Germanium phono . Who uses phonopreamp on germanium transistors ?

[vs music]

In the Soviet Union, several series of germanium transistors of the N-P-N structure were produced . .....I made Phono preamps with JFET transistors at the input ( 2SK117 , 2SK170 , KP303 , KP103 , KP307 , KPS104 and K504NT ( the last two are double assemblies produced in the USSR )). They sounded good and had a very low noise level . I still have one preamp with JFET transistors . But the germanium sound is different .

[vs music]

Hello everyone !  At the request of one of the forum participants-here is a diagram of the phono preamp that was shown in the message #13 .

[aboos]

Motivated by your positive reports on the sound qualities of a germanium based phono preamp, I also have build one based on the JBL circuit. This design avoids most of the design flaws you normally find with such circuits. I modified the circuit a little bit (increase open loop gain to have better NFB control at low frequencies which gets important if the overall gain is lifted to 40 dB or more) and I must say that this circuit sounds exceptionally good - it sounds better than an XONO clone it now replaces
The only drawback was the increased noise level - so I started experimenting with different germanium transistors and finally I found one which is on par with low noise silicon transistors:
ACY32 - per datasheet a low noise SIEMENS GE transistor designed for low noise low level input stages
It is difficult to find but it is really low noise. The only other one coming close is the AC150. All other types I tested (AC121, AC151, ACY12 and many others) showed a much higher noise level.

[vs music]

Motivated by your positive reports on the sound qualities of a germanium based phono preamp, I also have build one based on the JBL circuit. This design avoids most of the design flaws you normally find with such circuits. I modified the circuit a little bit (increase open loop gain to have better NFB control at low frequencies which gets important if the overall gain is lifted to 40 dB or more) and I must say that this circuit sounds exceptionally good - it sounds better than an XONO clone it now replaces
The only drawback was the increased noise level - so I started experimenting with different germanium transistors and finally I found one which is on par with low noise silicon transistors:
ACY32 - per datasheet a low noise SIEMENS GE transistor designed for low noise low level input stages
It is difficult to find but it is really low noise. The only other one coming close is the AC150. All other types I tested (AC121, AC151, ACY12 and many others) showed a much higher noise level.

Hello! I am glad to hear that you also liked the sound of germanium phono . I now have 7 phonopreamps , and among them there are devices with a very low noise level . But 3 germanium phonos - I like them for their sound . Thank you for sharing your experience with ACY32 transistors . I don't have such transistors, but maybe someday they will appear. My version , made according to the JBL scheme , showed a noise level of minus 69 db ( RMAA ) when measured ( this is an unweighted value )  , but this is without taking into account the influence of the cartridge . Probably with a cartridge - something around minus 63-65 db. But it 's still much less than the noise of a clean groove on a record .The sound of the germanium circuit is worth the effort .

[aboos]

Yes - the sound is really worth the effort. It's hard to describe but I like it very much. I compared it to silicon transistors (the JBL circuit can be used by simply exchanging the transistors and adapting two resistors to get the same biasing) - they sound different to germanium and germanium is better for mee.
As I'm an electronics engineer by education, I do not believe in vodoo or magic - there must be a reason. The only possible one I found so far is the lower bandwidth of these old germanium transistors. The open loop gain of the JBL circuit e.g. is 82 dB below 500 Hz and it starts falling by 6 dB per octave at 800 Hz. Combined with the falling response of the RIAA eq this leads to an almost constant negative feedback factor between 500 Hz and 20 kHz. Maybe this is one factor for the good sound.

[niclaspa]

Since I was curious how Germanium transistors would sound, I whacked together the Telefunken circuit mentioned as vs music's favourite.  However, since I don't need so much gain, I didn't use his schematic, instead I built the original:



After assembling it, I measured the frequency response with an anti-RIAA circuit.  I found that the bass started to level off at 100 Hz and reached -5 dB at 20 Hz.  Maybe, this was some rumble filter in the original circuit?  I don't know how to calculate the filter in this circuit, but I draw it up in the simulation program LTSpice.  The simulation showed the same roll-off in the bass.  I tinkered around with the component values for a bit and found some values that gave a fairly accurate RIAA compensation with less than 1 dB deviation between 20 Hz and 20 kHz.  I also noticed that the filter was affected by the input impedance in the following stage.  Therefore, I also added an output buffer:





It sounds nice.  It doesn't have the harshness that I am often disturbed by in solid state phono stages, but it is a bit too soft and laid back. The transients are not good enough for me. 

Anyway, it was a fun pastime and my curiosity was satisfied.   

[vs music]

Good job , Niclas !  This circuit does have a ramp at low frequencies ( as with all circuits on 2 transistors ) . In addition , there is a small nominal value of C7 at the output . Such schemes really need an additional buffer cascade - otherwise the slope at low frequencies will be even greater ( if the load is below 100 kohm ) . I also did the 3rd cascade.   I really like the sound of this preamp .

[aboos]

Most of these old 2 transistor schematics are not really high performance designs but compromises - just good enough performance with lowest possible cost (when these designs were created, transistors were really expensive). Major flaws are:
- the 2nd transistor needs to drive a quite low load at high frequencies as the feedback network will have a low impedance (only a few kOhms) which it is not capable of
- the 2nd transistor input impedance is quite low and loads the first transistor collector and such limits the open loop gain with the effect that the feedback factor is very low (almost non existent) for low frequencies and high for high frequencies
Overall result is exactly the sound experience described - missing sparkle in the highs, lack of transient attack ... - and this is not due to germanium transistors but due to circuit design flaws.

The JBL circuit (post#16 above) avoids these compromises to a great extend by
- driving the feedback network from the emitter follower unloading the 2nd transistor's collector and decoupling it effectively
- bootstrapping the bias resistor of the 2nd transistor from its emitter and thus effectively rising the input impedance not loading the first transistors collector too much
That's the reason why I realised the JBL schematic. Bass is well defined, has good punch and transients are perfect

[vs music]

Niclas and Andreas - thank you for your feedback !  A lot of germanium transistors are still sold in my country . The USSR produced them until 1991 . And among music lovers - there is an interest in amplifiers and preamps on germanium transistors . Some forums publish new circuits on germanium transistors . Here is a link to one of these forums . These are Sergey Toropov 's preamp schemes . https://hiend.borda.ru/?1-7-0-00000316-000-0-0

[vs music]

A scheme authored by A.I. Beslik was published in Russia . And many people have built such a preamp , and have given good reviews about its work . Several variants of the scheme have been published - with different values of the RIAA elements .  Here is one of these options . Germanium transistors of the USSR are used here ...

[vs music]

The Beslik circuit uses low-noise transistors P-28 ( П-28 ) . But these transistors have a limit voltage Uce of only 5 volts . If other transistors are used in the first two stages, you can increase the supply voltage of this circuit .

[niclaspa]

Andreas, interesting!     You mention that you had modified the JBL circuit a bit.  Could you give the details?  What is the gain of the circuit?

[vs music]

Here is an instruction and a diagram of the GATES M6169 broadcast phono preamp ( with germanium transistors ) - https://www.steampoweredradio.com/pdf/gates%20harris/manuals/m6169%20tt%20pre%20amp.pdf

[aboos]

Andreas, interesting!     You mention that you had modified the JBL circuit a bit.  Could you give the details?  What is the gain of the circuit?

Here is what I did:


Collector resistor of 2nd transistor is split into two and midpoint is bootstrapped by emitter follower output. This increases the AC collector impedance significantly and thus raises gain. I also moderatly increased the collector resistance of Q1, also increasing gain. With the closed NFB loop, overall gain is 40 dB at 1kHz (to lower gain, R2 can be increased. 120 Ohm will result in approx. 34 dB at 1 kHz). Resistor R15 (1kOhm in input) makes the circuit stable for all possible source resistances. Noise contribution is still pretty small when used with a standard MM cartridge.
The open loop gain (gain with no NFB) is about 80dB and thus gives at least 20dB feedback even at 50 Hz.
The C7+R16 combination reduces the gain of the second transistor starting around 800 Hz. This forms a dominant pole and avoids any tendency for RF oszillations and it avoids exessive feedback factors for higher frequencies (e.g. 35 dB at 10 kHz instead of >50 dB otherwise)
The LT Spice model of the AC151 (I used the ACY 32, but parameters are almost identical with the exception of a much lower noise factor) is quite accurate as the predicted frequency response (open loop and with NFB) matches perfectly my measurements and also DC operating points are spot on.
R7 (input resistor to ground) is 75 kOhm resulting in an effictive input resistance of approx. 68 kOhm. If you want to have the standard 47 kOhm, you may reduce R7 to 68k or 56k.
All electrolytics (10µF, 47µF, 470µF and 1000µF) are Panasonic FC.

Andreas

[niclaspa]

Thanks Andreas!