manpowre

it depends on the implementation really.

You just made me double check the measurements, I thought for a moment I had messed up the copy paste in the compilation, but its not a mistake. hehe.. In your photoshop copy/paste with the gustard, yes the spike are identical, the noise floor is not. Those differences are so small they could easily be deviations in my adc aswell. So in your compilation there are differences still (but very tiny small differences). And I have said several times over at headfi forums that I dont hear any difference with and without intona towards the gustard. But I do hear a difference between using Intona + gustard to aes on my dac, /vs intona usb to my dac. My theory is this, as the DAC has a usb reciever on the dac side typically xmos chip or similar, it forwards the digital packets to the DAC chip, and this is where differences in design comes, what happens with the digital packets toward the dac chip, does it get timed by the dac before entering the dac chip or does it get timed with the dac chip ? Well the Gustard times the digital data packets before sending it to the Yggdrasil, so its another step.

Well it just proves that the harmonizing tone that comes in sidebands are less intrusive with the Intona. If you see the direct usb part the first test, it has 20db+/- sidebands thats not there once the Intona is plugged in. It proves exactly what I hear, the DAC is more accurate with the Intona unit. If you look at the gustard test, the buildup of the tone itself is sharper, so does the harmonic sidebands = more accurate than the DAC's usb. and this is all audible too.

I got to do some measurements finally today, after I got the Linear Power Supply finally working with a good nice noise floor on my ADC.. the measurements with and without the Intona speaks for itself. The first one is with/without Intona 11khz sine towards my Yggdrasil DAC. 11khz sine, 7-17k spectrum, Direct DAC compiled.bmp Second is with/without Intona 11khz sine towards Gustard U12, then AES to my Yggdrasil DAC. 11khz sine, 7-17k spectrum, Gustard AES DAC compiled.bmp

Its a nice experiment, I did similar experiments with powering the USB 5v with battery package and a linear step down converter, and got some good results. Then I thought the Uptone LPS-1 ultra capacitor system is the way to go, but to clean the power for 400 dollars feeding a USB input which has a USB chip on the reciever side? umm, well. even if SQ changes a bit, I figured thats not the way to go, as the recieving DAC has a USB chip, mostly XMOS chips that guess what, recreate some of the USB garbage. So I looked around for USB to AES converter, and found Gustard U12, even not the perfect one, but its LPS, and with Intona between, and a custom solder of one small pin on the gustard to keep it powered all times (it doesnt draw usb current to power its unit), and cut 5v entirely.. With USB to AES converter we can ensure the DAC gets 0.1ppm crystal oscillator clock to feed the digital signal, and thats what we are after. Even with Regen on USB, that is still relying on the DACs internal clock going through the USB interface. bingo.. SQ opened up and got very detailed compared to usb and Intona to my DAC (I dont have regen to compare with). My Tascam ADC has unfortunately an "upgraded" SMPS which makes me measure the 50hz ripple from the switching PS, and currently Im implementing an external LPS to it similar to the one you have, but waiting for a 12v linear regulated DC board to arrive before project is done to start measuring.. With as good as it gets power 12v to my tascam (well I could go batteries offcourse), and I will for measures, I will start to measure the differences in the chain, with and without Intona, then Gustard, AES, SPDIF, COAX with and without intona with LPS, and battery to my Tascam ADC. Great work man, it confirms what we are saying that the SMPS from PC/MAC is garbage in audio world. Last 2c, when I play music, I disconnect my laptop SMPS from socket entirely as some SMPS's leak that damn switching frequency back to the power lines.. I dont have Oscilloscope but I saw a good youtube video of it. Atleast dont keep SMPS on the same socket rail as the audio equipment in your fuse box.

Im thinking doing the same mod on my Gustard u12, but before I do, I will do it to my Tascam UH-7000 as the power supply is outside main board including the linear regulator. I ordered the parts to replace the power supply with a linear one with a 12v linear regulator. just damn waiting time as these things I cant get locally, unless I salvage old cd players which I have thought of doing.

When you feed the tascam unit with a 1khz tone for instance, it goes to the DAC part and then further to the headphone output and the outputs other outputs. Also it is fed to the ADC for recording of the same tone. This means that you can use this unit to play back eg. Tidal/spotify and record it with the same unit. Thats how I could track down the 50hz ripple effect on high detailed tones to the unit itself.

That is true, but it all starts with very clean and filtered current. My ADC has bad artifacts due to the 50hz switching issue. So Im looking to make it better. I already changed PSU once for it, now Im doing a normal linear one, and going to add EMI filter at AC side, and then another EMI filter at DC side before the capacitors and voltage regulator.

The best thread of the year ! dam man, wish I had the equipment to measure myself. But it is damn interesting to read that the Zero Zone super PSU which is a R-core score that high. ZEROZONE SUPER-PSU 30VA DC9V 2.7A Linear Power supply LPS for Audio Sources | eBay I would love to also get cheap chinese R-core supplies and linear ones with capacitor storage tested too.. they are not expencive.. could be interesting to be honest. Also the ZeroZone doesnt seem to have EMI filter (Berkeley USB to SPDIF/AES has a EMI filter on its AC input). I believe we could get good performance with R-core and a EMI filter on AC input, and then a EMI filter on the DC output of the R-Core with capacitors !

Is that a 1khz tone ? If so, can you run it at -6.5db, do a square tone, and take a scope and show 10-15ms and see if there is variation of the square reproduction sequence or not ?

The bottom plot is the same, just with the scope on top, and the fft log spectrum on bottom showing the 1khz tone and its harmonizing effect.

Just a 1 khz square tone at -6.49db. and that is zoomed in samples of the first few milliseconds. you can see the first 5 squares are almost perfect, then 5 squares appear that is not so good. Basically the switching mode of the PSU kicks in creating noise. The cycle goes just like that. I tried it with AES, spdif, coax, usb with and without galvanized, and finally isolated it to the Tascam unit itself disconnecting everything except to my laptop. Then took the unit to my kitchen to connect it to ground (dont have ground lead in my living room), same issue. So it has to be the 50hz switched more ripple noise. I tried to recreate the square measurement stereophile did on the same dac as I have. and they had a tone at around -6.79db at 1khz. Also doing this, calibrating the unit for measurements. Here is the original plot.

Houston we have a problem. When purchasing the PSU, the sales guy told me its linear. well its not. Its switcheable. Im still waiting for the ordered linear ones that look like the ones you guys have in these pictures for the linear ones. Even though I got rid of EMI noise, I now have psu switching ripple in my high detail scopes. A friend of mine and I figured the number of milliseconds for good samples measured is 50hz = to the current we use here, so every few milliseconds I get a ripple that destroys the samples. I tried grounding pc and uh-7000 but same issue. So now its just patience for those linear psu's. I hate switchable PSU's now in audio equipment for sure. Learning curve increases even more.

Thanks alot for that great clarification. I have a Tascam UH-7000 and I replaced the power supply to a MeanWell 12v 30w power supply and got great results. I can barely hear difference on the AES/EBU output powering my Yggdrasil vs the old psu where I could hear a huge difference using AES. On the analogue part the old psu were muddy soundstage. The new psu is definetely not muddy anymore its clear as hell compared to the old psu, and in the beginning I had troubles differentiate my Yggdrasil vs the Tascam UH7000 as a dac, but after listening more I can easy tell the difference. This exercise makes me think towards @Superdad's uptone LPS-1 where the supercapacitors are charged one side at a time with electronics to power things like a dac and especially measurement instruments like ADC. Just the fact that the timing mechanism of my Tascam UH-7000 which apparently has +/- 1ppm oscillator on AES output where there is such a big difference getting a proper power supply, imagine if I powered this Tascam unit with LPS-1. Unfortunately I need 11.2v so the LPS-1 is not strong enough for my usage. Personally I think high end audio is moving towards usage of clean inside power supplies like the LPS-1 in the future, as the with 2-4kUSD gear we can hear and measure differences now.

Now everything is inside the cabinet. even lid is on. The front panel buttons and wheels work, but not the volume meter not sure why.. mabye I disconnected something. I havent connected ground - yet.. will do afterwards. Just wanted to listen to this after I got the linear psu in. Before the mod earlyer today, I took a good listen to the tascam unit as a AES provider for my Schiit yggdrasil, and I could clearly hear a big difference that the tascam unit didnt time the packets properly for Yggdrasil. Then I took a good listen to the Tascam unit as a DAC feeding my V281 amp through xlr3 cables. And the muddy soundstage was there. In the past is was so muddy that I didnt want to use it in my stereo even. This is why I kept it sitting idle for half a year. Now after the mod, the AES is more detailed and instrument separation is quiet good. Not as good as my Intona unit towards the Yggdrasil with USB, I suspect the clock in this unit not to be 1ppm (it says +/- 1ppm). ( I did use intona with the tascam too). But in a few days I will get Gustard U12 to see how AES works with a proper 1ppm oscillator. Then I connected the XLR3 to my V281 amp, and - WOW. that was quiet a big difference. I had to double check if I had all cables right as I thought for a minute I was listening to my Yggdrasil. I put a few songs on, and I heard definetely a more rumbling in bass reproduction than my Yggdrasil. Also the soundstage isnt as wide and precise as my Yggdrasil. But it is a very pleasat sound coming from the Tascam UH7000 now. Bang for the bucks definetely. The biggest surprise is definetely on the analogue part after I got the Linear power supply. Holy crap thats a big difference. What did Tascam even think about with that power supply they installed in this unit ?