Check and see if the 601 has a choke AFTER regulation, as seen in many of their other preamps.
I have owned a Canary CA801SE for exactly a year now and have tried out their CA906 for an eval.
I happened to have briefly owned CA-301 (someone borrowed it and blew it up), CA-309 (did not have quite enough juice to power the Eidolons), and I am currently using a highly modified (yours truly) CA-801SE...
The 801 circuit wise is very similar to the 903 except for chassis count and tube rectification. They are both based on the Wada Shigeru mod (SRPP instead of single triode follower) of the Marantz 7. The CA801SE handily displaced 3 preamps I had in my system (was in an evaluation frenzy at the time): My trusty Audio Research Reference 1 with Auricaps (huge soundstage buy no emotion at all), SP10 II (has its idiosyncrasies), and Cary SLP98P (gasses out and becomes muddy when challenged by complex passages).
After prolonged listening however, I find the Canary preamp lacking in textural richness of the mid highs, as evidenced by violin music. Violins are never steely or muffled as played through a lot of sub-par preamps but the Canary lacks richness and flavor despite various tube-rolling attempts. I am almost tempted to use the worlds constricted and compressed but that might be too negative. A friend who uses a CA903 setup just sold his because of the same reason. The CA801SE does something right with the midrange but most of the time there is a shrinkage of the body of mid highs that makes it uncharacteristic of a tube preamp.
This stock CA801SE also went to a friend's place with CA309 driving Genesis ribbon towers with the same results: Completely slaughtered by an MFA Lumi.
So I decided to roll up my sleeves and start swapping things out.
* First out goes the 1st 12AX7s interstage coupling cap, in comes a Jensen 0.01 copper oil.
* Then the 12AU7 (SRPP) stage coupling is swapped with a V-Cap reference (a bit of a waste of $$$ but thats the only 0.1uF I have)
* Then the Output caps with 4.7uF Mundorf Silver oil
* NFB loop DC blocker with 4.7uF Mundorf M Supreme
Then I checked the NFB loop and thought it could use some reduction in the amount of NFB so out goes the Feedback loop 25k resistor in favor of a 37k one. The sound became a lot more relaxed but at the expense of some bass ooomph disappearing. Still not enough improvement I think.
So, moving on to high Voltage:
The Sprague atom decouplers are swapped with ASC and Sprague Oil cans, that warmed up the sound quite a bit. Finally the violins are passable. I noticed that the biggest difference between the 801SE and the 903 is the power supply where the latter uses a tube rectifier, so I swapped out the 1N54XX diodes with Fairchild Stealth Hexfreds.
Just when I was getting pretty satisfied with the preamp, and think that it's getting close to the sound of an MFA Lumi (I must confess...I wish!), I got the chance to borrow the new Canary 906 to play in my system.
The 906 is somewhat like the Lumi line stage (2 6SN7 per channel), but I think the similarity is limited to tube config only. I think the circuit is quite different. The unit I borrowed was supposedly fresh off the production line. When I turned it on and played Midori-Sibelius violin concerto I was shocked. It *does* sound like my friend's highly modified Lumi! Good 6SN7 based preamps are not easy to make. While they are almost never thin or brittle, I have heard quite a few of them that are either noisy or easily "run out of gas" and convey a blurred soundstage when transients or complex passages are played, momentarily morphing from Hi definition (sorry to have borrowed Audio Research jargon) to AM radio. Absolutely not so with the CA906.
Out of the box the initial impression of the CA906 is a bit dark and buttoned down, but it has all the details and spatial information there is. Instruments in an orchestra are very well delineated and every tiny bit of nuance is passed with no exaggeration, squeaking, thinning, or coloring. These days when I audition any piece of gear I run through one or 2 of my favorite "reference CD tracks" and quickly jump to passages where the not-so-competent gear breaks down in various ways like compressed dynamics, muffled, hard piano notes, violin image unstable, thinning of certain violin notes, overly sibilant sounds, small background noises audible through ultra high resolution systems missing with the equipment in question. The Canary 906 passes each of these listening tests with ease, with everything, I really mean EVERYTHING better than my modded CA801SE.
"The story continues with the CA906. It continues to shine in my system which consists of a Sonic Frontiers transport and Processor 3, an Altavista NP220 premium gold power amp (sometimes I still use my VAC Phi 70/70, but recently I am more inclined to favor the NP220s more "honest" sound) and Avalon Eidolon speakers.
Strings are the area which the CA906 shines the most. With material like Gil Shaham playing Glazunov and Seiji Ozawa/Boston Symphony/Joseph Silverstein Vivaldi: Four Seasons, the CA906 made my system sing like never before. This preamp has a wicked way of presenting the crescendo of violins with unbelievable realism. It conveys the right texture of the strings regardless of pitch or loudness. Another area where it shines is jazz saxophone. A lot of preamps tend to make saxophones growl. A well-known challenging track is All Blues in the Reference recordings HDCD sampler. The CA906 not only was able to handle the passage when played loud, but the texture of the instrument is there and never sounds stressed.
Another thing worth mentioning is the over-engineered volume control. You need to dial it up/down 5 clicks to change the volume equivalent to 1 notch in a conventional volume control. This really helps me dial in the optimal listening volume for every track.
The CA906 FINALLY did get a chance to up against the Lumi for one evening. The setup also has a Sonic Frontiers Processor 3, but has a biamped setup with the preamp driving a pair of Tube Research Monoblocks and a pair of Rowland 1s. The speakers were Avalon Ascent IIs. The Lumi was the resident preamp of the setup, and the whole system sounded beautiful given the constraints of the bass extension and the top end of the Ascents.
We were a bit disappointed when we plugged in the CA906. First we had to deal with a hum problem where despite selectively floating each piece of gear with cheater plugs the hum never really went away 100%. But the biggest letdown was that while the CA906 did not sound anywhere close to the Lumi, and did not sound anything like it did in my setup. In that system the CA906 sounded more like my CA801SE before I did the mods. The size of Midori's violin shrank regardless of volume level, and things began to sound marginally thin and stressed. We believe that it might have something to do with the Lumi's capability to swing huge voltage into the additional load due to bi-amping. The CA906 is probably designed for driving 1 set of power amps only (preferably the Canary 300B amps perhaps?).
I ended up not buying the CA906 but neither did I continue to use the CA801SE as my reference preamp. So the 801 was left idle for a while until recently when I started looking into the regulator. I notice that post-regulation there is 2x100uF CDE lytics and then a choke before it gets to the main audio board where the anode resistors are. This is one heck of a strange design as a choke post-regulation would dramatically add output impedance to the regulator and render the regulator unable to sense the small voltage fluctuations on the load side due to current draw and compensate for that (caps are slow, regulators are fast). So I boldly took OUT the choke and the 200uF worth of lytics. I scoped it and I found that the ripple went from < 10mV PP (VERY IMPRESSIVE) to about 20mV PP (still OK in my books). Then I adjusted the trimpot of the regulator to make the output equal to the previous (the choke took about 12V out of B+ so I need to lower the new B+ by 12). This made the ABSOLUTE MOST dramatic change to the preamp, now the microdynamics are back and violins no longer sound borderline tinny. There is also a newfound correctness in phase and depth.
I keep wondering why Canary decides to configure its power supply like that. Is it for the sake of specs on the B+? It completely goes against the goal of having 0 impedance between regulator output and audio section.
There is mentioning of custom chokes on their webpage. IMO if they are placed before regulation they'd be a great thing to have, otherwise see if you can ask them to bypass the choke and use film or oil caps post regulation and not lytics. You can't go wrong with this change. Serious.
