The phenomenon is called "yes no teddy bear". It goes like this. If you are not sure of something, try to explain it to someone who doesn't know much about it. Encourage them to say "yes" or "no" as you go along, even though they don't fully actually understanding what you are saying. The best outcome is that you will have a eureka moment as you final work it out without any expert assistance.
The "yes no bear" was actually a real toy bear, see Schuco Bears
The "yes no bear" was actually a real toy bear, see Schuco Bears
Greg,
Your PCB's are spectacular! Well done!
Kean,
If you can explain amplifier operation such that I can understand, you will have succeeded where nobody else has been able to. I've ready every book about audio amplifiers that's been written since about 1945.
Yes, I'm particularly dense. But each author seems to find one (the same one perhaps) crucial concept to gloss over and lose me as a result. And, did I mention I am particularly dense? I'm still waiting for the perfect amplifier book.
..Todd
Your PCB's are spectacular! Well done!
Kean,
If you can explain amplifier operation such that I can understand, you will have succeeded where nobody else has been able to. I've ready every book about audio amplifiers that's been written since about 1945.
Yes, I'm particularly dense. But each author seems to find one (the same one perhaps) crucial concept to gloss over and lose me as a result. And, did I mention I am particularly dense? I'm still waiting for the perfect amplifier book.
..Todd
Pete,
When you've done that, you might think about a hybrid, tube front end, unity buffer SS output stage...... zero feedback.
Up for a challenge? No servos or current sources permitted
Todd,
You and I have conversed on amp design, but I never did find where you tripped up. Would you like to tell us, and I will have another go!!
Cheers,
Hugh
When you've done that, you might think about a hybrid, tube front end, unity buffer SS output stage...... zero feedback.
Up for a challenge? No servos or current sources permitted
Todd,
You and I have conversed on amp design, but I never did find where you tripped up. Would you like to tell us, and I will have another go!!
Cheers,
Hugh
Todd,
You and I have conversed on amp design, but I never did find where you tripped up. Would you like to tell us, and I will have another go!!
Cheers,
Hugh
Heck, I can't even figure out how a differential pair works.
It's sad and hopeless. 
..Todd
Oh yeah?
Hmmm, this could be interesting. I'm sure Bigun is still planning on going "tubey" with this kind of approach too. Make it a triple SS output and get some real fireworks going!
Pete, you might think about a hybrid, tube front end, unity buffer SS output stage...... zero feedback.
Up for a challenge? No servos or current sources permitted
Hmmm, this could be interesting. I'm sure Bigun is still planning on going "tubey" with this kind of approach too. Make it a triple SS output and get some real fireworks going!
Heck, I can't even figure out how a differential pair works.
..Todd
Hello Todd
I don't know what is your level of knowledges in electronics, but here is a nice page about amp, they explain differential pair near the end of the page.
Transistors
Bye
Gaetan
Last edited:
That would require modifying the power boards.
I got the "baby" down to .05% 200w !! THD20 and .003% at 1k normal 60w use while maintaining the EXACT harmonics of the "baby". Good enough for me..
Thanx a lot Keen, your "tricks" were invaluable.
OS
Pete,
When you've done that, you might think about a hybrid, tube front end, unity buffer SS output stage...... zero feedback.
Up for a challenge? No servos or current sources permitted
Cheers,
Hugh
I am very surprised you never made a "monster" from the "Baby". I am SO surprised what this little thing can do on so few parts!! I know it is only a sim ... but the cap multipliers on my big powerboard should allow for this performance in real life. Only 2 drawbacks for this classic... slew and gain (12-13v/uS and only 48db (no degeneration on the KSA1381. I used BOTH a CCS and the resistive tail source together so I could "dial in" for 40 or 70V rails and trim for perfect tail balance in either mode(halves the thd).
OS
Attachments
Andrew,
You simply measure voltage across the emitters of the input LTP. If the currents are balanced, then the currents are equal and the voltage drops across the two 22R resistors (from Ostripper's schemat) are identical. If such is the case, then the voltage across the emitters is exactly and precisely zero.
In fact, there is no requirement for a load resistor on the feedback device. It does nothing to trim 'balance'.
Yes, Pete, it's capable of outrageous power, and is one of the most scalable designs around.
Cheers,
Hugh
You simply measure voltage across the emitters of the input LTP. If the currents are balanced, then the currents are equal and the voltage drops across the two 22R resistors (from Ostripper's schemat) are identical. If such is the case, then the voltage across the emitters is exactly and precisely zero.
In fact, there is no requirement for a load resistor on the feedback device. It does nothing to trim 'balance'.
Yes, Pete, it's capable of outrageous power, and is one of the most scalable designs around.
Cheers,
Hugh
Hi,
that's where I measure balance of currents. The differential voltage across the emitters is very sensitive, if one has selected accurate emitter resistors.
Where do you measure balance of Vbe? I note an error of 1.53mVbe in the posted voltage readings. This seems to indicate that the selected pair have not been matched at operating current.
Where do you measure balance of Vce? Omitting the collector load in one half will create unbalance by increasing Pq and thus differential temperature of the "balanced" devices.
that's where I measure balance of currents. The differential voltage across the emitters is very sensitive, if one has selected accurate emitter resistors.
Where do you measure balance of Vbe? I note an error of 1.53mVbe in the posted voltage readings. This seems to indicate that the selected pair have not been matched at operating current.
Where do you measure balance of Vce? Omitting the collector load in one half will create unbalance by increasing Pq and thus differential temperature of the "balanced" devices.
Last edited:
True, Andrew, but the two devices of the LTP should be selected for identical Vbe at identical currents and voltage, of course.
Now, you mention pd of each device, and stress that balance is not possible if both devices do not have identical loads.
Let's look at the figures.
Assume a rail voltage of 36, that used on the AKSA and the Baby AKSA.
If we choose to use a VAS with no degeneration, and assuming that each input node of the LTP is at an identical voltage (as must be their conjoined emitters, or very close to it), then the Vce of the left (input) device will be (36 + Vbias + 0.6 -0.6) = 36.1V typically. The additive 0.6V is the voltage between base and emitter, while the subtractive 0.6V is the voltage across the base emitter of the VAS.
On the right (feedback) device, the Vce is (36 + Vbias +0.6) = 36.7V typically.
Now, if each device is passing 1mA (that is, in perfect balance, as designed) then the dissipation of the left input device is 36.1mW and that of the right feedback device is 36.7mW.
I put it to you that this difference is much too small to justify fitting a resistive load to the right feedback device. Even if one did fit an additional collector load resistor, its value would NOT be identical to the left input collector load, since this VAS driving side has current diverted to bias up the VAS.
Cheers,
Hugh
Now, you mention pd of each device, and stress that balance is not possible if both devices do not have identical loads.
Let's look at the figures.
Assume a rail voltage of 36, that used on the AKSA and the Baby AKSA.
If we choose to use a VAS with no degeneration, and assuming that each input node of the LTP is at an identical voltage (as must be their conjoined emitters, or very close to it), then the Vce of the left (input) device will be (36 + Vbias + 0.6 -0.6) = 36.1V typically. The additive 0.6V is the voltage between base and emitter, while the subtractive 0.6V is the voltage across the base emitter of the VAS.
On the right (feedback) device, the Vce is (36 + Vbias +0.6) = 36.7V typically.
Now, if each device is passing 1mA (that is, in perfect balance, as designed) then the dissipation of the left input device is 36.1mW and that of the right feedback device is 36.7mW.
I put it to you that this difference is much too small to justify fitting a resistive load to the right feedback device. Even if one did fit an additional collector load resistor, its value would NOT be identical to the left input collector load, since this VAS driving side has current diverted to bias up the VAS.
Cheers,
Hugh
Last edited:
I agree that the VAS voltage is all that differentiates between Vce of the two halves. It can typically be between 550mV and 1200mV depending on adoption of VAS degeneration.
Yes, the numbers are small, but are they insignificant? I believe that Os is finding that more precise balancing of Ie of the LTP pair is measurable. What if we can take this further? Many before him have said that precise balance of the LTP is not necessary, is not audible and is not measurable.
Are there more effects that are measurable, like balanced Vce and balanced Vbe?
Yes, the numbers are small, but are they insignificant? I believe that Os is finding that more precise balancing of Ie of the LTP pair is measurable. What if we can take this further? Many before him have said that precise balance of the LTP is not necessary, is not audible and is not measurable.
Are there more effects that are measurable, like balanced Vce and balanced Vbe?
Very close balancing of the LTP , even in the JLH "Ripple Eater" PSU add-on is audible, believe it or not! That is why in the schematic shown in the recent GB, there is an additional collector resistor, where JLH himself had none.
Hugh, the improvement is just as noticeable (more so?) in this topology as the added diode in your Soraya's CM.
Regards
Alex
Last edited:
I have 200R trimmers for "CCS ADJ." "tacked" on my 2 "supersyms" AND my 2 DX style baby amps. I would not of done this just to be "geeky". A very noticeable increase in depth and soundstage occurs in the "sweetspot". This was verified by my new scope as the voltage waveforms coincided the sound "sweetened". Very obvious cause and effect.
All my new amps will have this feature.
OS