I just would like to add that in the case of the buffer, it is NOT the case that if the input is exactly at zero volts, that the output is also at zero volts. The term offset voltage refers to an (DC) error in the opamp itself. So, for the buffer, even if the input is grounded, there will still be a DC offset output, which in this case is the 'input offset voltage' from the data sheet. This input offset voltage is always amplified in case the opamp circuit has a gain.
If you have an opamp spec'ed at 'input offset = 1mV' in a gain-of-100' circuit the expected output offset is 100mV.
As for the minimum gain for the 5534 to be stable, Beppe, why not check it yourself in the data sheet? You say you have little experience; keeping asking questions without trying to find it out and to check a data sheet (even if you understand less than half) doesn't really give you any more knowledge. If you want to learn you need to do it yourself. There are a lot of basic texts to be found on the 'net from which you really learn!
Jan Didden
If you have an opamp spec'ed at 'input offset = 1mV' in a gain-of-100' circuit the expected output offset is 100mV.
As for the minimum gain for the 5534 to be stable, Beppe, why not check it yourself in the data sheet? You say you have little experience; keeping asking questions without trying to find it out and to check a data sheet (even if you understand less than half) doesn't really give you any more knowledge. If you want to learn you need to do it yourself. There are a lot of basic texts to be found on the 'net from which you really learn!
Jan Didden
janneman said:
Totally agree!
I remember, many years ago, coming to an enlightenment on how to read and interpret datasheets. This represented a step-up function in confidence in designing and analysing analog circuits.
For example, Op-amp offset voltage specified with Rs = 10K.
Why??
Because that is how it is TESTED!
How does one then extrapolate this to other source impedances?
By looking at the bias CURRENT figures!
Perhaps a trivial point, but the unaware will be condemned to clutching at straws in the dark.
And don't talk to me bout unity gain stability ....
Yes, dual supply operation is preferred for audio I think.
I agree about 5534, it is high performance and excellent value for money and also I agree it needs compensation for stability below gain of 3 I think. It is available as a dual, 5532, that is stable at unity gain, no disadvantage for audio, both channels left and right in one chip.
Yes I agree you will see excellent square waves thru capacitor coupled audio circuits. This is because there is a standard for audio interconnection that is understood so sometimes not explained. The standard is that outputs are low impedance, typically between 10 and 100 Ohms, and inputs are high impedance typically between 10k and 100k Ohms. Your 5k pot is no problem. With this arrangement, low out and high in, the capacitor is transparent so far as square waves are concerned within the audio band. If you really want to have no capacitors and have no DC on the pot then you can use 5534 and use the offset null adjustment to obtain no DC offset at the output. A miniature trim pot is connected across the IC offset null terminals as shown in the application note.
Another way is to use DC servo, but this is not so simple.
You can put an opamp after the volume pot, unity gain if you like, or with gain. When there is a small DC offset at the opamp output, then, if your power amp has an input capacitor, you are OK. If it is DC coupled it will amplify the offset and your speakers will not be happy .
When you use a capacitor at the input you must provide a connection on one of the op amp inputs to ground through a resistor otherwise it will have no source for the input bias current, without the resistor to ground the output will drift away from zero.
When you use a capacitor in the signal path the value is chosen so that there is minimum attentuation at the lowest frequency of interest, say 1dB at 10Hz for example. There is a program online here http://www.st-andrews.ac.uk/~www_pa/Scots_Guide/experiment/highpass/hpf.html
that calculates the performance of a high pass filter, which is the proper name for the DC blocking capacitor.
I looked at your IC list and did not recognise many probably because I am old and have not been designing for many years. I have read about the AD797 it is very high performance, probably better than is needed for audio, which is quite slow really. The 5534/5532 was designed for audio many years ago and is still good. Many opamp ICs are designed to work inside carefully controlled precision equipment where conditions are carefully controlled. In audio things are simple and there is not much protection for the device, it can be accidentally shorted out and we expect it to survive, a 5534 will do that because the designers knew the kind of crazy stuff people do in audio, which is connected to music art and show business but not so much connected to precision instruments .
Hope this helps.
I agree about 5534, it is high performance and excellent value for money and also I agree it needs compensation for stability below gain of 3 I think. It is available as a dual, 5532, that is stable at unity gain, no disadvantage for audio, both channels left and right in one chip.
Yes I agree you will see excellent square waves thru capacitor coupled audio circuits. This is because there is a standard for audio interconnection that is understood so sometimes not explained. The standard is that outputs are low impedance, typically between 10 and 100 Ohms, and inputs are high impedance typically between 10k and 100k Ohms. Your 5k pot is no problem. With this arrangement, low out and high in, the capacitor is transparent so far as square waves are concerned within the audio band. If you really want to have no capacitors and have no DC on the pot then you can use 5534 and use the offset null adjustment to obtain no DC offset at the output. A miniature trim pot is connected across the IC offset null terminals as shown in the application note.
Another way is to use DC servo, but this is not so simple.
You can put an opamp after the volume pot, unity gain if you like, or with gain. When there is a small DC offset at the opamp output, then, if your power amp has an input capacitor, you are OK. If it is DC coupled it will amplify the offset and your speakers will not be happy .
When you use a capacitor at the input you must provide a connection on one of the op amp inputs to ground through a resistor otherwise it will have no source for the input bias current, without the resistor to ground the output will drift away from zero.
When you use a capacitor in the signal path the value is chosen so that there is minimum attentuation at the lowest frequency of interest, say 1dB at 10Hz for example. There is a program online here http://www.st-andrews.ac.uk/~www_pa/Scots_Guide/experiment/highpass/hpf.html
that calculates the performance of a high pass filter, which is the proper name for the DC blocking capacitor.
I looked at your IC list and did not recognise many probably because I am old and have not been designing for many years. I have read about the AD797 it is very high performance, probably better than is needed for audio, which is quite slow really. The 5534/5532 was designed for audio many years ago and is still good. Many opamp ICs are designed to work inside carefully controlled precision equipment where conditions are carefully controlled. In audio things are simple and there is not much protection for the device, it can be accidentally shorted out and we expect it to survive, a 5534 will do that because the designers knew the kind of crazy stuff people do in audio, which is connected to music art and show business but not so much connected to precision instruments .
Hope this helps.
I should clarify what I said about offset null. This will get rid of the output offset allowing an output capacitor to be not used.
It does nothing about input offset. The spec for 5534 input DC offset says typical 0.5mV with 50 Ohm source, but may be much bigger up to 5mV. Your source, the pot output, varies between zero and 1.25kOhms. The offset will be bigger as the source impedance rises. To avoid scratching noise when the pot is turned you probably need less than 5mV. We always used blocking capacitors to be safe but you only need one or two ICs with low input offset, you can select a chip with low offset from five or ten pieces. Fit an IC socket to your PCB so you can switch ICs. You need a meter that will read millivolts DC. Wait a minute for the IC to warm up before you test the offet.
Also what I said about needing an input bias resistor is when an input blocking capacitor is used. When there is no capacitor the pot provides the current path for input bias, no need for a resistor.
Ted
It does nothing about input offset. The spec for 5534 input DC offset says typical 0.5mV with 50 Ohm source, but may be much bigger up to 5mV. Your source, the pot output, varies between zero and 1.25kOhms. The offset will be bigger as the source impedance rises. To avoid scratching noise when the pot is turned you probably need less than 5mV. We always used blocking capacitors to be safe but you only need one or two ICs with low input offset, you can select a chip with low offset from five or ten pieces. Fit an IC socket to your PCB so you can switch ICs. You need a meter that will read millivolts DC. Wait a minute for the IC to warm up before you test the offet.
Also what I said about needing an input bias resistor is when an input blocking capacitor is used. When there is no capacitor the pot provides the current path for input bias, no need for a resistor.
Ted
I agree with ACD that OPA134 would be a good choice because it is FET input, very low input current so low offset, it is unity gain stable, it can be used with offset adjuster pot to eliminate output DC offset, it was designed for audio, it will drive your screened cable to the power amp, it's cheap and fast, seems ideal.
Ted
Ted
I will humbly submit that you should also take a look at the LM4562 released late last year by National. It comes in a standard size dip as well as soic and is easy to use in unity gain configurations. It also has very low input currents, although it is a bipolar, and unity gain offsets in the region of 100uV or less in typical use. (This from experience) It has excellent AC performance and represents just about the current state of the art in audio op-amps, best of all it is actually very transparent. (I am not a fan of op-amps, but I can live with this one.)
I use it in my Zhaolu DAC where it performs very well and soon will be installing some in my Sony SCD-777ES sacd player.
I use it in my Zhaolu DAC where it performs very well and soon will be installing some in my Sony SCD-777ES sacd player.
Originally posted by janneman
I just would like to add that in the case of the buffer, it is NOT the case that if the input is exactly at zero volts, that the output is also at zero volts.
The term offset voltage refers to an (DC) error in the opamp itself.
So, for the buffer, even if the input is grounded, there will still be a DC offset output, which in this case is the 'input offset voltage' from the data sheet
This input offset voltage is always amplified in case the opamp circuit has a gain.
If you have an opamp spec'ed at 'input offset = 1mV' in a gain-of-100' circuit the expected output offset is 100mV
Thank you very much for the very helpful reply Mr Didden.
I never looked at this input offset voltage.
Another data to check then.
But now that I understand that caps can be used in the signal path without particular problems I am less reluctant to use them
One at the input and another one at the output and offset problems are solved.
As for the minimum gain for the 5534 to be stable, Beppe, why not check it yourself in the data sheet?
You say you have little experience; keeping asking questions without trying to find it out and to check a data sheet (even if you understand less than half) doesn't really give you any more knowledge.
If you want to learn you need to do it yourself. There are a lot of basic texts to be found on the 'net from which you really learn!
Jan Didden
I promise you I will check for myself.
But if I need a buffer I think after all that best solution would be to selected a "native" buffer and not an opamp wired as unity gain buffer.
Thanks again a lot.
Kind regards,
beppe
I just would like to add that in the case of the buffer, it is NOT the case that if the input is exactly at zero volts, that the output is also at zero volts.
The term offset voltage refers to an (DC) error in the opamp itself.
So, for the buffer, even if the input is grounded, there will still be a DC offset output, which in this case is the 'input offset voltage' from the data sheet
This input offset voltage is always amplified in case the opamp circuit has a gain.
If you have an opamp spec'ed at 'input offset = 1mV' in a gain-of-100' circuit the expected output offset is 100mV
Thank you very much for the very helpful reply Mr Didden.
I never looked at this input offset voltage.
Another data to check then.
But now that I understand that caps can be used in the signal path without particular problems I am less reluctant to use them
One at the input and another one at the output and offset problems are solved.
As for the minimum gain for the 5534 to be stable, Beppe, why not check it yourself in the data sheet?
You say you have little experience; keeping asking questions without trying to find it out and to check a data sheet (even if you understand less than half) doesn't really give you any more knowledge.
If you want to learn you need to do it yourself. There are a lot of basic texts to be found on the 'net from which you really learn!
Jan Didden
I promise you I will check for myself.
But if I need a buffer I think after all that best solution would be to selected a "native" buffer and not an opamp wired as unity gain buffer.
Thanks again a lot.
Kind regards,
beppe
beppe61 said:
Just my opinion, but...
Most readily obtainable off the shelf unity gain buffer ics are intended to be used inside an op-amp feedback loop to boost output current, and are not very good performers on their own. There is the option of building a discrete unity gain buffer like the burson or buying something similar. Check other threads here - apparently someone on eBay is making a discrete unity gain buffer that is supposed to be reasonably good.
Op-amps make fine unity gain buffers and if you do not have the design chops to do your own this is a reasonable path to follow for good, predictable performance. Of course I would use a tube based white cathode follower or similar...
In fact your statement is pretty much the reverse of reality for an op-amp. AC and DC performance are always best at unity gain for a unity gain stable op-amp and degrade quite predictably as the device is called upon to provide gain. (Exception Norton/current amplifiers.)
tedr said:Yes, dual supply operation is preferred for audio I think
Thank you Sir.
First point fixed: dual supply.
I agree about 5534, it is high performance and excellent value for money and also I agree it needs compensation for stability below gain of 3 I think.
It is available as a dual, 5532, that is stable at unity gain, no disadvantage for audio, both channels left and right in one chip
I will check for this.
Thank you for the info.
By the way I will do more search on "native" buffer.
They seem easier to use for me.
No problem of gain selection. No feedback network to build.
Is it so bad this BUF04 I have at hand ?
With a cap before and one after.
I never heard of it by the way in a high end design.
Yes I agree you will see excellent square waves thru capacitor coupled audio circuits.
This is because there is a standard for audio interconnection that is understood so sometimes not explained.
The standard is that outputs are low impedance, typically between 10 and 100 Ohms, and inputs are high impedance typically between 10k and 100k Ohms.
Your 5k pot is no problem. With this arrangement, low out and high in, the capacitor is transparent so far as square waves are concerned within the audio band. If you really want to have no capacitors and have no DC on the pot then you can use 5534 and use the offset null adjustment to obtain no DC offset at the output.
A miniature trim pot is connected across the IC offset null terminals as shown in the application note.
Another way is to use DC servo, but this is not so simple
The two solutions appear difficult to me. Too difficult.
I will start with a real buffer I think.
You can put an opamp after the volume pot, unity gain if you like, or with gain.
When there is a small DC offset at the opamp output, then, if your power amp has an input capacitor, you are OK.
If it is DC coupled it will amplify the offset and your speakers will not be happy
my amp has an input cap so it is AC coupled.
I could do without the output cap after the buffer then.
When you use a capacitor at the input you must provide a connection on one of the op amp inputs to ground through a resistor otherwise it will have no source for the input bias current, without the resistor to ground the output will drift away from zero.
When you use a capacitor in the signal path the value is chosen so that there is minimum attentuation at the lowest frequency of interest, say 1dB at 10Hz for example. There is a program online here http://www.st-andrews.ac.uk/~www_pa/Scots_Guide/experiment/highpass/hpf.html
that calculates the performance of a high pass filter, which is the proper name for the DC blocking capacitor
I think my only possibility is to copy a working schematic
It is much more complicated than what I thought
I looked at your IC list and did not recognise many probably because I am old and have not been designing for many years.
I have read about the AD797 it is very high performance, probably better than is needed for audio, which is quite slow really.
The 5534/5532 was designed for audio many years ago and is still good. Many opamp ICs are designed to work inside carefully controlled precision equipment where conditions are carefully controlled.
In audio things are simple and there is not much protection for the device, it can be accidentally shorted out and we expect it to survive, a 5534 will do that because the designers knew the kind of crazy stuff people do in audio, which is connected to music art and show business but not so much connected to precision instruments .
Hope this helps.
It helps me a lot indeed and I thank you sincerely for all the kind advice.
I will stick with a proven schematic.
The issue is very complex indeed for a beginner.
Kind regards,
beppe
Originally posted by kevinkr
I will humbly submit that you should also take a look at the LM4562 released late last year by National.
It comes in a standard size dip as well as soic and is easy to use in unity gain configurations.
It also has very low input currents, although it is a bipolar, and unity gain offsets in the region of 100uV or less in typical use. (This from experience) It has excellent AC performance and represents just about the current state of the art in audio op-amps, best of all it is actually very transparent
Thanks a lot for your kind suggestion.
I will look at the datasheet for a buffer configuration.
(I am not a fan of op-amps, ...
May I ask you why ?
Kind regards,
bg
I will humbly submit that you should also take a look at the LM4562 released late last year by National.
It comes in a standard size dip as well as soic and is easy to use in unity gain configurations.
It also has very low input currents, although it is a bipolar, and unity gain offsets in the region of 100uV or less in typical use. (This from experience) It has excellent AC performance and represents just about the current state of the art in audio op-amps, best of all it is actually very transparent
Thanks a lot for your kind suggestion.
I will look at the datasheet for a buffer configuration.
(I am not a fan of op-amps, ...
May I ask you why ?
Kind regards,
bg
Originally posted by kevinkr
Just my opinion, but...
Most readily obtainable off the shelf unity gain buffer ics are intended to be used inside an op-amp feedback loop to boost output current, and are not very good performers on their own
May I ask from which point of view ? distortion may be ?
noise ? slew-rate ?
There is the option of building a discrete unity gain buffer like the burson or buying something similar.
Check other threads here - apparently someone on eBay is making a discrete unity gain buffer that is supposed to be reasonably good
thanks for the suggestion.
I will check these offerings.
Op-amps make fine unity gain buffers and if you do not have the design chops to do your own this is a reasonable path to follow for good, predictable performance.
Of course I would use a tube based white cathode follower or similar...
I guess you mean that tubes are better from a sonic point of view
In fact your statement is pretty much the reverse of reality for an op-amp.
AC and DC performance are always best at unity gain for a unity gain stable op-amp and degrade quite predictably as the device is called upon to provide gain. (Exception Norton/current amplifiers.)
this is very interesting as I do not need gain at all.
Just the power to drive properly my power amp.
Thanks again and kind regards,
bg
Just my opinion, but...
Most readily obtainable off the shelf unity gain buffer ics are intended to be used inside an op-amp feedback loop to boost output current, and are not very good performers on their own
May I ask from which point of view ? distortion may be ?
noise ? slew-rate ?
There is the option of building a discrete unity gain buffer like the burson or buying something similar.
Check other threads here - apparently someone on eBay is making a discrete unity gain buffer that is supposed to be reasonably good
thanks for the suggestion.
I will check these offerings.
Op-amps make fine unity gain buffers and if you do not have the design chops to do your own this is a reasonable path to follow for good, predictable performance.
Of course I would use a tube based white cathode follower or similar...
I guess you mean that tubes are better from a sonic point of view
In fact your statement is pretty much the reverse of reality for an op-amp.
AC and DC performance are always best at unity gain for a unity gain stable op-amp and degrade quite predictably as the device is called upon to provide gain. (Exception Norton/current amplifiers.)
this is very interesting as I do not need gain at all.
Just the power to drive properly my power amp.
Thanks again and kind regards,
bg
Hi Beppe,
I am a tube man... In a nutshell..
I have found that there are quite a few op-amps that sonically outperform the ancient 5532/5534 (~30yr old) which were really the first op-amps that were any good for audio. (I used them for a long time) I currently prefer the LM4562, but the AD8610 and 8620 are also very good as are a number of devices from BB like the 2132.
I am a tube man... In a nutshell..
I have found that there are quite a few op-amps that sonically outperform the ancient 5532/5534 (~30yr old) which were really the first op-amps that were any good for audio. (I used them for a long time) I currently prefer the LM4562, but the AD8610 and 8620 are also very good as are a number of devices from BB like the 2132.
beppe61 said:Originally posted by kevinkr
Just my opinion, but...
Most readily obtainable off the shelf unity gain buffer ics are intended to be used inside an op-amp feedback loop to boost output current, and are not very good performers on their own
May I ask from which point of view ? distortion may be ?
noise ? slew-rate ?
There is the option of building a discrete unity gain buffer like the burson or buying something similar.
Check other threads here - apparently someone on eBay is making a discrete unity gain buffer that is supposed to be reasonably good
thanks for the suggestion.
I will check these offerings.
Op-amps make fine unity gain buffers and if you do not have the design chops to do your own this is a reasonable path to follow for good, predictable performance.
Of course I would use a tube based white cathode follower or similar...
I guess you mean that tubes are better from a sonic point of view
In fact your statement is pretty much the reverse of reality for an op-amp.
AC and DC performance are always best at unity gain for a unity gain stable op-amp and degrade quite predictably as the device is called upon to provide gain. (Exception Norton/current amplifiers.)
this is very interesting as I do not need gain at all.
Just the power to drive properly my power amp.
Thanks again and kind regards,
bg
Primarily distortion, but often dc offset as well, (frequently temperature dependent as well) and imo when possible the best capacitor really is no capacitor..
beppe61 said:
Beppe,
As I explained, the opamp or buffer itself generates the offset. So even with an input coupling cap, you will have offset. Why not leave out the input cap and just use one at the output?
BTW, if you use an input cap you still have to provide a resistor from input to ground for the input device base current or gate voltage.
One more thing. Please don't call me 'Sir' or something like that. Makes me feel uneasy. My name is Jan.
Cheers,
Jan Didden
Originally posted by kevinkr
Hi Beppe,
I am a tube man... In a nutshell..
Ok Kevin. I understand.
I have found that there are quite a few op-amps that sonically outperform the ancient 5532/5534 (~30yr old) which were really the first op-amps that were any good for audio. (I used them for a long time)
I currently prefer the LM4562, but the AD8610 and 8620 are also very good as are a number of devices from BB like the 2132.
I will keep this LM4562 in my mind.
I will look for a suitable schematic.
Primarily distortion, but often dc offset as well, (frequently temperature dependent as well) and imo when possible the best capacitor really is no capacitor.. __________________
Kevin
Yes but with tubes you must use caps.
If the sound you get is fine, caps must be also fine.
Am I wrong ?
Thanks and regards,
beppe
Hi Beppe,
I am a tube man... In a nutshell..
Ok Kevin. I understand.
I have found that there are quite a few op-amps that sonically outperform the ancient 5532/5534 (~30yr old) which were really the first op-amps that were any good for audio. (I used them for a long time)
I currently prefer the LM4562, but the AD8610 and 8620 are also very good as are a number of devices from BB like the 2132.
I will keep this LM4562 in my mind.
I will look for a suitable schematic.
Primarily distortion, but often dc offset as well, (frequently temperature dependent as well) and imo when possible the best capacitor really is no capacitor.. __________________
Kevin
Yes but with tubes you must use caps.
If the sound you get is fine, caps must be also fine.
Am I wrong ?
Thanks and regards,
beppe
Originally posted by janneman
Beppe,
As I explained, the opamp or buffer itself generates the offset. So even with an input coupling cap, you will have offset.
Why not leave out the input cap and just use one at the output?
BTW, if you use an input cap you still have to provide a resistor from input to ground for the input device base current or gate voltage.
I think the only way viable for me is
1) select a good opamp to use as a unity gain buffer
2) find a working schematic
I was exceedingly optimistic
I thought that working with opamps were much easier
Instead they are more difficult than discretes.
One more thing. Please don't call me 'Sir' or something like that. Makes me feel uneasy. My name is Jan.
Cheers,
Jan Didden
Thank you very much Jan.
Bye
beppe
P.S. i am beginning to hate these little black chips called "opamps"
Beppe,
As I explained, the opamp or buffer itself generates the offset. So even with an input coupling cap, you will have offset.
Why not leave out the input cap and just use one at the output?
BTW, if you use an input cap you still have to provide a resistor from input to ground for the input device base current or gate voltage.
I think the only way viable for me is
1) select a good opamp to use as a unity gain buffer
2) find a working schematic
I was exceedingly optimistic
I thought that working with opamps were much easier
Instead they are more difficult than discretes.
One more thing. Please don't call me 'Sir' or something like that. Makes me feel uneasy. My name is Jan.
Cheers,
Jan Didden
Thank you very much Jan.
Bye
beppe
P.S. i am beginning to hate these little black chips called "opamps"
Hi Beppe,
The caps I use in tube circuitry are generally expensive, small value films. I now use teflon foil (tin) types almost exclusively, and you would cringe at what they cost. I also tend more recently to like transformer coupling despite all of the technical arguments against that approach.
Large value electrolytics tend to have lots of issues with their construction which can badly color their sound in some applications. These problems get worse as values go up, and relate to measurable parameters like ESR, ESL, DA, and DF. In some applications there can also be linearity issues that relate to the way the dielectric barrier behaves with ac across it, in particular if it becomes reverse biased during part of a cycle where there is little dc across it. (I've also had some very odd ringing problems in large value electrolytics in cathode bypass applications with high transconductance triodes. I avoid using cathode bias in these applications now for that reason.) I've found in most instances that paying for better el caps like some of the more modestly priced Black Gates (supplies now drying up) is audibly worthwhile.
Building a good performing op-amp based circuit is not nearly as difficult as doing the same thing with discretes, although much the same attention to detail will pay off.
You shouldn't be scared of op-amps or for that matter anything else operating at reasonably low voltages. Consider it an opportunity to learn and master your craft. You just need the right tools, and a few pointers now and then to keep you headed in the right direction. You do ask the right sort of questions so this is good.
You should have a scope, a signal generator, a good dvm, and the ability to solder. Everything else you can pick up along the way.. Most of all have fun, it's a hobby afterall..
The caps I use in tube circuitry are generally expensive, small value films. I now use teflon foil (tin) types almost exclusively, and you would cringe at what they cost. I also tend more recently to like transformer coupling despite all of the technical arguments against that approach.
Large value electrolytics tend to have lots of issues with their construction which can badly color their sound in some applications. These problems get worse as values go up, and relate to measurable parameters like ESR, ESL, DA, and DF. In some applications there can also be linearity issues that relate to the way the dielectric barrier behaves with ac across it, in particular if it becomes reverse biased during part of a cycle where there is little dc across it. (I've also had some very odd ringing problems in large value electrolytics in cathode bypass applications with high transconductance triodes. I avoid using cathode bias in these applications now for that reason.) I've found in most instances that paying for better el caps like some of the more modestly priced Black Gates (supplies now drying up) is audibly worthwhile.
Building a good performing op-amp based circuit is not nearly as difficult as doing the same thing with discretes, although much the same attention to detail will pay off.
You shouldn't be scared of op-amps or for that matter anything else operating at reasonably low voltages. Consider it an opportunity to learn and master your craft. You just need the right tools, and a few pointers now and then to keep you headed in the right direction. You do ask the right sort of questions so this is good.
You should have a scope, a signal generator, a good dvm, and the ability to solder. Everything else you can pick up along the way.. Most of all have fun, it's a hobby afterall..
kevinkr said:Hi Beppe,
The caps I use in tube circuitry are generally expensive, small value films.
I now use teflon foil (tin) types almost exclusively, and you would cringe at what they cost.
I also tend more recently to like transformer coupling despite all of the technical arguments against that approach.
Large value electrolytics tend to have lots of issues with their construction which can badly color their sound in some applications. These problems get worse as values go up, and relate to measurable parameters like ESR, ESL, DA, and DF.
In some applications there can also be linearity issues that relate to the way the dielectric barrier behaves with ac across it, in particular if it becomes reverse biased during part of a cycle where there is little dc across it. (I've also had some very odd ringing problems in large value electrolytics in cathode bypass applications with high transconductance triodes. I avoid using cathode bias in these applications now for that reason.)
I've found in most instances that paying for better el caps like some of the more modestly priced Black Gates (supplies now drying up) is audibly worthwhile
Good morning Kevin ! (here is 9.48 a.m.)
Thank you so much for all your kind and extremely helpful technical support.
First of all I must say that I have changed my opinion about caps in the signal path.
I am sure that they can be "audible" but when rightly selected they can be very good indead.
I understand now that lowering uF is better, as it is possible with tubes.
I read here and there that the simple replacement of an electrolytic output cap in let's say a cd player with a nice Black Gate Nx is very noticeable, for the better I mean.
For the transformers, there are designers more technical oriented (John Curl, Nelson Pass, ecc.) who do not use them.
Other high end tube designers swear for them to be good.
I really don't know.
Building a good performing op-amp based circuit is not nearly as difficult as doing the same thing with discretes, although much the same attention to detail will pay off
Kevin, after all these discussion I am not at all convinced that as you say "Building a good performing op-amp based circuit is not nearly as difficult as doing the same thing with discretes".
Some time ago I posted the following schematic useable as a line stage:
http://www.diyaudio.com/forums/showthread.php?threadid=51141
I did a prototype and it really works.
It is a "rib" of an mid '70 hi-fi preamp, well received by the audio press here in Italy. the Radford ZD22.
As you can see it is very simple.
On the basis of what I am learning now I am very hesitant to go on with opamps, unless I will find a proven schematic (that I will search of course).
Moreover there are so many designers who refuse to use them for audio purpose.
Nevertheless a lot of top dacs for instance use opamps in their output stages, like Benchmark, Theta and Wadia.
The requirements for a dac output stage and a line preamp are much the same, I think.
So if they works for them it could work for me.
But the schematic is mandatory.
Just a little cap can make stable an otherwise ringing opamp.
Little parts can have a big effect.
You shouldn't be scared of op-amps or for that matter anything else operating at reasonably low voltages.
Consider it an opportunity to learn and master your craft.
You just need the right tools, and a few pointers now and then to keep you headed in the right direction.
You do ask the right sort of questions so this is good.
You should have a scope, a signal generator, a good dvm, and the ability to solder
Actually this is my long term project.
But now I have problems with my job and I am about to migrate to Milan.
But what I find strange, really strange, is an opamp that is classified as buffer (LM4562) and in its datasheet I cannot find a tested schematic for a buffer, if I have seen well of course.
Is something like a product to make pizza without instruction.
By the way I will look for application notes.
I understand, for instance, that some power amps were built instead following schematics shown in the datasheets of transistors, and with very good results ( I think the Harman Kardon Citation 12 for example).
The schematic is the very first thing, after all.
Everything else you can pick up along the way..
Most of all have fun, it's a hobby afterall..
I have to say that if I had the time and the right tools I would think the same.
Now I am living in little critical conditions.
Thank you very much Kevin.
Kind regards,
beppe
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.