214 VDC on the plates, ~26 on the cathode for around 7.5 ma through the Hammond chokes (8ma max rating.) The circuit is still just a test mule, when time permits the Hammonds will be swapped for chokes capable of more current.
warning: the 6w6 is slightly microphonic in this application, not so much that I don't believe it can't be tamed with chassis design
warning: the 6w6 is slightly microphonic in this application, not so much that I don't believe it can't be tamed with chassis design
Attachments
Hi salas. No other real point of comparison, shunt reg was the design choice from square one. The thought of an almost no-cap (2x 0.1uF silver mica) PS without active gain devices was too tempting. Compared to the 10-turn based passive it replaced the sound is warmer and more detailed, completely nonsensical given the THD is at the limit of my 24 bit measurements mid-band.
Not sure if the tiny bit of hiss with my ear in the Tannoy's cone is due to the pre's open chassis. It's frequency response approaches the lower AM band. It measures clean. The main hurdle is inductive coupling to the chokes.
Not sure if the tiny bit of hiss with my ear in the Tannoy's cone is due to the pre's open chassis. It's frequency response approaches the lower AM band. It measures clean. The main hurdle is inductive coupling to the chokes.
dsavitsk said:Cool design, and nice CCS
They're gangbusters, I recommend them highly.
salas said:
So far just a single IXYS. The CF and glow tubes appear to me to be acting in push-pull as far as load current is concerned. The IXYS sees almost no dynamic load change on the output side, it's bored senseless. At this point noise performance is limited by packaging anyway, the extra isolation from a cascode doesn't buy much. That plus I keep blowing up cascodes.
Nice guess. I'm not sure exactly. It's up there but under 40ma. The CCS current was raised until the 0C3s glowed evenly around the entire perimeter when viewed from above. Not having a complete purple circle results in sub-sonic fluctuation/noise, almost like an instability which follows the flickering glow.
The base circuit design started with salas' 6V6 pre. Depending how the CCS is biased and how much voltage is dropped it accepts a few octals and different glow tubes. The 6W6 measured best and was unique in my experience in that, unlike most other CFs I've measured, the level and makeup of harmonic distortion stayed low and constant to the higher limits of audibility. The 0C3s brought current into line with the Hammonds. I suspect a much better option I haven't gotten around to is the Magnequest EXO-010 or EXO-001. Nice thing is biasing is a simple matter of selecting glow tubes.
Edit: corrected to proper glow tube! C, not D.
The base circuit design started with salas' 6V6 pre. Depending how the CCS is biased and how much voltage is dropped it accepts a few octals and different glow tubes. The 6W6 measured best and was unique in my experience in that, unlike most other CFs I've measured, the level and makeup of harmonic distortion stayed low and constant to the higher limits of audibility. The 0C3s brought current into line with the Hammonds. I suspect a much better option I haven't gotten around to is the Magnequest EXO-010 or EXO-001. Nice thing is biasing is a simple matter of selecting glow tubes.
Edit: corrected to proper glow tube! C, not D.
You expressed quite eloquently the distortion profile and that is why I ended up using it subjectively in a rather weird duty, instead of an output valve in some low power amp. I see my triode 6V6 line choice as a generous performance taste of big bulb line level DHT PITA. 6W6 is the same stuff in another package off course. Me I preferred to have X5.7 gain so I can use my Jfet phono design on it too. Not so low Zout but no 100% local feedback as in a follower. You lose some, you win some. How do you like the performance RDF?. You have that magnifying glasses called Tannoy coaxes if I recall correctly.
Can someone please explain how the choke loaded CF described by athos56 works?
I have two basic questions about it. First, wouldn't the 1K DCR cause the 5687 to draw lots of current? I would have expected a DCR ten or twenty times that. Second, how can the CF work without any DC bias voltage on the grid?
Thanks for any help.
P.S. I am thinking of using athos56's circuit as a CF driver for a 300B amp.
I have two basic questions about it. First, wouldn't the 1K DCR cause the 5687 to draw lots of current? I would have expected a DCR ten or twenty times that. Second, how can the CF work without any DC bias voltage on the grid?
Thanks for any help.
P.S. I am thinking of using athos56's circuit as a CF driver for a 300B amp.
salas said:
That's another trade off made with open eyes. The 6W6 in triode doesn't have much gain so feedback is low-to-moderate and, as you say, 100% local. It's an experiment to find that intersection between absolute minimum parts count and high performance. Power supply excluded it's a tube, choke, cap and resistor. By far the limiting factor is the Hammonds, below self-resonance the harmonic profile shifts towards push-pull and by ~100 Hz is pure odd, rising though still at low levels. I expect the Magnequests (or equiv quality) to do much better down low.
I like it very much but, as mentioned above, can't get a handle on what I perceive as extra warmth over the passive. By no means does it have the stereotypical feedback sound, if anything it veers towards the 'tubey' opposite. Into a 50K ohm test load the mid band distortion at 1 Vrms was in the 0.003% range almost all second, into the 300K front end on my amp it'll be lower still.
salas said:
My RCA black plates weren't good enough?
Thanks for the 6P6C tip, I've generally liked the Soviet equivs. The Hammonds' DCR don't bias enough current through a 6V6 though it showed a lot of promise. Once I get around to the Magnequests they'll be revisited. Otherwise biasing means two more resistors and an input cap with the current supply, a 10H choke input to limit current pulses to the lowest practical level. Did you mean replace the glow tubes with a Maida, or ahead of the CCS?Salectric, the tubes are biased by the choke's DCR.
rdf said:My RCA black plates weren't good enough?
The RCA is great sonically, but I mentioned those alternatives considering the microphony you discussed. I have found really quiet ones. Also the reissue 6V6 is something you want to have in your tone arsenal anyway. The 6P6C is rather more romantic, but you will love it for how coherently it does its trick. You will find some noisy and some silent ones in the bunch. Run 20-25mA with your Quests on 6V6 and you are in for a treat.
As for the Maida, I meant as a pre regulator before your CCS.
Re: Tube version of Pass B1?
So you are looking for a capacitor coupled 12AU7 cathode follower with a NJFET tail CCS? Off course its possible. But 12AU7 is colored, has been used CCSed before. Nothing to right home about.
Do it like in this thread like Athos and RDF IMO. Much better concept, tubes and sonics. B1 I have it, I did the capacitorless version in the Pass thread. Its nice, but not what the guys talk here. Here is double nice.
Skorpio said:
So you are looking for a capacitor coupled 12AU7 cathode follower with a NJFET tail CCS? Off course its possible. But 12AU7 is colored, has been used CCSed before. Nothing to right home about.
Do it like in this thread like Athos and RDF IMO. Much better concept, tubes and sonics. B1 I have it, I did the capacitorless version in the Pass thread. Its nice, but not what the guys talk here. Here is double nice.
Hi,
Sorry for my ignorance. Could someone be pls enlighten
me on the filiment supply pls. I notice that the supply is shared
between the rectifier & buffer tube but I don't understand
the grounding technic used. Is it to lower noise further & how
does one arrive at those resistor values ?
Thanks to all
Sorry for my ignorance. Could someone be pls enlighten
me on the filiment supply pls. I notice that the supply is shared
between the rectifier & buffer tube but I don't understand
the grounding technic used. Is it to lower noise further & how
does one arrive at those resistor values ?
Thanks to all
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.