Class-I
Hi Richard,
I've never published a paper on class-I and will never do so in the future as it's not my favorite solution for controlling the bias.
For more info see: US patent 4595883
http://www.diyaudio.com/forums/ever...-linear-audio-publication-41.html#post2799107
http://www.diyaudio.com/forums/solid-state/202684-class-i-siblings-17.html#post2861437
>are there any commercial products using this concept?
Yes, Pioneer has used it.
The class-I concept is most ingenious, but it also has some potential flaws. For this reason I prefer my one design, simply called AB-II.
>Have you done a paper .....
Please see my website: Auto Bias part II
Cheers, E.
Hi Richard,
I've never published a paper on class-I and will never do so in the future as it's not my favorite solution for controlling the bias.
For more info see: US patent 4595883
http://www.diyaudio.com/forums/ever...-linear-audio-publication-41.html#post2799107
http://www.diyaudio.com/forums/solid-state/202684-class-i-siblings-17.html#post2861437
>are there any commercial products using this concept?
Yes, Pioneer has used it.
The class-I concept is most ingenious, but it also has some potential flaws. For this reason I prefer my one design, simply called AB-II.
>Have you done a paper .....
Please see my website: Auto Bias part II
Cheers, E.
Controlled Iq, SOA etc
BTW, thanks for letting KC-P and others to reprint some of their Linear Audio articles. It's much appreciated by beach bums and I'm sure other impoverished people too.
_____________
In da late 70's early 80's, we did a lot of listening tests on SOA protection. (LEAK & Wharfedale were the same company under Rank). Most double slope and even badly implemented single slope SOA protection had what we called 'snap crackle pop' sound when triggered.
We concluded that only plain current limiting (or VERY shallow single slopes) was 'innocuous' from a sound viewpoint.
This wasn't subtle Golden Pinnae ****. But in dem days, devices were nowhere as good as today and hence quite aggressive protection was common.
It would be good to repeat some of these Listening Tests on modern stuff.
Alas, beach bums dun haf money to buy books .. even ones as good as yours
BTW, thanks for letting KC-P and others to reprint some of their Linear Audio articles. It's much appreciated by beach bums and I'm sure other impoverished people too.
_____________
In da late 70's early 80's, we did a lot of listening tests on SOA protection. (LEAK & Wharfedale were the same company under Rank). Most double slope and even badly implemented single slope SOA protection had what we called 'snap crackle pop' sound when triggered.
We concluded that only plain current limiting (or VERY shallow single slopes) was 'innocuous' from a sound viewpoint.
This wasn't subtle Golden Pinnae ****. But in dem days, devices were nowhere as good as today and hence quite aggressive protection was common.
It would be good to repeat some of these Listening Tests on modern stuff.
Last edited:
Alas, beach bums dun haf money to buy books .. even ones as good as yours
BTW, thanks for letting KC-P and others to reprint some of their Linear Audio articles. It's much appreciated by beach bums and I'm sure other impoverished people too.
_____________
In da late 70's early 80's, we did a lot of listening tests on SOA protection. (LEAK & Wharfedale were the same company under Rank). Most double slope and even badly implemented single slope SOA protection had what we called 'snap crackle pop' sound when triggered.
We concluded that only plain current limiting (or VERY shallow single slopes) was 'innocuous' from a sound viewpoint.
This wasn't subtle Golden Pinnae ****. But in dem days, devices were nowhere as good as today and hence quite aggressive protection was common.
It would be good to repeat some of these Listening Tests on modern stuff.
Yes, amplifiers like that were called "tweeter eaters". Very nasty pulses could be created into inductive loads, for example. I believe the Phase Linear 700 fell into this category.
Cheers,
Bob
OPS and bias for zero THD -
Ed--- Are you in the mood ? (
to add your OPS to what has been done so far here and start towards a completed topology and circuit design which can get built?
Seems your circuit will also melt in with the designs discussed here quit well.
Thx-RNMarsh
Ed--- Are you in the mood ? (
to add your OPS to what has been done so far here and start towards a completed topology and circuit design which can get built?
Seems your circuit will also melt in with the designs discussed here quit well.
Thx-RNMarsh
Ed--- Are you in the mood ? (
to add your OPS to what has been done so far here?
Seems your circuit will also melt in quit well with the designs discussed here and for the Current-Feedback Amp circuit being discussed. The CFA is awaiting a good OPS.
Thx-RNMarsh
Actually Phase Linear 700 was quite good in all its guises.
The really nasty amp was Crown DC300 in all its guises. The last big LEAK valve amps were better than DC300 into real speakers.
We concluded that the only thing DC300 was good for destroying speakers .. we used several in bridge for Death Testing. (DIN 45573 live test)
Actually I remember the Phase Linear 700 as a fairly nice sounding amp, but do remember they had a tendency to go up in smoke with strange impedance loads or just overheat when we used them for PA with multiple cabinets. I think the best of the best of the pro-audio stuff was by Spectra-Sonics.
Alas, beach bums dun haf money to buy books .. even ones as good as yours
You should make the effort for this volume. Not just the bias article but also Bob Cordell and other fine contributors.
Best wishes
David
Alas, beach bums dun haf money to buy books .. even ones as good as yours
BTW, thanks for letting KC-P and others to reprint some of their Linear Audio articles. It's much appreciated by beach bums and I'm sure other impoverished people too.
_____________
In da late 70's early 80's, we did a lot of listening tests on SOA protection. (LEAK & Wharfedale were the same company under Rank). Most double slope and even badly implemented single slope SOA protection had what we called 'snap crackle pop' sound when triggered.
We concluded that only plain current limiting (or VERY shallow single slopes) was 'innocuous' from a sound viewpoint.
This wasn't subtle Golden Pinnae ****. But in dem days, devices were nowhere as good as today and hence quite aggressive protection was common.
It would be good to repeat some of these Listening Tests on modern stuff.
My own implementation of SOA protection uses re-entrant switching of a speaker relay.
No pop crackle whatever until triggered.
And so far, in 4 years of almost daily use, only triggered once.
That was when I was changing cables and inadvertently touched the signal conductor.
It's all in the design
jan
Greetings everyone.
I appologise from the start if my question is trivial or seams unapropriate. I was reading @Mr Cordell's book ( a great one indeed ), at the beginning i found something that puzzeled me, but i've ignored it at firts, still getting to the more advanced chapters, that "something" bug's me more and more so i came here asking for help.
It is about the Slew Rate, more to the point, the actual SR of a given sine wave freq, it is stated in the book that for a 20Khz sine wave, the SR is about 0,125V/uS of peak volt. But i am not sure why it is so, for 20Khz the time is 50uS, so 1V is peaked in 12,5uS, and to find out the voltage slewed on 1uS i divide 1/12,5 and i get 80mV, where is the difference in my calculations from Mr. Cordells book? i know i am doing something wrong, but what is my mistake?
Again i appologise if any inconveniance is done by my posting this, i know it is a trivial thing but still it is important to me, i must understand where i go wrong and coreect my error.
All the best to everyone.
I appologise from the start if my question is trivial or seams unapropriate. I was reading @Mr Cordell's book ( a great one indeed ), at the beginning i found something that puzzeled me, but i've ignored it at firts, still getting to the more advanced chapters, that "something" bug's me more and more so i came here asking for help.
It is about the Slew Rate, more to the point, the actual SR of a given sine wave freq, it is stated in the book that for a 20Khz sine wave, the SR is about 0,125V/uS of peak volt. But i am not sure why it is so, for 20Khz the time is 50uS, so 1V is peaked in 12,5uS, and to find out the voltage slewed on 1uS i divide 1/12,5 and i get 80mV, where is the difference in my calculations from Mr. Cordells book? i know i am doing something wrong, but what is my mistake?
Again i appologise if any inconveniance is done by my posting this, i know it is a trivial thing but still it is important to me, i must understand where i go wrong and coreect my error.
All the best to everyone.
the average slope of your sinewave is the same as a triangle wave.
A sinewave does not have constant slope.
The top of each wave has zero slope, the zero crossing has the steepest part of the slope.
The difference of 125 to 80 is taken account by the extra steepness at the zero crossing.
Fortunately the mathematicians amongst us have worked out that using Pi in the Slew Rate equation predicts the maximum rate.
A sinewave does not have constant slope.
The top of each wave has zero slope, the zero crossing has the steepest part of the slope.
The difference of 125 to 80 is taken account by the extra steepness at the zero crossing.
Fortunately the mathematicians amongst us have worked out that using Pi in the Slew Rate equation predicts the maximum rate.
Ed--- Are you in the mood ? (
to add your OPS to what has been done so far here and start towards a completed topology and circuit design which can get built?
Seems your circuit will also melt in with the designs discussed here quit well.
Thx-RNMarsh
Hi Richard,
>what has been done so far
My latest design looks like fig.12 on this webpage. I still haven't got the time to build it (too busy with other things and not 'in the mood' to design the PCB art work).
BTW, as you can see, it uses a 'zero slope' current limiter in the OPS. Single or dual slope thingies will result in a negative output impedance (during limiting) which together with a reactive load, is asking for serious trouble.
Cheers, E.
It is basically a 'detector' circuit as Mike K has published, with two breakpoints.
But in his circuit, the 'output' of the 'detector' siphons away current from the driver transistors, thereby limiting output level to safe levels.
In my unit I use the detector output to drive a uPC1237. This chip drives the speaker relay. The uPC also does AC mains detection to avoid switch-on/switch-off transients, and has an option to retry the relay after an adjustable delay to see if the overload has cleared.
So there is a conceptual difference here: the 'normal' circuitry can be invasive and audible, while my circuit actually switches the speaker off.
That LOOKS like a greater nuisance, but in practice if it happens (and it shouldn't if the total design is OK) you just turn down the volume. I rather listen a bit less loud than having the protection coming in all the time.
It's a personal call of course.
The basic design was published by Elektor in 2008 IIRC although I have since made some improvements.
http://www.elektor.com/magazines/2008/may/protection-system-for-power-amplifiers.427862.lynkx
jan
Hi MarianB,
I'm glad you are enjoying my book. Your calculation would be correct if the waveform was that of a triangle with constant slopes. However, for a sinusoid, the slope is greatest near the middle zero-voltage point, and then decreases as the maximum voltage is approached. The voltage rate-of-change is thus greatest near zero volts, and is 0.125V/us per peak volt at this point for a 20kHz sine wave. We're basically looking at the derivative of a 20kHz sinusoid.
As an aside, I and others are conveniently loose in our use of the term "slew rate". We tend to use it to describe both the maximum rate of change an amplifier can supply AND as a property of the signal at any given time. The latter is properly the voltage rate-of-change (or current rate-of-change). I hope I am not creating confusion by pointing out this caveat.
I believe that the term slew rate was originally used to describe the maximum rate at which a gun turret could turn.
Cheers,
Bob
Greatings Mr. Cordell.
It really is an honor to get an answer from you cus i love reading your book. Verry serious health problems kept me years ago from going to highschool and college, ( thus, no electronics tutoring ) so any theory i needed, i had to learn it by myself or from some good people willing to get me out of the dark sometimes on some things.
Anyway, i do not wish to bother you with trivial stuff, i realise you must be verry busy so your answear is that much more appreciated. What i sayd above, was ment to explain somewhat my confusions, or at least theyr couse. I now know where i was wrong and i'm sort of ashamed that it was on somethins this basic...
Many thanks for your time and all the best.
Marian.
It really is an honor to get an answer from you cus i love reading your book. Verry serious health problems kept me years ago from going to highschool and college, ( thus, no electronics tutoring ) so any theory i needed, i had to learn it by myself or from some good people willing to get me out of the dark sometimes on some things.
Anyway, i do not wish to bother you with trivial stuff, i realise you must be verry busy so your answear is that much more appreciated. What i sayd above, was ment to explain somewhat my confusions, or at least theyr couse. I now know where i was wrong and i'm sort of ashamed that it was on somethins this basic...
Many thanks for your time and all the best.
Marian.
Last edited:
Thank you Marian, and remember that no question is ever too trivial. One of the purposes of this forum is to educate people from all levels of experience and knowledge. Often, raising such questions may help me by illuminating an area of my text where a couple of extra words could make something more clear to a broader audience. By the same token, the question and its answer can help others on the forum who did not raise the question.
For more basic learning of electronics leading up to the kinds of discussions here and in my book, there are 2 books that I recommend:
"Teach Yourself Electricity and Electronics" by Stan Gibilisco, McGraw-Hill, 2006.
"Practical Electronics for Inventors" by Paul Scherz, McGraw-Hill, 2007.
Don't be misled by the elementary sound of their titles. These books are packed with really good material. Look at their tables of contents on Amazon.
Cheers,
Bob
Bob, do you know how much inductance there would be from the die to the heatsink tab directly under it for TO-247 type packages? It strikes me that this is extremely important for RF behavior, because the thermal insulator forms a capacitor to the heatsink, which may have an inductive path to the amplifier PCB. Furthermore, it may be possible to take advantage of this capacitance and use it in the decoupling scheme. If very thin insulators are used it may be mandatory to pay attention to the connection between the heatsink and PCB, but if done cleverly the capacitance could be used in a positive way.
I believe that the inductance from the die down to the heat sink tab beneath would be very, very small. I'm not sure how much capacitance there would be between the tab and the heat sink across even a thin insulator. I would guess it to be maybe no more than 100pF.
Cheers,
Bob