Balanced to single-ended converter (XLR to RCA / BAL to SE)

Discussion in 'DIY' started by purr1n, Jun 27, 2020.

  1. Baten

    Baten Friend

    Pyrate
    Joined:
    Mar 18, 2018
    Likes Received:
    1,130
    Trophy Points:
    93
    Location:
    EU
    I just got this Radial isolator in together with a set of XLR>RCA. That's all I need, right? XLR cable to Radial TWIN-ISO, into the adapter cable into RCA input. Am I missing anything?

    Edit: so it works exactly like purr1n said. (well of course it does). Only thing I noticed, when I feed the radial a single ended source (audial TDA1541A) via RCA-XLR cable, the final output is out of phase and sounds really crappy, this Radial 1:1 will only work properly with "real" balanced sources! :)

    Thanks for all the info in this thread!
     
    Last edited: Sep 12, 2020
  2. Melvillian

    Melvillian Friend

    Pyrate
    Joined:
    Jul 3, 2017
    Likes Received:
    1,345
    Trophy Points:
    93
    Location:
    Los Angeles
  3. skem

    skem Friend

    Pyrate
    Joined:
    Nov 1, 2017
    Likes Received:
    1,911
    Trophy Points:
    93
    Location:
    Charles River
    Thanks to @Melvillian, I was able to compare three transformers
    My test chain is: DigiOne → Metrum Adagio (Zout = 640Ω @ XLR) → Pass Labs XA25 (Zin = 47kΩ) → B&W 801 Series III

    A little bit about the Haufe
    I got my hands on these because they are reportedly the favorite of Cyril Jones of Raindirk. Looking at the specs, we see the Haufe RK283 is clearly not for consumer audio as it is rated for 600Ω input and output Z. I wrote about what this means for consumer users here. Tl;dr: one should expect some audible defects unless compensated. Indeed, without compensation (straight into XA25's 47kΩ load), the RK283 sounds very crispy. Because the load is reflected through the transformer, I wrote Metrum to ask what kind of load the DAC/Pre can handle. An engineer wrote back and said the Adagio can drive 600Ω loads at 0.006%THD. They also said the smallest safe load is about 100Ω. That meant I could produce the required load merely by paralleling a resistor to the input of my power amp and dumping most of the signal to heat.

    If we do not load the Haufe at something close to 600Ω ohms, the frequency response isn't all that bad, but the sound isn't good (I tested it quickly). If I feed it a 20kHz square wave and take a peek at the output on oscilloscope, it shows the transformer's parasitic capacitance is causing it to ring at 1MHz, which is presumably driving some kind of IMD lower down in the audio band.

    uncompensated.JPG

    However, if I now load the output with 1kΩ, almost all the ringing disappears.
    Compensated.JPG

    There are many power amplifiers that won't reproduce a square wave this well, so I am pretty pleased. Putting this in parallel to the XA25 will result In a 980Ω net load. Let’s see how it sounds in this configuration.

    Listening Impressions
    The next step was to listen to the three transformers over the period of a week.

    Cinemag CMLI15-15B — The initial impression is "very clean." The sound is pleasant, detailed enough, I really could stop here. But I feel there is something just a little off with the transients. They are not quite a fast as if the amp were being driven directly from a honking transistor. Still, generally a very pleasant listening experience. My raw listening notes say:
    • Cinemags image well
    • Cinemags make tape hiss bad
    • Cinemags have a gentle sound
    Jensen JT-10KB-D — This is a 4:1 step-down transformer, so I have to turn the volume up quite a lot. I'm pegging the Adagio with better recordings that don't have lots of loudness/compression. The Jensen feels like it has a blacker background and the timbre sounds even better. But it has lost stage and sounds flat. My raw listening notes say:
    • Jensen might have more natural timbre and lower noise floor
    • Jensen doesn't image for crap. The image is just big and smeared
    Haufe RK283 — Here we have a surprise. The Haufe totally blows the other two out of the water on imaging. The sound seems to detach from the speakers and the sound field is uniform with exceptional instrument localization. I also think in some ways the timbre is the best, but there's a caveat—at 980Ω, the transformer still has a bit of sizzle. It could be because the 980Ω load is still 30% off the design spec, but what I am hearing also has a hint of odd-order distortion: everything seems extra fast and detailed, while I don't need to turn it down as I often do with odd-order distortion, it is not easy listening.
    • Haufe+Metrum has the best stage I've ever heard!
    • Haufe may have the best timbre; percussion/cymbals incredibly realistic without exaggeration.
    • Haufe is the hardest to listen to. Very analytical.
    Measurements
    The sound differences led me to take some measurements to see what was going on. Looking first at the frequency response and phase shifts:
    Transformer_Frequency_Response.jpg

    We see that the Haufe rolls off a touch in the bass, but at 0.1dB, this is totally inaudible. The better performance of the modern transformers is not surprising. The Cinemag stands out for it measures as being tipped up in the treble, which may explain why it sounds like it has detail but doesn't really seem to have good transients. Note that this uptick is NOT what the Cinemag spec sheet shows, but of course one must consider the particular circuit in which one is using the transformer and all the parasitic inductances and capacitances that abound—so it seems plausible.

    Transformer_Phase_Response.png

    The phase response is where things get really interesting. Here we see that the Haufe has by far the best phase performance and thus highest bandwidth. This may very well explain why the imaging was phenomenal for this transformer. I have long postulated in these pages that stage and timbre likely relate to phase linearity, more specifically to the relative phases of each instrument’s natural overtones, as this property is preserved at any arbitrary distance from the sound source. Unsurprisingly, the Jensen has the worst phase performance and the worst (virtually non-existent) imaging. I am a little surprised the imaging between Jensen and Cinemag are as different as they are given how close the phases are (psychoacoustics is a black science!) Note: don't know why the last datum is very low for Jensen and Cinemag—it may be a defect as the measurement approaches the Nyquist limit of the ADC in my measurement device.

    Next I did some distortion measurements at different input levels. Newbies, please read caveats about the meaning(lessness) of THD here. For reference, standard consumer level is -10dBV. I believe the Metrum is putting out a maximum of 4Vpp = +3dBV. Note the levels in the plot are for one side of the balanced signal, so add 3dBV to get the voltage across the transformer. Thus 0dBV on the plot is really +3dBV, which is the output of the Metrum. Unless specified, transformers are connected to Zout = 50Ω and Zin = 100kΩ.

    Cinemag_distortion.png
    Jensen_IsoMax_distortion.png
    Haufe_RK283_1K_distortion.png

    Here we see the Jensen has the lowest distortion—and that is likely audible as "more natural timbre" and blacker background, which I had initially mistaken for "noise floor." Cinemag does pretty well. Haufe... somewhere between acceptable and not so well. Those are definitely audible levels of distortion in the Haufe bass, and indeed I can hear the implications of it in the mid-treble. I do not understand why Haufe's distortion goes down as the volume goes up! This contradicts everything I know about how distortion is generated. If this is correct, it suggests that, ideally, this transformer is used before a volume control. It might also be a hint that this is something reactive and I need to lower the Z of the load a bit more.

    Not shown here—but actually much more important—is the even/odd character of this distortion. The Cinemag's distortion is balanced equal parts H2 and H3, hence the relatively clean sound. Jensen's distortion is H3 dominant, but its is at such a low level as to be totally inaudible and doesn't seem to have any effect on the sound. The Haufe is also H3 dominant, but it certainly does impose itself on the sound. It is almost too much for an H3-phobic like me, but I seem to be able to tolerate it. FWIW, I do feel it sounds better louder.

    Final remarks
    My partner likes the Cinemag CLMI15/15B the best, and I think I would recommend it to most people as for it is quite pleasant. It is definitely better than the Lundahl LL1540 that is built into the Metrum Adagio. But the Cinemag does lack a certain quickness and the stage is sad compared to the Haufe. Personally, I like the Haufe and I even have a soft spot for the Jensen. The Jensen is just clean, but imaging is too far gone for me to be interested in it for 2ch setups. As for the Haufe, I worry that I might tire of it given the sharp quality it has in its current configuration, but the sound stage brings things to life in a totally magical way for me. Paired with the right amp, or tweaked further, this transformer will be awesome.

    BONUS: Thanks to @dBel84, I have tried @MisterRogers's active balance/unbalance circuit. I feel mixed about it. It's very clean and of course measures far better than transformers, but I feel something is lost. It doesn't seem to have the clarity that the transformers had. I'm... at a lost to explain it.
     
    Last edited: Oct 31, 2020
  4. Baten

    Baten Friend

    Pyrate
    Joined:
    Mar 18, 2018
    Likes Received:
    1,130
    Trophy Points:
    93
    Location:
    EU
    Just a mini-review on the Radial TWIN-ISO set-up. Works perfectly if you need an easy to use 1:1 transformer. Sounds as inoffensive as it goes, I honestly can't hear any 'transformer' sound characteristic whatsoever. If you want to work around gimped SE outputs this is a perfect contender. Got mine used from an older German gentleman who only used it a short bit for stage duties.

    As of right now, Jensen branded Iso-Max PI-2XR are ~$269 (XLR to RCA), Radial Twin-Iso sets you back ~$259 (XLRs, needs an RCA cable) when bought new. I'm pretty sure they both use the same Jensens internally.
     
  5. Koth Ganesh

    Koth Ganesh Friend

    Pyrate
    Joined:
    Sep 26, 2015
    Likes Received:
    1,637
    Trophy Points:
    93
    Location:
    Chennai, India
    I'm confused. I thought the Radial line for converting balanced to unbalanced included the J-Iso (I own one) and the Pro-Iso. The Twin-Iso is really balanced in and out , isn't it?

    Add: NVM. I see you are using the XLR to SE cables on the output side.
     
    Last edited: Sep 25, 2020
  6. Baten

    Baten Friend

    Pyrate
    Joined:
    Mar 18, 2018
    Likes Received:
    1,130
    Trophy Points:
    93
    Location:
    EU
    Yep! Using RCA adapters/cables as outlined in post #1 here you can use it on the output side as a 1:1 XLR RCA transformer. Totally works, using it like that right now.
     
  7. Azimuth

    Azimuth FKA rtaylor76, Friend

    Pyrate
    Joined:
    Nov 9, 2015
    Likes Received:
    6,838
    Trophy Points:
    113
    Location:
    Nashville, TN
    Home Page:
    From my thoughts on the loner of the Iso-Max and the DIY Cinemag box built by @Melvillian:

    Chain: Metrum Ambre -> Theta DS Pro Basic II -> Bachelor Tube Amp -> HD650's / LCD-2C

    Particular note about the Theta - it has outputs like the Yiggy, meaning it has 2 dual output chips in it. One chip per channel, and one side for + and the other output for -. Only the balanced gets both outputs. No summing, but you get pure output out of both unbalanced and balanced and it was easy to compare to the unbalanced out.

    Iso-Max:

    Yeah, this sounds Jensen. Smooth low end, slightly rolled off highs, pure liquid. Kind of confusing soundstage. And super quiet (meaning, reduced volume output). It is too bad this is 4:1. I see why people like them, big bottom and sounds like a tube in the chain and the timbre might sound more natural. Although it sounds more "muted" and I feel like I am missing a huge amount details. Soundstage is definitely flatter. However, it is the resolution I am missing. I mean, it almost makes my 650's sound like PortaPros. Sorry Iso-Max, get outta my signal chain!

    Cinemag:

    Now we are talking! Details are there, nice stereo spread, maybe just a tad louder than the normal unbalanced output. The low end seems reduced a bit (as I can see that now with @skem's graphs, so thank you!). I think the transients here are fine, although they might be compressing a little as they saturate the transformer. The imaging is not quite as confusing with the Jensen, but it is WAY more closer to the unbalanced of the Theta than the Jensen. When going back to pure unbalanced, things are quiteter and maybe less stereo spread, but the details reach deeper with reverb trails and such. The top end can just get sloppy and maybe a touch harsh at times.

    I know this is not fair as if I had my choice, I would go with the unbalanced out of the Theta. But that is the advantage of the Theta I guess and I understand this is not an option for all. I guess I am just happy I don't have to build one of these boxes now. Anyway...back to the point...

    - if I had my choice between the Jensen and the Cinemag, it would certainly be the Cinemag. It is certainly the closest to the real unbalanced output for sure.
     
    Last edited: Oct 1, 2020
  8. Taverius

    Taverius Smells like sausages

    Pyrate
    Joined:
    Dec 27, 2017
    Likes Received:
    3,026
    Trophy Points:
    113
    Location:
    Rapallo, Italy
    Wondering how the 1:4 that's loaned sounds compared to the 1:1 jensens.
     
  9. Koth Ganesh

    Koth Ganesh Friend

    Pyrate
    Joined:
    Sep 26, 2015
    Likes Received:
    1,637
    Trophy Points:
    93
    Location:
    Chennai, India
    I have both the PC - 2XR and the PI 2-XR. To my ears, it's not a contest. The 4:1 is much quieter at any given volume level with a flatter soundstage and some missing detail. When switching to the 1:1, I immediately hear the detail and resolution that I miss out with the former box. My DAC is the Convert 2 which is a gorilla in terms of output voltage ( I always keep it set at the -14 dBFS level which corresponds to 18 dBU). The 4:1 does give me some peace of mind knowing that I might not damage the gear downstream. My other DACs are the Gungnir Multibit A1 and the Modius (IMO, similar to the Convert 2 in varying degrees). Since I always prefer the balanced outs from both Schiit DACs, the 1:1 is on duty all the time.

    Trust this helps.
     
  10. Taverius

    Taverius Smells like sausages

    Pyrate
    Joined:
    Dec 27, 2017
    Likes Received:
    3,026
    Trophy Points:
    113
    Location:
    Rapallo, Italy
    Yep! Now I can sort of triangulate and be happy with my own 1:1.

    Cheers!

    Nervosa executed with a bullet to the back of the head on a disused pier in a suspiciously artfully lit port once again!

    :pirate07:
     
  11. crenca

    crenca Friend

    Pyrate
    Joined:
    May 26, 2017
    Likes Received:
    3,827
    Trophy Points:
    113
    Location:
    Southern New Mexico
    Thank you @Melvillian for the opportunity to be part of the tour!

    Chain: Raspberry PI + PI 2 Design (coax out) > Gungnir MB A2 > Saga OG (passive) > DNA Starlett/Asgard 3 > Utopia/Verite Closed

    Iso-Max PC-2XR (4:1):

    Did not spend much time with this. There is a real synergy problem here with the Schiit box perhaps. No matter the volume the details, bass, stage, and everything else sounds flat, lifeless, and muted.

    Iso-Max PI-2XR (1:1) {personal box - not part of the tour}:

    This is my daily converter. Bass is textured, transparent, and present. Mids are clear with just a touch of wetness/liquidity. The highs are even with the rest of the FR and as transparent as the rest of the chain allows. If there is any significant roll off in the top octave, its not much but then I am 51 ;)

    Cinemag:

    The significant (i.e. beyond-placebo) difference between this box and the PI-2XR came down to the mids for me. The Cinemag did not have that slight wet/liquidity smoothness of the Jenson and as such is probably the more "neutral" of the two. I think @skem is correct, in that at first blush the Cinemag comes of as the more detailed, but this is only the first impression. With HP's the stage is a wash between the two IMO, as was instrument imaging/separation.

    Conclusion:

    I would be happy with Cinemag, though I prefer the Jenson PI-2XR ever so slightly. For me there just was not much between the two excepting my personal preference for a slightly "wet" sound.

    edit: I mistyped and had the PI-2XR as 4:1. It is 1:1, not 4:1 like the PC-2XR...



     
    Last edited: Oct 21, 2020
  12. Azimuth

    Azimuth FKA rtaylor76, Friend

    Pyrate
    Joined:
    Nov 9, 2015
    Likes Received:
    6,838
    Trophy Points:
    113
    Location:
    Nashville, TN
    Home Page:
    I am still trying to figure out the difference between the Jensen PC-2XR and the PI-2XR.

    When you look up the PI-2XR, information is buried in the PI-2XX series and does not tell you much about the XR. Are we sure it has 4:1 transformers? It is the Pro model and really designed for XLR-XLR for isolation.
     
  13. crenca

    crenca Friend

    Pyrate
    Joined:
    May 26, 2017
    Likes Received:
    3,827
    Trophy Points:
    113
    Location:
    Southern New Mexico
    Oops, I mistyped. The PI-2XR (that is, XLR to SE) is 1:1. The PC-2XR (again, XLR to SE conversion) is the 4:1 box
     
  14. tommytakis

    tommytakis MOT: E.T.A Headphones

    Pyrate
    Joined:
    Mar 22, 2018
    Likes Received:
    5,465
    Trophy Points:
    93
    Location:
    Los Angeles, CA
    Home Page:
    Big thanks to @Melvillian for loaning out his babies!

    Listening chain: Pi2AES > SFD2 > DNA Starlett | Elekit TU8800 > JAR650

    So first off Jensen 4:1, yeah this thing sucks ass... somehow made the sound more compressed and veiled. I didn't really enjoy this one at all, felt it made it much worse than the single ended from my DAC. Tonal synergy wasn't that great with my system as it seemed a bit warmer than without it. Majority of my listening time was done on the cinemag 1:1 version. I'm pretty sure it's the 4:1 because I heard @ChaChaRealSmooth 's 1:1 version not too long ago and I didn't think this badly of it.

    Cinemag 1:1 was much MUCH better in my experience. Tonally it's a bit dryer sounding, maybe a tad leaner? treble sounded slightly rougher than with single ended. I like the tonal synergy better for my system for both the Starlett and the Elekit. It didn't seem dynamically compressed like the Jensen 4:1. I think there's some minor improvement in staging and macrodynamics, where bass sounded a bit tighter with a bit more impact. Not sure if it's night and day changes, but I liked listening to it. Maybe it's more of a drastic change when using one of the Schiit multibit dacs like Yggdrasil and Gungnir Multibit.

    It's not a product I need right now, but somewhere down the line, if I do want to invest in want, I'd definitely try to make the one with the cinemag ones. Maybe I'll make a box out of the amazon cardboard boxes lol idk
     
  15. cameng318

    cameng318 Friend

    Pyrate
    Joined:
    Jun 14, 2020
    Likes Received:
    327
    Trophy Points:
    63
    Location:
    Wyoming
    Got the Cinemags and Neutrik connectors months ago, finally put it together over the weekend. It's a pain for me to work with metal. The mental barrier kept me from working on this. I dremeled holes on a mooncake box, and I'm glad my neighbors didn't call the police. I'm not showing the internals because it's quite a mess... flying wires and hot glue... In the end I'm quite happy with the result. My Gungnir A2 sounds way punchier on balanced. On the RCA out it was kind of mediocre, didn't impress me, and didn't offense me. I expected the transformers to give me colored sound, but the Cinemags do have minimal coloration that I can't tell immediately. Maybe after a few days I will understand it better.

    The thing that makes me a little concerned is Gungnir's impedance handling ability. It's not clearly rated for 600 ohms like the Yggdrasil. I'm paring it with a Magni 3+ with 25k ohm impedance, so it's safe right now. I'm in the midst of building a singled ended amplifier that benefits from low impedance. I'm trying to use a 10k pot, which would present a 5k load to each of the drivers. If I use JBL Nano Patch for volume control, it would be 2.5k load each. I guess I'll just try it out and see if they upset the output stage.
    [​IMG]
    [​IMG]
     
  16. drgumbybrain

    drgumbybrain Science Nut

    Pyrate
    Joined:
    Jan 9, 2017
    Likes Received:
    2,219
    Trophy Points:
    93
    Location:
    Lives in Fortaleza, Heart in Girona
    After a couple of problems with my Metrum dac, I decided to try those xls to rca solutions. I was specially interested when I discovered that my DAC did not have xlr to rca transformers and I have noticed that rca to amp was way behind balanced to amp in soundstage and micro details. I will post a pic in the end of the Metrum dac because the absence of lundhal transformers is very very strange. If someone with electronic expertise want to analyze @Hands @purr1n would be nice.
    First, I would like to sincerely thanks @skem that had a genuine brotherhood attitude and sent me his extra Haufe transformes across the globe and solder the first step in a board.
    He also sent me a drawing to the next step!!
    I really don't think I could find this kind of help anywhere but here.

    The initial impressions are very good. Those Haufe transformes really opened the sound especially with near field.
    The sound its a little brighter than direct Metrum xlr, but im really liking it. Will post more info in the future.
    I did not had silver solder or anything with the audio mambo jambo, and I was wondering if I should have bought cardas solder.
    But I didn't want start this OCD.... maybe in the future.

    good vibes from brasil
    IMG_9679.JPG IMG_9781.JPG IMG_9779.JPG IMG_9782.JPG
     
  17. skem

    skem Friend

    Pyrate
    Joined:
    Nov 1, 2017
    Likes Received:
    1,911
    Trophy Points:
    93
    Location:
    Charles River
    Glad that’s working for you. Don’t worry about fancy solder. That’s voodoo.

    I suspect the impressive near field response is because you hear more un-reflected sound and are therefore getting better imaging made possible by the RK-238’s incredible bandwidth. This also suggests to me that properly done room treatment should improve your far-field response. My system is in the same situation but I will wait for my next home before doing room treatment.

    As for the brightness, I think that’s the third-order distortion that I talked about. I don’t think they are actually tipped up in the treble, although the actual response you hear depends on the impedance of your load, which I don’t know and can’t test. I would love to find a solution with Haufe-quality imaging and bandwidth but less third order (which is inherent to balancing transformers by their design). OpAmps tested are clean but don’t seem to preserve the imaging. Maybe a discrete circuit? In the meantime, those Haufe are pretty special. The quest continues...
     
    Last edited: Dec 3, 2020
  18. AukePauke

    AukePauke Acquaintance

    Joined:
    Jul 23, 2017
    Likes Received:
    33
    Trophy Points:
    18
    Location:
    Netherlands
    Last week i made the Cinemag CMLI-15/15B box. Plugged it between my Gungnir MBa1 and the Sonneteer Alabaster with Grimm Audio TPR interconnects RCA/xlr Let it burn in for a few days. Forgot to clean the xlr contacts|\/|. So had to do it. Now i am very happy with it. It opens up the sound, the bass is a lot better, just a lot more fun. We'll spend money. Thanks for all the info!
     
    Last edited: Dec 9, 2020
  19. m17xr2b

    m17xr2b Friend

    Pyrate Banned
    Joined:
    Mar 26, 2016
    Likes Received:
    3,988
    Trophy Points:
    93
    Location:
    United kingdomland of fish and chips
    One of the production boards with Sowter 3575 input transformers. The middle connecter makes resistor loading for experiments a piece of for experimentation and solder free. While Sowter is great their delivery times are horrible, I'm still waiting for a pair I ordered months ago.
    [​IMG]
    [​IMG]

    This one has a case and was sold to @Walderstorn, I hope it's still working ok.
    [​IMG]

    [​IMG]
    [​IMG]
     
    Last edited: Jan 2, 2021
  20. m17xr2b

    m17xr2b Friend

    Pyrate Banned
    Joined:
    Mar 26, 2016
    Likes Received:
    3,988
    Trophy Points:
    93
    Location:
    United kingdomland of fish and chips
    My own BAL to SE converter is also a volume control, a sort of passive preamp.

    The Audio Note input transformer from the Stellaris and Slagle autoformers setup for 46 steps.
    [​IMG]

    The size difference between transformers:
    [​IMG]

    The AN inputs are worth their price and are just too good, the Stellaris performance isn't enough to push them to their full potential.

    My initial solution before adding the AN transformers:
    1. Get some vintage 50 year old silver plated military 15 way 4 pole switches bbm. Ebay auction for 15£
    2. With plenty of patience file extra grooves at equal distance and convert it to a 30 way rotary switch.
      [​IMG]

      [​IMG]
    3. Apply some silicone lubricant for smooth turns and plenty of contact cleaner
      [​IMG]

    4. Reassemble and marvel how 15£ worth of junk is transformer into the highest end of switches with a vintage feel and look.
      [​IMG]


      [​IMG]

    5. Buy an AVC, I got custom Slagle, highly regarded as some of the best.
      [​IMG]

    6. Solder 140 connections, in the right order in a makeshift sturdy frame case.
      [​IMG]
     

Share This Page