The simplistic Salas low voltage shunt regulator

[Salas]

Good news. Don't give up on the follower transistor yet, if it can fit in steadily, it boosts the performance. You may do a before and after listening test and decide better when finished. Can you upload a shot of your interesting total configuration?

 

[analog_sa]

Normally the R core gives better mains isolation due to completely separate primary and secondary

It seems that's rarely the case. From what i understand most R-cores are wound in a balanced manner.

 

[AndrewT]

That's what I have read when the manufacturer divulges his trade secret.

 

[pietjers]

A while ago I compared an R-core transformer with some toroid transformers in an experimental tube preamp project. Although R-cores have much lower capacitance between primary and secondary, there was not a big difference in sound quality between the R-core and a toroid of about the same VA rating. But there was a distinct difference when I compared bigger VA ratings. The bigger the transformer the better the sound! And this was not by a small margin. This pre was never built but I ended then with a 500VA (230V sec.) toroid with fast rectifiers and only about 50uF of smoothing capacitance before the voltage regulator and it sounded wonderful!

Peter

 

[mikelm]

Hi Peter,

Any theories about why the bigger transformers sound better ?

Does this extend to Power Amplifiers also ?

My only comparison of transformers was between similar specified torroidal & EI transformers when developing a power amplifier. They were about 200VA & 18VAC + 18VAC. In this A/B comparison the EI transformer sounded MUCH better.

These days I always used comprehensive filtration in all legs after the bridge and this makes the different types of transformer almost indistinguishable but I am interested to understand your findings more.

 

[pietjers]

Hi Mikelm,

Yes this also extends to power amplifiers. In power amplifiers it is better beef up your transformer than to enlarge smoothing capacitance (generally). I can not proof that with theory, but I think that it is an advantage to have fast charging times. So: big transformers, fast (soft recovery) rectifiers and just enough capacitance.

Peter

 

[mikelm]

Well, Peter that is an interesting perspective which I will definitely check out . . . . but it makes be laugh a little bit to hear this because it is so different than that which I would expect to be the case but I know from experience that I can believe your advice.

My personal PSU philosophy is more along the lines of " if the type of transformer can make a difference . . . . then the post transformer line regulation measures to filter out mains noise / diode switching / secondaries resonance etc . . . just aren't working well enough ".

Funny we should chat again today because only yesterday I was trying to implement a LM285Z-2.5 that you recommended to me when we were discussing LDR volume controls. Only in this case I could not get a good result from it - in fact in this case - a green LED gave much better measured result. I will PM to you - perhaps you can tell me if I made an error.

thx

mike

 

[RCruz]

. But there was a distinct difference when I compared bigger VA ratings. The bigger the transformer the better the sound! And this was not by a small margin. Peter

I do agree with you.... That corresponds to my findings

 

[orelli]

+1

 

[orelli]

also paralleling two secondaries help the sound (of the same voltage, of course)

 

[RCruz]

also paralleling two secondaries help the sound (of the same voltage, of course)

That way you double power output so you get double VA

 

[orelli]

Of course. I'm very agree with you an Pietjers; big txs are more important than big capacitance. My very personal theory about big txs, is that big VA have little resistance and
lower impedance. Like in the shunt reg, lower is the out impedance, better is.

 

[nicoch58]

Of course. I'm very agree with you an Pietjers; big txs are more important than big capacitance. My very personal theory about big txs, is that big VA have little resistance and
lower impedance. Like in the shunt reg, lower is the out impedance, better is.

is a fact

come sta' il Liga ?

 

[Salas]

Watch the loop area of the before reg cabling to be tight too.

 

[orelli]

@ nicoch.Sta una meraviglia . I listened his studio monitor and I made the RPG like diffusors for him.
@ Salas. Sorry , I don't understood.What's the loop area? Do you intend the cabling between tx,diodes bridges and filtering caps?

 

[AndrewT]

What's the loop area?

twist cable pairs to minimise loop area and balance radiation influence.

 

[orelli]

Thanks Andrew.

 

[AndrewT]

As a follow up:
Electricity can only flow in circuits. What goes out, MUST come back in.

In our circuits there is always a return route for a flow route.
Those flow and return make up a pair. That is the pair that should be low loop area.
It applies to mains cables. It applies to LV AC supplies from a transformer. It applies to LV DC supplies from rectifiers. It applies to signals, whether single ended or balanced. It applies to every circuit where electric current flows.

 

[orelli]

Very clear.

 

[Kazuo Ozawa]

Zobel of Reflektor +3.3V

1R 0.47u is quite different to listen to for instance. If stable on the scope as the other, you can see where it goes, if there is something to gain or lose in lower values or to stay where it was. Like a recognition shot.

After long journey I reached the goal to serch for Zobel of Reflektor +3.3V.
The story is following.(Of course no oscillation on all cases)

1) 2R2+2.2uF
Not bad, no harsh, but sound dim.

2) 2R2+0.22uF
Become brilliant, but harsh.

3) 0R47+0.22uF
Become decent, but still a little harsh.

4) 0R47+0.44uF(0.22u//0.22u)
Become great deep sound, no more harsh.
I also guess that it is very good for 1R0+0.47u as Salas recommends.

On this experiment I found a few principles about Zobel.
1 Zobel is strongly related to quality of sound.
2 If Zobel is not been optimized, sound become harsh or dim.
3 The optimal value is between the narrower area than I expected.

Next step, I build Reflektor +1.2V. If I get good result, will report.

Best regard