John Curl's Blowtorch preamplifier part II - Page 2314 - diyAudio
Go Back   Home > Forums > Member Areas > The Lounge

The Lounge A place to talk about almost anything but politics and religion.

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 16th May 2012, 02:29 AM   #23131
diyAudio Member
 
john curl's Avatar
 
Join Date: Jul 2003
Location: berkeley ca
I would like to say something here about components. What matters is WHAT WORKS. If we can't get the same quality active or passive parts that we were able to get relatively cheaply in the past, then we have to COMPROMISE. Compromise is NOT the best quality solution. Therefore we search the world for the parts we want and try in encourage others to make the parts we need, if it is at all, possible.
When I first started Vendetta Research, Toshiba had just introduced their P channel ultra low noise jfet, the 2SJ73, the complement to the 2SK146. Together, I realized that I could make a revolutionary product with these parts on the input stage. Today, my effort of about 30 years ago, placed #5 of the greatest preamps of all time in TAS, just this month. Just a phono preamp, done with a lot of care, and using 'breakthrough' jfets.
Unfortunately, these specific parts were discontinued about 20 years ago, and we had to resort to hand matching differential pairs, or using another, more noisy part, that we could parallel. In any case, WHILE THESE COMPONENTS REMAINED AVAILABLE, they were very cost-effective for the performance that they gave. Compared to a phono cartridge that might wear out in a year or two, these devices could go 20 years, without any problem.
Unfortunately, these jfets ceased to be produced, and with relatively short notice. Therefore, most of us, who were interested in continuing to use these parts grabbed as many as we could find, to continue existing production, and for future designs, if possible. Ayre, Constellation, and Parasound are 3 companies that have stockpiled these jfets, and we are trying to bypass them for some new designs, because they are so difficult and expensive to purchase, today. Yet, there is no real direct substitute, and so far, our audio quality will diminish if we don't continue with these parts.
To this day, I am not fully happy with ANY IC as an ultimate amplifying component. That is just a fact, based on my experience, so far. When I have to use IC's for audio reproduction, such as in the JC-3 phono stage, I completely protect the IC's from the power line, and from glitches that might be generated in the power supply. So the IC complement might cost about $20, OEM, but the retail price is still over $2000, given the support circuitry and case costs. Not much of an advantage, is it?
  Reply With Quote
Old 16th May 2012, 11:28 AM   #23132
Bonsai is offline Bonsai  Taiwan
diyAudio Member
 
Bonsai's Avatar
 
Join Date: Jul 2003
Location: Taipei, Taiwan
When I have to use IC's for audio reproduction, such as in the JC-3 phono stage, I completely protect the IC's from the power line, and from glitches that might be generated in the power supply.""

John, I would exepct that you do exactly the same for discrete designs as well. Shunt regs, cap mulitpliers and all the other good stuff we talk about here.

WRT JFET's, yes, they are great devices, but you have also produced what you yourself have described as a less than optimal sounding amplifier that used all the magic ingredients. The same applies top other device technologies as well. Its easy to take a great component(JFET, bip or IC) and product a bad product.
__________________
bonsai
Amplifier Design and Construction for MUSIC! http://hifisonix.com/
  Reply With Quote
Old 16th May 2012, 03:41 PM   #23133
diyAudio Member
 
Join Date: Nov 2008
Location: Oakmont PA
J.C.

So far on this thread we have discussed resistor distortion. All resistors distort. Most of it is third harmonic related to tempco but some have second and a linear tempco will produce less distortion that one that is lower but swings back and forth a bit. Then there are thermo-electric effects if the end connections of the resistor are heated non uniformly. Finally there is an issue with thermal mass. Even if you use two resistors in a divider that are low tempco, if one has a much greater thermal mass than the other there can be 9th harmonic distortion.

Power supplies were another issue. There really is DC on some AC power lines and over time this decreases the capacity of power transformers. Then there is the issue of power line noise. I can't give exact numbers on it as there is no standard for a noise generator. But using input inductors can help but add voltage sag. Using transformers that couple primary to secondary less helps, as does making sure the neutral goes to the windings closest to the core. Then using dual secondaries with full bridges is the best approach to reduce noise pass through. A capacitor directly across the line also helps. Diode noise may be reduced by diode selection and RC networks. Filter capacitors have to be selected to avoid resonances. It turns out the BJTs are most influenced by noise, Fets and those glowing thingies even less.

Then I mentioned capacitor distortion and everyone told me capacitors don't distort on a sine wave. True for any waveform where Dv/Dt =-(V)T. There are two ways to reduce this. One is to add a large constant to both sides of the equation, or as you and others have mentioned is that when there is DC on the capacitor the problem seems to be reduced. The other method is to use much bigger capacitance than would seem to be called for or just run higher impedances. Then the is the issue of V dC/dT. This is not an issue due to AC voltage but rather physical movement or vibration. This can be reduced by using more massive or larger parts, mounting them so the plates are not parallel to the excitation mode and using well made parts or prestressed parts.

Now the other issue is that of ASDR or the envelope that modulates the sine waves. The A (attack) can repeat from 5 times per second to once every few seconds. It can have a rise time as fast as hundredths of a second to a fraction of a second. The amplitude of the envelope is of course quite large. Although this may have some effect on capacitor coupling, my current OPINION is that this causes problems with the bias in an AB class amplifier. The usual compensation diodes have quite different time constants and delays for thermal issues. This is of course worse in BJT designs.

So I am currently showing under "Solid State" a low cost high power amplifier that address to at least some extent these issues.

The design of such an amplifier is of course not as refined as most designs here as they tend to be either subsets of existing designs or subtle refinements. The version I am showing is not really new but just a reexamination of forgotten techniques. I am deliberately basing it on BJT's as that seems to be the hardest case and should best illuminate problems. I suspect a fully developed version will use cross coupling, Mosfets or even those glowing thingies.

ES
  Reply With Quote
Old 16th May 2012, 05:00 PM   #23134
DF96 is offline DF96  England
diyAudio Member
 
Join Date: May 2007
I was thinking about capacitor distortion the other day, from dielectric non-linearity. Perhaps only a small problem, but here goes:
Assuming a non-poled dielectric, it will be odd-order unless a DC voltage is present. DC can add a much larger even-order component, as you get Vdc*Vsig^2 instead of Vsig^3. So to reduce this you need to avoid having signal and DC voltage across the same cap. Solution is a big cap to drop the DC and a small cap to handle the AC, the two caps in series with a high value bias resistor to the junction to set the DC conditions. Note that unless you need a particular low value cap (e.g. to set a rolloff) you can just use the big cap on its own as that will have little signal voltage across it so cannot generate distortion.

At this point someone will say:
1. it won't work because . . .
2. we have been doing this for years - well maybe you have but I have never seen it written down anywhere.
  Reply With Quote
Old 16th May 2012, 05:13 PM   #23135
1audio is offline 1audio  United States
diyAudio Member
 
Join Date: Mar 2004
Location: SF Bay Area
Blog Entries: 3
Interesting insight on capacitors. Experimentally, I usually see less distortion when there is a dc voltage across a cap. Most obvious with electrolytics but also with film caps. I figured that the dc bias would reduce the effective change across the cap from the AC signal.

Conceptually the same might be true for resistors. Especially for thermal issues. If the resistor has a bias and heat already the effects of the ac would be smaller. I know higher power resistors are more likely to be lower distortion. What is the time constant of the Tc of a resistor? How to measure it?

Maybe this is all naive misunderstanding of the issues?
__________________
Demian Martin
Product Design Services
  Reply With Quote
Old 16th May 2012, 05:26 PM   #23136
diyAudio Member
 
Join Date: Nov 2008
Location: Oakmont PA
Quote:
Originally Posted by 1audio View Post
Interesting insight on capacitors. Experimentally, I usually see less distortion when there is a dc voltage across a cap. Most obvious with electrolytics but also with film caps. I figured that the dc bias would reduce the effective change across the cap from the AC signal.

Conceptually the same might be true for resistors. Especially for thermal issues. If the resistor has a bias and heat already the effects of the ac would be smaller. I know higher power resistors are more likely to be lower distortion. What is the time constant of the Tc of a resistor? How to measure it?

Maybe this is all naive misunderstanding of the issues?

Demian,

Way back I showed how the third order distortion increases as frequency decreases, so you can actually calculate a thermal time constant. More importantly you can just pick a low frequency distortion that is significantly below all others. But you know all that as you sent me some resistors to compare!

The reduction in capacitor distortion is in the actual C dv/dt + V dc/dt =dq/dt as the dq/dt shrinks as a percentage.

There may be some electrostatic firming of the plate separation except that the V dc/dt term due to V is very small usually -140db or less.

the dc/dt from vibration is much larger. There is a baseline of 5 mG of vibration or more everywhere. (G is gravity)
  Reply With Quote
Old 16th May 2012, 05:33 PM   #23137
DF96 is offline DF96  England
diyAudio Member
 
Join Date: May 2007
Electrolytics are different because they are inherently asymmetric. A priori, DC could help or hinder. Other dielectrics are symmetric, unless poled - although I guess poling could happen during manufacture.

Adding a bias will only reduce distortion if it moves the device into a more linear part of its characteristic. This works for active devices, but for most passive devices I would expect either no change or things to get worse.
  Reply With Quote
Old 16th May 2012, 05:59 PM   #23138
diyAudio Member
 
Join Date: Nov 2008
Location: Oakmont PA
Quote:
Originally Posted by DF96 View Post
Electrolytics are different because they are inherently asymmetric. A priori, DC could help or hinder. Other dielectrics are symmetric, unless poled - although I guess poling could happen during manufacture.

Adding a bias will only reduce distortion if it moves the device into a more linear part of its characteristic. This works for active devices, but for most passive devices I would expect either no change or things to get worse.
It increases the 2nd harmonic distortion in resistors, but moves capacitors on non-sinusodal signals to a more linear part of the curve.
  Reply With Quote
Old 16th May 2012, 07:32 PM   #23139
DF96 is offline DF96  England
diyAudio Member
 
Join Date: May 2007
Why does the nature of the signal affect whether the cap curve is linear or not?
  Reply With Quote
Old 16th May 2012, 08:45 PM   #23140
bcarso is offline bcarso  United States
diyAudio Member
 
Join Date: May 2005
Location: Canoga Park, California
Hofer wrote in 2010 and again in 2012 essentially identically (from a pdf of an AES presentation, available I believe from Ap): "Resistor non-linearity is usually best modeled as having a simple dependence upon the absolute value of voltage: R(Vs) = Ro*(1 +/- kb * |Vs|)..." He adds: "Resistors also exhibit thermal modulation caused by the instanstaneous signel-dependent power dissipation and non-zero temperature coefficient (TC)... Non-linearity from thermal modulation can far exceed voltage coefficient effects."

I know that this was brought home a few years ago when there was a stubborn excess of distortion at low frequencies in a new Ap instrument, where surface mount resistors were being used. Duke Aguiar related showing Bruce the effect on the distortion of putting his finger on one of the resistors.

Hofer goes on: "Capacitors also have voltage coefficient effects, but these are much more difficult to characterize and model due to their inherent frequency dependence ---Some capacitors also exhibit a non-linearity related to current."

I recall Bateman said a single-package bipolar electrolytic had the lowest distortion, beating significantly a biased electrolytic and a biased series pair. This was reiterated by Jensen Transformers in an article they published on their site.
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off



New To Site? Need Help?

All times are GMT. The time now is 08:45 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2