The simplistic Salas low voltage shunt regulator

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Sorry i used a bc557.

In the meantime i have fixed the prereg, and replaced the bc557 with a branded bc560. The intermittent flickering stopped (for the moment, i havent done an endurance test). The bc560, like the bc550 yesterday gets verrrry hot... I burnt my finger on it!

Does the bc560 usually require a heatsink, o r do you think there is an error? Didnt see anything funny on the scope now....
 
diyAudio Chief Moderator
Joined 2002
Paid Member
All they can have across is the Vo-shunt Mosfet's VGS. All they can have through is the U$23 IDSS. Thus I would expect ~10mW in your (-)15V one if normal. They still cope without mini-sinks near 50V in some of my phono builds but with 2-3mA IDSS cascoding Jfets (V1.2R). What are your shunt Mosfets there?
 
Sorry i used a bc557.

In the meantime i have fixed the prereg, and replaced the bc557 with a branded bc560. The intermittent flickering stopped (for the moment, i havent done an endurance test). The bc560, like the bc550 yesterday gets verrrry hot... I burnt my finger on it!

Does the bc560 usually require a heatsink, o r do you think there is an error? Didnt see anything funny on the scope now....

Mine is V1.0, BC560 in my circuit is not hot. Vin is -21V, Vout -11.4V, Iout about 30 - 40 mA.
I don't know if this situatioin apply to your circuit as yours is BiB.
After the cicuit is on, Vout is -12.04, after about 5 minutes, it settle to -11.4V and never drifts again even after used more than 2 hours.

Regards
 
All they can have across is the Vo-shunt Mosfet's VGS. All they can have through is the U$23 IDSS. Thus I would expect ~10mW in your (-)15V one if normal. They still cope without mini-sinks near 50V in some of my phono builds but with 2-3mA IDSS cascoding Jfets (V1.2R). What are your shunt Mosfets there?

MOSFET = IRF620

But... I am using 2SK170 as U$23! That's then probably where the problem is, isn't it? (is K170 typically higher IDSS than K117?)

If I replace the U$23, I would probably have to replace the K170 on the other side as well, no? (sorry I don't fully understand the function of that one).
 
diyAudio Chief Moderator
Joined 2002
Paid Member
The multiplier is a filter, does not regulate anything. No matter how much harmonic noise it will cut, the intrinsic wide band noise of the regulator's output will be the same. The regulator itself has a CCS with enough rejection ratio for ripple and its error amp to regulate, not to find significant benefit beyond gross input ripple levels. The filters are good for eating noise to absolute minimum when situated last in the chain before very sensitive high gain low signal no PSRR circuits. Their Achilles' heel is high impedance.
 
diyAudio Chief Moderator
Joined 2002
Paid Member
VGS MOSFET should have been in the Volts region. It seems to me that the current follower is not open and the small BJT provides all the current it can to the load, that is why it gets hot. Can you confirm MOSFET orientation and PCB connections to original schematic please?
 

Attachments

  • (-)dtl.png
    (-)dtl.png
    26.2 KB · Views: 437
i v e tried cap multiplier with bjt s (though i have not tried 2n4401 and the likes) and in the end the music sounded a little dry. that effect was reduced a lot with a mosfet cap multiplier. Still the most analogue sound is with CLC type filters, L being air core and high gauge and C a combination of cap in multiple stages.
 
VGS MOSFET should have been in the Volts region. It seems to me that the current follower is not open and the small BJT provides all the current it can to the load, that is why it gets hot. Can you confirm MOSFET orientation and PCB connections to original schematic please?

You got it. The poor BC was doing everything alone, I had wired up another JFET wrongly! The heat/stability problem is fixed now!

When measuring aroung, I noticed that both the negative regs have much more noise than they should have though (0.4mVpp).... strange !
 
diyAudio Chief Moderator
Joined 2002
Paid Member
That's nice it got solved. :up:
When measuring noise with scope do a check with the crock end shorted to probe's nose. See what it does as a base line in the specific measurement layout and environment-time. Use 1X probe setting only. If its a DSO engage its bandwidth limit option also. That way you will know more of what's base and what's added. In general the NPN and NMOS semis are happier for self noise than their P counterparts.
 
Made some pictures, maybe this rings a bell?

This is the -15v reg with the tda1541a as load. Dont mind the small spikes (the fuzziness around the main shape), these are due to the clock... Scope setting is 20 mV/div
An externally hosted image should be here but it was not working when we last tested it.


This is the same reg, unloaded (the clock is no load on this reg)
An externally hosted image should be here but it was not working when we last tested it.


And here is the same reg with a 200r load, output voltage is still ok.
An externally hosted image should be here but it was not working when we last tested it.


And here, the same reg loaded by a 100r resistor (which draws more than thr ccs). Max Output voltage is now obviously less... 13v.
An externally hosted image should be here but it was not working when we last tested it.


Any ideas?
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.