Unbalancer folow pasive conversion which is resistor .
"The idea here is that a differential current-output DAC can work into two 60-ohm resistors, which will develop enough just voltage gain to allow a 12AX7-based differential amplifier to bring up to a useable voltage swing. The resistors act as I-to-V converters all by themselves. "
Brian have you try unbalencer , and compare to other output stages ?
I forgot to state that i use 6N23P-EV input tubes and 6CG7 output tubes. B+ is 250V (after R12).
Also, could anyone explain me how to reference the heaters to H+ voltage? I ask because i have quite an amount of hum in right channel and i read that could be due to non-referenced heaters... or it could be due to something else?
Hello, i finally came to an end building my Buffalo III + Unbalancer DAC.
I have to say its sound is really from another world: i never heard anything that could even come close to this, whatever the price.
I used 6N23P, that really please me for their definition and tridimensionality but lack a bit of mid/low warm presence, than switched to E88CC NOS tubes the like of Siemens, Amperex, Philips or Mullard at quite lower anode voltage (around 90V).
This way the sound has something magic and emotional that i couldn't ever find even in other high end sound systems.
Thanks to all for your useful suggestions!
I have to say its sound is really from another world: i never heard anything that could even come close to this, whatever the price.
I used 6N23P, that really please me for their definition and tridimensionality but lack a bit of mid/low warm presence, than switched to E88CC NOS tubes the like of Siemens, Amperex, Philips or Mullard at quite lower anode voltage (around 90V).
This way the sound has something magic and emotional that i couldn't ever find even in other high end sound systems.
Thanks to all for your useful suggestions!
I think that if the goal is to destroy all the positive characteristics of digital sound then this circuit un-balancer get this target!
The 100Kohm in series with a grid is incredible, it destroy slew-rate and smooth all.
Also the effect of the bottom U4 tube is a feedback and it give strange effects on the final sound.
Anyway if you like that sound I'm happy for you.
The 100Kohm in series with a grid is incredible, it destroy slew-rate and smooth all.
Also the effect of the bottom U4 tube is a feedback and it give strange effects on the final sound.
Anyway if you like that sound I'm happy for you.
I use 47k bass pass resistors and they don't destroy anything, slew-rate is still there (using good tubes) and yes, everything is smooth where it has to be smooth, but not at the price of loosing details or rough sounds where they should be.
If you consider sounding REALLY WONDERFUL and real-like a strange effect on the final sound, then yes, that's what Unbalancer does.
I have ss lovers audiophile friends and they all liked very very much my system. Be warned: one of them sold his more-than-10.000€-cost "perfect" ss amplifier after listening to my Unbalancer...
Fully agree, I have two sets of Unbalancers in my dsp setup and are building a
third for the remaining low end output. Replacing the well regarded and for
the task at hand optimated opamp output previously used in the midbass
section was a true revalation. That one was no wear near what the Unbalancer
delivers in SQ.
Best regards. Bengt Dahlberg
I have started to play around with mine again:
I really like the JJ 12BH7 platinum matched/goldpins as input and the 6h6p as output tube. I tried chinese 12bh7 from tubeampdoctor, but highly microphonic.
Right now I am playing with the input resistors...AN tantalum is out, Neoohm as well...currently tx2575 vs Rhopoint WW...kind of blind test: I have build in a switch and do not know which one is which...so I let them burn in and once I decided which one is best, I will measure which one is active...
I really like the JJ 12BH7 platinum matched/goldpins as input and the 6h6p as output tube. I tried chinese 12bh7 from tubeampdoctor, but highly microphonic.
Right now I am playing with the input resistors...AN tantalum is out, Neoohm as well...currently tx2575 vs Rhopoint WW...kind of blind test: I have build in a switch and do not know which one is which...so I let them burn in and once I decided which one is best, I will measure which one is active...
After long listenings, i chose to leave Rhopoint resistors, as z-foils were a little bit "screaming" in certain songs. The same for grid stoppers.
In anode position i found Nirvana using some old Neohm NOS Wirewounds: after quite a break-in, they deliver incredible soundstage, imaging and detail, much better than Tantalum, Amtrans or any other i tested. Just like lifting a veil...
As input tubes, i find Sylvania NOS 7963s to be well above any other, including Telefunken E88CCs or Siemens CCas, at an incredible fraction of the cost. They need to be properly adapted on a noval socket saver, an easy solder work. Be sure to have bass pass filter properly sized (47k res + 100 pF cap is ok), otherwise they will propel HFs to sibilance and screaming... Be careful with Ebay sellers that sell them without testing: their 7963s are often around or below 50% strength.
I like 12BH7s as output tubes too, especially current productions EHs and strong Sylvania/RCA NOS. To save money, strong RCA NOS 6GU7s are equally good if not slightly better and will cost just a bunch of coins.
If you use 7963s as input tubes, RCA NOS 6FW8s as output tubes will reequilibrate towards a sweeter and warmer sound.
Electronic Enterprises NOS 5687s (but no other brand's 5687s, including Tungsol NOS) are well worth a try too, very rich, wide, detailed and rock solid sound. But watch out: they need an adapter.
Last edited:
You can find some more infos about uncommon, good sounding tubes for Unbalancer/BCF here:
http://www.hifi4all.it/t633-valvole-preamplificatrici-sconosciute
Sorry it's in italian...
http://www.hifi4all.it/t633-valvole-preamplificatrici-sconosciute
Sorry it's in italian...
Last edited:
Thanks a lot...can you post a picture of the Neohm WW NOS Resistors ? Maybe I can find some on ebay...I am using typically Mills-NS series...but have the impression that wirewounds are very transparent, but give as well an edigy-metallic sound to the music, so the opposite to warm timbre.
I will make my shoot out later today on the z-foil vs. Rhopoint ww...but first I have modified my unbalancer so that I can now plug and play between SS and tube-rectifier. Going from Normal fast diodes to hexfred in my El34-power amp was a big step forward, but going towards gz34 tube rectifier was an eye opener ! I hope something similar will happen when tube rectification is applied to the Unbalancer. I will as well compare fullwave with ct vs. bridge (hybrid) and different rectifiers from rgn1064 to eml 5u4g to ez81 to gz34 and 5y3....curious what will happen...
in parallel, I am building the janus regulator, the maida v2 and the salashv2 shunt regulator...keep you posted...all with an electrolytic-free PSU.
I will make my shoot out later today on the z-foil vs. Rhopoint ww...but first I have modified my unbalancer so that I can now plug and play between SS and tube-rectifier. Going from Normal fast diodes to hexfred in my El34-power amp was a big step forward, but going towards gz34 tube rectifier was an eye opener ! I hope something similar will happen when tube rectification is applied to the Unbalancer. I will as well compare fullwave with ct vs. bridge (hybrid) and different rectifiers from rgn1064 to eml 5u4g to ez81 to gz34 and 5y3....curious what will happen...
in parallel, I am building the janus regulator, the maida v2 and the salashv2 shunt regulator...keep you posted...all with an electrolytic-free PSU.
Last edited:
You can find a pic attached. They need a while to break in, after that they sound very good, with incredible 3d image.
They're very open too, so they will reveal every source defect, included HF excess. But Mills WW are on a different, lower level in sound quality.
About rectifiers, i haven't found anything better sounding than Mullard GZ32. Organic sound but great equilibrium between all frequencies and speed, too, that can't be said about very natural sounding rectifiers like Brimar 5R4GY. I even find GZ32 to be better than GZ34, unless you spend 4 times as much buying a Mullard metal base GZ34 - that would be useless, as it's not substantially better sounding than a good GZ32!
Mullard GZ32 combines the speed and impact of black base GZ34 with the organic sound of Brimar 5R4GY, and doesn't cost any more than one of these two. It's a real steal!
Powered by a good tube rectified supply, Unbalancer is simply phenomenal...
Attachments
Last edited:
Smile...you are my man !!! I have just ordered some gz32...a co-incedence...
It took me more time than expected to convert the unbalancer to tube rectification this week-end...I have build it in a way that I can switch and test hybrid-rectification vs. ct-fullwave, 5v, 6.3v, 4v...and compensate for the voltage changes in b+, so that we really listen to the rectifier, not the bias.
I did only a short, first listening: But it is already like taking a veil/curtain aside. the micro-dynamics/ transient response is much better...and intersting enough not into the direction of soft, bit more into clear...crystal-clear...and that was only a cheap jj gz34...
On the i/v-resistor front I still have to compare more in depth...the rhopoints are burnt in now...much, much softer. But a certain coloration which the tx2575 do not have, they somd more natural right now...but i have tons of more resistors to compare...carbon comp etc.
I am thinking to advance the psu with a janus, salas or maida reg...and maybe another lc before the channel separation...maybe even dual-mono-psu ?
What other secrets did you try ? Tube rectification is a must imo...but you not only knew this already, but you have chosen they perfect rectifier already....so I womder where else you are miles ahead already.
It took me more time than expected to convert the unbalancer to tube rectification this week-end...I have build it in a way that I can switch and test hybrid-rectification vs. ct-fullwave, 5v, 6.3v, 4v...and compensate for the voltage changes in b+, so that we really listen to the rectifier, not the bias.
I did only a short, first listening: But it is already like taking a veil/curtain aside. the micro-dynamics/ transient response is much better...and intersting enough not into the direction of soft, bit more into clear...crystal-clear...and that was only a cheap jj gz34...
On the i/v-resistor front I still have to compare more in depth...the rhopoints are burnt in now...much, much softer. But a certain coloration which the tx2575 do not have, they somd more natural right now...but i have tons of more resistors to compare...carbon comp etc.
I am thinking to advance the psu with a janus, salas or maida reg...and maybe another lc before the channel separation...maybe even dual-mono-psu ?
What other secrets did you try ? Tube rectification is a must imo...but you not only knew this already, but you have chosen they perfect rectifier already....so I womder where else you are miles ahead already.
I don't know, i made so much tries and tests in a 2 years time that i no more remember them all...
Here are some clues:
- 7963 input tubes are a MUST, but you need a 33khz bass pass filter, as i said, and somewhat warmer output tubes. Anyway, i find that high transconductance tubes sound best at input, E88CCs (or CCas) are very good too. Keep 7963/E88CC input tubes voltage low (between 70 and 90 V) and current high to have a more relaxed yet very natural and wide sound
- I use a Janus PSU, to use GZ32s i had to adjust some cap values onboard. Snubber at every AC input (instead of caps) give some slight more balance
- R-core transformers are better sounding than toroidal or EI ones, upon my taste
- I filtered B+ supply using a pair of Tentlab electronic chokes, B+ is now absolutely more clean than using anything else (excl. real, but BIG chokes)
- Output caps are most important, i found Duelund to be the best, Audyn True Copper as second. I know Duelunds are very expensive, but you'll get a (many) thousand dollars sound using them
- LCR Polystyrene caps or Amtrans AMCHs are both very good in pass bass filter (C5 and C6)
- Mills resistors are very good as cathode resistors for output tubes
- For input tubes i use 2 x rechargeable D batteries as cathode bias, without resistors or CCS (1.35V are ok for 7963). IMHO the best sounding are Panasonic Cadmica, that are Ni-Cd; but if you prefer Ni-Mh ones, Imedion are very good too, a little harder but great energy
- Output tubes need grid stoppers if you want your Unbalancer to be silent even using high transconductance/mu tubes on BCF. You'll have to cut 4 traces to do that. 1k to 10k values are ok, use z-foil to avoid colorations.
- Grid stoppers have to be 1k on input tubes, too
- Better use Buffalo's own as volume control, instead of an input potentiometer/attenuator (unless you're ok to spend half a thousand on that at least). I added a battery powered Clone Note LCR passive between Unbalancer and power amplifiers to get rid of any unwanted noise (not needed using output tubes grid stoppers) and to keep ES9018 bit resolution at max even at low volume. I use a remote volume control/input switch board connected to Buffalo III, so that i can adjust volume and switch inputs using a remote; Clone Note sets the max volume i can reach. N.B. selecting inputs by remote is not so straight forward to implement even using a remote input select board: you need some OTTO II boards and some relays, as boards are made for switching between analog, not digital sources
- Battery ground tweaks can expand the space between upper and lower extremes, but they have to be well sized. I found 6V lantern batteries to be the best, coupled to 10000 uF caps and 1 uF Epcos metal grey MKP bypass. Effect is greater if you connect them to Buffalo outputs instead of Unbalancer outputs. They give great bass anyway!
- Buffalo's output grounds have to be connected to chassis using a DEN/DENO ground wire choke to keep noise to zero. No floating
- A 10/20mH common mode choke grounded via a pair of x7r caps helps reducing heaters ripple too. I don't use heaters elevation, as B+ is around 100V
- I find direct I2S input to Buffalo III from an I2S output player to be the best sounding solutions for input. Even better than good XMOS USB transports. I use a Raspberry PI connected to Buffalo's I2S inputs, with Squeezelite onboard, driven by a small HTPC running LMS. TPA's own Chronus/Rhea solution using BBB seems to be better still
Last edited:
As I said...you are a few steps ahead of me...
What I figured out so far myself:
- Duelund Cast as coupling caps is a must.
- No lytics in the PSU is a must. lytics add coloration and a false tone and softness. I use Mundorf tubecaps all over, but will experiment with others in the future.
- I tested the echoke (with its lytics on board) in my poweramp....I clearly preferred the sound of my amorphous chokes. More energy and transparency. same with its smaller brother mec50 etc. Killed the transients/resonsonaces, made it to sterile for me.
-I am using a separate heater-transformer, so no interaction of th dc-diodes with the core of hv.
- Currently I am using waveio-usb, which is feeding Ian's reclocker (a must)which feeds i2s into the biii...I am planning to build a beaglebone with hermes instead of my current setup (alix1d(mpdpuppy) with sotm usb-card and non-switching psu).
A few questions I have on your setup:
-I guess you have one Janus for both channels, correct ? What is the complete chain with rc/lc looking like ?
- How did you snubber ? I have not used snubbers on tube output as the resistance of my transformer is high enough...Do you used center taps transformers, so fw instead of bridge or hybrid-bridge ? Any source for good ct-r-cores-transformers ?
- for test purposes I included a Lundahl 1683 in my setup....but it buzzes like hell !!! It shares the AC-Input with three other transformers...but it buzzes even when there is no load on it ! Any hint there ? So the buzz comes from windings moving or the core, the voltage it outputs actually is great, kts sounds better than my EI-core....but the buzz is annoying.
- How does the battery and the dc-offset of the biii work together ? I understand that the current source normally sets the gridvoltage the right way, no matter what the dc might be ?
- The better the unbalancer and my whole setup becomes, the more I do have the feeling that I am still watching in bluray, not 4k....There is a certain graininess I try to get rid of...maybe the Janus is missing, maybe the kiwame Carbonfilm-resistors as the anode-load have been not such a great idea, dont know yet...
Best Regards
What I figured out so far myself:
- Duelund Cast as coupling caps is a must.
- No lytics in the PSU is a must. lytics add coloration and a false tone and softness. I use Mundorf tubecaps all over, but will experiment with others in the future.
- I tested the echoke (with its lytics on board) in my poweramp....I clearly preferred the sound of my amorphous chokes. More energy and transparency. same with its smaller brother mec50 etc. Killed the transients/resonsonaces, made it to sterile for me.
-I am using a separate heater-transformer, so no interaction of th dc-diodes with the core of hv.
- Currently I am using waveio-usb, which is feeding Ian's reclocker (a must)which feeds i2s into the biii...I am planning to build a beaglebone with hermes instead of my current setup (alix1d(mpdpuppy) with sotm usb-card and non-switching psu).
A few questions I have on your setup:
-I guess you have one Janus for both channels, correct ? What is the complete chain with rc/lc looking like ?
- How did you snubber ? I have not used snubbers on tube output as the resistance of my transformer is high enough...Do you used center taps transformers, so fw instead of bridge or hybrid-bridge ? Any source for good ct-r-cores-transformers ?
- for test purposes I included a Lundahl 1683 in my setup....but it buzzes like hell !!! It shares the AC-Input with three other transformers...but it buzzes even when there is no load on it ! Any hint there ? So the buzz comes from windings moving or the core, the voltage it outputs actually is great, kts sounds better than my EI-core....but the buzz is annoying.
- How does the battery and the dc-offset of the biii work together ? I understand that the current source normally sets the gridvoltage the right way, no matter what the dc might be ?
- The better the unbalancer and my whole setup becomes, the more I do have the feeling that I am still watching in bluray, not 4k....There is a certain graininess I try to get rid of...maybe the Janus is missing, maybe the kiwame Carbonfilm-resistors as the anode-load have been not such a great idea, dont know yet...
Best Regards
Last edited:
I'm much less advanced than you think! But i had time to make a number of tests, anyway. I have quite low electronics knowledge, but good ears.
I wouldn't say "a must", but certainly film caps are way better in PSUs. But they're vey big. I would suggest a mix of the 2 types, keeping film caps for lower value but sound influencing positions
No doubt electric chokes are better than electronic ones, but they're still too big to fit inside my chassis. Mec-50s are better than simple caps filtering, anyway. I don't find a lack of transients or a sterile sound when compared to caps filtering; but i couldn't test electric chokes in the circuit due to a lack of room to fit them.
One Janus for both channels. Janus is connected to Unbalancer in a conventional way, then B+ current leaves the board at R12, goes to Mec-50s tiny board, then re-enters Unbalancer at R12 as well.
No center tap transformer. A simple 100R resistor + a 100nF cap between AC windings will do. Snubbers wont' make sound better, but equilibrate some parameters.
For R-core transformers: HiFi Audio parts Sotre>>Amplifier kit > Tube Preamp > DAC kit >>shipping to worldwide. If you contact them, they will build r-core transofrmers for you upon your specifications. They're very good and very silent.
Never used Lundahl 1683, so i don't know... Have you tried it in other circuits?
Dc-offset is not due to cause any troubles with battery use. Battery doesn't provide current, but voltage reference (in facts, it never discharges in time, they will be charged by leakage currents too)
I don't use Kiwame in my Unbalancer. They sound good, but seem to lack a bit of resolution and "being there" (it's very difficult to explain this parameter) when compared to other resistors, whatever they claim.
Anyway, your question is very difficult to answer: a whole system is a complex patient to diagnose without even facing him...
I wouldn't say "a must", but certainly film caps are way better in PSUs. But they're vey big. I would suggest a mix of the 2 types, keeping film caps for lower value but sound influencing positions
No doubt electric chokes are better than electronic ones, but they're still too big to fit inside my chassis. Mec-50s are better than simple caps filtering, anyway. I don't find a lack of transients or a sterile sound when compared to caps filtering; but i couldn't test electric chokes in the circuit due to a lack of room to fit them.
One Janus for both channels. Janus is connected to Unbalancer in a conventional way, then B+ current leaves the board at R12, goes to Mec-50s tiny board, then re-enters Unbalancer at R12 as well.
No center tap transformer. A simple 100R resistor + a 100nF cap between AC windings will do. Snubbers wont' make sound better, but equilibrate some parameters.
For R-core transformers: HiFi Audio parts Sotre>>Amplifier kit > Tube Preamp > DAC kit >>shipping to worldwide. If you contact them, they will build r-core transofrmers for you upon your specifications. They're very good and very silent.
Never used Lundahl 1683, so i don't know... Have you tried it in other circuits?
Dc-offset is not due to cause any troubles with battery use. Battery doesn't provide current, but voltage reference (in facts, it never discharges in time, they will be charged by leakage currents too)
I don't use Kiwame in my Unbalancer. They sound good, but seem to lack a bit of resolution and "being there" (it's very difficult to explain this parameter) when compared to other resistors, whatever they claim.
Anyway, your question is very difficult to answer: a whole system is a complex patient to diagnose without even facing him...
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.