Eddie Current Studio 45 Custom Headamp Measurements

Discussion in 'Headphone Amplifier Measurements' started by purr1n, Oct 24, 2020.

  1. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    68,093
    Dislikes Received:
    112
    Trophy Points:
    113
    Location:
    Antarctica
    In light of this discussion where I was accused of being all like: "dude trust me as it’s surprisingly lacking in any sort of spec whatsoever from a site priding itself on measurements and objectiveness and the initial pitch feels intentionally vague" and requests from @Boops, @atomicbob and others out of curiosity in regards to this amp both subjective and objective, I figured I'd offer up some objective measurements and also discuss the history of this amp. Note that all variants of the EC Studio are sold out and it is unlikely that Craig will ever make one again.

    DSC00517.JPG

    I really won't get too much into how it sounds, other than maybe a few tidbits related on how this came to be. This amp came to be from a project Leviathan that never really took off, or at least took off in another direction. Basically one day, some guy bought some high-end Tango amorphous core interstage and output transformers and asked Craig to build a 2A3 amp out of it. The special thing about this amp was that an interstage transformer was used in lieu of a interstage coupling cap. There was an attempt to build this amp for Changstar folks, but it turned out that Tango went out of business (they are back in business again) and this project went kaput. The Studio amp came to be as a result of using Tribute nano-crystalline core interstages and Cinemag outputs with a tertiary winding (a McIntosh design from long ago).

    The OG Studio may have sounded too solid-state or even dry folks expecting a tube amp, especially for those who heard a Balancing Act. Some people were disappointed and sold their Studios. Others kept them. @zerodeefex loved it and often remarked that most people didn't understand what it was trying to do. I like to remind people that it was called Studio for a reason. The tonality of the OG Studio was like the OG Schiit Ragnarok, except clearer and better in every way. I think when it came to tube amps, most people did not expect this tonality (not warm, not wet, not bloomy).

    Over time, it became too much of a pain for Craig to obtain the nano-crystalline interstage parts from Tribute. Tribute did commit to making them, but delivery timeframes had been too difficult to nail down and communications were sporadic, and thus I explored a Plan B avenue by sourcing interstages from Monolith on behalf of Craig. The last few Studios sold were the Studio M, the M indicating the Monolith interstages. Personally, I preferred the Monolith parts as they possessed better microdynamics and added a small bit of wetness to balance the drier Cinemag outputs, but at the expense of ultimate transparency. Really, at the end of the day, the differences weren't huge. It was still an Eddie Current Studio. I don't think owners are missing out either way if they owned one or the other.

    People think I'm full of shit, but trust me, boutique parts are difficult to obtain, take a long time to build, and take forever to ship across continents. This not to mention how DHL or El Lay customs fucked up and sent a box of transformers back to Belgium only for Monolith to have to resend back to me. This is my last amp with boutique parts. Never again.

    I have another custom 45 amp, the one with the glowy mercury vapor rectifiers, using these same Monolith amorphous interstage transformers, but with double C-core Lundahl outputs and no feedback. This amp is a bit more forward sounding, like the OG Ragnarok, but with slightly less bass control because of lack of feedback. This amp was basically a development prototype with hacked holes in the top plate, messy internal wiring, etc.

    The Studio 45 Custom came to be because of coincidence. Craig had metal remaining for one more Studio amp, but none of the original parts. I had a last pair of Monolith interstage transformers. Since Craig was in the midst of the Studio Jr and Studio B builds, I simply asked: build me a custom 45 amp using the Studio B transformers. In essence, what I wanted was something more akin to the Leviathan with the Tango parts. A tube amp with some solid-state sensibilities, not a tube amp that sounded like a solid-state amp (OG Studio).

    Of course the problem is that the Studio B output transformers were designed with the 300B in mind, something Craig quickly reminded me. The 45 tube's plate resistance is twice that of the 300B! Well, let's double the 45s up in parallel half their plate resistance. It is a Studio after all which uses x2 paralleled output tubes. Duh.

    A side note on the Studio B output transformers: the prototypes of these were the Advanced 300B Output Transformer found on Mable Audio. Craig suspects these were based on ancient Western Electric parts back in the day and reverse engineered by someone in China. I think last someone checked, they were out of stock at Mable Audio. Craig bought a pair and disassembled these transformers, and sent them to an OEM in China, along with a Cinemag with the tertiary windings. The OEM responded that they actually knew about these Advanced 300B Output Transformers and thus could easily replicate them. Not only replicate, but also add the McIntosh tertiary winding to it. Another interesting tidbit is in regards to the laminations: no exotic materials used, but the core laminations are interleaved with different metals, permalloy and M-6 (don't quote me, I could have this wrong in respect to the actual materials) to counteract hysteresis (the memory effect of how a transformer hates changing signals).

    Now there are no guarantees that such an endeavor will work. The double C-core output transformers from China for the second run of the Super 7 amps ended up sounding worse than the prototype and the first run's Transcendar EI parts. However, in this case it did. I was super worried that it would end up sounding like shit, but I sat and listened to these parts, apples to apples, and was satisfied that they sounded the same. Most of amp development is an art and through listening.

    I breathed a sigh of relief. Finally, a project Leviathan type amp that sounded like a tube amp. Either way, everybody wins because Craig's built so many different kinds of amps, each with slightly different presentations. If it sounds like this is too much talk about transformers instead of tubes or design principles, that's because the transformer makes the tube amp, at least the SET with no or minimal feedback. This is why I tell folks to roll transformers instead of tubes. It's also why Craig's amps sound good even with cheap new production China tubes.

    Measurements in a bit. BTW, I suspect the measurements from the Studio B will be similar on the account that these amps use the same output transformer. In fact, the Studio B may measure better than this Studio 45 Custom because 300B has more power and the interstage transformers may add more distortion than interstage caps. I will try to grab a Studio B for measurements when one is available. Either way, it will be interesting to compare this TOTL tube amp with an entry-middle level one.
     
    Last edited: Oct 24, 2020
    Hammy, Xecuter, atomicbob and 27 others like this.
  2. SoupRKnowva

    SoupRKnowva Official SBAF South Korean Ambassador - Friend

    Friend
    Joined:
    Sep 26, 2015
    Likes Received:
    2,922
    Dislikes Received:
    6
    Trophy Points:
    93
    Location:
    Austin, TX
    I was actually gonna ask you about this amp soon, now I dont have too. I'd been reading some old threads and seen me poking you for details on other EC designs lately so i was getting inspired.

    Always love to hear how the decisions were made for these totl tube amps. Specially with Eddie Current ever since the 445 and 2A3mkIV with the dual input tubes and interstage transformers that disappeared after the studio.
     
  3. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    68,093
    Dislikes Received:
    112
    Trophy Points:
    113
    Location:
    Antarctica
    Eddie Current Studio 45 Custom
    0dbU or approx 0.775Vrms
    300-ohm load
    SE inputs, SE output
    upload_2020-10-24_22-8-40.png

    Eddie Current Studio 45 Custom
    -10dbU
    300-ohm load
    SE inputs, SE output
    upload_2020-10-24_22-10-34.png

    Eddie Current Studio 45 Custom
    -20dbU
    300-ohm load
    SE inputs, SE output
    upload_2020-10-24_22-11-33.png

    I know the measurement objectivists will scream holy cow, look at the AC mains harmonics, IMD, and other crap. To put things in proper perspective, below is the result of the amp that powers speakers which his holiness @ ASR exclaims as follows:

    Subjective Listening
    XXX It gives a smooth, very realistic sound impression, beating many "hi-fi" XXX ... And it runs out of power in a XXX

    Review Conclusions
    It is hard to believe such great performance can be had for so little money in the form of XXX. We get it because of research and measurements like I have been showing that resulted in the XXX. So I can definitely recommend the XXX on what good sound is like, and correlation between that and the measurements I am showing.

    [​IMG]

    If his holiness @ ASR can't hear this distortion from a class D solid-state amp in the LSR305, then he won't be able to hear the distortion of this tube amp. Keep in mind that this distortion is supposed to be pretty nasty with high levels of high order and odd order emphasis.
     
    atomicbob, JellyRhino, Sqveak and 2 others like this.
  4. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    68,093
    Dislikes Received:
    112
    Trophy Points:
    113
    Location:
    Antarctica
    So you guys know most tube amps will roll off in the lows or highs? Check this out. That's starting at 10Hz to 30kHz.

    Eddie Current Studio 45 Custom
    Frequency Response
    10Hz to 30kHz
    -10dbU
    300-ohm load
    SE inputs, SE output
    upload_2020-10-24_22-29-18.png

    Ever since I knew Craig, he was obsessive about this with his transformers and rejected many that couldn't do near full bandwidth. I was always like, why do you care? No one can hear that low or that high. Heck, he can't hear well past 12kHz anymore. Old habits die hard. He came from UREI. With pro-gear, one doesn't cut corners with FR because one pro component usually plugs into another and another and another.
     
  5. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    68,093
    Dislikes Received:
    112
    Trophy Points:
    113
    Location:
    Antarctica
    10 points to who can guess what those two spikes between 40kHz and 50kHz are?
    upload_2020-10-24_22-35-13.png
     
    atomicbob and zerodeefex like this.
  6. k4rstar

    k4rstar Done his time

    Friend BWC
    Joined:
    Jun 11, 2016
    Likes Received:
    4,948
    Dislikes Received:
    96
    Trophy Points:
    93
    Location:
    Toronto, Canada
    the 45 heater supply
     
    Jerry, Cspirou, zerodeefex and 3 others like this.
  7. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    68,093
    Dislikes Received:
    112
    Trophy Points:
    113
    Location:
    Antarctica
    So this is interesting and also might shed some light on measurement methodologies. The measurements above were with the volume knob turned all the way to the max. The reason this is done, and I am sure ASR does it this way, is to save time. It's sort of a minor major PIA dial in volume knobs to get desired output. Generally, it's easier to max out the knob and let the signal generator do its job. However, this does not reflect real world use.

    In normal use, we would of course dial in the volume knob to a comfortable listening volume. This measurement below is with the volume knob set near 1100 with the AverLAB outputting 14dbU for a resultant amp output of 0dBu or 0.775Vrms.

    Eddie Current Studio 45 Custom
    14 dBu input
    0 dBu output, adjusted via volume knob
    300-ohm load
    BAL input, SE output
    upload_2020-10-24_22-45-58.png

    BTW, this is how I did the Woo WA6SE measurements, by using the volume knob to adjust to get the desired output. Anyway, the main reason I wanted to do this is twofold: show how measurements can be different depending upon methodology, show the harmonic pattern of triodes, or at least vintage 45 tubes.
     
    Last edited: Oct 24, 2020
    will_f, atomicbob, Sqveak and 2 others like this.
  8. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    68,093
    Dislikes Received:
    112
    Trophy Points:
    113
    Location:
    Antarctica
    Yep, the high-frequency AC heater for the filament. It's kind of funky. The frequency is set by a few parts, and tolerances being tolerances, the left and right channels are off a bit. It doesn't really matter. The idea it so keep it well out of the audio band.

    The story is that the idea came from Andy Hefley (of GAS). I believe the original HF amplifier PCB in the BA was Andy's design. Supposedly aerospace designs took this approach, back in the day before transistors. At least that's the tale. For all we know, Voyager may have had a few vacuum tubes.
     
  9. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    68,093
    Dislikes Received:
    112
    Trophy Points:
    113
    Location:
    Antarctica
    Here is the distortion at 50Hz with 0dbFS output at the 1kHz signal. One thing about transform coupled SET amps is that as we move toward DC, we see more distortion, especially as we go up in level. We are seeing 10db higher D2 (-62.75db) at 50Hz compared to 1kHz

    Eddie Current Studio 45 Custom
    14 dBu input
    0 dBu output, adjusted via volume knob
    300-ohm load
    BAL input, SE output
    upload_2020-10-24_23-0-32.png

    And here is 20Hz below. Another 10db increase at D2 (-51.71db):

    Eddie Current Studio 45 Custom
    14 dBu input
    0 dBu output, adjusted via volume knob
    300-ohm load
    BAL input, SE output
    upload_2020-10-24_23-2-45.png
     
    atomicbob, Sqveak, crenca and 2 others like this.
  10. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    68,093
    Dislikes Received:
    112
    Trophy Points:
    113
    Location:
    Antarctica
    That pretty much wraps it up. I'm moving the amp from my test bench to my bedroom rack so I can give the Rocka Wavelight a good listen for the next few days.

    In closing, I would like to thank @m17xr2b for inspiring me to get off my ass to do a quick writeup and actually take measurements of an EC amp. Studio B measurements are still pending, but Craig decided that he wanted to build another Ultralinear amp, this time a VT-225 based one, instead of a Studio B. Therefore the Studio B measurements will have to wait. I think he just likes to build new shit out of curiosity instead of actually trying to make a living.

    VT225.jpg

    TBH, I already kind of knew how the EC amps would measure, having taken a few measurements of them before (on Changstar) and also seeing stuff on the scope while they were in development. I figured who cares. No one buys tube amps for their measurements. Hopefully this will satisfy the objective crowd and convince them that these things are a colossal waste of money and that they are better off with their SMSLs or Toppings.
     
  11. ChaChaRealSmooth

    ChaChaRealSmooth SBAF Gearmaster

    Staff Member Friend Gearmaster
    Joined:
    Dec 6, 2018
    Likes Received:
    5,992
    Dislikes Received:
    14
    Trophy Points:
    93
    Location:
    Avalon
    I'd laugh at the "shit" measurements except I own the EC Ultralinear and a DNA Starlett, where the former measures even worse.
     
  12. zerodeefex

    zerodeefex Grumpiest admin

    Staff Member Friend BWC
    Joined:
    Sep 25, 2015
    Likes Received:
    12,235
    Dislikes Received:
    57
    Trophy Points:
    113
    Oh man, these uber amps are what keep me dreaming.

    One day I'll find a way to afford the OG Studio. I love the flavors that have come out of Craig - he really has a good answer for all the different preferences.

    The original Studio was named aptly. It was not just the most SS, it was the cleanest, most resolving of the amps he's put out. I could see how many didn't like the sound, but damn, I loved that amp.
     
    Xecuter, obsiCO, Priidik and 3 others like this.
  13. Psalmanazar

    Psalmanazar Most improved member; A+

    Slaytanic Cliff Clavin
    Joined:
    Sep 27, 2015
    Likes Received:
    4,413
    Dislikes Received:
    134
    Trophy Points:
    113
    This looks awesome and better than most pro studio tube gear. The interstage transformers are key and get rid of the goo. They can turn ORFAS junk into stuff appropriate for classical recording.

    Most pro tubes were always what in playback chains I would call “lofi crap” meant to hide and color bad recordings. The design is intentional. Think a UA 610 on a shitty guitar amp. It just works. La2a to level and beef up a vocal. A Manley to goo up whatever with Maple syrup and glycerine. A drawmer 68/69 stuck on the end of a shitty digital recording to smooth it over.

    When you get to the good stuff, it’s legendary and will destroy and reveal everything in its path and you can’t even buy it. There was no readily available solid state preamps that could equal them in transparency until you hit late 80s-90s gear with modern production practices and laser trimming. The stuff before it that could was ultra limited manufacturing. If you even score something like the awesome Telefunken gear, you can’t use it easily as an English speaker because the guy in the USA who was repairing them is dead.

    Amir would shit on a V72 and Retro 176. He would tell a guy with money and inefficient mics to buy PGA2500 chip crap and a Behringer Composer. Even an SM7b scales.
     
    Last edited: Oct 25, 2020
    Cspirou and penguins like this.
  14. SoupRKnowva

    SoupRKnowva Official SBAF South Korean Ambassador - Friend

    Friend
    Joined:
    Sep 26, 2015
    Likes Received:
    2,922
    Dislikes Received:
    6
    Trophy Points:
    93
    Location:
    Austin, TX
    Come on now Marv, you know that what we all really want is The God Amp Pt.2: Electric Boogaloo
     
  15. dark_energy

    dark_energy Friend

    Friend
    Joined:
    Sep 27, 2015
    Likes Received:
    1,543
    Dislikes Received:
    24
    Trophy Points:
    93
    Location:
    Estonia
    I was there, 1000 years ago when this item was forged from special components and brought over the seas... No not really but I had a listen.

    This is a beast. I mean in comparison most amplifiers have like a recession in the mids and upper mids - detail and presence (not FR). This was its strength as I remember. The build and the looks are insane.
     
    Psalmanazar likes this.
  16. Cspirou

    Cspirou They call me Sparky

    Friend
    Joined:
    Sep 27, 2015
    Likes Received:
    5,691
    Dislikes Received:
    5
    Trophy Points:
    113
    Location:
    Northwest France
    This is the one DIY heater module I can't find anywhere. Pete Millet has released his own design, but no one sells PCBs or modules.
     
  17. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    68,093
    Dislikes Received:
    112
    Trophy Points:
    113
    Location:
    Antarctica
    Not necessarily. Let's do some apples and apples using a 32-ohm load. (With the volume knob at max for both.) The Ultralinear has high 60Hz hum, but the Triode has more AC mains harmonics. For a given signal, the Triode has more higher harmonics that go on and on, thus higher overall THD. When pushed hard, the Ultralinear has less even order. These are probably what accounts for the Utlralinear's more zippier and cleaner sound and the Triode's more rounded sound.

    EC Ultralinear 6L6 Proto
    0.55Vrms or -3dBu into 32-ohms
    [​IMG]

    EC Studio 45 Custom
    0.55Vrms or -3dBu into 32-ohms
    upload_2020-10-25_21-25-58.png


    Here is 10dBu output. Apples to apples (matching FFT window size).

    EC Ultralinear 6L6 Proto
    10dBu into 32-ohms
    [​IMG]

    EC Studio 45 Custom
    10dBu into 32-ohms
    upload_2020-10-25_21-32-43.png
     
    atomicbob, crenca, Sqveak and 2 others like this.
  18. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    68,093
    Dislikes Received:
    112
    Trophy Points:
    113
    Location:
    Antarctica
    ASR and the pure objective crowd are a victim of modern measurement gear that measures far far better than it needs to be. Given that SNR from a microphone / mic pre in a ideal situation will be no better than 85db SINAD, and the limitations of CD quality audio to 96db, it completely makes no sense taking measurements down to -150db.

    There's an art to visualization to showing what matters. This is because sometimes if we show what doesn't matter, it can be misleading. Let's do a realistic use case. I'll set the volume knob to 1100 on the amp. This means VERY loud with HD650s, a tick or two higher than I usually listen. And let's assume 14dBu output for 0dbFS (my DAC's output). Now given the limitations of microphones and the CD format, let's put a cap of -100db on the Y axis of the FFT measurement:

    Now observe the below result, which is much more sensible without the spurs, the grass, the skirts, the AC mains, and all the other shit that ASR likes to pick on which nobody can hear. The result below not perfect, it is a SET amp after all, but at least we have a presentation that's far more useful with less distractions.

    upload_2020-10-25_21-48-44.png

    And then ONE huge factor that ASR completely misses: music doesn't exist of single or double tones at 0dbFS. Heck, let's take Daft Punk's Get Lucky, which is quite a compressed recording, and we can see that the peaks (yellow) are about -15dbFS in the lows and -20dbFS in the mids. The geometric mean (red) is -20dbFS in the lows and -25dbFS in the mids. For the sake of simplicity, let's say the the frequencies of this track hit -20dbFS.

    upload_2020-10-25_21-52-14.png

    Using the information above, lets do a realistic analysis, the volume knob in the same 1100 position, let's feed the amp a -20dbFS signal (to simulate actual volume levels) in the lows and the mids using a 300-ohm load:

    upload_2020-10-25_21-57-17.png

    upload_2020-10-25_21-58-33.png

    Yup, things really aren't so bad when we take a realistic use case measurement, taking into account that music does not exist at 0dbFS at a single or double tone.
     
    Last edited: Oct 26, 2020
  19. Psalmanazar

    Psalmanazar Most improved member; A+

    Slaytanic Cliff Clavin
    Joined:
    Sep 27, 2015
    Likes Received:
    4,413
    Dislikes Received:
    134
    Trophy Points:
    113
    Exacto. And this is listening and not even taking into account which distortion is more audible. Imd in the mids and treble is much more audible than bass harmonic distortion. Neve and API sound less painful than most chip pres despite easily reaching insane distortion levels. The peaks aren’t going to be compressed and the fader ridden up to raise the distortion in listening either. It’s not tape noise. Even 80s-90s recordings that were self mixed by casette dub down from back in the day can have inaudible noise if the Dolby or dbx noise reduction was set up properly. Are people going to turn off Led Zeppelin II and III due to noise and distortion?

    This is how -80 something dB noise floor hardware can sound cleaner than anything that uses the TI or THAT chip amps for example. For German stuff. If anyone chooses an RME Octamic over a Telefunken V76 due to Amir type measurements, they’re just a deaf retard because only one of them doesn’t have massive imd in the treble and no truncation distortion. Only one of them will last 10 years too. The caps in the cheap switcher power supply (RME is spending that money on constant advertising) break down. Transformer hum doesn’t bleed into the signal path. Switcher spurs always do.
     
    Last edited: Oct 26, 2020
  20. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    68,093
    Dislikes Received:
    112
    Trophy Points:
    113
    Location:
    Antarctica
    IMD is almost always going to be related with THD, because higher harmonics mean more harmonics to interact with each other to form IMD! Duh. Tube amps, especially SET with their complex harmonics (or crappy plate amps in Amir's favorite JBL LSR305) will produce IMD. The noise floor will look like grass. The question is always how much is too much. Of course if we use the latest modern measurement gear such as the AP555 or AverLAB, we are going to see this stuff 100db down. Will we hear it? Nope.

    I did this measurement below for fun, Amir's favorite measurement, but using the same methodology and thought process I described above for a realistic use case. TBH, it's not realistic because musical spectrum does not look like that. Will we see much less in the highs and below 40-50Hz. (See two posts up for the spectrum analysis of Get Lucky).

    upload_2020-10-26_10-6-18.png

    The Aver won't do better than 32k window size for FFT so the floor in the lows is occluded, but we already know our ears are sensitive in the lows (Fletcher Munson). In the mids and highs, sure there is some IMD grass. Like a tiny fuzz. Can we hear this? No.

    What a lot of pure objectivists do not understand is that the reason modern recordings have much less noise is because of denoiser plug-ins, not because of modern "state-of-the-art" (another one of Amir's favorite phrases) chip technologies in the recording chain.

    Led Zeppelin wasn't post processed in ProTools with some random fucker clicking de-noise on the entire track. The microphones used today are the exact same ones used decades ago from the likes of Neuman. The good top notch microphone preamps are too. Some of the fancy light bulb compressors would measure "horribly" by ASR standards.

    Using the de-noise function is ultimately a fake way of removing noise. I hate it because actually removes fine detail as well and is one of the reasons modern recordings don't sound as good as noisy old ones from 50 years ago.

    The objectivists can be smart, but they lack experience.
     

Share This Page