My take on a discrete shunt voltage regulator

[HiFiNutNut]

Here is a picture of what I have built:

[ikoflexer]

hifinutnut, have a look at what John Curl had to say about the capacitor used in Vref and the output capacitor. You may find this useful.

link to post #2066

[HiFiNutNut]

Originally Posted by ikoflexer

hifinutnut, have a look at what John Curl had to say about the capacitor used in Vref and the output capacitor. You may find this useful.

link to post #2066

For me the Vref cap would not cause any problem like the output cap could. In v1 and v2 I see it providing a AC path as the input to the regulator.

He uses a 0.1uF at the output of the regulator.

I remember that you mentioned before that your regulator does not like a high Q film capacitor at the output. Have you tried it?

[Bonsai]

"He even provided a MOD to his circuit so that for those who stubornly insisted on using local bypass could use it without having the resonance."

Hifinut, I did quite a bit of simulation on the Jung SR a few years ago and came to the conclusion that although it may have superb noise and transient response performance, it quickly suffers from phase margin problems when the output is loaded with a cap. I thing a simple series reg followed by a shunt is the way to go.

[HiFiNutNut]

Thanks, Bonsai. My subjective experience matches your simulation.

[HiFiNutNut]

I replaced the output electrolytic capacitor with 0.5R + 4 x 2.2uF (MKT) ||. The sound further improved. The difference is not small.

Tomorrow, I will try 0.5R + 1.1uF (MKP). If it does not work, I will increase the uF until it works, otherwise,, I will reduce the R to 0.2R. If it still works, I will remove the R. If it still works (I don't think so), it will stay that way. The cap ESR is only 0.003R. Some of these will first be tested with no load because I worry about a severe resonance may kill the expensive OPA627 (or not?)

[HiFiNutNut]

I will have 2-3 v2 regs for my XO/EQ circuits housed within the same chasis.

For raw DC, one transformer, one bridge (shottky diodes rated at 3A), 3300uF || 0.1uF ceramic, then for each regulator, 5.1R output. In front of each regulator, 2200uF || 0.1uF MKP infront of CCS-shunt. This is effectively a CRC for each regulator but the first C is shared, and the last C is installed on the regulator board (instead of the raw DC board) to provide shorter AC loop.

What do you think?

[HiFiNutNut]

Ikoflexer, Could you explain the purpose of C3 in your schematic? I am wondering if it has a sonic signature on the regulator or not.

[HiFiNutNut]

I have just done the 1.1uF (Vishay Blue box MKP) + 0.5R output cap experiment.

I did not expect any difference from the MKT but the result was otherwise.

The regulator now sounds as good as I can hope for. It is very HIGH-END! I am not sure if I can find any fault with it. So the 1.1uF stays. I may still push it to the limit by trying different Rs to get the optimal value.

[thanh1973]

Good Work
I have some wima mkp10 2.2uF and 10uF caps laying around. Do you think that upping the value will be better or worse or no different?

I have one other idea. Have you tried putting a silver mica cap in parrallel with this cap.
http://australia.rs-online.com/web/s...r+mica&x=0&y=0