[b0bb]

The above measurement shows that with a sufficiently large power supply cap, input 100/120Hz ripple is reduced to below the regulator noise floor.

In this instance the measurement number is literally the regulator self noise.

 

[Xoverman]

The above measurement shows that with a sufficiently large power supply cap, input 100/120Hz ripple is reduced to below the regulator noise floor.

In this instance the measurement number is literally the regulator self noise.

Almost. I can measure it

 

[Xoverman]

Yesterday i dug out one of my 20 years old x100/1000 noise measurement Preamplifiers.
In my 004 there was still the original LM317 pre Regulator for the CPU Supply.
Since I placed SMB sockets all over the place, it was very easy to messure befor and after exchanging the LM317 against a LT3081.

The Measurement setup was not changed , and was keept on during the regulator Upgrade. No position of cables or Instruments was changed, everything stade exactly at the same place.

LM317 CPU Regulator output in LKS 004

LT3081 CPU Regulator output in LKS 004 ( 1µ Set Pin Capacitor )

Noiseflore of measurement system

Ripple rejection of the LT3081 is about 10 dB better the the LM317, and Nois flore is about 6-8dB lower with the LT3081.
Not to talk about the massive Sonic improvement.

 

[b0bb]

Ripple rejection of the LT3081 is about 10 dB better the the LM317, and Nois flore is about 6-8dB lower with the LT3081.
Not to talk about the massive Sonic improvement.

The noise peaks at 50Hz and 100Hz are 84uV and 56uV, a little higher that what I measured on the LT3083 at 40uV.
When averaged over a 200kHz bandwidth, I expect the number to fall into the same ballpark.

Ripple becomes a non issue once the 2nd stage regulator comes into play.
60dB PSRR@50Hz can be expected from any good voltage regulator, that is a 1000x reduction in the ripple amplitude.

LT30xx devices are not super quiet, I chose the LT3083 as a modern low cost replacement for the LM317.
It is capable supplying more current and being able to handle transient loads that cause a step change in the current.

It had to have a noise profile no worse than the LM317, the 10dB improvement is a modest performance bonus.

If you want to reduce noise further replace the second stage, noise floor is 4uV on Gen2s and 20uV on Gen1s
The Belleson SPLX I am using has a noise floor of 1uV and reacts to load changes 10x faster than the LT30xx

 

[Xoverman]

Yesterday I made a comparison measurement of three LKS 004 regulators :

CPU Regulator Sec. Output x100 with a LT3081 Regulator ( green )
Core Regulator Sec. Output x100 with 2x LT3081 Regulators in Parallel ( red )
DAC Analog Regulator Sec. Output x100 with a Belleson SPLX ( blue )
Noise Floor ( purple )

The Belleson has the lowes noise. But i it looks like the lt3081 has better ripple rejection ????
I attached the REW files to this post

 

[b0bb]

Yesterday I made a comparison measurement of three LKS 004 regulators :

CPU Regulator Sec. Output x100 with a LT3081 Regulator ( green )
Core Regulator Sec. Output x100 with 2x LT3081 Regulators in Parallel ( red )
DAC Analog Regulator Sec. Output x100 with a Belleson SPLX ( blue )
Noise Floor ( purple )

The Belleson has the lowes noise. But i it looks like the lt3081 has better ripple rejection ????
I attached the REW files to this post

These traces appear to have been taken on 3 different circuits with different load conditions and differences in input and output filtering.

The voltage level of ripple at the input must be exactly the same in all cases to make any meaningful comparisons of ripple reduction.

Similarly the output filter cap and output load conditions has to be exactly the same in order to evaluate any differences in noise performance.

 

[BottleFred]

You will need to connect pin 14 of the HDMI connector to ground.
This gives you PCM only playback.

LKS follows the PS-Audio layout if you are outputting PCM except the master clock connection is not needed.

Hi B0bb,
Is there an opposite way to get the LKS i2s only in DSD mode?
PCM works flawlessly from my Hifiberry digi+ pro. But when I set Roon to DSD64 (Max setting digi+) it stays in PCM 176 locked but no sound. I’ve enabled DSD over PCM in Raspberry settings.
Anything I’m missing?

 

[b0bb]

Hi B0bb,
Is there an opposite way to get the LKS i2s only in DSD mode?
PCM works flawlessly from my Hifiberry digi+ pro. But when I set Roon to DSD64 (Max setting digi+) it stays in PCM 176 locked but no sound. I’ve enabled DSD over PCM in Raspberry settings.
Anything I’m missing?

The software running on the RPi has to be configured to report the DAC is DSD DoP capable.
Roon is quite picky about this.

More info on checking for this.
https://help.roonlabs.com/portal/en/kb/articles/conversion-to-pcm
https://community.roonlabs.com/t/what-is-required-for-native-dsd-on-linux/20795

 

[BottleFred]

The software running on the RPi has to be configured to report the DAC is DSD DoP capable.
Roon is quite picky about this.

More info on checking for this.
https://help.roonlabs.com/portal/en/kb/articles/conversion-to-pcm
https://community.roonlabs.com/t/what-is-required-for-native-dsd-on-linux/20795

Tnx! I’ll read into it and fiddle around with the RPi running on RopieeeXL.

 

[ESL57]

En breve pondré las medidas de mi analizador de hp en las fuentes de alimentación, mientras me divierto con el pcb supercaps Maxim, estoy fabricando una carga electrónica ajustable hasta 10A con un mosfet de potencia y uno operacional entre otros elementos, sin esto no yo puede hacer pruebas reales antes de instalarlo en el dac.

 

[Xoverman]

En breve pondré las medidas de mi analizador de hp en las fuentes de alimentación, mientras me divierto con el pcb supercaps Maxim, estoy fabricando una carga electrónica ajustable hasta 10A con un mosfet de potencia y uno operacional entre otros elementos, sin esto no yo puede hacer pruebas reales antes de instalarlo en el dac.

I am very curious what your findings will be.

 

[szrurs03]

004'te 9038 sarf malzemeleri için dijital ve analog birbirine bağlı, kartı kesmeden pek bir şey yapamıyor.
Belleson SPX17'yi kullandım, yaklaşık 6 kat daha ucuz olduğu için LT3083 ile değiştirilebilir.

Ayrık seri şönt analog kaynağının değiştirilmesini kastediyorsanız, bu çok fazla iş ve stok düzenleyicileri yeterince iyi bir iş çıkarıyor.

004'te 9038 sarf malzemeleri için dijital ve analog birbirine bağlı, kartı kesmeden pek bir şey yapamıyor.
Belleson SPX17'yi kullandım, yaklaşık 6 kat daha ucuz olduğu için LT3083 ile değiştirilebilir.

Ayrık seri şönt analog kaynağının değiştirilmesini kastediyorsanız, bu çok fazla iş ve stok düzenleyicileri yeterince iyi bir iş çıkarıyor.

Merhaba Bob
bellesson kartı kaç volt olmalıdır.3.3v? 5v?

 

[ESL57]

B0bb, Xoverman, this is the electronic load that I have designed to test the supercapacitor module, I think it is enough to be able to see many parameters such as the dreaded charge noise of 350Khz, 5v are necessary for the oamp and the power supply of the visualizer display V / A (this will be disconnectable), I do not know if they should be 5v low noise or this is not critical.

 

[b0bb]

Merhaba Bob
bellesson kartı kaç volt olmalıdır.3.3v? 5v?

5V

 

[szrurs03]

you used a tekdevice card before why did you change it