The simplistic Salas low voltage shunt regulator

[albertob]

Hi friends,

I adjusted the shunt reg. as you said. He had a bad connection problem in Q1, that's why so hot Q4. It is fixed.

When you turn 24.3 V and I have a minute I have 24V. In the next half hour does not change. Is the maximum time I've tested.

I placed a load resistance of 390 ohms to simulate a consumption of 60 mA (2 x B1). Q4 does not heat anything. Q1 warms a bit. I placed a small sink.

I wonder if everything is correct before you try it with my B1. Here is the final scheme and measures to see if everything is OK.

An externally hosted image should be here but it was not working when we last tested it.

An externally hosted image should be here but it was not working when we last tested it.

Best regards!

 

[Salas]

Looks OK now. With 390R load q4 was not warm because you run about 74mA and you spent 61mA on the load, so very little for it to burn. You can use maybe 22R instead of 27R1 so to have a bit more current for q4 when with b1, it helps subjective quality. Place a small sink on q4 too.

 

[merlin el mago]

Congratulations Albertob, welcome to the club, let us know your subjective opinions about how works?

 

[albertob]

Thanks for the help.
I will try the B1 and see differences with the LM7805.

 

[fred_com]

Today I did experiment on simple man's 1.2 dead bug, so someone can practice easier and learn.

Hi Salas, how to set Vout in this reg? C3 is ceramic NP0 as in previous versions? And of what quality should be non-polars?

Thank you.

 

[Salas]

You must be more diy wise with this one so to keep it minimum parts. So you measure J2's IDSS with a 9V battery, then you roughly think of it at 70% because it will lose mA under 0.6V only at circuit's position. Then you multiply that by some R9 resistor value that gives your target Vout -0.6V. More or less its in the ballpark. After you measure your final Vout you can adjust the resistor's value more surely. Use 1W low ppm there.
The C3 in the experiment is silver mica I had handy, but yes NP0 are great too.
Those NP you see are Lelon major bad. Maybe those green Nichicons I see on parts vendors are better. Can be polar too.

 

[merlin el mago]

Mind you 1.2 needs good layout practices. Tight. If it is done wrong it will show 250mV 12.5MHz oscillation. You will understand something is wrong even if without a scope because it is going to choke the music in the phono, will sound distant with hum everywhere in the spectrum. If the signal is as strong and clean as before without new noises, its a success.

Ok will wait until someone can offer/sell/trade me a good pcb layout for v1.2

 

[pidesd]

is it possible to easily have a bjt instead of jfet cascode in the schematic at post number 2722? what would be the resistors value if it s doable?

 

[Salas]

Yes easy. 1 bc550, 1k emitter resistor, a 2V green led, and 1k2 resistor for it. If you want cascode bjt then 2 bc550 and 2 620R for the LED feeding bases from midpoints is the changes. You know the schematic.

 

[pidesd]

thanks will try it

 

[pidesd]

in sslv1.2, what is r10 for Vout=26v? i have 10k right now. without load everything seems fine but with load it idles at 8v. i ve look at pretty every transistor and they look fine. i m clueless.

 

[pidesd]

also,in the schematic at post 2722, is it possible to easily have 1.2v Vout? maybe by eliminating the led?

 

[Salas]

in sslv1.2, what is r10 for Vout=26v? i have 10k right now. without load everything seems fine but with load it idles at 8v. i ve look at pretty every transistor and they look fine. i m clueless.

The resistor develops voltage by the amount of current that runs through it. See why you lose current. You don't give info about how you got the jfet underneath. With trimmer? I would use 27k so to ask for less current at 26V. Idling low when on load smells oscillation when connected to it by the way.

 

[Salas]

also,in the schematic at post 2722, is it possible to easily have 1.2v Vout? maybe by eliminating the led?

Limited. BD140 needs Vbe 1V alone. Maybe with D45H8.

 

[pidesd]

Limited. BD140 needs Vbe 1V alone. Maybe with D45H8.

d45h8 has Vbe of .5V i think. dont know if it s enough...

 

[pidesd]

The resistor develops voltage by the amount of current that runs through it. See why you lose current. You don't give info about how you got the jfet underneath. With trimmer? I would use 27k so to ask for less current at 26V. Idling low when on load smells oscillation when connected to it by the way.

sorry i meant r6 with the jfet cascode. should it be 4.3k or 10k? the rest of my part list is what is recommended for up to 22v out on page 190. by the way just checked my zobel and it s 10uf + 1 ohm...should correct that first.

 

[pidesd]

ok zobel was fine. didnt notice 2 smd were sandwiched. vref cap=10uf.

symptoms: when i power up with load, Vout=8v. slowly but steadily(3 minutes), the voltage rises to around 17v where it rises suddenly to my desired 26v. weird. any clue?

 

[Salas]

Initial oscillation, backs up with thermal rise. Check your values and layout, show photo. R6 even 10k fine.

 

[pidesd]

i just tightened my layout but nothing changed.

fortunately, i m holding a BBQ party friday ( you are all invited of course ) and i ll be able to take a couple of pictures from someone s camera. you certainly deserves a handful of beautiful pictures, salas

 

[albertob]

Hi friends,

I adjusted the shunt reg. as you said. He had a bad connection problem in Q1, that's why so hot Q4. It is fixed.

When you turn 24.3 V and I have a minute I have 24V. In the next half hour does not change. Is the maximum time I've tested.

I placed a load resistance of 390 ohms to simulate a consumption of 60 mA (2 x B1). Q4 does not heat anything. Q1 warms a bit. I placed a small sink.

I wonder if everything is correct before you try it with my B1. Here is the final scheme and measures to see if everything is OK.

An externally hosted image should be here but it was not working when we last tested it.

An externally hosted image should be here but it was not working when we last tested it.

Best regards!

Hi,
To continue with the shunt voltage regulator ...
The tests made with an input capacitor (after the bridge rectifier) of 2200uF, capacitor unbranded green, and adjustable output (trimmer) to more than 30V. Set to 24 volts and plugged it into my B1.

- Less noise, more warmth, better midwoofer and timbre in the voices.
- Loss of brightness in the treble and a little less space in the sound.

Overall a smooth and sound with a little less bright. With a sound Lm8718 is hard and rough but more detailed.

I mounted the circuit in a welded plate with an input capacitor of 4600uF Nichicon. The output now reaches beyond 21V (adjusting the trimmer), is that correct?.
I adjusted to 18 volts and ...

He has won the detail and luster it had lost earlier. A gentle and relaxing sound in detail, without fanfare. I recovered the spatiality and say who has won presence. The depth of the soundstage is now a little older.

I'm happy with the result.

Some questions:
Is this normal controlled the descent of volts?, 5k trimmer with no more than 21V up to B1 connected.
Q4 does not heat anything, almost cold, however Q1 gets quite hot and has 2 sink. R1 is 22.7 Ohm/2w
If I measure across jfet (after Trimmer) is 0V. between S and G of Q4 is also 0V.
If you logged out - to gnd is an unbearable humming. If I connect in -B1 from the buzz is minimal, only hitting the ear listening to the speaker. I do not understand, the difference is only 20 cm connection cable (unshielded).

The input transformer is 18-0-18. I connected the two poles (18V)of the bridge rectifier and 0V pole is free. The first capacitor is 4600uF Nichicon. C1 and C2 is Panasonic 150 Panasonic UF (not 100uF). 470uF C4 is known unbranded and 0.1 uF capacitor C4 I deleted.

Excuse my English

Greetings and thanks to all