If I'm understanding all of this; the goal is a high slew rate limit?
Looking at 1 example: an amp that belongs to a friend advertises a slew rate of 22V/uS. Would this be a slew rate limit and does that correspond to the maximum frequency of the amp?
Looking at 1 example: an amp that belongs to a friend advertises a slew rate of 22V/uS. Would this be a slew rate limit and does that correspond to the maximum frequency of the amp?
I have discussed this in detail in this thread. I really have no intention to copy my own text 😎
Jan
Jan
The goal is a high enough slew rate. If your big signals are clipping before the slew rate becomes a problem, then it's high enough.
Without knowing the designed power output (i.e., the max voltage swing) there's no way to know how much slew rate you really need. Using the already-mentioned rule of thumb (>1V/uS per volt of output swing), 22V/uS should be fine for 60W into 8-ohm loads. If your friend's amplifier is designed to give 25W, great. If it claims 250W, he may have a problem.
-JS
I believe that these old Bob Cordell writing on Transient Intermodulation Distortion should answer your question well.
CordellAudio.com - Another View of TIM
Sorry to borrow this thread. How about offset voltage?? Does it matter?? I am asking this because the OP177G used in the S300i has a slew rate of just 0.6v...but it has very low offset voltage. Why would an amp use just an op amp?? But it has very low offset voltage.
Offset can increase the even-order distortion of the connected loudspeakers. I haven't a clue how much offset you would need for this to become a problem.
Is that OP177G in the forward path or in a DC feedback loop? In the latter case its slew rate doesn't matter.
Is that OP177G in the forward path or in a DC feedback loop? In the latter case its slew rate doesn't matter.
No. The goal is accurate input signal tracking. 🙂
Too low a slew rate limit will be a problem and too high a slew rate limit could be a problem due to the way the circuit functions. Designers usually have two ways to boost max slew rate: increase currents and decrease capacitances. These can have adverse side-effects like increasing distortion and reducing stability. So a higher slew rate limit is not always a good sign.
5V/us is the theoretical slew rate of a 40V peak 20kHz sinewave, eg: a 100W average power. Music you listen to from digital sources will rarely if ever require this.
It is better to make goal for lowest distortion at 20kHz full power. Usually, if slew rate amplifier too low, it is difficult to make 20kHz distortion low. Bob Cordell advocate this.
In DIY, you should try yourself, how slew rate limit affect to the sound. You can over compensated your amplifier and compare it with optimal compensated.
In DIY, you should try yourself, how slew rate limit affect to the sound. You can over compensated your amplifier and compare it with optimal compensated.
Can Slew rate be improved by using paralleled transistors in the input stage LTP?
Lets assume a bipolar LTP input stage, using BC547C.
Would application of 4 transistors, instead of 2 transistors (paired up in each of the long tail branches so that they have their independent emitter degeneration resistors, say 200R, instead of the single 100R in the case of a pair of single transistors), and then the collectors joined and cascoded .... provide any added value to the overall parameters that could be achieved by an amplifier using such concept?
Just curious.
I believe that I have seen constructs somewhere, in which the input devices were paralleled. But what would be the benefit thereof?
Lets assume a bipolar LTP input stage, using BC547C.
Would application of 4 transistors, instead of 2 transistors (paired up in each of the long tail branches so that they have their independent emitter degeneration resistors, say 200R, instead of the single 100R in the case of a pair of single transistors), and then the collectors joined and cascoded .... provide any added value to the overall parameters that could be achieved by an amplifier using such concept?
Just curious.
I believe that I have seen constructs somewhere, in which the input devices were paralleled. But what would be the benefit thereof?
Assuming that the total tail current is also kept the same, so each new pair on half the original tail current, I can't think of any way that could improve the slew rate. It might degrade it, if the extra capacitance of the extra transistors necessitates a stronger frequency compensation, but most likely it will not have any effect on slew rate at all.
Transistors are often paralleled to reduce the thermal noise of their base resistances (resistance of the doped semiconductor out of which the base is made, that is), or to improve power handling.
Transistors are often paralleled to reduce the thermal noise of their base resistances (resistance of the doped semiconductor out of which the base is made, that is), or to improve power handling.
But in case you double the sum of LTP current SLR may increase, depending on the details of the circuit.
Indeed, but you can also do that by just doubling the tail current of a single pair of transistors - provided they can handle the current.
Thank you Gentlemen for your replies.
So basically: if current to remain unchanged - no apparent benefit.
I will stick to the original schematic then (if it works - don't fix it).
Speed in amplifiers seems to be critical in audio performance .
Audio amplifier must be fast , but ...... for every tone and overtone - at every frequency simultaneously , at every amplitude , at every load ..etc
SR says nothing about this .
Can be useful for designing puposers , but not for audio qualities .
BR
Audio amplifier must be fast , but ...... for every tone and overtone - at every frequency simultaneously , at every amplitude , at every load ..etc
SR says nothing about this .
Can be useful for designing puposers , but not for audio qualities .
BR
Amplifiers are very complex and the input differential pair is just one small cog in a large machine.
I bet the speed bottleneck lies downstream.
I bet the speed bottleneck lies downstream.
Speed in audio amplifiers is like acceleration and speed in cars.
The speed of cars are not tested with an explosion of 100 ton of TNT in is bottom , it could reach the moon , but is unreal .
The speed in amplifiers must be lineal and specially accurate to the input signal , at every condition and step of voltage and load .
That is quite impossible in electronic devices that amplify the signal.
The lack of speed (not infinite) only can be replaced with attenuation . The human ear seems to be more tolerable to these changes.
Don't worry if you cannot reach the moon.
BR
The speed of cars are not tested with an explosion of 100 ton of TNT in is bottom , it could reach the moon , but is unreal .
The speed in amplifiers must be lineal and specially accurate to the input signal , at every condition and step of voltage and load .
That is quite impossible in electronic devices that amplify the signal.
The lack of speed (not infinite) only can be replaced with attenuation . The human ear seems to be more tolerable to these changes.
Don't worry if you cannot reach the moon.
BR
CD goes up to 20kHz. That is not very fast. 
The claim that an amp needs to be very fast needs explanation.
The claim that an amp needs to be very fast needs explanation.
- Status
- Not open for further replies.