George,
Perhaps the cleaning of a record with alcohol leaching the plasticizers is analogous to what Christophe has said about hearing all the little bubbles in the surface. This could be through an out-gassing process over time with the probably DOP or some other plasticizer migrating out of the vinyl? I have never used alcohol so I can't comment on that. but a very small amount of detergent which is loaded with surfactants has always done the trick on a really dirty album for me. You just have to make sure you rinse the surface well to remove all the remaining detergent. DI water or distilled works just fine for that.
Perhaps the cleaning of a record with alcohol leaching the plasticizers is analogous to what Christophe has said about hearing all the little bubbles in the surface. This could be through an out-gassing process over time with the probably DOP or some other plasticizer migrating out of the vinyl? I have never used alcohol so I can't comment on that. but a very small amount of detergent which is loaded with surfactants has always done the trick on a really dirty album for me. You just have to make sure you rinse the surface well to remove all the remaining detergent. DI water or distilled works just fine for that.
It turns out that it's difficult to extract the nonlinearity from the published curves. But it's easy to show that it's absolutely insignificant. If we assume that there's no cascoding (which, as has been helpfully pointed out, nearly eliminates the variation of Cin with signal), assume 5V Vcb, and assume 40dB first stage gain, the variation of input C for a 2N4401 is about 0.14pF for a 0.1mV input. Whether linear or not, that's hilariously low.
John's argument about nonlinear input capacitance does not seem to be well thought out.
No SY, it is not 'well thought out'. And there is no SPICE simulation either. No, I just build successful designs and measure them at high levels and high frequencies to see if there is any extra distortion from secondary sources such as non-linear capacitance. Then I build hundreds of them to be enjoyed by users around the world. Works for me! '-)
I might say also, that I hope to remain free of any technical burdens that I do not wish to partake, including arbitrary questions of how I design.
I might say also, that I hope to remain free of any technical burdens that I do not wish to partake, including arbitrary questions of how I design.
Pano, I answer the questions that I have conveniently at my disposal to answer.
The non-linear input capacitance is a moot point, because it takes about 1Kohm or more to even measure it with the devices used. Since the SCP-2 never sees more than 50 ohms source impedance, I am in safe territory. Second, it requires graphical inspection of the non-linear input capacitance to initially determine if there will be any problem. It might be noted that the Vendetta stages are DELIBERATELY operated in the most linear region with regards to non-linear capacitance.
Third, I took care, decades ago, to measure the distortion at frequencies from 5KHz (my normal standard) to well beyond 20KHz. IF I had seen a significant departure from the Class A predicted distortion curve with level or frequency, I would have fixed the problem, or at least known there was a problem. I just did not want to put up the graphs on the devices, at this time. Maybe later.
The non-linear input capacitance is a moot point, because it takes about 1Kohm or more to even measure it with the devices used. Since the SCP-2 never sees more than 50 ohms source impedance, I am in safe territory. Second, it requires graphical inspection of the non-linear input capacitance to initially determine if there will be any problem. It might be noted that the Vendetta stages are DELIBERATELY operated in the most linear region with regards to non-linear capacitance.
Third, I took care, decades ago, to measure the distortion at frequencies from 5KHz (my normal standard) to well beyond 20KHz. IF I had seen a significant departure from the Class A predicted distortion curve with level or frequency, I would have fixed the problem, or at least known there was a problem. I just did not want to put up the graphs on the devices, at this time. Maybe later.
OK, now for what I promised Joshua G, yesterday. A simplified commentary on the Vendetta SCP-1, now at least 30 years old.
If we go back to the question of how we can achieve simple and elegant biasing of an input stage, this is a prime example.
Because the input jfets are self biasing (they are on already), they will draw current by just attaching a voltage across the Source-Drain connections and that of course, is our first example of a jfet amplifier.
However, in this case we use complementary matched jfets working with opposite voltages, and this makes this circuit a push pull design for distortion, but a parallel design with regards to noise. All you really need is 2 matched complementary jfets and perhaps a couple of load resistors and you are in business.
The balance pot between the two sources cancels out any residual harmonic in the input stage, as well as provides a certain amount of current control with temperature, and perhaps a best operating point. In future, it might be a 5 ohm pot, just to reduce the input noise further, even though, noise-wise, the 10 ohm (pot) looks like a 2.5 ohm noise source. (more later)
If we go back to the question of how we can achieve simple and elegant biasing of an input stage, this is a prime example.
Because the input jfets are self biasing (they are on already), they will draw current by just attaching a voltage across the Source-Drain connections and that of course, is our first example of a jfet amplifier.
However, in this case we use complementary matched jfets working with opposite voltages, and this makes this circuit a push pull design for distortion, but a parallel design with regards to noise. All you really need is 2 matched complementary jfets and perhaps a couple of load resistors and you are in business.
The balance pot between the two sources cancels out any residual harmonic in the input stage, as well as provides a certain amount of current control with temperature, and perhaps a best operating point. In future, it might be a 5 ohm pot, just to reduce the input noise further, even though, noise-wise, the 10 ohm (pot) looks like a 2.5 ohm noise source. (more later)
Last edited:
Evan if shot on film arn't all cinemas these days digital, DLP or similar based, so the film will have been digitised.
With todays new sensors and digital darkroom software, you can manipulate an image to get your desired results, and the increasing pixel count means you can get softer graduations that give (dare I say it) a more film like feal to a captured image.
With todays new sensors and digital darkroom software, you can manipulate an image to get your desired results, and the increasing pixel count means you can get softer graduations that give (dare I say it) a more film like feal to a captured image.
I asked a very basic question about phono stage design- why an FET rather than a bipolar for the first stage of a state-of-the-art unit. So, based on reality rather than stories, why an FET rather than a bipolar for an MC input stage? So do you actually have a technical reason which stands up to basic analysis?
As can be readily seen, the self biasing characteristic of the jfet makes it even easier to use than would be a bipolar device.
The bipolar device(s) would have to be cap coupled or differentially coupled. Both cause potential problems due to the added parts, that can only add noise.
The complementary jfet input stage LOWERS noise by 3 dB, and linearizes the output.
In the total gain stage, no coupling caps either input or output, are needed.
The bipolar device(s) would have to be cap coupled or differentially coupled. Both cause potential problems due to the added parts, that can only add noise.
The complementary jfet input stage LOWERS noise by 3 dB, and linearizes the output.
In the total gain stage, no coupling caps either input or output, are needed.
Now, for the non-linear capacitance issue:
Here are two complementary devices similar to the devices used in the Vendetta. Just multiply the capacitances by 2.5 or so and you get the original devices. Similar curves I believe. Note that I operate the input jfets at 9-12V either Vds or Vdg.
Here are two complementary devices similar to the devices used in the Vendetta. Just multiply the capacitances by 2.5 or so and you get the original devices. Similar curves I believe. Note that I operate the input jfets at 9-12V either Vds or Vdg.
Attachments
There's a lot of incorrect stuff on that site, but they are, after all, trying to convince you to buy their service. Now, archiving analog material digitally IS a good idea, vinyl DOES deteriorate with playing (so does tape), but the the stuff they say about cleaning and how vinyl works is not accurate, starting with the notion or surface lubricants.
Yes, John, I used those graphs for both the FETs and the bipolars to determine that the nonlinear capacitance claim is nonsense. Perhaps you want to take a moment to run the numbers about linearity at 10mV collector/drain swings before making your next set of claims? Or is this an ex cathedra exercise rather than engineering?
At what Vdg do you see an increase in noise from hot carrier? Speaking of nonlinearity, you did notice that the capacitance graph is shown as a function of log voltage, right? The "linear" region is anything but.
- Status
- Not open for further replies.