The simplistic Salas low voltage shunt regulator

Status
Not open for further replies.
Jump to Latest
#6,762
PS for Pi?

Not sure if this is appropriate for this thread or if I should start a new one. My question concerns using the low voltage positive supply section of the black board for a linear power supply for a Raspberry Pi/DAC combo. The Pi recommends a 2A PS and the DAC sits on top of it and gets its supply from the Pi, so I am thinking maybe 3A max @ 5V. BJT or IFR? Any other issues to consider? I already have board (I used other two parts for +/- 25V.)
Thanks
 
#6,764
Salas said:
BIB as is is not reliable for >1A. You would need slower big MOSFETS like IRFP9230/230 on big sinks to run 3A. But from what I see here https://learn.adafruit.com/embedded-linux-board-comparison/power-usage the Pi demands are lower. Measure your DAC demands also. A 0.1R in series with its now PSU feed could indicate by voltage drop across it and Ohm's law.
Click to expand...
I have the Pi2 which is a quad core @ higher clock rate and more RAM than A/B and I have read reports that it "stumbles" on less than 1800mA (without the DAC) so I guess I'll have to look elsewhere.
Thansk
 
#6,765
For Vref, I finally decided put a 1,6k resistor bypassed with a 150uF Rifa PEG124. I will listen to the result in a few days.

My CCS LEDs were all burned when I tried turning on the reg today and I had no idea why. I replaced them and everything is fine now.
 
#6,771
As a very-latecomer to this thread, could someone point me to posts in which the noise performance may be discussed? Thanks in advance.

Clearly, if everything is working properly and in the absence of parasitic oscillations, the zener noise should dominate.

Brad
 
#6,773
Last edited:
#6,774
LEDs as references

One remark: Many LEDs have superbly low noise. They also have a significant temperature coefficient of the forward voltage Vf, which for some is a decent match to the Vbe of a silicon bipolar. This has been exploited by many in current generators, with a resistor in the emitter and the base tied to the LED, with the resulting collector current fairly constant with collector voltage and temperature. However the noise in the current is limited by the current noise in the emitter resistor, along with the shot noise in the base current and the thermal voltage noise of the base spreading resistance. The delta V, Vf -Vbe, is a bit low for best noise performance.

The quietest voltage reference I've found to date is when it is derived from a cascoded or multiple-cascoded JFET arrangement as a ~zero-tempco current generator. The reference voltage can then be developed across a parallel R-C and buffered. The low-frequency excess noise in particular can be exceedingly low, much better than any commercial voltage reference.

The somewhat-impractical aspect of such a reference is due to the wide parameter variations in JFETs, which require one to test a sample and make adjustments for low tempco.
 
#6,779
Hi folks,

I've laid out a 70x36mm SSLV 1.1 PCB for 5V@20-40mA duty (~200mA CCS). Would appreciate another pair of eyes on it. I had to settle for 4.7uF MKS2 + 100nF MKP1837 for the Zobel due to size limits. The references are bypassed with 1uF MKP416.

Thanks.
 

Attachments

  • SSLVMini.png
    SSLVMini.png
    140.4 KB · Views: 312
#6,780
The loop areas aren't tight and long distance bypassing isn't productive, but it should work electrically. As for correct routing I am not sure, can't be following it in detail with a fast look. Just check everything twice.

P.S. Where are the Leds for CCS? Using one Vbe transitor current control voltage?
 
Status
Not open for further replies.
Top Bottom