Yes, see the proven and documented circuits of Didden and Jung in AudioXpress, which have been duplicated by thousands of builders with great satisfaction. Excellent noise performance, excellent load and line rejection, they simply do what a DC supply is supposed to do, supply DC and NOT have a sound.
OTOH, if you're of the school which wants the electronics to have an aural signature (and there are many out there who do), you will not be happy with them.
Joshua_G, I am surprised by so many experienced engineers turning on you like this.
First, English, which you write very well, is not your first language. Little changes in the language are expected.
Second, definitions are perhaps the heart of the problem. It seems that these engineers are hung up on absolute 'definition' rather that the idea you are trying to convey.
Let me give you an example:
Back in 1976, Dr.Matti Otala, excellent speaker and writer in at least 5 languages, was put to task by an AES peer reviewer over the difference between 'overdrive' and 'overload' when a circuit falls out of any reasonable sort of non-linearity. The difference is subtle, but that did not stop the reviewer from going on about it. I read his criticisms, myself. It was excessive.
Another example:
About 1984, I first contacted Bob Pease, an application engineer at NSC, regarding his paper on dielectric absorption in caps. Walt Jung, Scott Wurcer and I were doing some research along that line, and I wanted to talk to him about it.
I misused some term about how to describe my test pulse, something like 'peak-period' and he had the same attitude to me, as these engineers have to you. He was insulting! Wow, an application engineer, on the job, insulting a caller! Wow!
Well, I think that these prime definitions are burned into many of these critical engineers at a very early level in THEIR core courses in engineering. IF they made a mistake in class, they might be marked down, so it became important to them. Unfortunately, for me, my first 3 years of college had NO electronics, and I had to learn it on my own or on the job as a technician. My 'definitions' are just about as loose as yours, but so what?
These guys can't even understand what we just built together, because it DOESN'T fit THEIR definition of a voltage regulator, or a buffer, or anything else. They can't think 'outside the box'!
Why these engineers are allowed to attack you, yet if I attack them with the same force, I get binned, I don't know. Kind of one sided, don't you think?
First, English, which you write very well, is not your first language. Little changes in the language are expected.
Second, definitions are perhaps the heart of the problem. It seems that these engineers are hung up on absolute 'definition' rather that the idea you are trying to convey.
Let me give you an example:
Back in 1976, Dr.Matti Otala, excellent speaker and writer in at least 5 languages, was put to task by an AES peer reviewer over the difference between 'overdrive' and 'overload' when a circuit falls out of any reasonable sort of non-linearity. The difference is subtle, but that did not stop the reviewer from going on about it. I read his criticisms, myself. It was excessive.
Another example:
About 1984, I first contacted Bob Pease, an application engineer at NSC, regarding his paper on dielectric absorption in caps. Walt Jung, Scott Wurcer and I were doing some research along that line, and I wanted to talk to him about it.
I misused some term about how to describe my test pulse, something like 'peak-period' and he had the same attitude to me, as these engineers have to you. He was insulting! Wow, an application engineer, on the job, insulting a caller! Wow!
Well, I think that these prime definitions are burned into many of these critical engineers at a very early level in THEIR core courses in engineering. IF they made a mistake in class, they might be marked down, so it became important to them. Unfortunately, for me, my first 3 years of college had NO electronics, and I had to learn it on my own or on the job as a technician. My 'definitions' are just about as loose as yours, but so what?
These guys can't even understand what we just built together, because it DOESN'T fit THEIR definition of a voltage regulator, or a buffer, or anything else. They can't think 'outside the box'!
Why these engineers are allowed to attack you, yet if I attack them with the same force, I get binned, I don't know. Kind of one sided, don't you think?
john curl said:
John,
That has surprised me too. I suspect it is only because they felt a softer target on your side 🙂 . However these attacks on definitions etc. look silly. I would notice thought that in some respects Joshua_G has shown considerably more sense regarding audio than many of his attackers.
Cheers
Alex
john curl said:....
Why these engineers are allowed to attack you, yet if I attack them with the same force, I get binned, I don't know. Kind of one sided, don't you think?
SY said:
That implies that all the attacks to Joshua have been because of some technical errors? If so, it could have been done in a more decent way, dont you think so?
Also, attacking Joshuas design attempts as being an insult to John is nothing different than insulting Johns design. He clearly stated that he used this design approach for a certain application, so....
Going back to the voltage stabilisator (or whatever it has to be called according to the resident "police of correct semantics") of Joshua, these open loop designs are perfectly OK with folded cascode stages (which are drawing a pretty constant current) or balanced stages (also drawing quite constant currents). I really don't think that "regulation factor" would be anything to worry about for these applications.
Also, small thermal voltage drifts will not upset such stages, no way.
Tino
That's been the origin, yes. A stubborn refusal to understand some basic engineering concepts and principles, along with a condescending attitude toward engineers with far more accomplishment and experience than his own, has exacerbated the issue.
Tino and Alex, thanks for your independent evaluation of the situation.
For me, it is the case of an enthusiastic audio designer trying to learn something from me. Not a bad concept, and IF I can also teach others through this example, then it will have been worth it.
Now, Joshua_G has asked me a few questions by E-mail, but I would rather answer them here, because more people 'might' learn something useful.
For one thing, paralleling jfets seems to be relatively easy, compared to bipolars or even mosfets. This is because all you have to do is measure Idss of each part, and only parallel similar Idss parts.
Idss is easily measured by anyone with a 9V battery and a cheap multimeter. Just short the gate to the source, put a 100 ohm resistor between the drain and the battery (note polarity of the battery). Attach the other end of the battery to the source gate short. Measure across the 100 ohm resistor with the multimeter to measure the Idss, by converting the measurement to current. For example 10 ma Idss would measure 1V, 20ma 2V, etc. Sort the devices and parallel 2 or 3 as necessary.
Since these parts have a negative tempo they will adjust and not individually run away from each other like bipolar transistors might.
I have seen this circuit oscillate once when a Marsh designed ultra high frequency cap was used instead of the RT Rel .1uf cap that I originally designed in. The extra long leads necessary with the oversize Marsh cap created the added inductance to make the circuit oscillate.
For me, it is the case of an enthusiastic audio designer trying to learn something from me. Not a bad concept, and IF I can also teach others through this example, then it will have been worth it.
Now, Joshua_G has asked me a few questions by E-mail, but I would rather answer them here, because more people 'might' learn something useful.
For one thing, paralleling jfets seems to be relatively easy, compared to bipolars or even mosfets. This is because all you have to do is measure Idss of each part, and only parallel similar Idss parts.
Idss is easily measured by anyone with a 9V battery and a cheap multimeter. Just short the gate to the source, put a 100 ohm resistor between the drain and the battery (note polarity of the battery). Attach the other end of the battery to the source gate short. Measure across the 100 ohm resistor with the multimeter to measure the Idss, by converting the measurement to current. For example 10 ma Idss would measure 1V, 20ma 2V, etc. Sort the devices and parallel 2 or 3 as necessary.
Since these parts have a negative tempo they will adjust and not individually run away from each other like bipolar transistors might.
I have seen this circuit oscillate once when a Marsh designed ultra high frequency cap was used instead of the RT Rel .1uf cap that I originally designed in. The extra long leads necessary with the oversize Marsh cap created the added inductance to make the circuit oscillate.
How about 'virtual battery' Tino?
(I guess this definition is already taken!) Could not Technics have just used a normal voltage regulator? (Oh, they did, in fact, just like me)
(I guess this definition is already taken!) Could not Technics have just used a normal voltage regulator? (Oh, they did, in fact, just like me)syn08 said:
I wouldn't bother to reply, but some sort of civic duty feeling makes me though raise the :bs: flag.
Are you sure? 😀
Andre Visser said:
I wouldn't bother to reply, but some sort of civic duty feeling makes me though raise the :bs: flag.
Are you sure? 😀
Yes.
zinsula said:
No. JC stated nowhere, nor did anyone else, that JC's design
approach was using a standard circuit off the net, mistreating
it in > 10 ways, publishing the circuit each time and then biting the
feeding hand.
syn08 said:
Then you will have to teach me, where are that current flowing that you suggest?
Andre Visser said:
Short of building the circuit yourself, take a look at the AD797 datasheet http://www.analog.com/static/imported-files/data_sheets/AD797.pdf page 14. The maximum capacitive load supported in a follower configuration is about 20pF. The datasheet also has a clue on a configuration that would be stable in larger capacitive loads (fig. 42).
Set aside the stability, look at page 9 fig. 27 on how the small signal follower step response looks like. It displays precisely the kind of ringing it was claimed to be avoided by an open loop "regulator"design.
The feedback network loads the opamp. The opamp output current flows in the load and the 120ohm feedback resistor in parallel. Basic Kirchoff law.
Syn08, I see a buffer cct that is supposed to keep a DC level on it's output, besides the cap on the output isn't specced but is shown as a small one.
How can the 120 Ohm feedback resistor load the op-amp when only driving the high impedance neg input?
How can the 120 Ohm feedback resistor load the op-amp when only driving the high impedance neg input?
gerhard said:
sleep a night over it.
(Hint: this is not an inverter)
Oh, I'm sorry, you are right, in this configuration the feedback network does not load the opamp. Still, the current flows in the load. The circuit will oscillate as hell but will probably take the abuse (depending on the load).
My fault, and I apologize to everybody for the confusion. I should always take a second look before posting, this is what happens when heavily multitasking.
Your email was not clear to me - I though you mean the input current, not the output.
Juergen Knoop said:talking about a opamp feeding a cap?
Sorry I'm lost!
regards
There is an additional load assumed to be in par. with the cap.
That's the circuit that has to be powered.
The cap itself is contra productive because it resonates with the
output of the op amp that is usually inductive. This challenges
the op amp's stability. A _small_ capacitor directly across the load
might help, however, at high frequencies where the op amp
runs out of steam. It may have to be isolated by a resistor, which
is not welcome in a power supply.
regards, Gerhard
- Status
- Not open for further replies.