Hi ,
for learning and testing i make a "Play" PCB with all I found around the Amp Topics.
One Part is the DC Servo.
I found two ways to realise them.
Each of them says it was the better 🙂
Buit wat was your expiriences ?
What is better A or B. ... I ask about the 2 Diodes 🙂
You see in simple block draft as Picture... So only about the 2 varinats of connecting DIodes.
Of Course has a block picture not all in it
Kind regards Michaela
for learning and testing i make a "Play" PCB with all I found around the Amp Topics.
One Part is the DC Servo.
I found two ways to realise them.
Each of them says it was the better 🙂
Buit wat was your expiriences ?
What is better A or B. ... I ask about the 2 Diodes 🙂
You see in simple block draft as Picture... So only about the 2 varinats of connecting DIodes.
Of Course has a block picture not all in it
Kind regards Michaela
Last edited:
Might I suggest reading Self and Cordell on the subject. Then give various plan a try in Spice. Be sure to look at initial power up. Simulate it.
There are about 10 more ways to do it. Do note, usually the resistor values are very high, so op-amp selection is critical. FET usually.
There are about 10 more ways to do it. Do note, usually the resistor values are very high, so op-amp selection is critical. FET usually.
hi 
what do mean with : Self and Cordell. The books ??? Which one Expecilly.
But Iam a teeny and the Books are to complicate to read, also in english
I need here a more young & pupil better understandable help and text 🙂😛
Sure that there are more ways , but these 2 are the most used ... so I ask about that
experiences.
Sure LT-SPice is fine, but in real life are better to ask about real experiences. 😎
Then SPice and experiences would be the choice for the way 🙂
what do mean with : Self and Cordell. The books ??? Which one Expecilly.
But Iam a teeny and the Books are to complicate to read, also in english
I need here a more young & pupil better understandable help and text 🙂😛
Sure that there are more ways , but these 2 are the most used ... so I ask about that
experiences.
Sure LT-SPice is fine, but in real life are better to ask about real experiences. 😎
Then SPice and experiences would be the choice for the way 🙂
Last edited:
Doug Self Audio Power Amplifier Design Handbook
Bob Cordell Designing Audio Power Amplifiers
Marshall Leach Introduction to Electroacoustics and Audio Amplifier Design
Then there are a series of books from Lancaster and Jung on op-amps etc.
Self and Cordell wrote the definitive books on amplifier design from a practical stand point. Leach did a more detailed mathematical approach.
There are similar books on tube design.
There is no substitute for the basic learning. A few hints on a forum does not provide a degree in engineering. It can help with a hic-up here and there!
LTSpice is more than fine. it lets you play and measure easily and quickly. Then of course you prototype and measure. It is a lot easier finding your phase margin in Spice, than the expense of output transistors!
As far as your two circuits, I have usually seen a circuit using two op-amps. Model them. Do full tolerance simulations. Thermal drift, and as I mentioned, start up and power down. The last thing you want is an op-amp to latch up at the rail! ( hint; diodes)
Bob Cordell Designing Audio Power Amplifiers
Marshall Leach Introduction to Electroacoustics and Audio Amplifier Design
Then there are a series of books from Lancaster and Jung on op-amps etc.
Self and Cordell wrote the definitive books on amplifier design from a practical stand point. Leach did a more detailed mathematical approach.
There are similar books on tube design.
There is no substitute for the basic learning. A few hints on a forum does not provide a degree in engineering. It can help with a hic-up here and there!
LTSpice is more than fine. it lets you play and measure easily and quickly. Then of course you prototype and measure. It is a lot easier finding your phase margin in Spice, than the expense of output transistors!
As far as your two circuits, I have usually seen a circuit using two op-amps. Model them. Do full tolerance simulations. Thermal drift, and as I mentioned, start up and power down. The last thing you want is an op-amp to latch up at the rail! ( hint; diodes)
So in SPice I see that A is more critical in disortation.
With the same Testparameter the Varionat B is more better in disortation.
This is what I see.
Can anybody confirm that B is better in disortation as A ?
..... so now i must do some homework for school, over and out for today 🙂
.... Have a great evening or morning depending on where you live and write here ...
With the same Testparameter the Varionat B is more better in disortation.
This is what I see.
Can anybody confirm that B is better in disortation as A ?
..... so now i must do some homework for school, over and out for today 🙂
.... Have a great evening or morning depending on where you live and write here ...
Could you draw a larger part of the circuits? The way they are drawn, both schematics show op-amps working open loop, so both should distort like crazy.
Distortion of any reasonable cap on the input is many orders of magnitude lower than that of the 3886. So why would you?
As you apparently want us to guess, I guess you are trying to make a non-inverting integrator circuit that will be in parallel to the feedback resistor. The capacitor connection in circuit B is a mistake, it actually has to go to ground. There is another capacitor between the output and the negative input of each op-amp and a resistor between the main amplifier output and the positive op-amp input, and a resistor from the op-amp output to the feedback input of the main amplifier.
The diodes are meant to protect the op-amp inputs against overvoltage, and they may or may not be meant to prevent the polarity reversal that some FET op-amps have when you go outside their common-mode range.
Both circuits are likely to protect against overvoltage, but if polarity reversal is a concern and if +15 V and -15 V are the same as +Vcc and -Vcc, then circuit B won't prevent issues due to polarity reversal. Circuit A may prevent polarity reversal if the ratio between the resistances of the undrawn resistor to the positive input and the resistor from the negative input to ground is large enough to always keep the voltage in the specified common mode range of the op-amp.
By the way, I have no experience with either circuit. I much prefer circuits like this:
You can dimension it to have a nice and flat second-order Butterworth high-pass response. The op-amp is well protected, it only has to handle very limited currents and its output impedance has almost no impact on the in-band amplifier gain. Disadvantages: people who dislike coupling capacitors are bound to dislike it for its coupling capacitor, and people who dislike DC bias loops for its integrator.
The diodes are meant to protect the op-amp inputs against overvoltage, and they may or may not be meant to prevent the polarity reversal that some FET op-amps have when you go outside their common-mode range.
Both circuits are likely to protect against overvoltage, but if polarity reversal is a concern and if +15 V and -15 V are the same as +Vcc and -Vcc, then circuit B won't prevent issues due to polarity reversal. Circuit A may prevent polarity reversal if the ratio between the resistances of the undrawn resistor to the positive input and the resistor from the negative input to ground is large enough to always keep the voltage in the specified common mode range of the op-amp.
By the way, I have no experience with either circuit. I much prefer circuits like this:
You can dimension it to have a nice and flat second-order Butterworth high-pass response. The op-amp is well protected, it only has to handle very limited currents and its output impedance has almost no impact on the in-band amplifier gain. Disadvantages: people who dislike coupling capacitors are bound to dislike it for its coupling capacitor, and people who dislike DC bias loops for its integrator.
Marcel, why a blocking cap plus a servo?
Btw, I don't share a common misunderstanding that a servo is outside the signal chain. After all, the servo becomes integral part of the transfer function of an amp. How could an element be outside the signal chain and still be part of the transfer function?
Btw, I don't share a common misunderstanding that a servo is outside the signal chain. After all, the servo becomes integral part of the transfer function of an amp. How could an element be outside the signal chain and still be part of the transfer function?
I would recommend against attempting to design a composite amplifier with DC servo, as what appears to be essentially your first audio electronics project.
Neither do I.
I wanted to have an AC coupling capacitor at the input so a malfunction of the preamplifier could not possibly cause a large DC voltage at the amplifier output (it can with just a servo, when the DC at the amplifier input is large enough to overdrive the servo). As my main amplifier is a discrete amplifier that could theoretically have up to 2.5 V of offset at its output, I wanted a servo to get rid of that. Combine them and you can make a nice second-order high-pass response that is much flatter in the audio band than a first-order response with the same cut-off frequency.
Ha!! tested out the like feature. Very concise and easily understood Marcel. . So the capacitor is for failure of external input DC, while the servo is there to maintain that low level DC via servo, including during the most likely and/or severe failure mode and at the same time, with a bonus effect of a better response of the combined filter. Sorry, i'm just putting it in my own words to make sure I understand, not understand, understand your thematic. I hadnt compartmentalised and combined the 2 methods like that. Personally i'm onboard with a capacitor sometimes being the most elegant solution, but there is no arguing with the quality of a well designed servo. Using the 2 like that as a measured brute force approach to the individual challenges like that is clever. i'm interested in thought processes as much as circuits. So with your preferred circuit you offend everyone equally. Its the fairest way.
Last edited:
It can't keep the output at zero when the main amp fails. For example, a shorted output transistor will pull the output to a supply rail no matter what the DC loop does. I have a circuit that monitors the main amplifier's error voltage and opens a relay to protect against issues like that.
The 2.5 V DC offset that I wrote about and that the loop has to get rid off, is the worst possible offset when the main amplifier is in perfect working order. Otherwise you understood precisely what I meant.
The 2.5 V DC offset that I wrote about and that the loop has to get rid off, is the worst possible offset when the main amplifier is in perfect working order. Otherwise you understood precisely what I meant.
Last edited:
Hi ,
I did a test on a punched tape PCB
I make Variante B, then I connect to an China Board with LM3886.
Keep in mind, above showed was just a block picture.(missing e.g. Cs)
Results:
Without an Audio input (on input of amp ) the DC on the LM3886 Output ( Speaker out) was 3mV
Connecting my DC Servo (DC Blocker) to the board, the DC goes down to 0,4mV ( m. by Fluke )
So I see, something works.
Also I see if I play with signalgenerator.
100 Hz----- High Gain and High ampltute at outside DC Servo
5000 Hz --- Low Gain and very Small amplitute at outside of DC Servo
It behaves at frequencies like a "low pass" with increasing gain as frequency decreases
The question:
So this was i see and measure was the right function of the DC Blocker, named DC Servo , right ?
I did a test on a punched tape PCB
I make Variante B, then I connect to an China Board with LM3886.
Keep in mind, above showed was just a block picture.(missing e.g. Cs)
Results:
Without an Audio input (on input of amp ) the DC on the LM3886 Output ( Speaker out) was 3mV
Connecting my DC Servo (DC Blocker) to the board, the DC goes down to 0,4mV ( m. by Fluke )
So I see, something works.
Also I see if I play with signalgenerator.
100 Hz----- High Gain and High ampltute at outside DC Servo
5000 Hz --- Low Gain and very Small amplitute at outside of DC Servo
It behaves at frequencies like a "low pass" with increasing gain as frequency decreases
The question:
So this was i see and measure was the right function of the DC Blocker, named DC Servo , right ?
Last edited: