logo elektroda
logo elektroda
X
logo elektroda

Why the popularity of germanium transistors in audio, better than silicon?

andreyatakum 4197 74

TL;DR

  • Germanium transistor audio amplifiers are examined as a retro alternative to silicon and tube amps, especially among audiophiles chasing a “germanium sound.”
  • The key argument is that the sound comes from circuit topology and distortion—often transformer-coupled classic schemes—not from germanium itself.
  • Germanium makes up about 0.0007% of Earth’s crust, while silicon is around 20%, and germanium parts had to stay below 75–85°C during soldering.
  • Conclusion: germanium transistors are not better than silicon in audio; they just sound different, and similar effects can be achieved with silicon designs.
  • They also suffer from high reverse current, low operating frequency, lower gain, and high temperature sensitivity, which made them hard to source and use.
AI summary based on the discussion. May contain errors.
ADVERTISEMENT
Treść została przetłumaczona polish » english Zobacz oryginalną wersję tematu
📢 Listen (AI voice):
  • Modern electronics can distort audio and video perception

    #61 21908172
    viayner
    Level 43  
    Posts: 10628
    Help: 1563
    Rate: 2051
    Hello,
    Fellow @elukam according to your definition I am a caveman, your arguments are not devoid of sense, but you are too quick to draw conclusions (favourable to your theory), consider a few cases:
    - Something was weak and now it is supposedly better, here I would disagree, current equipment is saturated with electronics, audio tracks are sometimes more complicated than not one simple computer from a few years ago, why? To delight the unsophisticated listener, DSP creates the impossible from the non-existent and you don't really know what you are listening to.
    - Equipment used to be weaker - yes - but this was due not to the fact that tubes or Ge transistors were used, but to the technology available in general, all the electronics, transformers, loudspeakers transmitted everything much less faithfully. Therefore, if we were to use, for example, electron tubes nowadays, the situation would be different.
    - Old TVs - ok, they had poor resolution, but that's the impact of technology, today's TV is more of a computer with a screen than a TV, the power of the processors is huge and they even add AI, just what for? Processing even 8k doesn't require that much processing power, here we have again the creation of the impossible from the non-existent, the enhancement of images, it's the electronics that know what's supposed to be on the screen and it's not necessarily still the actual image - but it's stunning!
    Regards
  • ADVERTISEMENT
  • Nostalgia drives claims of germanium audio superiority

    #62 21908191
    James596
    Level 30  
    Posts: 1866
    Help: 140
    Rate: 571
    Col @elukam has captured the essence of this situation quite well. Something that has been forgotten by the general public, due to its flaws or features, is suddenly discovered by someone who looks at it with sentimentality, spreading an aura of "coolness" around it

    A good example is cassettes. What do they associate the typical bread-eater with? Not with poor sound quality and inconvenient handling?
    Suddenly there are a handful of audiophiles who make a "discovery" about how cool it was and add an ideology that the cassette plays better than the CD, digital roughness, analogue pleasure and other hackneyed slogans. Others pick it up and it goes on.

    If the same group of people had mentioned the wonderful sound quality of a cassette 10 years ago, they would have knocked their heads together. And today? Look at how much metal-type tapes go for.

    Exactly the same pattern applied to Unitra in our country some 10-15 years ago, which can be seen even in the threads of the electrode.
    Years later, you can see a slight reversal of the trend and the dropping of the "eye flaps" :) Strongly in price, however, went the old computers and analogue cameras.

    I'm just waiting for people to start mentioning the superiority of AM radio over FM, what noise and interference is natural to the ear, not like that sterile purity of UKF. :D Because you know it used to be, right?
  • ADVERTISEMENT
  • #63 21908201
    sigwa18
    Level 43  
    Posts: 12019
    Help: 1188
    Rate: 3341
    In addition, you can hear the storm approaching.
  • #64 21908204
    CHCl3
    Level 9  
    Posts: 37
    Rate: 13
    My colleague @James596 would like to add, out of spite, that AM has one advantage over FM; range. I can't listen to the Turks, Chinese or even Germans on FM, but I can on KF.
    Of course you can also over the internet, but that's not it .
  • ADVERTISEMENT
  • #65 21908212
    bratHanki
    Level 39  
    Posts: 5722
    Help: 452
    Rate: 1394
    Beijing still broadcasts in Polish?
    In July and August there are super far-field propagations on UKF. More than once I managed to listen to Serbs, Albanians and even Greeks. As far as I remember it was around 96-97 MHz. An analogue tuned radio is best suited for this purpose.
  • ADVERTISEMENT
  • #66 21908216
    CHCl3
    Level 9  
    Posts: 37
    Rate: 13
    >>21908212
    They are broadcasting, around 7 MHz, if I remember correctly. Yet the last time I listened was last summer; and it wasn't very interesting.
  • #67 21908222
    Janusz_kk
    Level 39  
    Posts: 5858
    Help: 227
    Rate: 1472
    James596 wrote:
    A good example is cassettes. What do they associate the typical bread eater with?

    In my opinion, turntable records are a better example, it's the same symptom. Poor sound versus CD.
  • Wavelength, not modulation, determines range and power

    #68 21908347
    andreyatakum
    Level 15  
    Posts: 831
    Rate: 1155
    CHCl3 wrote:
    in spite of the fact that AM has one advantage over FM; range.

    Here it is rather not AM-FM that plays a role, because these are just types of modulation. The role is played by the wavelength. Medium waves require a high output power, but their propagation depends on the time of day and does not exceed several hundred kilometres. Short waves require less output power and their propagation is almost unlimited, but depends on many factors - time of day, solar activity, season etc. In these ranges they usually use amplitude modulation (at 2-4 Mhz railway stations use FM... i.e. frequency modulation). On the ultra-low waves they use FM. But there are also exceptions. E.g. AirBand (118-136 MHz) uses AM.
  • Modern electronics already exceeds human hearing limits

    #69 21908439
    elukam
    Level 18  
    Posts: 420
    Help: 11
    Rate: 195
    viayner wrote:
    - Something was weak and now it is supposedly better, here I disagree, current equipment is saturated with electronics, audio tracks are sometimes more complicated than not one simple computer from a few years ago, why? To delight the unsophisticated listener, DSP creates the impossible from the non-existent and you don't really know what you're listening to.
    I wasn't referring to the effects of DSP, only to the mechanisms of faithful recording and playback. As far as they are concerned, after the justifiable end of the era of tubes, germanium transistors and transformers in high frequency paths, we have reached a level in popular equipment that exceeds human perceptual abilities. In other words, it is no longer necessary or worthwhile to do better.
    Everything is more complicated because it costs practically nothing to increase quality with additional electronic components. So there is nothing wrong with that. It is not, as some audiophiles imagine, that a simple circuit is a good circuit. It is exactly the opposite.
    And when it comes to DSP, why not. Since the signal is transmitted digitally between modules and the source is digital too, there is no reason not to process it with DSP instead of complex, unreliable, uneven crackling and humming analogue circuits. The sound in a room strongly depends on the speakers, their positioning. A small change in geometry has the same effect as changing some parameter in the DSP settings. So why not.
    Quote:
    - Equipment used to be weaker - yes - but this was due not to the use of Ge tubes or transistors, but to the technology available in general, all the electronics, transformers, loudspeakers carried everything much less faithfully. Therefore, if today we would use, for example, electron tubes, the situation would be different.
    This was due to the fact of using tubes and Ge transformers or transistors. And today nothing would help them if more modern germanium tubes or transistors were produced.
    Quote:
    - Old TVs - ok, they had poor resolution, but that's the impact of technology, today's TV is more of a computer with a screen than a TV, the power of processors is huge and they even add AI, just for what? Processing even 8k doesn't require that much processing power, here we have again the creation of the impossible from the non-existent, the enhancement of images, it's the electronics that knows what's supposed to be on the screen and it's not necessarily still the actual image - but it delights!
    This is a completely different topic. But the video signal today practically only exists in digital form. There is no option to return to the technology based on and consequent to the construction of the widicon. When solid-state CCD sensors were developed and when they reached a resolution inadequately high in relation to existing television technology in a b. short time, that was the moment to behead it with a guillotine.
  • CCD resolution growth drove the shift to IP cameras

    #70 21908983
    TechEkspert
    Editor
    Posts: 7277
    Help: 17
    Rate: 5661
    This is an interesting observation with CCDs and their resolution. In CCTV, cameras started to reach more lines than PAL accepted, e.g. 700TVL. This pushed the development of DVRs with analogue inputs into AHD and HD-SDI/TVI/CVI. Inexpensive high-resolution FHD cameras operating on similar installation principles as PAL cameras offered the possibility of easy installation upgrades.

    But CCD resolution was increasing and the adaptation of 4MPx and 8MPx and 16MPx cameras in CCTV enabled the move to IP cameras and packet streaming instead of analogue signals in a dedicated path. This also changed the approach to installation design as switches could aggregate streams from multiple cameras in a single link and the medium could be changed from copper to fibre or even wireless or even from LAN to cameras reached over the internet.

    This is the result of the development of many parallel technologies, video streams are sent over the Internet, which used to be accessed via 56Kbps modems, a camera consuming 5W compresses video using the H.265 codec, PCs used to use MPEG2 acceleration cards because the CPU was too weak, copper ethernet on a twisted-pair cable and RJ45 plug (8P8C) accelerated from 10Mbit/s to 100 and later 1000 and even 2500 and 10Gb/s, and a 1GB/s switch costs PLN 60.
  • #71 21909007
    CHCl3
    Level 9  
    Posts: 37
    Rate: 13
    >>21908347
    Of course it is the wavelength.
    I used AM as a mental shortcut. In civilian radio, AM is used on lower frequency and longer range wavelengths, and FM, as it is known on UKF, with coverage, most often, local.
  • Simplicity and nostalgia favor germanium audio circuits

    #72 21911343
    viayner
    Level 43  
    Posts: 10628
    Help: 1563
    Rate: 2051
    Hello,
    i've come up with a few more suggestions, sentimentality about the 'freebies' in question:
    1) Note that the glory days of lamps or Ge elements were associated with the youth of quite a few users of this forum, these were times of fond and good memories, so we may have a desire to return to something associated with these good memories. It is often the case that "what a shell soaks in..."
    2) Simplicity of layouts, I personally feel that complication makes no sense or need. In the past, a lot of circuits could be made on a single transistor or tube, compare the old radio receivers containing only a few tubes and still providing very good reception (I am not considering the resonant circuits and their selectivity), nowadays such a radio is dozens of transistors enclosed in a few integrated circuits, and with the old resonant circuits not providing special quality or selectivity.
    3) Coming back to the audio track, the simpler the track, the less we add extra "colours", acutely I appreciate it.
    Regards.
  • Complexity improves selectivity, distortion, and robustness

    #73 21911394
    Fimek
    Level 16  
    Posts: 229
    Help: 3
    Rate: 355
    Hi,

    viayner wrote:

    (...) these were times of fond and good memories, so we may have a desire to return to something associated with those good memories(...)


    Here the consensus :)

    viayner wrote:

    Simplicity of the layouts, I personally feel that complication makes no sense or need (...)


    Well, unfortunately, it is simply impossible to agree with this statement. Lack of selectivity means, especially in an environment with the presence of interference, lack of sensitivity and lack of resistance to intermodulation. Resonant circuits are expensive to manufacture (because actuation) and not robust to shocks - if you can replace them with something, it's only better. Tubes are expensive, large and unreliable - unlike integrated circuits.

    viayner wrote:

    Returning to the audio track, the simpler the track, the less extra 'colour' is added, an ac I appreciate.


    You are very much mistaken :) a short feedback loop, involving, for example, only one component, is the cause of additional distortion appearing, not a way not to add it.

    If I had to encapsulate this in one bracket, I would write that the old circuits may have been understandable, simple to reproduce and required little 'technology' (just right for an amateur), but that doesn't mean they were better - quite the opposite. Complication of electronic circuits is not done 'because it is': because everything is becoming more complicated and harder to service, so audio or radio must be like that too. The exact opposite is true - usually the reasons are very rational and strictly engineering, for example: avoiding time-consuming operations in the manufacturing process, minimising distortion, improving selectivity, improving resistance to cross-modulation, low current consumption, miniaturisation, resistance to harsh environmental conditions and many factors.

    I, too, recall with fondness my beginnings as an electronics engineer, when it was possible to make a shortwave receiver based on the BF966 and LM386 - but it works differently and better now. And that the design of an analogue receiver in modern technology requires more knowledge, e.g. knowing the basics of DSP, FPGA coding, Altium/Kicad etc. - well, that's hard, you may need more knowledge, but it's easily accessible (because of the Internet and lots of open projects) and easier to learn (you can tailor it to yourself, because ChatGPT will explain everything).

    Apart from that, decent engineering was also complicated in earlier times, e.g. in the 70s, 80s - only consumer engineering in the People's Republic of Poland was simple and easy to understand - just compare Unitra, for example, with Kenwood from those days.

    Greetings,
    Fimek
  • Preference for simpler audio paths over added coloration

    #74 21911836
    elukam
    Level 18  
    Posts: 420
    Help: 11
    Rate: 195
    viayner wrote:
    Returning to the audio track, the simpler the track, the less we add extra 'colour', just the way I appreciate it.

    The typical audiophile ignorance of the subject, as deep as an artesian well, emanates from here :)

    Moderated By Felini:

    3.1.11. Publishing posts that do not contain substantive content, containing harmful advice, advice that does not attempt to solve a problem and publishing identical posts that follow or repeat information included in the discussion.
    3.1.9. Disseminate content that is ironic, mocking or malicious, constituting a sign of disrespect towards other Users or third parties.

  • Germanium transistors used for fuzz pedal distortion

    #75 21935236
    buszyl33
    Conditionally unlocked
    Posts: 771
    Help: 58
    Rate: 185
    buzerek wrote:
    mkpl wrote:
    Germanium transistors are used in music to generate distortion in guitar pickups.

    A long, long time ago, I made a FUZZ guitar pedal for my brother – a foot-operated one, based on an AF428 transistor, I think. The musicians were delighted.

    Back then, we made do with whatever we had. I made a fuzz and a duck too. And we played rock. But the German amps are gathering dust in boxes, and now Si reigns supreme. But, but, someone here mentioned valves. Now that’s a topic! It’s worth talking about. They really are a joy to listen to.
    As a former user, I recommend :)
📢 Listen (AI voice):

FAQ

TL;DR: Germanium junctions drop about 0.2–0.3 V, and one poster summed up the real trade-off as "they just sound different." This FAQ is for audio builders and vintage-hi-fi readers who want to know whether germanium transistors truly beat silicon, or mainly add nostalgic distortion and transformer-era color. [#21905241]

Why it matters: Many buyers confuse vintage character with higher fidelity, so knowing what germanium actually changes can save money, parts-hunting time, and design mistakes.

Feature Germanium transistor amps Silicon transistor amps
Junction drop Approx. 0.2–0.3 V Approx. 0.6–0.7 V
Thermal behavior Very temperature-sensitive More stable
Leakage current High, rises with heat Lower
Typical appeal in thread Vintage tone, softer character Better parameters, cleaner design
Best summary Different sound Better technical platform

Key insight: The thread’s core conclusion is simple: germanium is not inherently higher-fidelity than silicon. Its appeal comes from vintage circuit behavior, transformers, and distortion that some listeners find pleasant. [#21905241]

Quick Facts

  • Germanium parts in the thread were said to require soldering care below about 75–85°C, because heat quickly shifts their parameters and can damage the crystal. [#21905241]
  • The cited junction-voltage contrast was 0.2–0.3 V for germanium versus 0.6–0.7 V for silicon, which directly affects biasing and crossover behavior. [#21905606]
  • Germanium’s crust abundance was given as about 0.0007%, versus roughly 20% for silicon, which helps explain why germanium devices became expensive and scarce. [#21905241]
  • A modern nostalgia benchmark appeared in the thread: the Analog Audio Design TR-1000 reel-to-reel machine was quoted at about US$26,000. [#21905652]

Why have germanium transistors become popular again in audiophile audio amplifiers if silicon transistors have better technical parameters?

They became popular again because listeners value vintage character, scarcity, and nostalgia more than raw measurements. The thread repeatedly argues that audiophile demand follows rarity and fashion, much like tubes and reel-to-reel decks. In that view, germanium sells because it sounds different, looks retro, and is harder to source today than ordinary silicon parts. [#21905471]

Germanium vs silicon transistors in audio amplifiers: which is actually better for sound quality and why?

Silicon is technically better for accurate sound reproduction because it is more stable, quieter, and easier to design around. The thread’s main article states that germanium devices are “no better than silicon transistors — they just sound different,” and ties the preferred character to distortion and old circuit practice rather than to the semiconductor material itself. [#21905241]

What is meant by the so-called "germanium sound" in amplifier designs?

“Germanium sound” means a vintage tonal character created by circuit distortion, not a magic property of germanium itself. In the thread, that character is linked to classic amplifier topologies, transformer coupling, lower-band distortion products, and nostalgic listening preferences. The article explicitly says the goal is often not perfect Hi-Fi accuracy, but a sound that evokes older equipment and earlier listening memories. [#21905241]

How do output and coupling transformers change the sound of classic germanium transistor amplifiers?

They shape the sound by limiting bandwidth and adding their own tonal color. The article states that retro-audio fans especially value classic schemes with coupling and output transformers, because those transformers largely introduce the timbre they like. That means part of the “germanium” signature often comes from iron and circuit architecture, not only from the transistor type. [#21905241]

Why were early germanium transistor audio amplifiers so often built with transformers instead of transformerless output stages?

Designers used transformers because early germanium transistors were expensive, unstable, and inconsistent, so matching complementary pairs by hand was slow and wasteful. Transformer-coupled circuits let builders avoid that selection work. The article says this mattered when production technology was immature and similar transistor pairs had to be picked from several samples to hit usable amplifier parameters. [#21905241]

How can you recreate a germanium-like vintage sound using silicon transistors and transformer-based circuits?

You can recreate much of it by using silicon transistors in vintage-style transformer circuits that intentionally shape distortion and tone. The article directly says the so-called “germanium sound” can also be achieved on silicon transistors through appropriate design choices, and it even shows a transformer-coupled silicon amplifier from magazine "Радио" No. 1/1981 as an example. [#21905241]

What makes germanium transistors thermally unstable, and why did older parts need careful soldering below about 75-85°C?

Germanium is thermally unstable because its energy gap is smaller than silicon’s, so temperature changes strongly alter current and device behavior. The article says this made germanium parts easy to damage at higher temperatures, and even soldering required protection so the crystal stayed below about 75–85°C. That is far less forgiving than modern silicon handling. [#21905241]

Why do germanium transistors have high leakage current and a wider spread of parameters than silicon devices?

They show high leakage and wider spread because germanium technology was less stable and more sensitive to temperature and production variation. The thread says reverse current is high and rises with heat, while early manufacturing often forced builders to hand-select matched pairs from several pieces. That combination made leakage, gain spread, and bias drift persistent design problems. [#21905241]

What is popcorn noise in germanium transistors, and why is it a problem in low-noise audio circuits?

"Popcorn noise" is low-frequency burst noise that appears as random crackles or popping events, a key characteristic that makes quiet transistor stages sound dirty. One poster lists popcorn noise alongside pink noise, low gain, and huge leakage as major germanium weaknesses. In a low-noise preamp or quiet audio stage, those bursts are far more objectionable than a merely colored tone. [#21905356]

What does pink noise mean in the context of transistor audio amplifiers, and how is it different from ordinary hiss?

"Pink noise" is broad-spectrum noise whose energy tilts toward lower frequencies, a key characteristic that makes it sound fuller and less sharp than high-frequency hiss. In the thread, pink noise is named as a serious germanium drawback. Ordinary hiss is perceived as more top-end noise, while pink-noise-heavy behavior raises the audible noise floor across more of the listening band. [#21905356]

How does the lower junction voltage of germanium devices affect crossover distortion in class B and class AB amplifiers?

It can reduce crossover distortion because germanium begins conducting at a lower voltage. A poster states germanium junctions conduct at about 0.2–0.3 V, versus 0.6–0.7 V for silicon, so the dead zone around zero crossing is smaller in simple class B stages. Another poster notes this matters less in properly biased class AB designs, where pre-bias already reduces the effect. [#21905606]

Why do some designers claim silicon transistor amplifiers generate unwanted high-frequency harmonics at the zero crossing, and is that really audible?

They claim it because transformerless silicon designs can produce high-frequency artifacts at zero crossing, but the thread’s main article argues those products sit above about 10 kHz and are inaudible to most listeners. The author explicitly doubts the audibility claim and says he does not hear it. In other words, the mechanism may exist, but the listening impact is disputed. [#21905241]

Can listeners really hear the difference between a germanium amplifier and a silicon transistor amplifier in normal music playback?

Some listeners may hear a difference, but the thread does not support a universal, reliable distinction in normal playback. One participant directly asks whether anyone can truly detect a germanium amplifier versus a silicon one, while the article’s main position is that the audible character comes from distortion profile and circuit choices, not from germanium alone. That makes blind identification uncertain unless the designs differ substantially. [#21905651]

Which vintage Polish and Soviet germanium transistors were mentioned for audio use, and how hard are TG50, TG70, GT107, GT108, GT308, P27, P28 and MP39B to find today?

The thread mentions Polish TG5, TG50, TG70, plus Soviet GT107, GT108, GT308, P27, P28, and MP39B as notable germanium types. It also says special low-noise grades existed, but “did not survive the era of silicon transistors” and are now very difficult to find. So the short answer is: these parts are historically relevant, but sourcing good originals today is hard. [#21905241]

How do you test an old germanium transistor like a TG51 or AF428 to check whether it is still usable in a fuzz pedal or amplifier build?

Use a simple pass/fail bench check first. 1. Measure it with a transistor tester and confirm the device is still recognized, as shown for a working TG51 in the thread. 2. Check whether leakage and gain look sane for the circuit. 3. Install it in a low-risk build such as a fuzz stage, like the old AF428 pedal example, and listen for excess crackle or instability. [#21905506]
AI summary based on the discussion. May contain errors.
ADVERTISEMENT