The simplistic Salas low voltage shunt regulator

[Salas]

Originally Posted by jkeny

Am I right in thinking that LEDs + diode can be substituted for the zeners? This seems to be more flexible, less noise & less hassle, if true!

If the zener isn't low noise, by all means yes.

[ikoflexer]

Originally Posted by jkeny

Am I right in thinking that LEDs + diode can be substituted for the zeners? This seems to be more flexible, less noise & less hassle, if true!

For low voltage out, why not. It'll do great. If I recall you wanted 3.3V out? The zener won't be your problem, that can be easily replaced with a LED and a diode, or two 1.7V LEDs. Your problem will be with the shunt mosfet, since 3.3V is right in the Vgs threshold. Start looking for for mosfet with low Vgs threshold. I found some NEC and Fuji mosfet with low Vgs threshold (1-2V) in some old CRT displays. On top of that you need them p-channel... I don't know. Can our DAC work with higher than 3.3V?

[disco]

Originally Posted by Salas

No, hopping on my chair. But we don't have the proper smiley.

Anyway, you are right, that is why I enjoyed the sarcasm. Too much trash online. We have all got our share.

Ah, hopping. Clear
Still waiting for those A94 to arrive. It's only four weeks now, why complain?

[Salas]

Originally Posted by ikoflexer

For low voltage out, why not. It'll do great. If I recall you wanted 3.3V out? The zener won't be your problem, that can be easily replaced with a LED and a diode, or two 1.7V LEDs. Your problem will be with the shunt mosfet, since 3.3V is right in the Vgs threshold. Start looking for for mosfet with low Vgs threshold. I found some NEC and Fuji mosfet with low Vgs threshold (1-2V) in some old CRT displays. On top of that you need them p-channel... I don't know. Can our DAC work with higher than 3.3V?

I would make a robust 10V shunt and then put an RC filter on the output, directly on dac, claculated to end up at 3v3 with the dac draw. If the dac does not demand better than a large good cap's output impedance, it will be beneficial decoupling and noise kill.

[jkeny]

Why does email notification not work sometimes, I got no notification of any posts from this page!

Anyway, thanks for the replies, all. My point is that there seems to be so many issues when choosing zeners - fakes, noise, getting the correct value, etc that substituting LEDs for the zeners seems like a good approach. Do you have to be careful with choosing LEDs - are they noisy also?

I would make a robust 10V shunt and then put an RC filter on the output, directly on dac, claculated to end up at 3v3 with the dac draw. If the dac does not demand better than a large good cap's output impedance, it will be beneficial decoupling and noise kill.

Good idea, but I'm not confident enough to rely on an RC to knock down the V based on current draw - I don't know the DAC's current draw & the 3.3V goes straight to the PS pins of a EZUSB-FX2 IC & Xilinx FPGA IC which I'm afraid I would kill with over-voltage!

I intended to try this for a 12V Tripath TA2020 amp which also could use a 5V supply, so I might just try the 5V supply first - no special shunt mosfet considerations here. What about the LEDs - what's the determining factor here - current draw? Again, I dont't know the current requirements for the 5V supply of the Tripath TA2020 & it's not in the datasheet as it's normally supplied internally from pin 30 & routed to pins 2 & 8.

[Salas]

Generic yellow, green and red leds are very low noise. No issues of erraticallly noisy samples.

[jkeny]

Originally Posted by Salas

Generic yellow, green and red leds are very low noise. No issues of erraticallly noisy samples.

Great, so why aren't these the preferred choice rather than zeners? Are there other issues ?

[Salas]

From the start in the phono thread I have told people that the Vref is anybody's choice. It was that I had good zeners and I prefered them for 36V since stringing many leds up to there could even top a single good zener. It kept it simple, no many leds matching, compact, and Vout predictable value. Now that I see that bad zeners are about, as Rcruz got ''woosh'' I remind you that Vref is down to your choice and testing. In James's board there was a connector instead of a Zener so to put whatever there, as a subsystem. Its that not everybody has followed thoroughly such long threads, naturally. We have talked, leds, resistors, tempco hard to get zeners, diodes, everything really.

P.S. Nobody made it on James's board. You should, because its the best in provisions and grounding, plus it flips to -Vout too. Has very intelligent pads for reversing the elements.

[jkeny]

Sorry, I didn't read that far back - I can see the number of LEDs being cumbersome & a hassle for maybe above 12V, OK. So what are the low noise zeners you use? What would be considered a low noise figure if looking up datasheets? Something available from sources, like Mouser, should be trusted, I guess?

[Salas]

Best are tempco DO-7 1N46XX zeners from Microsemi. They are labeled low noise. Quoting 1uV to 5uV up to 6.2V. They don't sell them online, at least one by one. Then the 500mW glass bodied red color 1N47XXA ones are good enough. Those can be found. But up to 10-15V I would go Leds. As I did with the DCB1.
For 25V to 38V simplistic phonos, one good zener is tempting, but as you saw Rcruz got one so so and a wooshing one, plastic cased. A large filter cap across, filters much though. And there is TL431 in the end, that goes up to 36V.