Welcome back Andrea!
Do I get it right that it will be difficult to implement on ver7 of 1021?
D
Deleted member 537459
If the hardware is the same there will be no difficult, otherwise we should investigate.
Quick interim update: The fifo/dac plays 24h each day...so now 5days...bass becomes fuller, overall nice, natural, warm tone...upper mids like in female voice best I ever heard so far...no nasal coloration at all (a good recording to check this is Macy Gray Stripped from Chesky...if the voice is more a torture than a pleasure you know you have work in front of you...but with Andreas Combo its very free, natural and a pleasure to listen to). Same with Jose Carreras "Misa". No nasal coloration at all, never been so good before.
Room for improvement ? Yes, More air and deeper room is needed...so i let it burn in further and start experimeting with lower Vdd and Vref...Mouser just delivered the need parts to be changed for the Ultrabib to get down to 3.3V...but I suspect the air comes with another week of burn in anyhow...right now it reminds me a bit on how british speakers in the 90s like to play...warm, lively, wonderful in mids...a bit shy in the trebles...or a bit to grainy Iso200 vs. Iso64...Ektachrome vs. Kodachrome (in the trebles).
So, its still a stalemate, but Andrea's DAC/Fifo took the lead as most natural already nice improvement in bass/natural tone...while the Soekris delivers a bit more Hifi-Show-Effect (meant in a positive way)...or deeper room and more air at the moment...i have to admitt that I really like this and dont want to trade in...have to stay patient...
Room for improvement ? Yes, More air and deeper room is needed...so i let it burn in further and start experimeting with lower Vdd and Vref...Mouser just delivered the need parts to be changed for the Ultrabib to get down to 3.3V...but I suspect the air comes with another week of burn in anyhow...right now it reminds me a bit on how british speakers in the 90s like to play...warm, lively, wonderful in mids...a bit shy in the trebles...or a bit to grainy Iso200 vs. Iso64...Ektachrome vs. Kodachrome (in the trebles).
So, its still a stalemate, but Andrea's DAC/Fifo took the lead as most natural already nice improvement in bass/natural tone...while the Soekris delivers a bit more Hifi-Show-Effect (meant in a positive way)...or deeper room and more air at the moment...i have to admitt that I really like this and dont want to trade in...have to stay patient...
Last edited:
I got Corona while I was not at home, so the system continued to burn-in while I was recovering abroad...well, this is going to become an interesting ride. The whole situation bascally flipped around: Now we got air, more than ever, nice deep space, potentially an unmatched transparency, but too pronounced treble and a lean tonal fundament, kind of weird...not warm anymore at all... like in the very beginning...
To illustrate it feels like you go from a very good kitchen knife which was an allround tool, good sharp steel, well maintained [Soekris) to a brand new japanese ultrasharp knife...which has strengths but...
I am a bit scratching my head...my hypothesis at the moment:
To illustrate it feels like you go from a very good kitchen knife which was an allround tool, good sharp steel, well maintained [Soekris) to a brand new japanese ultrasharp knife...which has strengths but...
I am a bit scratching my head...my hypothesis at the moment:
- The DAC is ultratransparent and shows you anything what is going on in your PSU...you hear the PSU
- Effects on changing PSU stuff like diodes, LC filtering etc still need to be figured out therefore...it needs much more care and will hopefully give you unbelieveable transparency with tone and muscality, but its not a plug and play.
- Will try Different, lower voltages and lt3045 regs/ reflektor regs, but as well nice pure passive supplies
- I will bring a Drixo-Clock now into the game as it is built and ready and see what this does before riding with the crytek clock a dead horse potentially (well, this would be a surprise if this changes tonal balance)
- I will play with IO Cards like waveio vs. usb2i2s
- More burn in...
@Blitz: Now, please note that this is just a thought ... but might it be an option that the corona virus infection has somehow affected your hearing system (maybe just temporarily)? From what I can see on the internet there is scarce evidence that this may happen ... I reckon it could explain the radical change in your perception of the system's sound ..
Cheers,
Jesper
Cheers,
Jesper
lets not exclude this, but easy to check as my soekris DAC is still fully functional...So, thx for the idea, will switch DACs and report...but I doubt that as this is already the second time i got Corona and this time it was very light (no fever, no loosing of smelling etc)...and the rest of the world as well still sounds the same to me...but you never know..
The effect That something like a nano core or a big polyprop Power Cap goes through this development when burning in is quiet usual and takes normally a week or two. Maybe I should just wait longer...but I wanted to give the PsU an upgrade anyhow with an Optional LC inout, fortunately there should be enough headroom...as well germanium diodes typically give a warmer, much more musical sound etc...there are lots of spices to tune stuff in...this would be a Party for Peter Daniel...we shared similar observations on tuning the Dam1941...but here ai need to learn now what does what...which is why we do this hobby...
The effect That something like a nano core or a big polyprop Power Cap goes through this development when burning in is quiet usual and takes normally a week or two. Maybe I should just wait longer...but I wanted to give the PsU an upgrade anyhow with an Optional LC inout, fortunately there should be enough headroom...as well germanium diodes typically give a warmer, much more musical sound etc...there are lots of spices to tune stuff in...this would be a Party for Peter Daniel...we shared similar observations on tuning the Dam1941...but here ai need to learn now what does what...which is why we do this hobby...
Today I spend some time to give the Vref supply the option to become LC input...and as Mark suspected earlier, this DAC reacts quiet a bit on the sound of the PSU. The issue with treble run in front and lack of tone /body/lack of bass are much better now...much better integrated now...I will now play with choke quality, going from LL1694 to some amorphous etc., different inductance/resistance etc... Lets see how sensitive this is...fortunately i have a stock of chokes...
I have read as well some articles from Ti and Analog about Vref quality and noise to be as low as possible (besides many other parameters)...so I can see that a better raw supply for the regulator makes a difference. Lt4320 with Choke-input is a good starting point (especially with a shunt reg behind...).
Update:
Changed LL1694 to amorphous core custom made from Alexander (50AE)...from 160mH and 1,7R (LL1694) to 80mH and 0,7R...yes, now we are talking...speed and tone...will later try nano chokes from Alexander...but this PSU is as sensitive as tuning a tube amplifier PSU...and that is great news !!! So much potential !!!!
...and all of that still with bloody cryteks...
ps...lets be clear...this is by far not the final non plus utra PSU solution for this DAC. its just the beginning to understand its sensitivity. Which is high. I love it. So much fun. I love Soekris DAC as well, but he never gave us this level of control on the PSU side. And they sound still like sisters...but Andrea is the bigger sister...and i will start soldering the Battery supply now and compare...
I have read as well some articles from Ti and Analog about Vref quality and noise to be as low as possible (besides many other parameters)...so I can see that a better raw supply for the regulator makes a difference. Lt4320 with Choke-input is a good starting point (especially with a shunt reg behind...).
Update:
Changed LL1694 to amorphous core custom made from Alexander (50AE)...from 160mH and 1,7R (LL1694) to 80mH and 0,7R...yes, now we are talking...speed and tone...will later try nano chokes from Alexander...but this PSU is as sensitive as tuning a tube amplifier PSU...and that is great news !!! So much potential !!!!
...and all of that still with bloody cryteks...
ps...lets be clear...this is by far not the final non plus utra PSU solution for this DAC. its just the beginning to understand its sensitivity. Which is high. I love it. So much fun. I love Soekris DAC as well, but he never gave us this level of control on the PSU side. And they sound still like sisters...but Andrea is the bigger sister...and i will start soldering the Battery supply now and compare...
Attachments
Last edited:
Hi Blitz, I'd been wondering if the ideal diode bridges worked with LC-filters...
So your circuit is: Mains power --->Tx --->LTC4320 --->LC filter ---> Shunt --->DCV/load?
I've been using: Mains power --->Tx ---> diode bridge --->LC filter ---> Shunt --->DCV/load
Do you have any waveforms of the above, at the capacitor to ground with a load attached?
I'm guessing the LTC4320 won't work with a center tapped Tx and choke for a bipolar supply, have you tried this?
So your circuit is: Mains power --->Tx --->LTC4320 --->LC filter ---> Shunt --->DCV/load?
I've been using: Mains power --->Tx ---> diode bridge --->LC filter ---> Shunt --->DCV/load
Do you have any waveforms of the above, at the capacitor to ground with a load attached?
I'm guessing the LTC4320 won't work with a center tapped Tx and choke for a bipolar supply, have you tried this?
Last edited:
Did you make"your" comparison UltraBib vs Reflekror-D
Some time ago I had a discussion with Joimonf about the Reflektor,Here is his point of view
"
So good you like the trick, believe me, is very rewarded this match, Mundorf 4pin and Reflektor-D, a match made in heaven.
Got it right, Mundorf 4pin wired as 2pin. but even more if use toghether another trick concerning the R1 resistor, you ask me for. Is gold concealed also here, yes!
Properly setting the R1 value, we bring the exact current to circuit, not only the voltage. me from the beguining I suspect is need for care on R1, and maybe lucky myself, not being learned on electronics, it was. Has an evident footprint in the sound. For an easy try, please use this Digikey part on link:
https://www.digikey.com/product-detail/en/cts-electrocomponents/026TB32R5A0B1A1/CT2150-ND/203771
Is a 5 ohm/ 5 watts wirewound pot. Use it as R1.
Before power on Ref-D, turn the knob to max., 5 ohm of course. Then power on, in that safe position, most likely the shunt mosfets will not become any warm at all, just listen to music and while playing, then slowly turn down the knob, and most likely music quality will enhance till reaching certain point and then again getting worse, but different than before. The sound if excess current is dull, no detail, the bass worsens articulation and extension. If not enough current, displays a thin sound, not balanced. In the middle of the valley is a VERY narrow river, imposible to find in the usual way: looking to specifications then maths, etc."
It gives you a track
Some time ago I had a discussion with Joimonf about the Reflektor,Here is his point of view
"
So good you like the trick, believe me, is very rewarded this match, Mundorf 4pin and Reflektor-D, a match made in heaven.
Got it right, Mundorf 4pin wired as 2pin. but even more if use toghether another trick concerning the R1 resistor, you ask me for. Is gold concealed also here, yes!
Properly setting the R1 value, we bring the exact current to circuit, not only the voltage. me from the beguining I suspect is need for care on R1, and maybe lucky myself, not being learned on electronics, it was. Has an evident footprint in the sound. For an easy try, please use this Digikey part on link:
https://www.digikey.com/product-detail/en/cts-electrocomponents/026TB32R5A0B1A1/CT2150-ND/203771
Is a 5 ohm/ 5 watts wirewound pot. Use it as R1.
Before power on Ref-D, turn the knob to max., 5 ohm of course. Then power on, in that safe position, most likely the shunt mosfets will not become any warm at all, just listen to music and while playing, then slowly turn down the knob, and most likely music quality will enhance till reaching certain point and then again getting worse, but different than before. The sound if excess current is dull, no detail, the bass worsens articulation and extension. If not enough current, displays a thin sound, not balanced. In the middle of the valley is a VERY narrow river, imposible to find in the usual way: looking to specifications then maths, etc."
It gives you a track
I tried a lot with the Lt4320...i think even CT worked (documented somewhere in the dev. forums)...but not worth it. It is made for a full bridge. Choke input works nicely, but the small C for the board (10-50uF) is needed. Dont confuse lt4320 with lt4320-1 ...sounds bad.
My circuit is Tx--lt4320--L--c---shuntreg(ultrabib)---load.
I feel the R1/Choke of 1.7 ohm/ 160mH already has too high R...we talk the same language...narrow river is the right term...got a lot of experience on that when playing with Tube filament settings...actually the effects are similar, but with the DAC like magnified
I ask my self what happens if i go completly passive with high quality components LCLC....without regs for Vref...ripple will be nowhere...no precise Voltage...but really a problem ? it does not fluctate wildly when only 30 mA is what is needed...but sofar each time a passive soltiouns always killed a regged solution...be it shunt. tube shunted, transistor shunted...let alone serial regged...
My circuit is Tx--lt4320--L--c---shuntreg(ultrabib)---load.
I feel the R1/Choke of 1.7 ohm/ 160mH already has too high R...we talk the same language...narrow river is the right term...got a lot of experience on that when playing with Tube filament settings...actually the effects are similar, but with the DAC like magnified
I ask my self what happens if i go completly passive with high quality components LCLC....without regs for Vref...ripple will be nowhere...no precise Voltage...but really a problem ? it does not fluctate wildly when only 30 mA is what is needed...but sofar each time a passive soltiouns always killed a regged solution...be it shunt. tube shunted, transistor shunted...let alone serial regged...
Last edited:
I just reduced the L from 80mH /0.7R to 20 mH /0.175R...and i like it a lot better...this is on an amorphous core...i will now use this as the orientation to go to nano...but low R and high quality cores...wow...this DAC is a killer...frightening...tone and transparency come together
Imassume that the Drixo will make a difference...and not a small one...
but do you remember how you eat your meals when you have been a kid ?
The most deliscious pieces at the end.
But seriuosly...I want to settle the DAC and its PSU in...so you appreciate the clock change...and than the battery change (or vice versa). I dont like to to change too many parameters at the same time as you dont know whaT does what.
Sofar we can summerize: The Soekris is phantastic...the leader in price/performance if you are on a budget.
Nevertheless: Andreas's DAC is the leader.
Period.
Of any DIY-Dac available. A pure friendship nice gesture / present for the money he asks for as it destroys any other DAC (yes, got the Ess, 4499 etc here..amd listened to commercial R 2 R DACs) with those 30 Euro Crystek clocks. You dont need the big Drixo to make it the reference for a life time FIFo/DAC.
And than comes his clocks...and his battery supply...which i did not listend to yet...intentionally. Get the basics right first (PSU).
Imassume that the Drixo will make a difference...and not a small one...
but do you remember how you eat your meals when you have been a kid ?
The most deliscious pieces at the end.
But seriuosly...I want to settle the DAC and its PSU in...so you appreciate the clock change...and than the battery change (or vice versa). I dont like to to change too many parameters at the same time as you dont know whaT does what.
Sofar we can summerize: The Soekris is phantastic...the leader in price/performance if you are on a budget.
Nevertheless: Andreas's DAC is the leader.
Period.
Of any DIY-Dac available. A pure friendship nice gesture / present for the money he asks for as it destroys any other DAC (yes, got the Ess, 4499 etc here..amd listened to commercial R 2 R DACs) with those 30 Euro Crystek clocks. You dont need the big Drixo to make it the reference for a life time FIFo/DAC.
And than comes his clocks...and his battery supply...which i did not listend to yet...intentionally. Get the basics right first (PSU).
Last edited:
Ak4499
Regarding:
"Tx--lt4320--L--c---shuntreg(ultrabib)---load.": To be precise, The L takes the position of R1 on the Ultrabib board and the Lt4320 Mini board (from Prashant) instead of the Diodes.
Last edited:
That was indeed a DAC i did not build, a Gustard a22, still using it but not for serious listening. But lets not go there, this gets off topic.
My point was that I build a lot of stuff, at least 10 DACs and many different output stages etc. Started with Sheldons Pcm63 tube dac, Erno Borberly 1704 and stages, Rakk DAC Kevin Carters DACs, Doedes Dddac, Buffalos 2,3, Soekris and now This one...Solid state output, Tube output, now DHT.(anyone needs a DAC?...just kidding)
Amdrea's might become the end of the journey. That was my point.
My point was that I build a lot of stuff, at least 10 DACs and many different output stages etc. Started with Sheldons Pcm63 tube dac, Erno Borberly 1704 and stages, Rakk DAC Kevin Carters DACs, Doedes Dddac, Buffalos 2,3, Soekris and now This one...Solid state output, Tube output, now DHT.(anyone needs a DAC?...just kidding)
Amdrea's might become the end of the journey. That was my point.
Having tried a few DACs, I can say that a FIFO makes every one better and properly implemented, your source can be hot boiled trash - as long as the bits can limp up to the FIFO you’re in great shape.
If you ever have some time to kick back after climbing the mountain, Blitz, I would be very interested to hear how you find the older DACs when fed the full domain-isolation capabilities of a FIFO. It’s a lot of fun.
BTW, I squeeze out a really quiet, floating 5VDC, high current line from a 9VAC TX winding with a silicone full bridge into a LC filter, a subsequent (much, much smaller) LC stage for hash and a discrete “super” shunt to program i(circuit total)>i(critical for first LC) and stabilize voltage, which feeds approriate down-stream regs. Salutary benefits are a graceful fold-down into and out of a short, fast startup with no inductive voltage spike at power on (programmable), and very low noise and impedance over a wide range of supplied currents with good efficiency.
I’m hoping the saved voltage from the LTC4320 would let me “add” an aggressive 60hz notch circuit to snub the residual diode commutation spikes instead of relying on the regs. I’ll have drag out a scope but I’m pretty sure I didn’t get rid of those entirely with a snubbed (quasimodo’d) TX and increased current.
Is all this better than a wall wart and a LT3042/pass transistor? Probably not, but it’s a lot more fun. I would imagine if you did nail down a low impedance, quiet LC supply it would sound a lot like a big battery hooked up the same integrating capacitor ie you’d be “listening” to the capacitor network. All of which I’ve tried, all of the different approaches made me appreciate the others, and all of it was fun and rewarding, caveat being that time/value left the equation a long time ago.
It’s great to see an enthusiast’s passion yield excellent results, congratulations and enjoy the tunes…
If you ever have some time to kick back after climbing the mountain, Blitz, I would be very interested to hear how you find the older DACs when fed the full domain-isolation capabilities of a FIFO. It’s a lot of fun.
BTW, I squeeze out a really quiet, floating 5VDC, high current line from a 9VAC TX winding with a silicone full bridge into a LC filter, a subsequent (much, much smaller) LC stage for hash and a discrete “super” shunt to program i(circuit total)>i(critical for first LC) and stabilize voltage, which feeds approriate down-stream regs. Salutary benefits are a graceful fold-down into and out of a short, fast startup with no inductive voltage spike at power on (programmable), and very low noise and impedance over a wide range of supplied currents with good efficiency.
I’m hoping the saved voltage from the LTC4320 would let me “add” an aggressive 60hz notch circuit to snub the residual diode commutation spikes instead of relying on the regs. I’ll have drag out a scope but I’m pretty sure I didn’t get rid of those entirely with a snubbed (quasimodo’d) TX and increased current.
Is all this better than a wall wart and a LT3042/pass transistor? Probably not, but it’s a lot more fun. I would imagine if you did nail down a low impedance, quiet LC supply it would sound a lot like a big battery hooked up the same integrating capacitor ie you’d be “listening” to the capacitor network. All of which I’ve tried, all of the different approaches made me appreciate the others, and all of it was fun and rewarding, caveat being that time/value left the equation a long time ago.
It’s great to see an enthusiast’s passion yield excellent results, congratulations and enjoy the tunes…