I forgot:
- I/V resistor have to stay below or equal to 50R. Using medium mu tubes like E88CC or 7963 you even 20R resistors can be used
- Using film caps across I/V resistors as Broskie's suggests in some Unbalancer-for-current-dacs schematics enhances sound and smoothes dac output
Janus 1k5 resistor value cannot be changed: it has been calculated thoroughly and changing value would make noise rejection less effective, with hum appearing. You have to provide enough secundary voltage taking into count the voltage drop across it.
On Unbalancer i have 0,68 uF film + 2x520uF electrolitic + 2x1uF film bypass; r12 resistors are replaced by 2xmec50 chokes, each with its 2x47uF electrolitics.
I went a quite abundant to be sure to block any minimal ac noise.
Well, have you found this: Tube DIY Asylum
...I guess you canchange tubes and as well values...will do more research...
I have found a great intro on duncan's psu designe from dhtrob.com...trying to calculate a psu which is fast with small uF and good hf isolation...as the janus would have killed the ripple already...
...I guess you canchange tubes and as well values...will do more research...
I have found a great intro on duncan's psu designe from dhtrob.com...trying to calculate a psu which is fast with small uF and good hf isolation...as the janus would have killed the ripple already...
Last edited:
What is said in Audio Asylum thread van be said about any unregulated psu used to power tube circuits, but they work very well and better than regulated ones in my opinion. Not everything can be espressed as ripple, dip, spikes and the like.
I tried ss ps with Unbalancer, but Janus sound has that "something special".
N.B. i tried other values for output resistor, but all of them led to heavy hum. 1k5 is the only one that seriously kills noise.
I can see why...this resistor needs to be 1/gm, so indeed when not changing the tube as well, no chance to change it. Although some people did simulations and reported that the Janus is anyhow quiet sensitive if not the right operating conditions are met in current, voltage, transformer regulation etc.
My plan currently is the following:
- I will have the tube rectifier on a seperate small experiment board, so I can switch types of rectifier to play with...e.g.Mesh RGN1064 or rgn 364 or 6d22s etc.
- I will design an Choke-input PSU, so omitt the first cap on the unbalancer pcb and have a choke instead of the first resistor of the first crc. I have always used there a choke anyhow but with 30uf-20h-47uf). Need to calculate now if it is good enough for choke input (has currently 20h with 550 ohm and 100mah capabilty, but not sure if is gapped).
- than no resistor for the voltagesetting/channel separation of the unbalancer, but two cables go from the pcb where usually the last R of the normal CRCRC of the unbalancer goes. This prefiltered B+ now goes into two Jani, for each channel separately, but only for the reg part obviously. no Input cap and no output cap necessary as this is all on the unbalancer already. Only very few oarts are left then and you could build a Janus again on a small experiment board.
I guess this will be it. Voltage drop is than only half, so 27mA*1500ohm=40V. As the unbalancer not only reduces ripple, but as well isolates the psu from the swing of the music signal, we have now as well nearly a dual mono psu. The choke will fight a lot of the HF noise which the unbalancer would not be able to fight. the choke inout will be a rectifier saver and has lowest hf-garbage. And still no lytics or big caps needed. The max size will be 47uF of Mundorf Tubecaps (as already today)...and I can play with all kind of rectifiers, even a 274 !!!
Only open question left is if the regulator tube of the Janus needs in any case DC-heater or if AC is sufficient. I have a Janus pcb with and one without the Dc-reg, not sure if the DC heater was only optional for a preamp or if the reg tube already has been DC-heated as well.
My plan currently is the following:
- I will have the tube rectifier on a seperate small experiment board, so I can switch types of rectifier to play with...e.g.Mesh RGN1064 or rgn 364 or 6d22s etc.
- I will design an Choke-input PSU, so omitt the first cap on the unbalancer pcb and have a choke instead of the first resistor of the first crc. I have always used there a choke anyhow but with 30uf-20h-47uf). Need to calculate now if it is good enough for choke input (has currently 20h with 550 ohm and 100mah capabilty, but not sure if is gapped).
- than no resistor for the voltagesetting/channel separation of the unbalancer, but two cables go from the pcb where usually the last R of the normal CRCRC of the unbalancer goes. This prefiltered B+ now goes into two Jani, for each channel separately, but only for the reg part obviously. no Input cap and no output cap necessary as this is all on the unbalancer already. Only very few oarts are left then and you could build a Janus again on a small experiment board.
I guess this will be it. Voltage drop is than only half, so 27mA*1500ohm=40V. As the unbalancer not only reduces ripple, but as well isolates the psu from the swing of the music signal, we have now as well nearly a dual mono psu. The choke will fight a lot of the HF noise which the unbalancer would not be able to fight. the choke inout will be a rectifier saver and has lowest hf-garbage. And still no lytics or big caps needed. The max size will be 47uF of Mundorf Tubecaps (as already today)...and I can play with all kind of rectifiers, even a 274 !!!
Only open question left is if the regulator tube of the Janus needs in any case DC-heater or if AC is sufficient. I have a Janus pcb with and one without the Dc-reg, not sure if the DC heater was only optional for a preamp or if the reg tube already has been DC-heated as well.
Last edited:
First calculation would lead me to a new transformer (i wanted to try your r-core recommendation anyhow, right ?) with 400-0-400V with a resistance of 400ohm to get the right Q of the following filter which would be for 53mA 8H at about 70ohm and a 100uF Mundorf-tubecap as the input for the Jani.
I have still an amorphous core choke double-c with the right specs which is oversized for 200mA, caps are in the current setup already...so need to solder the second board next and order the transformer...and need to figure out if I need Dc-heating for the regtubes...and than this thing should blow me away.
I have still an amorphous core choke double-c with the right specs which is oversized for 200mA, caps are in the current setup already...so need to solder the second board next and order the transformer...and need to figure out if I need Dc-heating for the regtubes...and than this thing should blow me away.
I started to do some more listening before changing the system again....and the tube rectification is a game changer...
I listened to it with the tx2575 as i/v-resistors...and it sounded extremely mellow, soft, nice, but not enough "air", Not less than before, but as the resolution is now improved, you ask for a new level of performance....so, I switched to the Rhopoint ww back, which before I found a bit metallic and annoying...nope...it is absolutely not annoying at all and is polishing the windows you look through....compared to an already pretty high performing tx2575 !!!
It all goes back to what The founder of Linn always stated: Start with the source. I will get the psu of the DAC right now and than revisit my component choices...but the tube rectifiers take a big source of error out of the equation...next will be the choke-input vs. cap-input etc.
I listened to it with the tx2575 as i/v-resistors...and it sounded extremely mellow, soft, nice, but not enough "air", Not less than before, but as the resolution is now improved, you ask for a new level of performance....so, I switched to the Rhopoint ww back, which before I found a bit metallic and annoying...nope...it is absolutely not annoying at all and is polishing the windows you look through....compared to an already pretty high performing tx2575 !!!
It all goes back to what The founder of Linn always stated: Start with the source. I will get the psu of the DAC right now and than revisit my component choices...but the tube rectifiers take a big source of error out of the equation...next will be the choke-input vs. cap-input etc.
Yes, i forgot to say that rhopoint resistors need come more time to break in than z-foil ones. Neohm NOS WWs need some more time yet.
Choke input is surely better than cap input, no doubt about it!
For my experience and imho, fully broken in Rhopoints are slightly better than fully broken in Neohm upw-50s. But your taste could be different, as it's not a neat improvement.
About long break in, anyway, i was talking about Neohm NOS WWs to be used as anode resistors: they really require a long time to sound at their full potential; during that time they sound metallic and overbright, even if much much detailed.
Buy Now RGL0833 from Lintech IS09001:AS9120 Trusted Distributor
...The one type shown in your photo seems to be a croster resistor and a film type actually...have you ever slaughter one and looked into them ? Where did you get them from ?
The other one rgl0879...not even google nows this one...
...The one type shown in your photo seems to be a croster resistor and a film type actually...have you ever slaughter one and looked into them ? Where did you get them from ?
The other one rgl0879...not even google nows this one...
Last edited:
I purchased them some years ago from an italian guy whose father was a diyer but was dead. The guy was not a diyer so he put on sale some components he found at home, including a pack of mixed values NOS Neohm resistors, probably from the 80s/90s (not many, just 15 or 20). He said the pack had a label stating "Wirewound".
I bought the pack for a bunch of euros as i thought they should be interesting. I was right that time.
I found them very good and i tried to find others on sale, but i had no luck here in EU. I guess a stroke of luck is needed to find any more, or a more serious effort...
Having so few of them, i never thought about breaking one to see what's inside...
Hi guys,
I've just finished my unbalancer which is set as I/V 'current' Dac (4xPCM1704) converter.
I am considering selecting small value (470pF) capacitor for the low pass filters (C5, C6).
What type of capacitor would you recommend in this position (filtering high frequencies): silver mica, C0G, polypropylene, polystyrene. etc?
TIA
bern
I've just finished my unbalancer which is set as I/V 'current' Dac (4xPCM1704) converter.
I am considering selecting small value (470pF) capacitor for the low pass filters (C5, C6).
What type of capacitor would you recommend in this position (filtering high frequencies): silver mica, C0G, polypropylene, polystyrene. etc?
TIA
bern
In my setup, removing bass pass filter caps (33khz cut) causes unpleasant HFs rendition, some sibilance and too much presence of some HF instruments (for ex. drum hats or cymbals). Instead, using proper resistors and caps, bass pass filters don't have a negative impact in sound.
But leaving them onboard using high transconductance output tubes (that sound the best) causes slight hum and noises on output.
I'm experimenting with various solutions to fix that: moving filters in another position (before input tubes), introducing output tubes CC grid stoppers, using some output attenuation, etc...
I still can't find a perfect solution, but i'll keep you adviced.
Last edited:
By the way: I made some progress on thetube rectification front...took more time and effort to make it switchable between bridge and fw and change kt so that accepts as well old rectifiers etc.
Result:
- Fullwave (two diodes) sounds extremely musical and nice. The right tone is definetley there. At the same time a bit less dynamic than
- Bridge with two solid state diodes (mur860) and tube rectifier. Sounds a bit like turbo charged. More power, air, dynamics. But at a price. There is a slight mechanical signature, which impresses from a hifi side, but degrades tone, so female voices are a bt artificial in comparison, individual notes of a piano are very pronounced, but the instrument as a whole gets lost a bit etc. It is not huge difference, but notable.
That was with a 30uF input cap and a Mullard gz34. Changed to an 4uF in Oil and plugged in the Psvane 274B we replica. Woooooooowwwwww !! The thing is worth every penny. I was shocked. So detailed, so smooth and liquid. Now I understand those guys in the head-phone forum which praised it so high.
But it changes over time, is in a burn in right now, s need to wait where it will settle down, some reported 400 hours of burn in...
I will change the psu again this weekend to make it switchable between choke and cap input with same b+ coming out of it, have a transformer with 250-0-250 and 350-0-350. Lets see.
I ordered as well a bunch of tv damper diodes like 6ax4 etc, which tomas meyer from vinylsavour likes in his project and will build a complete tube bridge...maybe we can have both: The musicality of the pure tube FW rectification and the dynamics of the hybrid bridge ? That would be a 10/10.
It will be as well interesting to see if cheap damper diodes can be on par with the 274b king of the hill.
Result:
- Fullwave (two diodes) sounds extremely musical and nice. The right tone is definetley there. At the same time a bit less dynamic than
- Bridge with two solid state diodes (mur860) and tube rectifier. Sounds a bit like turbo charged. More power, air, dynamics. But at a price. There is a slight mechanical signature, which impresses from a hifi side, but degrades tone, so female voices are a bt artificial in comparison, individual notes of a piano are very pronounced, but the instrument as a whole gets lost a bit etc. It is not huge difference, but notable.
That was with a 30uF input cap and a Mullard gz34. Changed to an 4uF in Oil and plugged in the Psvane 274B we replica. Woooooooowwwwww !! The thing is worth every penny. I was shocked. So detailed, so smooth and liquid. Now I understand those guys in the head-phone forum which praised it so high.
But it changes over time, is in a burn in right now, s need to wait where it will settle down, some reported 400 hours of burn in...
I will change the psu again this weekend to make it switchable between choke and cap input with same b+ coming out of it, have a transformer with 250-0-250 and 350-0-350. Lets see.
I ordered as well a bunch of tv damper diodes like 6ax4 etc, which tomas meyer from vinylsavour likes in his project and will build a complete tube bridge...maybe we can have both: The musicality of the pure tube FW rectification and the dynamics of the hybrid bridge ? That would be a 10/10.
It will be as well interesting to see if cheap damper diodes can be on par with the 274b king of the hill.
Unfortunately I need low pass filter since I use PCM1704s usually in Pure NOS (NonOverSampling) mode. So without digital filters with standard 44.1kHz content the aliasys and images are close to audio bandwidth. Sometimes I use integer upsampling but I prefer listening to in pure NOS mode.
Thanks in advance. I look forward your findings
. In the meantime, I will try to experiment with silver mica and NP0/C0G.I never tested Psvane 274B replica in my system; but i tried Brimar 5Z4GY, that in the same forum is reported to be even better.
I didn't find it any better than my Mullard GZ32, even if they also report it to be ahead of GZ32 instead. No deeper bass, no smoother sound.
But i have an RCA 5R4GY from 1940, that my brother sent me from USA as a gift, and that is slightly better than my GZ32, deeper bass mainly. Somebody says it's on par with real WE 274B for a much lower price... I don't know as i never tested a real WE 274B until now.
I tested other (more recent) 5R4GYs, Brimar included, and i prefer GZ32 overall - but that 1940 babe is a different story.
Mullard's plastic base GZ34 are one SQ level lower than GZ32 IMHO.
Anyway, i'm arguing if the improved SQ you experimented could be partly due to the Oil input cap after the rectifier: that wouldn't sound strange to me. So i'll try one of them as soon as i can and that would be good for my RCA 5R4GY too, as it likes no more than 4uF as first following cap.
What oil cap did you use in that position? It was not bypassed, was it? I heard Mundorf Supreme EVO Oil short cylinders to be very good as filter caps, ever tried them?
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.