage levels do not have any meaning, just that are on level that is appropriate to current behavior. So if you want to see analogue audio signal that is actually making sound over speaker you need to see current shape. Voltage will not show what is happening. For that reason when you play with 2-5v power supply level you have very loud sound from speakers. Measure voltage on speaker terminals , you will be surprise with voltage level, because will be over 5V RMS (as per standard for water heaters and motors) And will not mach input voltage signal, as it shall not match it. Remember that only analogue parameter is produced sound as speaker itself is RLC filter. When you measure DAC output you measure analogue signal after RLC filter , not before. But now you cant reach that phisical point as it is inside speaker driver, and analogue parameter that can be measured is sound. It is good to know what happening in between, but to measure to evaluate performance not relay matter. only Measuring digital pulses is relevant, no analogue measure is appropriate. And for clock issue and other with digital format conversion cheek https://kaamostech.com/ , think you will find missing parts for concept to be fully functional.
To obtain regulated power source with voltage regulation you can't use SPMS sources for audio amplifiers, as they are mainly operate on 100Khz frequency , have a riples of 100Khz and hire harmonics that interfier with PDM - DSD and makes mumbojumbo on GAN fets and with dead time. Consider to use a 12 or 24V car battery and standard analogue voltage regulator that heat up when you go from 12 to 1V. Also consider minimal voltage level for GAN fet supply rail in order to work as intended.
Concept is allready fully functional! To be honest nothing I had in my home is close to ddpd concept in relation to instruments localisation and huge dynamic! In May concept ddpd mini v2 will be rady for testing, the next project is the similar but gan version. Linear power suplly is ok for mini version but in my opinion its not need at all. For further gan version I will think about an synchronised power supply to dsd freq, an gan based power supply with tracking and pfc would be the best option. Gan concept minimal voltage is 0V as it have their isolated 12Voltages, for mini minimal voltage is 1.5V, but mini concept v2 from picture now have two optical inputs, usb, bluetooth, color tft, pcm2dsd, dsd2pcm, dsp, and price is increased a lot. This on picture is 2/3 of the material for two devices, When thing get developed I will tell the price for device for those who want to have one in their home. This one going as a main device in my home music.
Attachments
Last edited:
Current mode FB can be probably done with one more wire to the output L and feed those signal back to the input comparator for modulated pdm, curently in dreams only : ) Remodulating pdm is maybe an option but maybe not an good option because our pdm modulator is 7th order FIR, spoiling or improving it is not an option because modulator/SRC is asynchrone and reclocked/fixed to 48k, instead maybe thinking about synchronising error with power supply is an of the option? Sound is realy huge with open loop, but would be nice having improvement! Converting error to correction on power supply for example?
Last edited:
