[jezz]

"Son of Zen" is a different amp than Zen (article here). Pass calls it "balanced single-ended."

 

[wdiabc]

Quote:

"Son of Zen" is a different amp than Zen (article here). Pass calls it "balanced single-ended."

Thanks for the article. I guess they are almost identical. Then this is a BJT version for Son of Zen. 

 

[wdiabc]

Updates: 
 
So far I start to realize that this amp rolls off at the high end, near 15-20Khz, because the stax headphone impedance drop to around 50K to 100K near that frequency and given the huge output impedance of this circuit ( near 50K), certainly the 20Khz end will roll off, which is OK to me and that really made Stax Lambda sound like my ortho, haha. 
 
I will try to play with idle current but does any one know what idle current stax is using for some of their high end amps? I am referring to T1 T2 323 etc. I have looked at some other amp designs, and many of the tube versions simply used a 50K load resistor for the output stage, which is identical to my design, so I feel the output impedance is really the same, I dont know how these amps deal with the high frequency roll off. 
 
Please help. 

 

[kevin gilmore]

sr-007 and sr-009 are j66k at 20khz
 
srm727-2 is 8.5 ma per output driver, and is 3db down at 100khz with a standard 120pf load
all the other solid state amps are similar. Remember that there is up to 20db of feedback to
keep the frequency response flat.
 
the tube designs are 50k plate resistors, and 3db down at about 50khz. Also 20db of feedback.
 
The T2 is 20ma of bias per output tube and is 3db down at about 230khz.
(lots of other tricks going on here)

 

[wdiabc]

sr-007 and sr-009 are j66k at 20khz

srm727-2 is 8.5 ma per output driver, and is 3db down at 100khz with a standard 120pf load
all the other solid state amps are similar. Remember that there is up to 20db of feedback to
keep the frequency response flat.

the tube designs are 50k plate resistors, and 3db down at about 50khz. Also 20db of feedback.

The T2 is 20ma of bias per output tube and is 3db down at about 230khz.
(lots of other tricks going on here)

Thank you so much for the info. Now I am really having a hard time if i want to boost up the idle current or not. I will try another layout to handle this roll off issue

 

[wdiabc]

updates, using this new layout can fix the high freq roll off, sorry guys I used negative feedback, but it helps to make the FR flat: 
 
It is a little trick that discovered, this transistor is 'tube like', because its now biasing itself, and also the changes in the 'plate voltage' really cause the transistor to change its Ic, which means now this transistor is like more like a resistor. So it does two things, 1) achieved negative feedback for a better frequency response and 2) lowered its dynamic resistance, and increased output.
 
I always try to avoid capacitor when applying feedback, my oscilloscope told me these inductors and capacitors really makes the oscillation happen. 
 
btw, the two small Re sets the gain of the stage, and the variable resistor sets the idle current. Some math here: 
V of LED: 1.8v - 0.6V( Vbe) = Idle Current * Re+2*Idle Current *Rvar
 
For example:
1.8v-0.6v = 3ma * 200ohm + 6ma * 100ohm, so we need the var resistor to be 100 ohm for 3ma idle current. of course this number is definitely off, but not by much, since we have to consider the Ibe
 

 

[wdiabc]

now I am happy, time to ditch these testing resistors, I am going to replace them with carbon Comp resistors, my all time favorite! 
 
 

 

[wdiabc]

Quote:

Surprisingly, this is a cheap inverter from HK and its a 50khz fast switching circuit. If you take a look at this circuit, 50hz switcher will need a huge transformer. The efficiency of this circuit maybe around 80-90%. The to220 transistor does not get hot at all.

with extensive tests, this little power unit is capable to handle 19volts DC input, with an output of 320V continuously at 12ma. However, it will need a soft start circuit, in order to allow this thing to start oscillation, otherwise with the heavy load it won't start to oscillate and the power supply voltage will not go up. I simply used a transistor for that circuit, I broke up another mpsa42 during experiment and it died because of this imperfect 'soft start circuit', so i am not posting a misleading schematic for it. 

 

[wdiabc]

Update:
 
I killed one channel of the fiio e10 during testing....Now I need to fix it. But I can still use the lineout, I need to be more careful next time. 
I will soon post a new circuit which allows this amp to be used by unbalanced inputs. 

 

[jcx]

Quote:

...The T2 is 20ma of bias per output tube and is 3db down at about 230khz.
(lots of other tricks going on here)

has this changed? - schematic claimed 10 mA ccs

 

[kevin gilmore]

original T2 was 20ma. Which is why it ran at stupidly high temperatures.
 
Diy T2 runs at less current. And still gets very hot.

 

[spritzer]

"Stax SRM-T2, fry some eggs while you listen"... 

 
I've been working on possible ultra simple Stax amps for a few months now but but I'm so frustrated by the disappearance of high voltage transistors that I'm going with all tubes for this one.  That has problems as well but I'm fairly sure ECC83's will remain in production for decades to come. 

 

[wdiabc]

Ecc83 can be another great option. Maybe you can consider a cascaded hybrid circuit. Simply cascade two tubes on top of the two transistors then you will have enough gain.

"Stax SRM-T2, fry some eggs while you listen"... 

I've been working on possible ultra simple Stax amps for a few months now but but I'm so frustrated by the disappearance of high voltage transistors that I'm going with all tubes for this one.  That has problems as well but I'm fairly sure ECC83's will remain in production for decades to come. 

 

[spritzer]

That is possible but one has to thing about the filament-cathode voltage limit in that regard.  Gain is not really the problem with these amps though as most are run without feedback.  Add feedback and it becomes an issue. 

 

[wdiabc]

Quote:

That is possible but one has to thing about the filament-cathode voltage limit in that regard.  Gain is not really the problem with these amps though as most are run without feedback.  Add feedback and it becomes an issue. 

The design with tubes ususally comes with no gloabl feedback, and with a 50K plate reisistor, I was wonderng how does the circuit deal with high frequency roll offs? ( plate resistor is big and can signiciantly cause a roll off in output when working at 20Khz, since the stax headphone has about 50K resistence at 20Khz)