I was wondering what the current thinking is on this.
is -3db 10 megHz better that 5 meghz closed loop bandwidth?
is -3db 20 megHz better that 10 meghz !!!
at what frequency does any further increase become irrelevant ?
should we aim at dB of feedback being constant throughout the audio bandwidth ?
anyone know about this ?
mike
is -3db 10 megHz better that 5 meghz closed loop bandwidth?
is -3db 20 megHz better that 10 meghz !!!
at what frequency does any further increase become irrelevant ?
should we aim at dB of feedback being constant throughout the audio bandwidth ?
anyone know about this ?
mike
no jokes .. 
I have Transformer volume control so my audio bandwidth is upper limited to about 100khz
broader bandwidth within in the amplifier has other benifits.
lower distortion at high ( sonic ) frequencies.
also better PSRR at high ( sonic & ultrasonic ) frequencies.
not sure if the lower distortion will be noticable but for me anything that reduces noise is good news.
I just wonder if anyone has experience in this area
mike
I have Transformer volume control so my audio bandwidth is upper limited to about 100khz
broader bandwidth within in the amplifier has other benifits.
lower distortion at high ( sonic ) frequencies.
also better PSRR at high ( sonic & ultrasonic ) frequencies.
not sure if the lower distortion will be noticable but for me anything that reduces noise is good news.
I just wonder if anyone has experience in this area
mike
mikelm said:
You do not mention if it is about pre amplifiers or power amps.
And so we do not know at what voltage gains we are.
If we take a power amp with gain x20 and -3dB at 20 Mhz
it sure has got some GBW = gain x bandwidth.
And you sure need our GOOD LUCK
to build that one ...Pre amplifiers with moderate and lower gains
can have high upper freq limit.
But the point at where you get more troubles and distortion
than you get anything useful out of it
is in reality for audio much lower, than most of us like to think it is.
I agree with AndrewT here.
And I would be very pleased if I build a power amp
- with 20-50 Watt into 4/8 Ohm
- with -3dB at 50.000 hertz
- with granted stability into most normal real life loads
I probably would not feel too bad if I only managed 30-40kHz as long as distortion at 1 kHz stays low and nice.
And I do not get any ringing disturbance, when see square waves at my oscilloscope testings.
You can simulate and maybe even build power amps for 1 MHz output.
But a piece of advice:
do not use it with loudspeakers at the output!
Stick to using low capacitance low inductance pure resistive
absolute linear power resistors for load.
And enjoy the quiet, pure distortion free and very laidback sound of such resistors!
The lovely sound for extreme test data audio freak monsters.
Here is a topic where me and another guy share some of opinions and experiences
on upper band frequency limits for real audio applications.
Solid State >input filter for microphone preamp
http://www.diyaudio.com/forums/showthread.php?postid=1091772#post1091772
http://www.diyaudio.com/forums/showthread.php?postid=1093017#post1093017
Regards & New Year!
lineup
have ALL HIS HEARING limited to below ~12.000 hertz==============================================
* to test your ears upper bandwidth limit:
1. Use your CD-player & your Headphones
2. Calibrate for a normal listening volume (level) at 1.000 hertz sinus signal.
3. Without changing Volume try 10 > 11 > 12 > 13 >14 kHz
Until you can not hear just about nothing at all.
Remember that this result IS NOT -3dB .... more likely -20dB / -30 dB (my guess)
for Your Ears.
The actual -3dB frequency is considerably lower than this
==============================================
Re: Re: What is the optimum -3db bandwidth for amplifiers with feedback ?
Thanks for your reply
I speak of power amps and I find 26dB gain is sufficient
JLH simple class A goes to about 1meg and stability is fine if care is taken and many people love the sound of this amp. I normaly played this amp through speakers .. .
I wonder has anyone had experience with wide bandwidth power amps ???
cheers
mike
lineup said:
Thanks for your reply
I speak of power amps and I find 26dB gain is sufficient
Originally posted by lineup
If we take a power amp with gain x20 and -3dB at 20 Mhz
it sure has got some GBW = gain x bandwidth.
And you sure need our GOOD LUCK
You can simulate and maybe even build power amps for 1 MHz output.
But a piece of advice:
do not use it with loudspeakers at the output!
Stick to using low capacitance low inductance pure resistive
absolute linear power resistors for load.
JLH simple class A goes to about 1meg and stability is fine if care is taken and many people love the sound of this amp. I normaly played this amp through speakers .. .
I wonder has anyone had experience with wide bandwidth power amps ???
cheers
mike
Bandwidth should be limited by input RF filter and/or output inductor.
If there is no input filter and you ask about bandwidth BEFORE output inductor, then 200kHz is more than mandatory in my opinion (-3+0 dB).
Some power amps, including mine and I believe MikeB's or Bob Cordell's and so on... easilily exceed 1Mhz in this criterion, but still not in 5-10MegHz range. Op-amps and small signal gain stages are faster.
What for?
More feedback at 20kHz using simple compensation, lower TIM/PIM, sometimes flat feedback factor across audio...
regards
If there is no input filter and you ask about bandwidth BEFORE output inductor, then 200kHz is more than mandatory in my opinion (-3+0 dB).
Some power amps, including mine and I believe MikeB's or Bob Cordell's and so on... easilily exceed 1Mhz in this criterion, but still not in 5-10MegHz range. Op-amps and small signal gain stages are faster.
What for?
More feedback at 20kHz using simple compensation, lower TIM/PIM, sometimes flat feedback factor across audio...
regards
thanks for your reply,
Yes my i/p bandwidth is limited by TVC < 100khz
and with this kind of amp a very small ( damped ) inductor at o/p is needed
I'm not suggesting I try to amplify 1meg + and send it to speakers !
rather I'm interested in the advantages & possible disadvantages of amps with high internal bandwidth.
I just wonder has anyone compared 1mhz with 5mhz internal bandwidth - even if input is filtered to 100khz or less
in other words, does flat, low distortion to 10 - 20khz and extra high frequency PSRR make an audible difference in others experience ?
cheers
mike
Yes my i/p bandwidth is limited by TVC < 100khz
and with this kind of amp a very small ( damped ) inductor at o/p is needed
I'm not suggesting I try to amplify 1meg + and send it to speakers !
rather I'm interested in the advantages & possible disadvantages of amps with high internal bandwidth.
I just wonder has anyone compared 1mhz with 5mhz internal bandwidth - even if input is filtered to 100khz or less
in other words, does flat, low distortion to 10 - 20khz and extra high frequency PSRR make an audible difference in others experience ?
cheers
mike
The distiction between the three is reasonably clear. ( though I'm a bit foggy of the presise definition of gain bandwidth product - can it be used to define closed and open loop situations ? )
my original question refers to HF -3db point closed loop gain - there is no confusion here.
The open loop unity gain of such an amplifier would be very high, 50 - 100meghz perhaps.
I guess I will have to try for myself and find out.
I agree that the lower HF distortion may be difficult to distingush but I have always found in the past the lowering noise at all frequencies reaps sonic rewards so I think if it is do-able the extra PSNRR will be an asset.
mike
my original question refers to HF -3db point closed loop gain - there is no confusion here.
The open loop unity gain of such an amplifier would be very high, 50 - 100meghz perhaps.
I guess I will have to try for myself and find out.
I agree that the lower HF distortion may be difficult to distingush but I have always found in the past the lowering noise at all frequencies reaps sonic rewards so I think if it is do-able the extra PSNRR will be an asset.
mike
Bandwidth
When determining the bandwidth of both amps and preamps, I use the rule of 10: For the upper limit, I go 10X above and for the bottom end, I go 10X below. So, I shoot for -3dB at 2Hz and 200KHz. This gives a nice, flat (~ -0.1dB) response from 20-20K.
Following this rule, first mentioned by an electronics professor in college, I have always had great preamp and power amp performance and no HF oscillation or RF problems.
This is used, of course, in conjunction with good designs, good PCB layouts, sufficient open-loop gain, HF compensation, etc.
When determining the bandwidth of both amps and preamps, I use the rule of 10: For the upper limit, I go 10X above and for the bottom end, I go 10X below. So, I shoot for -3dB at 2Hz and 200KHz. This gives a nice, flat (~ -0.1dB) response from 20-20K.
Following this rule, first mentioned by an electronics professor in college, I have always had great preamp and power amp performance and no HF oscillation or RF problems.
This is used, of course, in conjunction with good designs, good PCB layouts, sufficient open-loop gain, HF compensation, etc.
i have an amp design in the works, with a GBW of about 2Mhz, open loop. one must also take certain phase shifts (due to propagation delay in the components) into account when feedback is applied to set the gain at a useable level (such as about 30 for a 100w amp), because some of this propagation delay is going to show up at the - input of your diff amp and make your amp susceptible to oscillation. this happened with my design, and it took a bit of capacitive modification of the frequency response curve (to a -3db rollover of 30khz) to keep the amp stable when clipping is reached. when you run an ac transfer sim of your design, if you have any peaking of your response curve, you've found an instability. it usually occurs at the rollover freq or higher on the down slope of the frequency response curve. your rollover should be smooth. because of junction capacitances (or interelectrode capacitance in tubes) it is possible in ALL amp designs.
in my design, the peak was at 1Mhz, and the bugger would oscillate at 1Mhz at clipping. after adding some capacitance in the right place, the amp settled down, and is ready for prototyping.
in my design, the peak was at 1Mhz, and the bugger would oscillate at 1Mhz at clipping. after adding some capacitance in the right place, the amp settled down, and is ready for prototyping.
i don't think there is such a thing
as a general optimum -3dB upper bandwidth
it depends on the actual circuitry of this amplifier
it also depends on chosen transistors
very much, I think it depends on how fast the output stage transistors are
Ft, as have got it, is the frequency, in MHz,
where a transistor does not provide any gain no longer
Most used power transistors have Ft like 3 or 4 MHz
there are some faster Bipolar with 6, 10 and even 40-50 MHz
But for 90% of our usual diy designs it would be <=6 Mhz
========================
There are 3 different levels, where we would add some capacitance
to tune our power amp = restrict upper frequency:
1. input filter
it is good to remove such noise and things that are not actual audio information,
BEFORE it enters amp.
Because once inside, they would make feedback create leftover from the correction job.
2. within and between stages
here we can adjust and slow done some fast transistors
so they do not feed frequencies into slower next stage
such signals can not be handled very good by slow devices
and will also only give extra correction work to global feedback
if they even can get through ...
To this category we count also gate stoppers for MOSFET
which reduces the speed of those power FET, a bit.
3. in the global feedback network itself
this sets the overall bandwidth of the amplifier
defines the band where any any gain will be performed
===========================================
There is no such general -3dB optimum, that would suit every, or even most, power amplifiers.
However for each amp + load setup, there is a maximal upper frequency, that is possible.
As many amplifiers use similar transistors ( monkey see - monkey do )
and the usual topology ( Op-Amp style )
we end up in somewhat same area, for -3dB overall bandwidth:
.... something like 50-200 kHz.
The General Rule is to try to trim and make the amplifier
as fast as possible in conditions at hand
and still do a performance we can be satisfied with.
A too fast amplifier will have more distortion,
than if it is tamed a bit.
There is no good reason to slow down an amplifier,
just for the sake of making it slow.
lineup
as a general optimum -3dB upper bandwidth
it depends on the actual circuitry of this amplifier
it also depends on chosen transistors
very much, I think it depends on how fast the output stage transistors are
Ft, as have got it, is the frequency, in MHz,
where a transistor does not provide any gain no longer
Most used power transistors have Ft like 3 or 4 MHz
there are some faster Bipolar with 6, 10 and even 40-50 MHz
But for 90% of our usual diy designs it would be <=6 Mhz
========================
There are 3 different levels, where we would add some capacitance
to tune our power amp = restrict upper frequency:
1. input filter
it is good to remove such noise and things that are not actual audio information,
BEFORE it enters amp.
Because once inside, they would make feedback create leftover from the correction job.
2. within and between stages
here we can adjust and slow done some fast transistors
so they do not feed frequencies into slower next stage
such signals can not be handled very good by slow devices
and will also only give extra correction work to global feedback
if they even can get through ...
To this category we count also gate stoppers for MOSFET
which reduces the speed of those power FET, a bit.
3. in the global feedback network itself
this sets the overall bandwidth of the amplifier
defines the band where any any gain will be performed
===========================================
There is no such general -3dB optimum, that would suit every, or even most, power amplifiers.
However for each amp + load setup, there is a maximal upper frequency, that is possible.
As many amplifiers use similar transistors ( monkey see - monkey do
)and the usual topology ( Op-Amp style )
we end up in somewhat same area, for -3dB overall bandwidth:
.... something like 50-200 kHz.
The General Rule is to try to trim and make the amplifier
as fast as possible in conditions at hand
and still do a performance we can be satisfied with.
A too fast amplifier will have more distortion,
than if it is tamed a bit.
There is no good reason to slow down an amplifier,
just for the sake of making it slow.
lineup
Thank for your answer lineup - and happy new year, I'm still 2006 here
your points are thorough and all make sense to me.
but my question is this
If we have an amplifier that can comfortably handle a very high bandwidth, what is the highest bandwidth above which we do not notice any benefit.
I just wonder if someone out there has done this stuff and come to some conclusions.
Anyway I guess I will have to build a few amps and find out for myself ...
your points are thorough and all make sense to me.
but my question is this
If we have an amplifier that can comfortably handle a very high bandwidth, what is the highest bandwidth above which we do not notice any benefit.
I just wonder if someone out there has done this stuff and come to some conclusions.
Anyway I guess I will have to build a few amps and find out for myself ...
What do you mean by "notice any benefit"?mikelm said:
Sonically or in measurements?
Theoretically what one wants to achieve with an amp is a perfectly flat response over the range 20Hz to 20kHz. Since perfection is not of this world, we are happy with some roll-off at both ends. If you want -0.1dB at 20kHz you need -3dB at 200kHz or so. If you are willing to compromise and accept -1dB at 20kHz you just need -3dB at 30kHz or so.
Can you hear the difference between the same amp, if you just change this single parameter, between -0.1dB at 20kHz and -1dB at 20kHz? I can't, and I doubt anybody can reliably ABX the difference in sound, even when playing pure tones; so the difference in sound would be rethorical.
In practical terms, the rule-of-thumb that DCPreamp wrote above is a good one, actually it's even high-end. It was quite impossible to achieve with BJT output stages just a few years ago but nowadays you can aim that high with FETs. But it's definitely not a requirement for or a guarantee of amplifier sound quality.
There is a very simple experiment you can build on a protoboard in less than an hour and check for yourself the benefits - or lack of - of very high bandwidth amps.
Just build yourself one of the CMoy variants with a modern op-amp and FET output buffer. These little amps have > 1 MHz closed-loop bandwidth usually. Use a good Sennheiser headphone and just add a switch for a -3dB low pass filter at 30kHz and at 200kHz. Just a warning: these little things tend to oscillate...
And HAPPY NEW YEAR!
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.