Unfortunately, I report it as it is, rather than what is convenient. It would be like saying that fast food should be better than old fashioned home cooking, because it is more 'hi tech'.
Tubes are as much competition to me as any solid state design. This is not because of the fact that they have been available about 100 years, but because of their inherent characteristics.
Tubes are as much competition to me as any solid state design. This is not because of the fact that they have been available about 100 years, but because of their inherent characteristics.
Hi John,
Sorry to hark back a number of posts, but here goes ....
As a technician who performed a great deal of warranty work, I was able to see the best and worst circuitry first hand. You get to know what is unstable and what isn't. In my experience, making unstable circuits, stable, has always improved the sound quality. This is both subjectively and through hard measurements. Yes, some are pretty bad. The very worst examples of design I have ever seen come from "engineers' or "designers" who "design by ear". It's gotten to a point where I hate that phrase.
I guess what I am pointing out is that it is difficult for a master to explain step by step how he goes about his work. You tend to forget about the many small steps you whiz through without even thinking of it. You do design following your own design rules, and you make an effort to listen to it too. The listening sticks out for you.
-Chris
Sorry to hark back a number of posts, but here goes ....
Nor do I as I can not be objective. However, just because an engineering practice is called 'good engineering' doesn't mean it is so. If we ignore things like this, we are in agreement. You know how to assemble circuits so they do not spontaneously ignite. This you do on an unconscious level, it is part of your process. That was my point entirely.
As a technician who performed a great deal of warranty work, I was able to see the best and worst circuitry first hand. You get to know what is unstable and what isn't. In my experience, making unstable circuits, stable, has always improved the sound quality. This is both subjectively and through hard measurements. Yes, some are pretty bad. The very worst examples of design I have ever seen come from "engineers' or "designers" who "design by ear". It's gotten to a point where I hate that phrase.
I guess what I am pointing out is that it is difficult for a master to explain step by step how he goes about his work. You tend to forget about the many small steps you whiz through without even thinking of it. You do design following your own design rules, and you make an effort to listen to it too. The listening sticks out for you.
-Chris
Hi David,
Hi John,
There are some servos that do not mess up the audio, but any that throw the diff. pair out of balance will create havoc with the signal. The same holds true for the old DC offset control. I'd much rather see 30 mV DC offset at the output than throw the diff. pair out of balance to achieve less than 5 mV at the output terminals. You can vary the current in the predriver stages to achieve the same DC offset correction as messing with the diff. pair balance. If you have good matching and tracking, you can do this with a trimmer and skip servo circuits all together.
Sometimes I think that the only reason servo circuits came to be originally was to eliminate an adjustment point and possible warranty claims. Since then, I feel servos are used to cover up poor design work as well. You should be able to remove a servo in a design and have a stable DC offset of some value. If not, I think the circuit needs further work.
-Chris
Transistor matching is very important! Especially the voltage amp stages, also extending from predriver to output stages. The less the differential amp has to do, the better job of signal correction it can do. Also, the better circuit balance you have, the less drift in DC offset with temperature you will have.
Hi John,
There are some servos that do not mess up the audio, but any that throw the diff. pair out of balance will create havoc with the signal. The same holds true for the old DC offset control. I'd much rather see 30 mV DC offset at the output than throw the diff. pair out of balance to achieve less than 5 mV at the output terminals. You can vary the current in the predriver stages to achieve the same DC offset correction as messing with the diff. pair balance. If you have good matching and tracking, you can do this with a trimmer and skip servo circuits all together.
Sometimes I think that the only reason servo circuits came to be originally was to eliminate an adjustment point and possible warranty claims. Since then, I feel servos are used to cover up poor design work as well. You should be able to remove a servo in a design and have a stable DC offset of some value. If not, I think the circuit needs further work.
-Chris
anatech said:
Generally I think dc servos are a good thing, when designed and applied properly. Suppose you start with an amplifier that employs a 100 uF non-polar electrolytic in the feedback return path to achieve a 2 Hz 3 dB point. If you substitute a first-order servo that also causes the amplifier to have a 2 Hz 3 dB point, you will not have compromised the low-frequency or transient response. I like dc servos that employ a dual op amp and only require a single integrating capacitor, perhaps on the order of 1 uF. It is quite economical to employ a high-quality 1 uF capacitor. Dual op amps that are of high quality are also relatively inexpensive.
The servo does indeed introduce a high-pass filter function into the audio path, but so did the simple electrolytic in the feedback return leg.
You have now removed an evil 100 uF electrolytic that would have been bad for the sound even if it was bypassed by a smaller film capacitor. On paper the complexity is higher, but cost and space for the same quality is lower. The output of the servo can go to the amplifier's inverting input node through a fairly high-value resistor, as it needs to only inject enough correction current to overcome the maximum anticipated input-referred offset error. The large resistor tends to reduce the ability of the servo and its op amp to impact sonics via noise or distortion in the servo.
In a proper design, there is no need for the operation of the servo to upset the balance of the input differential pair. Indeed, in designs where the input pair is loaded by a current mirror, the current mirror usually enforces the balance condition.
Of course, a servo is more complex than a single non-polar electrolytic, so there will always be opportunities for less experienced designers to screw things up with one, I suppose.
Finally, there are often more coupling capacitors in a system signal path than necessary. One often may have one at the output of the preamplifier, and yet another one at the input of the power amplifier. In principle, if the power amplifier has a dc servo, the input coupling capacitor of the amplifier can be bypassed, knowing that the output of the preamp is probably at or very close to zero anyway. Any small dc offset present at the input of the amplifier will just be handled by the dc servo. In this case, an overall improvement in LF transient response will have been had by the use of a servo.
Cheers,
Bob
Very well said, Bob. I find servos a necessary evil, that I try to keep out of the way of the audio signal as much as possible. I did not always feel this way, but now, IF I want a controlled low frequency rolloff above 5 Hz or so, I will use a quality series cap. However, the series cap that I would choose, would be large and expensive, and not cheap or convenient. I don't like the price of quality polystyrene or teflon caps either, but they both work and measure better than virtually any alternative.
However, IF I can set my servo response at 1Hz or significantly below, then servos are not much of a problem.
What Bob is referring to, is also what I have gravitated to, over the decades. I once thought that the 'non-inverting' servo with 2 identical caps was OK, but now I prefer the dual op amp, 1 cap 'inverting-inverting' servo, as the price of dual IC's is lower, and the combined offsets is much better than in the past.
At the AES, Dimitri and I saw lots of dipped TANTALUM caps used for all kinds of things! UGH!
However, IF I can set my servo response at 1Hz or significantly below, then servos are not much of a problem.
What Bob is referring to, is also what I have gravitated to, over the decades. I once thought that the 'non-inverting' servo with 2 identical caps was OK, but now I prefer the dual op amp, 1 cap 'inverting-inverting' servo, as the price of dual IC's is lower, and the combined offsets is much better than in the past.
At the AES, Dimitri and I saw lots of dipped TANTALUM caps used for all kinds of things! UGH!
Hi, Mr. Curl,
I've read the paper about capacitor distoriton. Tantalum (and ceramic) generate bad distortion (when used as audio coupling), but they can work in higher frequencies than elko or plastic capacitor. I see this as benefit of tantalum (or ceramic), and often it is a needed property that cannot be replace by elko or plastic capacitor, for example at decoupling position (to decoupling Mhz). Today's transistor and opamp are getting faster and faster, and capacitor type that can match their speed are ceramic and tantalum.
So maybe ceramic and tantalum at certain positions are necessary evil.
But this makes a new dillema. If we trace current loop from +/-rails to opamp to output, then it will be aparent that decouling caps has to be the same quality as the output/signal decoupling cap, because they are carrying the same current.
I've read the paper about capacitor distoriton. Tantalum (and ceramic) generate bad distortion (when used as audio coupling), but they can work in higher frequencies than elko or plastic capacitor. I see this as benefit of tantalum (or ceramic), and often it is a needed property that cannot be replace by elko or plastic capacitor, for example at decoupling position (to decoupling Mhz). Today's transistor and opamp are getting faster and faster, and capacitor type that can match their speed are ceramic and tantalum.
So maybe ceramic and tantalum at certain positions are necessary evil.
But this makes a new dillema. If we trace current loop from +/-rails to opamp to output, then it will be aparent that decouling caps has to be the same quality as the output/signal decoupling cap, because they are carrying the same current.
Hi Bob,
Most servos operate by applying a DC bias to one to the gates/bases. This upsets the DC balance of the diff. pair, no?
If you an design an amplifier without the feedback capacitor to ground I think you are further ahead. This causes gain at DC, but you should be able to buck that out with the servo. A trim pot also works fine as long as your circuit stays in balance with changing temperature.
If I'm wrong here, can someone explain to me how this works? I'm serious about this.
My view is that those coupling capacitors are really required. Some amplifiers do not even have DC offset detection / protection (think Bryston 4B for one). The chances of equipment being changed out is too high to argue that a specific situation is safe. It may not be safe in 5 years.
Now, if we had to make a choice as to which capacitor to keep, I vote for the cap on the input of an amplifier. That way you have a DC block for sure,
-Chris
Most servos operate by applying a DC bias to one to the gates/bases. This upsets the DC balance of the diff. pair, no?
If you an design an amplifier without the feedback capacitor to ground I think you are further ahead. This causes gain at DC, but you should be able to buck that out with the servo. A trim pot also works fine as long as your circuit stays in balance with changing temperature.
A "proper design" is rare from what I have seen.
I don't think that is true. To correct for a DC offset, the diff. pair will shift the static currents so that one conducts more than the other. This would shift the operating point to one side. Most DC servos are designed to amplify the DC error (and filter out audio), and apply that to one side of the pair. This does the same thing as the dif. pair would do without the servo, but to a greater degree.
If I'm wrong here, can someone explain to me how this works? I'm serious about this.
I have to disagree with you on this Bob. When a product like an amp or preamp is manufactured, there is no guarantee that the other piece of equipment is safe from DC offsets. Krell did this. Explain to the customer why a DC offset in teh preamp output caused a big bang and force the amp into protection. The servo op amp in the preamp had high DC offset all by it's little self. The muting transistors can also fail DC, outputting on heck of an offset as well. Those 2SC2878 transistors often had about 15 VDC reverse E-B bias. Even though these were designed for that, a relatively high failure rate exists.
My view is that those coupling capacitors are really required. Some amplifiers do not even have DC offset detection / protection (think Bryston 4B for one). The chances of equipment being changed out is too high to argue that a specific situation is safe. It may not be safe in 5 years.
Now, if we had to make a choice as to which capacitor to keep, I vote for the cap on the input of an amplifier. That way you have a DC block for sure,
-Chris
Hi David,
Hi John,
-Chris
True, but anything way up there is noise. You would want to filter those things out at the input. In fact, most amplifiers us a ceramic or mica cap to ground across the input.
Hi John,
Another Bryston 4B? In my view, omitting the power switch is a poor decision. I disagree with this very strongly. Do you at least have a standby mode where the bias is shut off and the voltage amp is 0 mV for sure?
-Chris
You have now removed an evil 100 uF electrolytic
that would have been bad for the sound
even if it was bypassed by a smaller film capacitor.
I would rather replace one 100 uF electrolytic, like for example Black Gate electrolytic capacitors,
with one 10 uF or 4.7 uF Polypropylene Capacitor.
Even if no longer be able to hear those lovely musics below 30 Hertz 😕
this is true hi-fi quality.
Adding
op-amp + supply current + filter film capacitors + resistors and a servo feedback extra loop rail
vs.
one good Polyproylene.
Price is no issue compared to chassis, transformer, potentiometer etc.
Like Nelson Pass put it in another topic, to explain why he does not like the use of servos:
Servos are to replace one capacitor with another filter capacitor, effectively within the signal path.
that would have been bad for the sound
even if it was bypassed by a smaller film capacitor.
I would rather replace one 100 uF electrolytic, like for example Black Gate electrolytic capacitors,
with one 10 uF or 4.7 uF Polypropylene Capacitor.
Even if no longer be able to hear those lovely musics below 30 Hertz 😕
this is true hi-fi quality.
Adding
op-amp + supply current + filter film capacitors + resistors and a servo feedback extra loop rail vs.
one good Polyproylene.
Price is no issue compared to chassis, transformer, potentiometer etc.
Like Nelson Pass put it in another topic, to explain why he does not like the use of servos:
Servos are to replace one capacitor with another filter capacitor, effectively within the signal path.
Lumanauw, it might be necessary to use high quality glass or ceramic caps, but that is not what is used today for bypassing. Cheap, high DA, high distortion caps are used because they are cheap. However, we may have to borrow from microwave technology to get quality caps in future, and speeds increase. The caps that I was referring to have been around for more than 40 years and I was rather surprised that they were still used in modern design.
john curl said:
Definitely! 😀
I have no preamp ON-OFF switch, and the preamp is turned on as soon as the main switch for 6 mains outllet adapter is on. Then, when power amp is turned on, the preamp output has already settled 😀
My colleague is solving sequential turn-on circuit for components of audio chain.
here's one of his thoughts.lineup said:
I disagree.
High Fidelity requires accurate reproduction of the whole sound.
Not rolled off from ~300Hz and 3dB down @ 30Hz.
For many years I used coupling caps. In the 1950's the caps I listened through were mostly made of wax paper. They often leaked over time and heat. They had a smooth sound, probably due to high DA, but they did have low harmonic distortion at all frequencies.
In 1963, I got my first Dyna amp and preamp. It really was better, and I think the caps might have been Mylar, but I am not sure. It was the first time that I heard real high fidelity, especially with stereo headphones.
When I worked at Ampex Audio Division, in 1968, we used tantalum coupling caps throughout with DC bias on them. They seemed OK at the time. Mylar was just too small in practical values for a good coupling cap, but we did use them for filters and EQ.
In the early 70's we used tantalum coupling caps of high value in the feedback loops of power amps and preamps. We also sometimes used them as input and output coupling caps, where necessary. This seemed practical at the time, but not necessarily audio perfection. We kept our tube stuff for the best quality sound, in many cases. Only when we wanted extra power or some special function, did we use solid state by preference.
In 1973, when I started working with Mark Levinson, he had converted to polycarbonate caps of high quality, and he convinced me that they were really better than tantalum coupling caps, and I then learned about ceramic cap distortion while visiting Tektronix in 1974, so we switched to mica or polystyrene for small caps, polycarbonate up to 10uf, but we still used tantalum caps for higher values up to 1000uF. This is how we made the JC-2 preamp in 1974.
By 1978, Richard Marsh started talking about dielectric absorption in audio. I knew a lot about DA, because a friend of mine, and former classmate, did his masters thesis on this subject and he sent me the equations and technical articles in 1974, but I could not see, at the time, how this could be very important to audio. At the same time, in 1977, I started measuring tantalum caps and their low frequency distortion. It was VERY bad, and it was difficult to get around without servos. So in 1980, I decided to servo my audio circuits, if I could, in order to remove the coupling and feedback caps entirely. This became the JC-80 preamp. Until now, most of my cap discoveries are based on what we did with the JC-80. That is: Direct coupled, servo offset controlled, and polystyrene bypass and EQ caps. I have never regretted designing with this approach. I have removed ceramics and tantalum caps completely, using Mylar only for servo caps.
I am writing this long history to show the actual evolution of my experience with caps. I have no ax to grind, I don't make or sell caps, and I don't just jump on the 'latest' cap design, either. I just use what works.
In 1963, I got my first Dyna amp and preamp. It really was better, and I think the caps might have been Mylar, but I am not sure. It was the first time that I heard real high fidelity, especially with stereo headphones.
When I worked at Ampex Audio Division, in 1968, we used tantalum coupling caps throughout with DC bias on them. They seemed OK at the time. Mylar was just too small in practical values for a good coupling cap, but we did use them for filters and EQ.
In the early 70's we used tantalum coupling caps of high value in the feedback loops of power amps and preamps. We also sometimes used them as input and output coupling caps, where necessary. This seemed practical at the time, but not necessarily audio perfection. We kept our tube stuff for the best quality sound, in many cases. Only when we wanted extra power or some special function, did we use solid state by preference.
In 1973, when I started working with Mark Levinson, he had converted to polycarbonate caps of high quality, and he convinced me that they were really better than tantalum coupling caps, and I then learned about ceramic cap distortion while visiting Tektronix in 1974, so we switched to mica or polystyrene for small caps, polycarbonate up to 10uf, but we still used tantalum caps for higher values up to 1000uF. This is how we made the JC-2 preamp in 1974.
By 1978, Richard Marsh started talking about dielectric absorption in audio. I knew a lot about DA, because a friend of mine, and former classmate, did his masters thesis on this subject and he sent me the equations and technical articles in 1974, but I could not see, at the time, how this could be very important to audio. At the same time, in 1977, I started measuring tantalum caps and their low frequency distortion. It was VERY bad, and it was difficult to get around without servos. So in 1980, I decided to servo my audio circuits, if I could, in order to remove the coupling and feedback caps entirely. This became the JC-80 preamp. Until now, most of my cap discoveries are based on what we did with the JC-80. That is: Direct coupled, servo offset controlled, and polystyrene bypass and EQ caps. I have never regretted designing with this approach. I have removed ceramics and tantalum caps completely, using Mylar only for servo caps.
I am writing this long history to show the actual evolution of my experience with caps. I have no ax to grind, I don't make or sell caps, and I don't just jump on the 'latest' cap design, either. I just use what works.
- Status
- Not open for further replies.