Sound Quality Vs. Measurements

Status
Not open for further replies.
diyAudio Member RIP
Joined 2005
[snip] So, I still don't understand the criteria for selecting R5, the resistor between the CCS and the diff pair emitter resistors? Is it just managing the voltage across Q5 so I could use a lower voltage part? Is it a tradeoff of imperfect parts, who makes more noise, the CCS transistor or the resistor? It is about half and half designs that have this part or not.

It's not entirely clear. In principle it's not needed. It reduces dissipation in the current source transistor, may make the input stage slightly harder to break with huge overdrives, may isolate the collector capacitance of the transistor from the diff pair at high frequencies --- that is, the impedance at the diff pair 47 ohm resistors never gets below 10k in parallel with a small resistor self-capacitance.

As far as a noise generator, its own self-noise is mostly ineffectual since in series with a fairly high current source output impedance, at least at audio frequencies. The noise in the current source is due to various mechanisms: thermal and excess noise in the emitter resistor, shot and excess noise in the transistor base current, transistor voltage noise changing the voltage and hence the current across the CCS emitter resistor, ditto the voltage noise of the two series diodes. Not to mention ripple in the power rail modulating the diodes' current and hence their voltage.

Haven't done the inventory to determine which is the most significant, and one unknown is the voltage noise of the transistor due in part to the (maybe known/tabulated somewhere but unknown to me) base resistance Rbb'. With no power supply effects, probably the thermal of the emitter resistor is dominant, followed by the transistor voltage noise (for example, if the latter is 3nV/rt Hz, the effect on the current would be about 4.4pA/rt Hz; the current noise of a 680 ohm R is about 4.9pA/rt Hz at room temp).
 
Thanks for the response TVR are you going to snub that supply ....?

I may, just to see if it makes a further difference. I was hoping for more high harmonic suppression, 240 and up. How much of that is actually through the PS, and how much induced I need to play and see. I really don't have the experience to know what to expect.

I can clearly measure the difference, but can I hear any difference? For my bench amp, the measurement is important. For music, is this the best place to spend money? I bet not the first. It was not a particularly noisy amp to start with. The Z.3 is.

What was that on Einstein's wall, "Everything you can count doesn't count, and everything that counts you can't count"

Right now I am celebrating inputting the 120's IPS into Simetrix and getting my first plots. (UGG-LE) I have about, oh 388 pages to read and only one more malt based liquid diet supplement cold. I am mad at those guys in S.F. They got me hooked on their porter and then raised the price!
 
It's not entirely clear. In principle it's not needed. It reduces dissipation in the current source transistor, may make the input stage slightly harder to break with huge overdrives, may isolate the collector capacitance...

Brad,
Wow thats a very impressive and comprehensive explanation. Thanks for the insight.


Yes. Like, "Let's assume idle current for this stage 1 mA, for next stage 10 mA, for 3'rd stage 100 MA, then calculate precision values of resistors needed" :D

Anatolly

Doesn't the math and engineering also dictate the choice of currents and number of stages etc.

Thanks
-Antonio
 
Hi,

What about PSU with fast acting diodes and snubbers, like, dis-like ......? maybe TVR can give it a go in his Hafler...

While covering HV rectification, here is a site showing some easy to apply improvements to ordinary bridge rectifiers:

S5 Electronics K-12M Tube Amp

These techniques make a modkateers life much easier.

Ciao T
 
Hi,

Thanks TL, interesting stuff and it leads me to part two, if so, why not more amplifiers with FR diodes and snubbers ..... ?

Don't ask me. At AMR I only use Fast Diodes and Schottkies plus usually "interesting" circuits. I also tend to regulate the supplies for the "Power" part of the Amplifier. But it is not common practice.

You need to ask the other designers.

Why do some swear by standard block rectifiers .... ?:confused:

Well, some swear by K-Mart Beef, some swear by Kobe Beef... Some swear at those who swear by Kobe beef, especially those who never tried it and insist that all beef tastes the same anyway.

Ciao T
 
Many designers just don't know any better. I didn't for years, myself. It was embarrassing when I found that I had overlooked something important like hi speed diodes in quality audio gear. IF I had been more open minded 25 years ago, I could have avoided certain problems with my Vendetta Research SCP-2, that were audible.
It took my former business partner (now deceased) to get me to try it 20 years ago, and I have never since used standard diodes for quality audio, when it is possible to avoid them. The Parasound JC-1, JC-2, and JC-3 all use high speed rectifiers in the power supply circuits. My cheaper stuff, maybe not.
 
Hi,

Off hand, John, do you think a power line filter BEFORE the power transformer(s) and hence diodes in the rectifier, would improve matters?

It hits -70 dB at 1 MHz.

I am not John, but let's apply some logic here.

The mains transformer has comparably high leakage inductance and resistance.

The actual diode switching induced glitch is finished in around 20uS.

The lower frequency effects from resonances in the mains transformer are quite low in frequency, with several milliSeconds period.

The switch-off transients circulate in a tank circuit formed around the mains transformer secondary (including parasitics), reservoir capacitor (including parasitics) and rectifier.

They are caused by (respectively) depleting the charge stored in the diode junction of the Diode when it is reverse biased and the effect of hard switching of quite high currents from an inductive source.

Does placing a filter prior to this tank circuit prevent any of the effects within this tank circuit or the release of energy stored in the transformer inductance?

Does placing a filter prior to this tank circuit prevents the effects of the Tank circuit or the release of energy stored in the transformer inductance from "escaping" into the mains power supply?

Does changing the rectifier to diodes that have a minimised or no switch-off energy reduce the effects of the switch-off transients?

Does placing a snubber across the secondary of the mains transformer help disspate energy circulating in this tank circuit or the released energy stored in the transformer inductance?

Does placing an RC or LC filter between multiple reservoir capacitors help attenuate the energy circulating in this tank circuit or the release of energy stored in the transformer inductance from the actual audio circuit?

Does designing a power transformer for high isolation and self-damping help to reduce the energy circulating in this tank circuit or the released energy stored in the transformer inductance?

Do any of these measures affect the coupling of RF energy present in the mains supply?

Ciao T
 
One can mitigate "snapping" or high speed reverse recovery which is due to drift and gradient of minority carriers' concentration during the depletion mentioned
by means of proper dumping oscillation in question: snubbing,
by choosing devices with soft reverse recovery or the ones that do not store minority carriers: Schottky diodes,
by RF decoupling: CM chokes, feed-through caps, shielding etc.
or
by the combination of thereof.

Seems like ingredients are well known but not the recipe;)

P.S. Use of enigmatic terms like "power line filter" is not a solution
The former BTW serves either to bring conducted interference, caused by device connected to the grid, into compliance to whatever standard applicable or just an opposite - to prevent interference from the grid getting into the device or both...
dvv, which one did you have in mind?
 
As the service level agreement for the grid is to power light bulbs and motors. Any other use thereof is up to the designer to make compatible. A fuse and a ceramic cap does not cut it. There is regrettably no standards for preventing signal injection; X10 counts on it. A good designer would be a good citizen, provided the accountant lets him.

In the IT industry we spend a lot of money to make the power suitable for electronic devices. See Libert, APC, etc. Shame on us for trusting the power company, or in trusting all of our equipment designers. I am NOT propping up mega-buck boxes with flashing lights. My Tripp-Lite strips are just fine, thanks.
 
One can mitigate "snapping" or high speed reverse recovery which is due to drift and gradient of minority carriers' concentration during the depletion mentioned
by means of proper dumping oscillation in question: snubbing,
by choosing devices with soft reverse recovery or the ones that do not store minority carriers: Schottky diodes,
by RF decoupling: CM chokes, feed-through caps, shielding etc.
or
by the combination of thereof.

Seems like ingredients are well known but not the recipe;)

P.S. Use of enigmatic terms like "power line filter" is not a solution
The former BTW serves either to bring conducted interference, caused by device connected to the grid, into compliance to whatever standard applicable or just an opposite - to prevent interference from the grid getting into the device or both...
dvv, which one did you have in mind?

Exactly which part of "power line filter" is enigmatic?

I would have thought the description is self-explanatory. It's quite simply a filter which cleans up the incoming power, the device being placed between the wall outlet and whatever you wish to obtain filtered power from it.

Since the filter is of the floating type, it has no connection to the ground, except for the indicator LED (if it's shining, the filter is functional, if it's not shining, the filter has blown a fuse).

Ultimately, there's a photo of it a few pages back.
 
Power lines are usually dirty. So, we use power line filters to clean up the power line. Many simple passive filters are designed to contain the RFI generated by other electronic equipment like faxes, computers, etc, and keep it out of the power line. These filters can be problematic with audio equipment. I generally avoid using them as such. I usually make my own.
 
Status
Not open for further replies.