let me define horrible for a minute ...
beyond spectrum analyzer and scope on a variety of loads there is always a listening session for an amplifier ...
To put you more in the picture it is my job to repair amplifiers...today i repair an average of 400 per year ...so to keep thing in control i have a near field set of speakers ( quite bright) specific music that i know very well ( with some variety ) but what is really important is that for almost 5 years the rest of the system including sources is always the same .
This as a practice alone i suggest to anyone that wants to evaluate condition A to condition B which only consume time and will cost no money only thing you need to do is consume time yo learn your music and system very well .
Expanding to a huge variety of music and 4 sets of speakers or sources produces way too many variables and results that are not usable for A to B tests .
So every time an amplifier is on a listening test things like, sound stage,bandwidth,depth, speed ,distortion (s) , guts, coloration ,discoloration and many other information that the ear can ""see"" about sonic behavior that instruments canot present as graphic images ...
kind regards
sakis
beyond spectrum analyzer and scope on a variety of loads there is always a listening session for an amplifier ...
To put you more in the picture it is my job to repair amplifiers...today i repair an average of 400 per year ...so to keep thing in control i have a near field set of speakers ( quite bright) specific music that i know very well ( with some variety ) but what is really important is that for almost 5 years the rest of the system including sources is always the same .
This as a practice alone i suggest to anyone that wants to evaluate condition A to condition B which only consume time and will cost no money only thing you need to do is consume time yo learn your music and system very well .
Expanding to a huge variety of music and 4 sets of speakers or sources produces way too many variables and results that are not usable for A to B tests .
So every time an amplifier is on a listening test things like, sound stage,bandwidth,depth, speed ,distortion (s) , guts, coloration ,discoloration and many other information that the ear can ""see"" about sonic behavior that instruments canot present as graphic images ...
kind regards
sakis
Apart from it being highly unlikely anyone would hear such obvious tonal difference between modern high Ft audio power transistors, you don't show the application, type, schematic or conditions of comparison.
Bear in mind that any obvious differences will be in circuits that are not optimal for the particular device or possibly fakes which are quite common now due to our careless bargain buying practices. You can't be certain what you are comparing unless buying from the manufacturer or their certified distributor.
Actually, with most modern audio output BJTs there is a lot going on in the emitter structure that makes direct comparison under 'same conditions' difficult, at best. At the very least, even keeping output topology exactly the same, there will be differences in optimal standing current, and it is I think beyond discussion that getting this right indeed changes the sound of an amp. 2SC5200/2SA1943 are rather different from 2SA1216/2SC2922 in this respect, and that's only one point of difference, actually one of the simpler ones. For instance, 2SA1216/2SC2922 will happily operate without external (note!) emitter resistors and with proper compensation remain thermally stable.
Things get far more involved once changes of beta with current, changes of Ft with current and chnages of capacitances with almost everything are taken into account.
Things get far more involved once changes of beta with current, changes of Ft with current and chnages of capacitances with almost everything are taken into account.
Agree.
Output transistors, being most often used as unity gain followers (and if not, as unity gain complementary pairs, which ends up being equivalent) by definition have the most extended trequency response they can provide.
In this case, the transition frequency is given as 30 MEGA Hz.
And they are typically used not by their own, but within complex, lots of NFB designs, which would certainly *swamp* any audible difference, at least something so gross as to be considered "muffled", which implies *many* dB loss at audible frequencies.
Which is not the case.
Output transistors, being most often used as unity gain followers (and if not, as unity gain complementary pairs, which ends up being equivalent) by definition have the most extended trequency response they can provide.
In this case, the transition frequency is given as 30 MEGA Hz.
And they are typically used not by their own, but within complex, lots of NFB designs, which would certainly *swamp* any audible difference, at least something so gross as to be considered "muffled", which implies *many* dB loss at audible frequencies.
Which is not the case.
East electronics,
I read your post with great interest where you mentioned the 5200 as rough and dull .
My tech and i were sorting out a replacement pair for my sansui amplifier which blow one pair of 1386 /3519 pair of sankens .
Yesterday he told me he used toshiba 5200 transistors with a driver 1306 (the green ones) and he was praising about the 52s.
I know this is about perspective and how much experience you are with sound quality of numerous amps. But he has a lot of amplifier experience and owns some of the best vintage sounding amps .
So it drove my inquisition and id like to ask what driver transistors were you using back in 2012 with those 5200 that made the sound bad.
I read your post with great interest where you mentioned the 5200 as rough and dull .
My tech and i were sorting out a replacement pair for my sansui amplifier which blow one pair of 1386 /3519 pair of sankens .
Yesterday he told me he used toshiba 5200 transistors with a driver 1306 (the green ones) and he was praising about the 52s.
I know this is about perspective and how much experience you are with sound quality of numerous amps. But he has a lot of amplifier experience and owns some of the best vintage sounding amps .
So it drove my inquisition and id like to ask what driver transistors were you using back in 2012 with those 5200 that made the sound bad.
2sc3858 Silicon NPN Triple Diffused Planar Transistor (Complement to type 2SA1494) Application : Audio and General Purpose
and
2sc2922 Silicon NPN Epitaxial Planar Transistor (Complement to type 2SA1216) Application : Audio and General Purpose
what is the difference between these technologies triple diffused vs epitaxial ? Which of them is better in same circuit for example krell ksa50 ? Maybe one of them gives less distortion?
and
2sc2922 Silicon NPN Epitaxial Planar Transistor (Complement to type 2SA1216) Application : Audio and General Purpose
what is the difference between these technologies triple diffused vs epitaxial ? Which of them is better in same circuit for example krell ksa50 ? Maybe one of them gives less distortion?
As I know, the main difference is the value of transit frequency (ft) - check out post #1 under
https://www.diyaudio.com/community/...amilies-for-audio-power-output-stages.147656/
Toshiba was/is king of the triple diffused types, Sanken the king of epitaxial planar. But not all Toshiba triple diffused types were ”TMe“ or “MET” , and not all Sanken epitaxial planar were “LAPT”. Consult the documentation to find out what types were the good ones - they typically ran higher price. Toshiba made epitaxial planar and Sanken made triple diffused too, but they didn’t seem to be anything special.