Tangent Young-Jung power supply

[tangent]

Have you been cleaning the board?

[schubert]

Originally Posted by tangent

Have you been cleaning the board?

Well, no, I hadn't been, though it was on my list of things to try. So I gave it a thorough scrubbing, but there was no change.

I measured all the voltages I could think to measure and put them on the schematic. I included voltages I had measured with no op amp installed. I also included a few I had made with the op amp installed but not powered - I temporarily replaced Q2 with a new part (no change) and when I reinstalled the old one I didn't get the pin soldered to the top trace that led to op amp pin 7, so the op amp had no power for a while (it was soldered to the bottom trace, though). I've looked over this quite a bit, but without knowing the voltages for a healthy unit it's hard for me to try to pinpoint the problem.
YJPS schematic p2.JPG
If nothing sticks out as being obviously of culprit nature I'll start pulling parts one at a time. One thing I noticed, which may or may not mean anything - when first plugged in D2 stays lit for about as long as it would take to charge a capacitor. Could a capacitor be causing problems? Perhaps I'll replace C8-11 one at a time.

I realize all these numbers don't exactly add up right - my meter is no great shakes.

[sachu]

I could really use a board for a power supply build for my Scott Nixon USB DAC. I want to move away from using the SLA batteries and put in a nice solid power supply. This seems perfect.

An idea on the parts cost for this project?

Can you get a 5 V output from this design?

[Beefy]

I don't think that this is available yet. Have you considered a σ11?

[sachu]

Originally Posted by Beefy

I don't think that this is available yet. Have you considered a σ11?

I have..i just don't think the sigma11 is good enough for a digital supply.

[Beefy]

Originally Posted by sachu

I have..i just don't think the sigma11 is good enough for a digital supply.

Why not? Tangent reports 7µV noise. AMB reports less than 12µV. Both are far superior to commercial designs.

[sachu]

Originally Posted by Beefy

Why not? Tangent reports 7µV noise. AMB reports less than 12µV. Both are far superior to commercial designs.

I was thinking more on the lines of using a programmable shunt regulator on the likes of TL431, a transistor, a few resistors and coupling caps to get a good enough PSRR . I need a rock steady PSU at 5V with very low noise as this is the supply to the receiver and the interpolation filter on my DAC...


I understand that the error amplifier in the sigma11 is basically what is governing the PSRR.The other supply I am considering is COlin's regulator that is being discussed here.

I don't need a super high current regulator design, 300mA would be plenty enough. I would even settle for half that.

The parts cost on the sigma11 is not too bad, abt 20$ considering that it would go into my existing DAC case...

[tangent]

Originally Posted by sachu

An idea on the parts cost for this project?

Fully loaded, with case and all, something like $90. I haven't really priced it out...that's just based on the old STEPS cost of $70 for something similar, plus the additional costs here. Might even be closer to $80.

Can you get a 5 V output from this design?

Sure. You need to change the voltage reference, probably the op-amp, and maybe the gain setting resistors. You need an op-amp that will itself run well on 5 V, and for the rest I recommend a 2.5 V reference and a gain of 2. This is covered in the YJPS docs, but I haven't tried it, so you'd be pioneering. Jung covered this in the original 1995 series of articles, and recommended the AD848 for the op-amp.

[sachu]

Originally Posted by tangent

Fully loaded, with case and all, something like $90. I haven't really priced it out...that's just based on the old STEPS cost of $70 for something similar, plus the additional costs here. Might even be closer to $80.

thanks for the reply tangent.

The parts cost is inclusive of the transformer I presume? I already have a transformer with a separate secondary winding of 0-6V..after rectification I should get around 7V DC. Would that be enough to use as input to the design?

Originally Posted by tangent

Sure. You need to change the voltage reference, probably the op-amp, and maybe the gain setting resistors. You need an op-amp that will itself run well on 5 V, and for the rest I recommend a 2.5 V reference and a gain of 2. This is covered in the YJPS docs, but I haven't tried it, so you'd be pioneering. Jung covered this in the original 1995 series of articles, and recommended the AD848 for the op-amp.

Do you have any more boards left ? Does p-2 p wiring have any effects on your design?

[tangent]

Originally Posted by sachu

The parts cost is inclusive of the transformer I presume?

Fully loaded, yes. Nothing left out.

I should get around 7V DC. Would that be enough to use as input to the design?

No. The Jung 2000 design is a "high drop-out" regulator. To get 5 V out, you might need as much as 10 V unregulated. This makes it less than 50% efficient at such low voltages, I know. If you want efficiency, there are many other designs to choose from.

Do you have any more boards left ?

No. I'm still tweaking the design. What with that work ongoing and manufacturing time, we're a month or two away from seeing final boards.

Does p-2 p wiring have any effects on your design?

Are you asking what would happen if you made it on a perfboard? It's a fairly "fast" design, so it benefits from a good, tight PCB layout. If you switched to a slow op-amp, it could tolerate a big sloppy layout, but then you'd lose a lot of the performance advantages.

[sachu]

Originally Posted by tangent

Fully loaded, yes. Nothing left out.

No. The Jung 2000 design is a "high drop-out" regulator. To get 5 V out, you might need as much as 10 V unregulated. This makes it less than 50% efficient at such low voltages, I know. If you want efficiency, there are many other designs to choose from.

Yeah, I was hoping to use the separate winding on the transformer for further isolation between the analog and digital supplies. is there any detriment in using the same secondary taps for both digital and analog sections of the DAC?

Originally Posted by tangent

No. I'm still tweaking the design. What with that work ongoing and manufacturing time, we're a month or two away from seeing final boards.

Dang..this is the limitation that I am facing with the Toole regulator boards on diyaudio.com too.

Originally Posted by tangent

Are you asking what would happen if you made it on a perfboard? It's a fairly "fast" design, so it benefits from a good, tight PCB layout. If you switched to a slow op-amp, it could tolerate a big sloppy layout, but then you'd lose a lot of the performance advantages.

yes..that is exactly what I asked. It is interesting that a few years back I had read one of the many Analog Devices application notes which said that most prototyping they did for fast designs were point to point.
With the minimal use of wires I wonder if point to point would be fast enough.

[tangent]

Originally Posted by sachu

is there any detriment in using the same secondary taps for both digital and analog sections of the DAC?

Oh, probably, but I'm not the right guy to ask. I've done nearly zero digital audio DIY.

I had read one of the many Analog Devices application notes which said that most prototyping they did for fast designs were point to point.
With the minimal use of wires I wonder if point to point would be fast enough.

The issue is parasitics. The document you're talking about probably referred to ratsnest type builds, where you take a copper-clad board, tie that to some solid ground, then build your circuit in the air above it, with connections to ground as needed, and everything else held up above the board only by the stiffness of the wires. You still have to keep details in mind like minimizing the distance of bypass caps from the nodes in the circuit that need to be bypassed. Other than that, this keeps parasitics very very low: thick wires for low L's and R's, and air dielectric makes for low C's.

This sort of build doesn't hold up well to vibration. It's good for prototyping and hackery.

[sachu]

Originally Posted by tangent

Oh, probably, but I'm not the right guy to ask. I've done nearly zero digital audio DIY.

The issue is parasitics. The document you're talking about probably referred to ratsnest type builds, where you take a copper-clad board, tie that to some solid ground, then build your circuit in the air above it, with connections to ground as needed, and everything else held up above the board only by the stiffness of the wires. You still have to keep details in mind like minimizing the distance of bypass caps from the nodes in the circuit that need to be bypassed. Other than that, this keeps parasitics very very low: thick wires for low L's and R's, and air dielectric makes for low C's.

This sort of build doesn't hold up well to vibration. It's good for prototyping and hackery.

That is exactly what I am referring to The use of a copper board to build the circuit.

I built a diode ring mixer with trifilars a couple of quarters back using a copper board and the rest of the stuff suspended in the air....wa quite fun and worked lke a charm..agree with you about the vibrations and such.

[tangent]

Perfboard fixes the structural problems, but adds a lot of parasitics. This is what I was warning against above. It's not a guarantee of failure, just a risk.

[PJPro]

Did schubert ever sort out his problem?