Good....
I understand that in order to avoid noise I should let the jfets run free (Rs = 0) but I need to lower the gain so I can get lower THD.... (jfets are quadratic and produce 2nd harmonics) I need resistors.Why did you suggest using wirewound ? I know these have the lowest ppm specs but does it mean these are quieter ? (I must read Horowtiz´s book again
)Do you detect any issue using a 33r wirewound as Rs for this purposes ? I have another build where I degenerated AC using a EL cap // Rs and it lowers the noise also increasing gain but it leaves a nasty signature killing the nuances... Can the coil produce any effects there ?
You're talking the opposite, V's = rectifiers no problem, pV or nV no physical process has contact potentials that low. Belief is different than, "I have seen no compelling argument to justify an extraordinary claim".
Good....
Why did you suggest using wirewound ? I know these have the lowest ppm specs but does it mean these are quieter ? (I must read Horowtiz´s book again
Do you detect any issue using a 33r wirewound as Rs for this purposes ? I have another build where I degenerated AC using a EL cap // Rs and it lowers the noise also increasing gain but it leaves a nasty signature killing the nuances... Can the coil produce any effects there ?
Hi RCruz,
Something sounds just not right in the behavior of the circuit you are describing. Apart from using faulty or crappy resistors, you should not be seeing that much difference in noise.
Are you measuring the noise with an instrument, or are you just hearing it? Are you looking at the output with a scope? If you use wirewounds, they must be truly non-inductive. Any chance you may be getting high-frequency parasitic oscillations that may make audible noise as a symptom?
Cheers,
Bob
I meant potential barriers. Please Ed do some research on work functions (they are all in VOLTS for metals). You are possibly seeing some spot heating effects or electro-migration of extremely small contact areas or any one of several other things that in another dicipline are well known.
i will look into it.... I am starting a new build using the same parts as in my silent one.... I will report back.
Last edited:
Hi salas
I changed the input jfets in the silent build and got the same results (Silence) so I guess it must be the resistors. I am not using any RC filtering V+ on the first stage and already experimented with severall different shunts and the results are the same (Silence).
All other builds use the same kiwame for Rs and they are noisier.
The silent version does use special caps on the second stage RC V+ filter but I guess that this is not so important.
Scott,
I understood what you meant and it looks like we are actually not in disagreement! Where I think we differ is in the physical implementation of real circuits. Since you brought up the Boltzmann constant, I think we both understand it is from statistical mechanics and is derived from ideal gas models.
From Planck, Boltzmann etc. the Thermal Voltage (Vt) is well known for semiconductors. I don't know how many digits of accuracy to which this has been determined or the variance.
The question is does Ohm's law extend linearly to very small voltages or is it really a series of steps, or even a gap and then change?
Or in complex terms, does the statistical model hold at the extremes?
ES
I would like to focus on a subtle argument here that underlies a significant portion of SUCCESSFUL audio design. This is the inclusion of coupling capacitors, and more subtly, the inclusion of SERIES contacts in the audio path.
First, a little history: Coupling Caps have been with us for perhaps 100 years, since the first vacuum tube, and are used as such, even today in many designs. Now, why did they do this? The PRIMARY reason was to remove the DC DIFFERENCE between the input and output of the tube, transistor, or fet that is almost always there, in order to facilitate proper operation.
Even in the vacuum tube era, perhaps 50-60 years ago, serious effort was taken to REMOVE the coupling caps where possible. For example, a Dyna MK2, 3, or 4 power amp has only one SERIES cap in the circuit. Quite a feat!.
This not only saved money, but helped to reduce a tendency to low frequency oscillation, and extended the low frequency response.
The first coupling caps for the first 50 years, were mostly made of WAX PAPER with a foil sandwiched between the paper insulators. They worked OK, but gave a 'warm' sound because they had so much dielectric absorption, or DA. Some musical guitar, etc preamps still demand this sort of cap for best subjective performance.
About 1960, MYLAR caps were introduced, and even kept a company secret by companies like Marantz. They had a more ideal capability and were more reliable, as well. Still the Military, used better caps in many applications, such as polystyrene and Teflon, especially for LOW DA applications, such as sample and hold, and ANALOG COMPUTERS.
Well now what? As we switched from tubes to bipolar transistors, our circuits changed a little, BUT input and output coupling capacitors came back in masse, as we removed the output transformers, and could not SELF BIAS the bipolar transistors. Also, the MYLAR caps were just too BIG in solid state circuitry, so we went to smaller electrolytic caps, first aluminum, then to Tantalum. By 1970, almost everything used aluminum or tantalum coupling caps. (more later)
First, a little history: Coupling Caps have been with us for perhaps 100 years, since the first vacuum tube, and are used as such, even today in many designs. Now, why did they do this? The PRIMARY reason was to remove the DC DIFFERENCE between the input and output of the tube, transistor, or fet that is almost always there, in order to facilitate proper operation.
Even in the vacuum tube era, perhaps 50-60 years ago, serious effort was taken to REMOVE the coupling caps where possible. For example, a Dyna MK2, 3, or 4 power amp has only one SERIES cap in the circuit. Quite a feat!.
This not only saved money, but helped to reduce a tendency to low frequency oscillation, and extended the low frequency response.
The first coupling caps for the first 50 years, were mostly made of WAX PAPER with a foil sandwiched between the paper insulators. They worked OK, but gave a 'warm' sound because they had so much dielectric absorption, or DA. Some musical guitar, etc preamps still demand this sort of cap for best subjective performance.
About 1960, MYLAR caps were introduced, and even kept a company secret by companies like Marantz. They had a more ideal capability and were more reliable, as well. Still the Military, used better caps in many applications, such as polystyrene and Teflon, especially for LOW DA applications, such as sample and hold, and ANALOG COMPUTERS.
Well now what? As we switched from tubes to bipolar transistors, our circuits changed a little, BUT input and output coupling capacitors came back in masse, as we removed the output transformers, and could not SELF BIAS the bipolar transistors. Also, the MYLAR caps were just too BIG in solid state circuitry, so we went to smaller electrolytic caps, first aluminum, then to Tantalum. By 1970, almost everything used aluminum or tantalum coupling caps. (more later)
Last edited:
Around 1970, due to the lower prices and availability of PNP bipolar devices, we tried to eliminate the input and output caps by switching from single polarity supplies to +/- supplies. A few power amps pioneered this change, but it led to another problem, ADDED capacitor non-linear distortion, due to where the caps were relocated which was in the feedback loop of the amp and preamp, usually operating at approximately 0 DC Volts. This added distortion was difficult to measure, at first, because SMPTE IM distortion normally used 60 Hz as the lowest measuring frequency, and this was the most sensitive test normally available, at the time, for measuring audio products. Tantalum caps were the worst in this regard, much worse than aluminum, for some reason. I published the distortion results with Tantalum caps in my IEEE paper in 1978.
Today, we try to SERVO, in order to remove all the coupling caps, either: input, output, or feedback, if possible (more later)
Today, we try to SERVO, in order to remove all the coupling caps, either: input, output, or feedback, if possible (more later)
I have read the advertisement, but I have trouble with a 2W carbon film resistor having less excess or even regular noise than a metal film of the same value. But I have not measured it.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.