Good Regulator for 7 - 8ma load

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Trans,
thanks for the post18 link. I had missed that.

Your redraw of the Jung Shunt Reg omits the sensing "Bridge". This is important for good regulation. The remote sensing terminals are also good for regulation at the load.

For those reading the post go to the Jung link to see correct implementation.
 

iko

Ex-Moderator
Joined 2008
The shunt reg can safely pass another 10mA to ground continuously, think of it as reserve current ready to be given to your load when it needs it. So you can set it up for a total current draw of about 20mA. It will have really low output impedance, and outstanding psrr. Maybe it's over the top for your need, but you're bound to learn some good things building it. Then again, you can just use an LM chip and call it a day. You might never be able to hear a difference, because the regulator might not be the weakest link in your chain. They're all suggestions, and you can learn a lot by implementing two or three and comparing them. Beauty of diy :) sky's the limit. Well, and what the wife permits.
 
Thanks all for your continue comments and suggestions. I'll be ordering Linear Audio vol. 4 to check out the regulator bake off. I found a few of the performance plots online at tech-diy.com or something like that. Looks like the Jung performed the best, though obviously I haven't yet read about the listening folks did.

Yes this is likely way overkill for my application. I probably would be just fine without a regulator at all. Or was also thinking about an a basic LC arrangement as that is supposed to prove reasonable regulation as well, "if target inductance is achieved.".

But I think I'm finally settling on an attempt at implementing the Jung design just because. Though I'm interested if someone can describe the benefit of running the Jung as a shunt regulator instead of how its is normally described and configured?

Is it that my idle current is to low to achieve the best of what the regulator has to offer?

-TJ
 
Member
Joined 2011
Paid Member
I'm interested if someone can describe the benefit of running the Jung as a shunt regulator instead of how its is normally described and configured?
Jung himself attempts to do so, on page 5 of the audioXpress article linked in post #18. For example,
... made a valuable point related to power supplies, which is that a shunt-type regulator localizes audio dynamic currents. Going further on this theme, in contrast, a series-type regulator returns dynamic audio currents all the way back to the raw DC supply! The latter type uses a long-series loop, as opposed to the short loop of a shunt. For audio, there are some profound implications to these differences ...

... In stark contrast (any) shunt-type regulator provides inherent isolation between the audio stages and a dirty raw supply, by virtue of a high-z series input leg consisting of either a current source (or a choke), and a shunt regulator stage that regulates its own terminal voltage.
and many paragraphs in between
 
I'm curious what you mean by higher harmonics. I'm not challenging the claim, just want to know what you mean.

A couple months after the article was published one correspondent suggested that we take a look at the harmonics if a 1kHz (for example)current was impressed upon the power supply rail. Mirabile dictu, the rankings in the listening test seemed to correlate better with the harmonic content than PSRR, noise or Zout.

Anyone can do it with a decent sound card and fft program.

My guess is that the error amplifier in the commonly seen shunt designs are less than perfect...
 
With load of just 8 mA you can place a 22 - 100 Ohm resitor along with a 330 - 1000 uF capacitor to form a RC low - pass filter on the output, which will virtually eliminate any noise.
Like this :

|''''''''''''''''''''''|
| 7815 |------[|100R|]-------------------OUT
|.............| __|__
| _____ 1000uF
| |
| |
GND GND
 
Same goes for ripple rejection...an RC filter causes roll-off at all frequencies above fc = 1/2pi*R*C. The higher the C and the R, the faster the roll-off...
I think I can relate this to a mechanical setting...
Think of the output voltage as a crank or a wheel set to a determined position. Any ripple or noise causes the crank to slightly wobble to left and to right. Now, the higher the C in the RC filter, the bigger the crank is, so "wobbles" have less impact on it's position - a big C responds slowly to voltage change. The higher the R, the harder is the crank to rotate, which suppresses the wobbling further. But, the higher the R, the harder it is to set the output voltage (rotate the crank to a position, or charge the output capacitor), and output current - which goes for output impedance. Same thing goes for the crank size - a crank with higher diameter rotates easily, or the bigger C size gives less output impedance - but only until it discharges, or the crank goes to full position. It's a crude way to explain, but that's how I got the idea after some thinking:)
 
Guys sorry I dropped off this thread for a bit. I think I'm going with the Jung shunt. Downloaded a bunch of articles I need to read first though. What would folks suggest as a first step. Identify the parts or should I go with what Jung recommends in his last article not on the shunt version but the standard Jung Super Reg.

Just acquired an HP 35665A with full feature set enabled from John Bau of Precision Services. Excellent price...so have a bunch of reading to do on understanding how to use the FFT to look at different aspects of the power supply and my amps.
 
Guys sorry I dropped off this thread for a bit. I think I'm going with the Jung shunt. Downloaded a bunch of articles I need to read first though. What would folks suggest as a first step. Identify the parts or should I go with what Jung recommends in his last article not on the shunt version but the standard Jung Super Reg.

It's easy enough to modify the series regulator to the shunt. I don't think you can make an a priori decision -- try both and see which actually sounds better.
 
Member
Joined 2011
Paid Member
I'd suggest you arrange the circuit (shunt or series) so there's at least 3 mA of dc current flowing into the opamp's output pin; this (1) keeps you away from opamp crossover nonlinearities, and (2) operates the Jung level shifter at a reasonable current.

In the Jung AX interview article's schematic, (VBE / RB) sets the dc current flowing into the opamp output, and D8+C5 perform the level shift function.

Since your output current demand is so low, you could install a lower power, higher fT pass transistor like the 2SA1930 in the QPNP position. Whereas Jung's D45H11 specs are 10 amps & 70 MHz, the 2SA1930 is spec'd at 2 amps and 200 MHz.
 

Attachments

  • jung_shunt_AX.png
    jung_shunt_AX.png
    25.3 KB · Views: 139
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.