Any interest in GC input buffer and power supply regulator pcb's?

[moving_electron]

cjd,

It looks like folks are happy with you reg. power supply layout as there have not been any comments for a few days.

I a not clear on the current status. Does the current layout implement actual part sizes?

If so and you think it is done (or at least can call it good for now) would you be so kind as to post a link to a pdf version of the artwork for the view (with parts) from the top view as well as etchable artwork as viewed from the copper side?

[cjd]

I've been digging up part specs and these are part of the result of that. In places where cap ranges are specified, you'll note I have allowed for three sizes to accomodate that. Where resistors are specified, I'm looking at a few seemingly common types people use and am trying to allow for all of those (within reason).

First up, half the power, use two for full setup. Mirror is do-able though I do sort-of question the value. The G/V markings are for the + side. I did spend just a little time so far working to get these combined on a single PCB with a combined ground at that point but it hasn't happened for me yet. I included an option to use a single huge cap though it goes off the board. Not sure if there's any value in this or not.



Next, LM3886 PCB for use with a regulated PS. A cap in the mute circuit would need to go under-board on this layout. Everything else could go up top if the mute resistor is the right type to fit between the chip pins. Ground connection is back to being jumpered so I could run traces to both V+ pins on the chip. Not sure why I missed that before.



The LM3875 PCB.



And finally, a go at a board for the OPA627 buffer, thanks to Carlos's schematic. This could be full of problems still.



I'm working on a JFET buffer layout as well but have discovered some data sheets are missing information OR I'm missing knowledge to find the information on the data sheets. Probably the latter.

As normal, critiques are more than welcome.

C

[r!sc!]

JFET buffer, working OK inside my amp. Pedja's circuit.

[Pedja]

One thing may be important (overlooked earlier... I am not into the non-inverted GCs). Feedback network, though it won’t carry currents higher than 1mA would be better connected to the central ground point independently of the signal ground. Might not be that hard to solve in the LM3886 case rotating R3.

I am not sure if the path by which + reservoir cap is charged (currents may be as high as 10A) is too close to the regulator... Also, what is C3 on the reg board?

[cjd]

To make sure I understand:

R3 on the chip-amp boards should not go to signal ground but to power ground then? That may actually be easier on the 3875 board if it gets mounted behind the board. 3886 will probably require re-shuffling the components some again. Is a slightly long lead on one side OK here? (Preferable, I am sure, to a more convoluted output path)

C3 on the regulator board, I have no idea?! I'm going to go update that pic, so everyone else'll probably get all confused by your question. I know I had started playing with leaving room for small poly caps both from in+ to in- at the chip pins, and from V+ to G and V- to G, again at the chip pins. I'm assuming your comment on the + cap trace and regulator apply to this board? I'm not clear on this comment at all. As in physically too close? (should it just get moved up, or over some?)

r!sc! - that layout helps me a lot. Thanks!

And Pedja - inverted boards are in the works. They're required if these are to live up to my idea of being "stackable" to build paralleled and/or bridged setups, and I also know lots of folks still prefer that layout (especially with buffer, etc.)

C

[r!sc!]

Keep good work

Boban

[Pedja]

Originally posted by cjd
R3 on the chip-amp boards should not go to signal ground but to power ground then?

Let's say so.

That may actually be easier on the 3875 board if it gets mounted behind the board.

But isn’t the Rfb already there? Also, I think a bit of care is needed as to where to put the pad for the resistor there, the chip leads will be quite close to the board. It may be good to put them in the line with the corresponding pins. Or anyway to make them just to be the pads, no holes.

3886 will probably require re-shuffling the components some again. Is a slightly long lead on one side OK here? (Preferable, I am sure, to a more convoluted output path)

I think it is more important to have a short feedback path than to save 2 or 3cm of the output path which is anyway long.

I wouldn’t use a bypass PSU cap but would move electrolytics closer.

I'm assuming your comment on the + cap trace and regulator apply to this board? I'm not clear on this comment at all. As in physically too close? (should it just get moved up, or over some?)

Yes, too close physically. I am saying what I see like a possible cause of the problems, I don’t know if it indeed will make problems or not. Maybe not. Maybe it will…

Pedja

[Sheldon]

Pedja,

I built your ultimate reg. Made a few mistakes along the way (didn't fully discharge the caps before trying to resolder something, tested the validity of 2amp fuse by inadvertantly shorting the output, etc.). Anyway, got it together with 27 v zeners, 3055/2955 substituted for bd 912/911, and 150 ohms for R1 and R2. I get 24.1v output on the neg side and 24.4 on the pos side. On the pos side I get 850mv across R1, which means I could probably lower the value some. On the neg. side, I get no voltage across R2. Any ideas on what to check first? Replace the J309?

Somewhat related question; will the j309 substitute ok for the j310 in your buffer?

Thanks, I'm having fun with this,
Sheldon

[Pedja]

Originally posted by Sheldon
On the neg. side, I get no voltage across R2. Any ideas on what to check first? Replace the J309?

Sheldon,

No voltage should appear across 1k gate resistor. Did you, by mistake, measure that one? Or none of these resistors show voltage drop and you have 0V measured directly between J309’s gate and source? If it is so, i.e. you have 0V there, check R2 itself (for short – you in fact can do this first of all).

Eventually you could report about the actual bias through that J309 if you connect some resistor of known value (something like 100R) in series with its drain.

Somewhat related question; will the j309 substitute ok for the j310 in your buffer?

Yes.

Pedja

[Sheldon]

Ok, checked my circuit and found the error. Not surprisingly, it actually makes a difference whether or not the gate and drain connections are swapped. Anyway, I now get 5 volts across the 150 ohm resistor in the neg channel (33ma). Guess I fried the little ******. I'll try another j309.