[00940]

I just had a look at the schematic. Why is the output cap (C11) (and way less importantly the led D7) connected to the sense ground (RGND) ? In Jung's articles, it's connected to the "dirty" ground (CGND).

Maybe I missed it, but I couldn't find a pic of the board layout ?


I must confess (sorry for being blunt) that the idea of spending a fair amount of money to get the lowest possible output impedance and then throw it away by putting the regulator far away of the powered circuit (and throwing connectors into the mix) is looking a bit strange to me. AOS or ALW way of making the reg as small as possible to put it very near of the amp attracts me more.

 

[error401]

Quote:

Originally Posted by 00940 /img/forum/go_quote.gif I must confess (sorry for being blunt) that the idea of spending a fair amount of money to get the lowest possible output impedance and then throw it away by putting the regulator far away of the powered circuit (and throwing connectors into the mix) is looking a bit strange to me. AOS or ALW way of making the reg as small as possible to put it very near of the amp attracts me more.

I think that making use of the remote sensing should negate the effects, but I have to agree. Regulators really belong with the components they're powering, though I can't argue with an external preregulator to keep as much 60Hz as possible away from the audio.

 

[00940]

You're of course right, remote sensing should negate the bad effects of long wires... if it doesn't throw the regulator into oscillation. With an oscillator, it can be checked. Without one, I'd stay away of remote sensing in the case of a two-boxes system.

 

[n_maher]

Just in case anyone was interested in using one of these in a project, there's one listed in the FS forums.

 

[funch]

Soooooo ...... whatever happened to this PS? Still under development, or
DOA?

 

[rds]

I hope it is being made available. It is very nice.

 

[tangent]

I was waiting for those who got the v1.0 boards to give some feedback. *Crickets*

I guess silence is an answer, of a sort, to the question, "does it work?"

 

[rds]

It worked great for me. No issues with the performance at all.
If the wish list requests from this thread were implemented then I think it would be completely ready for a production run.

 

[rds]

Any news on the YJPS production boards?

 

[espressogeek]

Yea I have been considering one , well, just to have. Another option was the omega 11, I think that is what it is called.

 

[tangent]

I've been interleaving v1.1 work with other stuff I have going. It's nearly done, but progress is slow because I don't get back to it very often. I'll try to get it done this weekend.

EDIT: v1.1 will be another proto run, because so many things have changed. If v1.1 works, v1.2 will be the production run.

 

[qwerty870]

I'll get a couple whenever they become available.

 

[tangent]

Finally (!) I have something I'm willing to present in public again. Schematic and board layout:



(Click board to see larger version.)

Changes since v1.0:

- Diode bridge and transformer are farther away from the board edge.

- Added a third unregulated filter cap

- Line filter choke can now be the big ELF 850 series I used on the first versions of the STEPS, or you can use the inner four holes for the same chokes as were used on STEPS v1.2. Larger board allows this while still keeping plenty of space between the primary and secondary sides of the supply.

- Line filter caps have more space around them than on STEPS. I have some caps here that have the right specs but don't fit on the STEPS, so this fixes that.

- Unlike on STEPS, you can now parallel the secondaries on the transformer, not just the primaries.

- Transformer pads are labelled now, to make it easier to use wire-leaded toroids

- Still plenty of breathing space around the transformer. I briefly played with allowing 35 VA transformers, but on running the numbers, there's actually no way to use that much power with this design. Removing that allowed me to move the transformer closer to the upper board edge, away from the diode bridge and filter caps.

- Moved fuse clips farther apart to allow use of the type with ears. (I fixed this on the STEPS long ago, but was in the YJPS v1.0 because it was forked from the STEPS design before the fix.)

- Heat sinks have mounting holes and better outlines

- Changed board mounting holes so they have explicit clearance around them for bolt heads

- Output pads now use a Molex KK footprint instead of wirepads.

- Output voltage is now adjustable, with the default schematic values giving a 18 to 30 V range, centered on 24 V. (I yield, I yield!)

- Remote sensing pads now shown on the board and schematic as jumpers, to make it clear that they need to be connected somehow, whether it's by an actual jumper or at the circuit being powered to get remote sensing.

- Split the unregulated section of the ground plane away from the control section, so one can play with decoupling them with a small resistor or inductor.

- Test points allow larger wire loop test points. (This may change again.)

- Most of the test points collected together at the board edge now, rather than scattered around the board.

- Various part name changes (just to say, don't go comparing old docs and schematics to the board and expect things have stayed the same)

I'm allowing some time for public comment and for me to sleep on the recent changes before I send this off for prototyping. I may do it on Monday, but no promises.

 

[MASantos]

Tangent, I can't see the image you posted above. Do other members see it?

 

[n_maher]

Here's the image, Tangent's url just got cut off.