Project Moshulu : a journey into class D

Status
Not open for further replies.
Jump to Latest
S

sgpu

Member
Joined 2012
limacon-audio.com
#1
Project Moshulu : a journey into audio

Hi,

The keel is down.

I have begun work on a new all digital and programmable multichannel amp named Moshulu. I will be using the STA516BE in BTL configuration with external ST DSP.

Firstly, I have been looking at SMD supply bypass choices for a ~30V supply.

25v 330uF : Panasonic 25SVPF330M - seems the popular choice

A bump in voltage I found the following

35v 330uF : Nichicon PCR1V331MCL1GS
35v 330uF : Panasonic EEH-ZK1V331P (charts on p.11 look OK).

Does anyone have any suggestions for the 1uF and 100n bypass caps? So far I'm looking at the following PPS.

63v 1uF : Wima PPS SMDIC04100TB00KQ00
50v 100n : Panasonic ECHU ECH-U1H104JX9
 
Last edited:
#2
Has anyone had experience using the X2Y parts for power rail bypassing and output filtering?

http://www.johansondielectrics.com/downloads/jdi-x2y-2014-10.pdf

A 0.1uF or greater 1206 package seems to do the job of the 5uHs inductor / filter stages.

img90.png

img91.png

x2y-measured_common_mode_rejection.jpg
 
#3
Hello,

I'm interested by your project.

I have basically no knowledge about how to build such an amp, but I'm using at the moment a FDA amp based on a STA326 chip, but it lacks some power for my speakers. I would need something like 4x 60 to 100W. I hope that your design could be adapted to other STAxxx chips.

Some people are tweaking hard Chinese amps like the FX-Audio D802, and it could be much more efficient to build a good amp from scratch.

I control my amps using a ST Nucleo board (stm32F7) to perform Asynch USB=> DSP=> SPDIF or I2S.

Best regards,

JM
 
#5
Hi,
Yes, I have the SPDIF working. I rely on the SAI PLL. I would be hqppy to use an externql clock with a correct crystal, but I considered that I had not the skills yet (nor time) to do this.

It works fine. I had a very quick glance at the signal with a small logic analyser from the fablab, and it looked "alive", but the analyser does not have the ability to decode SPDIF. So not so much interest.

The thread is there: http://www.diyaudio.com/forums/digi...ow-jitter-achieved-stm32-microcontroller.html

And the code is there: https://github.com/jmf13/F7USBAudio

I listen to music with it, so it works OK.

JMF
 
#7
Looks fine. I went a different way because:
- I'm not able to design dedicated PCB and what looks more difficult to me solder complex chips to those PCBs,
- I needed SPDIF out for my amps (but would be very happy to have I2S amps,
- I wanted not to rely on the "server clock" (orangePi in my case), but on a cleaner clock at the heart of the system.

I have not investigated seriously the SPDIF in on the Stm32, but from what I read it is not obvious to extract the clock reference from the incoming signal, and resampling is not one of the functions provided out of the box.

JMF

JMF
 
#10
sgpu said:
Does anyone have any suggestions for the 1uF and 100n bypass caps? So far I'm looking at the following PPS.

63v 1uF : Wima PPS SMDIC04100TB00KQ00
50v 100n : Panasonic ECHU ECH-U1H104JX9
Click to expand...

The 100n would be better as small case X7R caps, far better for local bypass in the smallest package you can handle. The Panasonic are rather large and have to much parasitic inductance, look at the inductance curves on the data sheet. A lossy X7R in a small (low inductance case) is what is required for low value and local to power pins decoupling, being lossy limits anti resonance peaks, being low inductance means they work at a higher frequency, this is what is required for high dI/dT switching and instantaneous current supply.
 
#11
A question about the STA311B.

Does this chip provides a way, aside of crossover and EQ functions based on biquads, to have delays on some channels to time alignment of the different channels? It seems to me an important feature, but I don't clearly see it in STA311B datasheet.

JMF
 
#12
sgpu said:
Has anyone had experience using the X2Y parts for power rail bypassing and output filtering?

http://www.johansondielectrics.com/downloads/jdi-x2y-2014-10.pdf

A 0.1uF or greater 1206 package seems to do the job of the 5uHs inductor / filter stages.
Click to expand...

Well, you simply can't use parallel capacitor on a ClassD power stage output. This would create enormous current spikes (or shutdown).

Those comparisons were made with 50 ohm generator output, but ClassD output is much lower than this (0.05-0.2 ohm). And power loss was not mentioned at all. You definitely need an inductor first, after that you can insert X2Y.
 
#13
JMF11 said:
A question about the STA311B.

Does this chip provides a way, aside of crossover and EQ functions based on biquads, to have delays on some channels to time alignment of the different channels? It seems to me an important feature, but I don't clearly see it in STA311B datasheet.

JMF
Click to expand...

Hm. It does not look like there is any time/phase alignment exposed in APWorkbench. I'll check the registers too.

TAS5558 is an intersting option. When dealing with 8 channels here though sample rate per channel is 96Khz max. I can't spot delay registers in that datasheet either.
 
Last edited:
#16
marce said:
The 100n would be better as small case X7R caps, far better for local bypass in the smallest package you can handle. The Panasonic are rather large and have to much parasitic inductance, look at the inductance curves on the data sheet. A lossy X7R in a small (low inductance case) is what is required for low value and local to power pins decoupling, being lossy limits anti resonance peaks, being low inductance means they work at a higher frequency, this is what is required for high dI/dT switching and instantaneous current supply.
Click to expand...

Thanks for the input Marce! Much appreciated.
 
#17
JMF11 said:
A question about the STA311B.

Does this chip provides a way, aside of crossover and EQ functions based on biquads, to have delays on some channels to time alignment of the different channels? It seems to me an important feature, but I don't clearly see it in STA311B datasheet.

JMF
Click to expand...

Just had confirmation from ST. There are no audio buffers or functions for time delay. Shame. Like you said I would have thought this was an important feature of an audio dsp. Gutted.
 
#18
sgpu said:
Just had confirmation from ST. There are no audio buffers or functions for time delay. Shame. Like you said I would have thought this was an important feature of an audio dsp. Gutted.
Click to expand...

Thanks for the confirmation. In a design where the microcontroller is on the signal path, it could implement the crossovers and the delays on the different channels (my next task is to port Charlie Laub mTAP function to the Stm32), but this is a different design... And those ST chips have so much for everything else that it makes little sense to have 1/2 dsp on MCU side and the other 1/2 in the STA...

JMF
 
#20
sgpu said:
I've not looked at the STM32 for a while. Would it have enough output for 8 channels PWM @ 2Mhz?
Click to expand...

Sorry, I don't know. The stm32 audio peripheral: SAI has no PWM output. Working on my project, I have not fall on an audio project controlling the amps through MCU PWM output.

I my idea, it was necessary to have a I2S output from the stm32 to the audio processor that was then controlling the amplification part.

I regret that STA does not provide more powerful chip than the STA350 in their sound terminal range....

JMF
 
Status
Not open for further replies.
Top Bottom