BorderPatrol DAC SE

Discussion in 'Digital: DACs, USB converters, decrapifiers' started by k4rstar, Sep 9, 2019.

  1. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,778
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    @Hands, All of this TDA1543 stuff is janky. Mind you I'm not talking about the TDA1541 (which if double-crown chips may not necessarily be bad), but almost anything based off of the TDA1543 is janky. The Teradak brand probably seems jankier because of HF'ers who performed all sorts of tweaks and after each one announced how much better it sounded when in reality, it just sounded different or yielded a 1% subjective improvement.

    These freakshow or more politely "gimmick" designs date back the past 15 years, where every five years, some dude rediscovers this POS low-end garbage chip on HF, HIYA, etc. And quite frankly, I'm sick of it. This is not to take away from others' positive impressions of the Border Patrol DAC or other TDA1543 based DACs. That's fine. I myself prefer vintage 2A3 tubes with higher distortion than new stock Shuguang 2A3s. I've liked headphones with odd mids from Audio-Technica. Audio is personal.

    What I seek to do is offer balance, a counterpoint. If a poster waxes poetic on the Border Patrol DAC, then I am certainly going to highlight the problems I see with its design, starting from the use of the TDA1543. Again, this doesn't take away the validity of one's personal experience. But SBAF has a wider audience, has always had standards for good sound, therefore I feel that it's necessary to warn people away from this DAC.

    First of all, let's take a look at the TDA1543 THD specs. The bottom line is the THD vs. frequency for a 0dbFS signal.
    upload_2019-9-15_9-32-47.png

    So what we are seeing is about 75-80db THD (SINAD, which is THD+N is 70-75db per datasheet). Now per my reasonable "not-shit" SINAD requirements, and to not move goalposts, that figure is pretty good, nevermind that it doesn't capture the full 16-bits of Red Book / CD specs. However, as a DA chip, which serves as the source, the core, the engine, the heart of a DAC, it's garbage. The performance here is its theoretical maximum as its output will be subject to additional distortion imparted by passive or active components by the time it reaches the DAC's RCA outputs, and further down the chain. There's just not a lot of margins here.

    Did we ever wonder why so many TDA1543 designs stacked 4, 8, 16, or 32 chips? It's because the performance of a single TDA1543 is so horrible that it requires stacking chips to linearize the errors. The idea is that averaging a bunch of random non-linearities will result in better overall linearity. There is some risk to this with smaller chip stacks, e.g. four chip stacks, because what if the four chips used were all bad in the same way? This would result in performance that is worse than a single chip! The idea is that the chips would all be randomly bad in a different way, so all the differences average out to something better.

    Bottom line: the TDA1543 is broken. It was already a low-end low-cost part in 1989.

    To be continued.
     
    Last edited: Sep 15, 2019
  2. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,778
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    Going back to the "gimmick": there was sort of a fad 10 years ago to do the direct out from DA chip thing in DIY circles. Less active parts in the chain are always better right? It's like if a DAC or phonostage is able to drive a power amp directly, then in most instances it's better to go with a passive volume controller instead of a preamp with additional active circuits, especially one if is a purist. I recall quite a few experiments on various DA chips with voltage outs. Resistors, capacitors, small transformers even. Heck, I tried them myself with the AK4393 chip. Except sometimes, this approach is going one step forward and two steps back.

    The problem with a single TDA1543 as implemented in the Border Patrol DAC, and confirmed by the Stereophile measurements is the chip's inability to provide any sort of current to drive line-inputs with ease. Unless we stack of a bunch of TDA1543s, or provide a single TDA1543 with a suitable output buffer, we are going to run out of gas. The 0dbFS signal into a 100k-ohm load (per the Stereophile FFT) looks like behavior near the onset of clipping. The -3dbFS signal is much better but still worrisome given the fact that most preamps (and the Border Patrol DAC most certainly will require an active preamp) will have a volume pot that perhaps presents a load as low as 10k-ohms, which will ask for ten times more current than the 100k-ohm load.

    This is going to sound relaxed at best and soft and mushy at worst. If people like this sound, more power to them. It's not high-fidelity though.

    Also, since the Border Patrol DAC doesn't implement an LPF - the output to the DA chip is cap-coupled to prevent DC - we will have ultrasonics. In some instances, these ultrasonics may present issues to systems (amp and transducers) that can reproduce ultra-high frequencies. Amps can oscillate or become unstable and tweeters can blow. These circumstances will be rare, and while a known risk with DIYers, the approach is questionable for a commercial product. Personally, what I think may be worse are the high-frequencies in the audible range beating with the sampling frequency, eventually resulting in IMD at lower frequencies. The Stereophile high-frequency two-tone measurements present this issue quite well.

    Finally, the last thing, one I can probably buy off on, is the tube rectification. I've experimented with various diodes in my place of tubes in various tube amps (yes, you can actually do this), and the sound is different. Even with a choke or two chokes in the power supply, the sound is different. It's neato!

    My question is why spend the effort on tube rectification, thus adding to the cost, while not addressing the above deficiencies in the first place? Or is it the chase to the "Analogue" sound that takes primacy - which is actually a fallacy of sorts.

    More discussion on this...
     
    Last edited: Sep 15, 2019
  3. Azimuth

    Azimuth FKA rtaylor76, Friend

    Pyrate
    Joined:
    Nov 9, 2015
    Likes Received:
    6,839
    Trophy Points:
    113
    Location:
    Nashville, TN
    Home Page:
    I thought the TDA1543 chips were stacked because it could drive loads without a buffer. Most designs I saw with this had a single resistor as an I/V, which is not ideal. Most TDA1543 equipment used an op-amp for I/V (slash) buffer with a single cap for HPF and a LPF of some kind.

    Not sure how the design is in this DAC, but I agree that trying eek out the best performance out of a 30-year old obsolte chip is a very niche product in 2019.
     
  4. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,778
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    I believe the BP takes the signal from Vout pins of the TDA1543 to a cap to RCA. No LPF, no IV, no buffer.
     
  5. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,778
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    This is troubling. When people say analog, what they mean is vinyl (or tape). The problem with "analog" or vinyl? It all sounds so different, depending upon the era. The golden standard of decades back, the sprung Linn LP12 sounds dated in contrast to modern heavy tables from Nottingham, Clearaudio, Kuzma, etc. Tables with cheap plinths like certain Regas, Project, even some low-end VPIs, are on another level of garbage not any better than the lightweight plastic consumer tables from Technics in the 70s and 80s. My Classic 4 TT sounds like a Dangerous Music Convert-2 with Schiit Yggdrasil resolution and Crane Song Solaris soundstage. It doesn't sound anything at all like how the Border Patrol is described. Go figure.

    This is another reason why I hate these "analoguey" DACs. Typically, their sound gravitates toward the shitty sounding turntables of a bygone era or cheap tables today, but without the resolution and bass performance of modern DACs, or even the resolution of good modern vinyl setups (playing good material using simple recording and mixing chains).* These kinds of DACs to me are mere facsimiles of vinyl. Imitators without the advantages of digital or the advantages of vinyl. So what's the point?

    *Note that I do not recommend vinyl if the majority of your music collection or preferences is from 1985 onward. At best there will be no difference. At worst, the vinyl will sound worse. More vinyl re-releases of older recordings are coming from digital preservation masters, not the analog tape masters, so no point to it. Other vinyl-heads will disagree because they feel the presentation of vinyl is so different and euphonic. Vinyl can be superior at times to modern digital releases because the mastering is better, but this seems to be less the case as vinyl becomes less niche and less effort is being put into vinyl mastering for pop music.
     
    Last edited: Sep 15, 2019
  6. Hrodulf

    Hrodulf Prohibited from acting as an MOT until year 2050

    Pyrate
    Joined:
    Sep 28, 2015
    Likes Received:
    4,276
    Trophy Points:
    113
    Laymen judge the unknown by aesthetics. So, you'll get optics-driven minimalist designs, because minimalism feels easier to understand and looks better.
     
  7. Hands

    Hands Overzealous Auto Flusher - Measurbator

    Staff Member Pyrate MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    12,285
    Trophy Points:
    113
    Location:
    Colorado
    Home Page:
    @purr1n I meant all of the Teradak stuff looks janky, regardless of what they're using. I'm well aware of the TDA1543's paper specs (or worse, if not using an active output stage).

    Granted, looks can be deceiving. Maybe Teradak has good designs. But I think you can find more modern, cleaner implementations of these sorts of vintage ICs from random eBay sellers other than Teradak.

    And that's actually probably the real root of what I meant. I'm not sure Teradak has updated their designs in the last several years or worse. We could probably find better if folks really wanted to test a cheap 1543 DAC with active output stage.
     
    Last edited: Sep 15, 2019
  8. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,778
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    I'd rather we not. I mean I'd be OK with TDA1541 double crown, for $250 a pop, but then we are getting into PCM1704, PCM63, and super niche territory again. Let the past lie die. Kylo Ren is right.

    --

    One more thing on "Analogue" in regards to the "natural" 3db roll-off at 20kHz depending upon chip pinouts to a load or the effects of a NOS LPF. This is absolutely not "analog" in the sense that MC carts actually have a rising response in the top octave which every MC cart owner knows or at least can sense. Again, the term "Analogue" with respect to digital is more often than not associated with bad TT setups: records and needles worn to shit.

    Where turntables struggle most to keep up with digital is bass. Why design DACs with turntable like bass? It's asinine.

    --

    Niche actually measures acceptably.
    Freakshow measures like shit.
    Both will have their charms and appeal.
     
    Last edited: Sep 15, 2019
  9. 9suns

    9suns [insert unearned title here]

    Pyrate
    Joined:
    Jan 19, 2017
    Likes Received:
    1,754
    Trophy Points:
    93
    Location:
    EU
    @purr1n I don't know if this is the adequate thread to ask or not, but what do you think about the TDA1547 chips? Some "highly regarded" DACs use it, such as Museatex Bidat, Prism DA1 or Marantz SA-1 SACD player...it definitely seems interesting and people doesn't talk too much about it.
     
  10. Hrodulf

    Hrodulf Prohibited from acting as an MOT until year 2050

    Pyrate
    Joined:
    Sep 28, 2015
    Likes Received:
    4,276
    Trophy Points:
    113
    Just let's not get into the DAC chip = DAC fallacy. The chip does dictate some things, yet it can be messed up easily.
     
  11. Psalmanazar

    Psalmanazar Most improved member; A+

    Pyrate Slaytanic Cliff Clavin
    Joined:
    Sep 27, 2015
    Likes Received:
    5,345
    Trophy Points:
    113
    @purr1n the exception is adcs designed to saturated like tape. Very useful when recording drums, vocals, and clean guitars. Lavry and Apogee are the biggest names and Apogee the only mass produced one. I’m not sure anyone should buy them for that feature either when tons of interface pres have pads now. I don’t know what anyone would want a DAC that does that. Then there are the insanely priced, memed Burl and JCF gear.
     
    Last edited: Sep 15, 2019
  12. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,778
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    That stuff serves a function, a purpose. Anyone who has recorded even a YT video knows what a pain in the ass it is to set input levels for digital. I just went through this exercise this weekend as my son was wondering why the dialog replacement he was working on was playing back staticky. We set an input level we thought would fine, backed it off a bit, and then still realize a few parts still clip. No biggie, it's for a YT video anyway.

    Setting levels was so much easier with tape (but tape is horrible in 20 more ways than digital). Another purpose could also be to bring back the old transformer coupled with Class A circuits sound. Last time I checked, audio engineering is still an art, not a game of 0.0000001.

    Function is different from freakshow. Burl's DAC with a dose of D2 is just that, a dose. Sure, they are using an old AK chip, the AKM4699, arguably their best sounding one even today. The DAC doesn't measure as good as the X-Sabre Pro, but again, pro gear doesn't do the 0.000001 thing, other than Benchmark. But we know that Benchmark doesn't make real pro gear anyway since none of it is rack-mountable. At least Burl didn't resort to the Vout pins from TDA1543 for their "analogue" vision.

    P.S. Looking back, I should have picked up the Burl in place of the Solaris for the DAC reviews.
     
  13. Psalmanazar

    Psalmanazar Most improved member; A+

    Pyrate Slaytanic Cliff Clavin
    Joined:
    Sep 27, 2015
    Likes Received:
    5,345
    Trophy Points:
    113
    Yeah reaper doesn't show you the volume gain from the summing in the waveform preview so you can easily clip internally if you don't check your levels by rendering the track. Other DAWs are even easier to mess up in.

    The Burl sounds exactly what you're imagining it sounds like but nothing else will sound like whatever the hell they're doing with the discrete, class a opamps and the transformers. A freakshow DAC will mostly be useful for converting for a loopback of analog processing and effects, not really for clean mixing. Pros and semipros are often just happy with something that sounds cool for that. If you're running something into some analog gear, you might as well make it sound great. The Apogee and Lavry DACs and then the saturation functions of the ADC are baked into so many records. But then again so is the awful Alesis ADAT.

    I've never seen any guitar or metal dudes use the Benchmark converters. No Myteks either except for the 8 channel 192 ADDA, which uses the AD1955. The absolute garbage you used to see were Apogee Duets. Now the really bad stuff is Behringer, Audient, and UAD. RME is way better than all of these. They all suck.

    The Solaris is backordered everywhere. I don't see why people are still so hyped up over a clean converter when there are so many clean enough, multichannel antiseptic interfaces going around that will give you a ton of channels (RME, Lynx sortof, Focusrite Claret/Rednet, Presonus Quantum) and other stuff that actually sounds good and is clear enough (SPL, Steinberg multichannel, Dangerous, Apogee, MOTU). I guess it's just the hype train. Apogee is a bit on the freakshow side but you know what you're getting and they drank the Apple koolaid and made it retard proof for bandcamp and logic "producers"
     
    Last edited: Sep 16, 2019

Share This Page