SicPhones (Kickstarter): Zen-style class A amp with Silicon Carbide JFET

Page 1 of 3
2 3
  1. zzffnn
    This headphone amp is currently a Kickstarter project:
    The following description was copied from Tyll at innerfidelity:
    "This is an amp that will show up on Kickstarter soon, and is a single-ended, single stage, constant current source amplifier based around a newly introduced Silicon Carbide depletion mode trench JFET manufactured by SemiSouth. It's currently planned as a sub-$500 amp, and it sounded clean as a whistle and fast, fast, fast without being bright to my ears."
    Some interview-type Q&As between the designer and Srajan@6moons, indicating that this amp was inspired by Nelson Pass's design:
    I looked at the designer's web site and his resume/experiences, but can not extract much information.  
    I may pledge $55 to get its PCB and with Silicon Carbide JFET. Does it sound like a good idea?
    You can get a finished product by pledging $279. 
  2. zzffnn
    I do not have knowledge or skills to DIY this amp (anyone helps me build one, if this project lives?). But I read about it a bit.
    Many at diyaudio, including Nelson Pass himself, seem to like the low distortion characteristics of Silicon Carbide FETs. Whether a particular implementation can take the advantage of Silicon Carbide FETs (e.g., ensure linear operation, ect) remains to be determined.
    Edit: in case you are wondering, measurements are not available yet. I asked the designer (through kickstarter) for the following:
    Amplifier Gain
    Damping Factor (in XX Ohm Load)
    Output Power to XX Ohms
    Frequency Range (-/+ XX dB)
    Crosstalk @ XX kHz
    Noise (A-wtd)
    Dynamic Range (A-wtd)
    I hope the designer has done some measurements. Or maybe Tyll will provide more measurements? 
  3. Avro_Arrow
    So what your saying is this is a commercial product and not a DIY project...
  4. zzffnn
    You can buy its PCB with Silicon Carbide JFET and part list/sources (by pledging $55), then DIY from there. I can not tell exactly, but other components shown in the photos do not seem uncommon to me. 
    Edit: I would like experienced DIYers to comment on its potential, if possible. For example, would you invest $55 to give it a try?
    Edit 2: The designer replied quickly but remains vague about measurements. I assume he has done sufficient measurements during design stage.
  5. nikongod
    DIY Audio is not an investment unless you are a MOT. 
    No part of audio is an investment. 
    I would absolutely spend $55 to try it out. It sounds like its $55 for a PCB, and FETs - nothing else. So its not $55. 
    It could be really cool, but why not post a schematic. The translation from reality to English does not look like it was done correctly. There are a few terms in the description that are frequently misused or misread. 
  6. zzffnn
    Thanks for the comment!
    There is surely lack of information. I could not find a schematic though, nor do I know where to dig it out (maybe on diyaudio?). I understand that I need to buy more components to make a working amp, which will cost a lot more than $55 (likely more in the $100-200 range).
    Please specify what you meant by "There are a few terms in the description that are frequently misused or misread."? Thanks. I am here to learn.
  7. qusp Contributor
    those Jfets dont come cheap, so if they are matched... not a bad deal just by itself. I take it that this is the R85? he doesnt say which. they appear in the photos to be to247 so cant be the new to220 R055 (which would be good, nice gate resistance figures) thing is there are no 'audio grade' depletion mode (normally off) devices listed at semisouth yet and the ro85 and co are bloody expensivev (more than your 55 just for one part), is he gathering everyones money to buy them and get the price down? he has good taste in parts, mills resistors, nichicon caps, 2oz copper boards. i'm tempted myself but I need more info
    ahh I see, hadnt seen the kickstarter thing yet, interesting, long way to go though
  8. qusp Contributor
    lol just watched the video, the guy plugs it in, and immediately the experience is amped and he notices more detail as hes while hes talking, quite impressive hahaha, that kinda stuff turns me right off. its not the first semisouth based headphone amp, a number of people on diya have done it, maybe the first he knows about. i've even played with a basic common gate follower using r550
    not much heatsink either, bias cant be too high
  9. nikongod
    The description says that this is a constant current source amplifier. Im curious what is meant by constant current source amplifier - it could be several things. 
    The coolest is the following:
    The vast majority of headphone & speaker amps are VOLTAGE amplifiers. 
    Transconductance amplifiers (Current amplifiers - sort of)  have some unusual characteristics (very high output impedance) that make them very interesting with certain headphones, and total crap with others. When you find a happy pair of headphone & amp they sound outstanding. 
    It could also be a simple voltage amplifier with a CCS. Not quite as cool. 
    Plenty of things are far too good to be labeled as "audio grade." 
  10. qusp Contributor
    I know, thus my inverted commas, the 'audio grade' parts I speak of in this case are probably just parts that are binned as not good enough for their primary range (but perfectly good enough for the easier task of audio). I mentioned the audio grade parts (with the A like R100A, R065A etc) simply because they are the only ones that are priced anywhere close to what could get you 2 parts for less than 55 bucks
    I think its most likely the second type of CCS, simply a dynamically adjusted aleph style current source driving the fets, the former as you say, would be great with some headphones and crappy as hell with others. making such an amp would severely limit the market. but then again on his website he appears to favor fullrange speakers, which can often be better driven by gM type amps
  11. zzffnn

    ^ True. I had a F2 (a transcoduntance amp) and understand what you said.

    The designer replied me and said that the output impedance depends on the load. He suggested an output impedance of less than 1 ohm without feedback. He also said he tested the amp with Grados and got good results. He did not test with BA IEMs.
  12. zzffnn

    Aleph J has feedback though (along with 0.4 ohm of output impedance) per FirsWatt. This Sicphones amp was said to be without feedback AND with low output impedance.

    How does a "Zen-style amp" work? I heard about its name at diyaudio but do not know about the design.
  13. qusp Contributor

    fixed... no such thing as a no feedback amp, I really wish designers wouldnt use the term.
    well its not a current source on the output then with that outputZ, the whole point of a current source is it has high impedance.
    i'd take the aleph CCS over Zen, I couldnt hope to explain it more eloquently than Nelson himself has
    his articles are excellent, recommended reading if you plan to build the amp
    edit:  I suppose a passive transformer based 'amp' could be said to have no feedback, but certainly nothing that has fets, degeneration, or gain, otherwise the amp would have infinite bandwidth and wouldnt have a reference to compare the signal to, not exactly a recipe for stability. its a fashionable, but misrepresented term
  14. zzffnn
    I got that "constant current source" wording from Innerfidelity. There seems to be some confusion.

    Edit: I quickly read Pass' description of the Zen amp. It has an output impedance of around 1 ohm and one gain stage and is a "constant current source". Not sure about global feedback as I don't have knowledge to understand it.
  15. qusp Contributor

    no you have misunderstood, the constant current source in the Zen amp biases the output fet, which has local feedback and the amp has an output Z of 1ohms. having an output impedance of 1ohm and being a current source are mutually exclusive; a current source by definition has very HIGH, in the ideal infinite impedance. in Zen the current source does not drive the output, the fet does. the current source is between the positive rail and the fet, the fet drives the output
    ha glad I explained that right
    in Nelsons words, from the article I linked
Page 1 of 3
2 3

Share This Page