if you vertically piggyback/stack them, I can't see how the 2 can be effectively sinked.

one of them, per channel, runs hot enough as it is!

---

[Quoting linux-works]

could there be a nice 2nd board (mezz board) that can 'do the gamma1 trick' (so to speak!) and give us the funct that way?

---

I've already designed as much expansion support into the board as I personally want: mounting holes at the tail of the board for a Modified Linkwitz Crossfeed board, and perfboard in one corner. My idea of expansion is a collection of independent boards that can be hookup-wired together. Pluggable daughter boards and such don't interest me, personally. I get that that doesn't please you, and I'm willing to tweak the design if you can give me something concrete to think about. Good design doesn't come out of lack of desire. You have the desire, so you're going to come up with a better design than I will.

---

[Quote]

as long as the board can plugin via some location (a well defined interface, like the feedback loop insert point as well as the audio in/out/passthru block - then the mezz board could be designed to mate with the bottom main board.

---

Let's take feedback loop access as an example. In that case, I'd suggest you just replace R4 on the PIMETA board with some type of connector, moving R4 up to the daughterboard. Don't ask for pads you don't really need.

---

[Quoting tomb]

When using a TREAD for charging/line power, the batteries will backfeed the TREAD's LED if the wall supply is diconnected.

---

---

[Quoting bdhornback]

Would it be possible to put 1206 pads for the resistors as well?

---

The thing is, though, I added the SMT pads to the caps because there are desirable parts you can get in SMT that you can't get in equivalent leaded form: MLCCs with low parasitics for the bypasses, and PPS film caps for the input coupling caps. Is the wish for SMT resistors just because you like SMT, or is there some special SMT-only resistor that I haven't heard of yet? Last time I checked, SMT resistors were pretty staid. All the good stuff was leaded.

---

[Quoting masantos]

one of the features of the v1 was the hability to stack buffers and get the extra current output.

---

Notice that I'm running traces on the opposite side of the board under the buffers. If I did put these additional footprints in, I'd have to pessimize those traces.

if you could reserve a round hole toward the middle of the pimeta board, that might be useful for when you need some support of a partial (half length) mezz. if the board happens to only need enough coverage for an edge to the middle, say, having a screw post hole in the middle area would be all you'd need to at least suspend the mezz.

just a thought. it might make the middle of the board look ugly with a center hole but it could provide utility nonetheless.

for the top board, the crossfeed is the more important part. bass boost only really needs access to the 'r4' area and so its easily done at the pot or switch, itself. not a big deal to omit this if space is tight.

for the crossfeed, though, the parts that get 'in the way' on your current crossfeedboard are those 2 large caps followed by the 2 smaller ones. I wonder if that board is flipped over (like the amb/mrx gamma1) and those caps are pointing downward, would it be possible to juggle the bottom layer parts so that they don't bump into each other?

in inverted crossfeed sandwitch might be pretty cool.

and on the crossfeeds that I've built, I've only selected a 'hi' or 'low' setting but never allowed a switch between them. that saves some pads and 2 R's and some routing. I'd also want to put one of the tiny relays I've been using on that board to allow truely remote wire selection of the crossfeed.

I can try to do a layout but I'm REALLY new at this. you may not want the layout I come up with

[Edited by linux-works on 02-16-2009 at 04:13 PM.]

Different presentation this time so you can evaluate the copper more easily. Again, I'm not showing the inner two layers: audio ground and V+; they're pretty much solid, excepting deliberate holes around the op-amps.

Key changes:

- All the Rs and most of the Cs have SMD pads now

- The layer stack-up is flipped yet again, putting V- on the bottom along with all of the SMDs now, and most signal traces on top.

- Split the ground ring in perfboard area so four of the previous group of 9 are separate, leaving only 5 hard-connected to audio ground.

- Added headphone guy

Notice how lacy the V- layer is now, with all those SMD pads. I had to allow the copper pour to go down to 8 mils to prevent it from falling apart.

Still-open question: is there something specific needed on this board to support daughterboards?

I tried adding the third diode requested by TomB, but couldn't quite fit it. I could have done it with a 1N4148, but I worry about the inrush surge current blowing it up.

[Edited by tangent on 03-02-2009 at 04:05 AM.]

---

[Quoting tangent]

- Added headphone guy

---

Awesome!

---

[Quoting tangent]

I tried adding the third diode requested by TomB, but couldn't quite fit it. I could have done it with a 1N4148, but I worry about the inrush surge current blowing it up.

---

The SMD resistor and capacitor pads are inviting.

the times that I wanted the cap, I had to xacto knife that bit of trace out and that 'ruins' the board if you are not very clean about it. and I hate playing with knives (lol).

surely there will be 2 clipoff leads left by the time the builder is ready to deal with the c1* places?

Regarding the question of "why", in the past it was found to be a source of confusion. So many people recommend leaving these caps out that the chances are high that the average newbie will follow that advice but forget to install the jumper, then wonder why the amp was making no sound. It seems that "leave it out" is advice more likely to be followed correctly than "jumper it". Could be wrong on that...maybe there's a bunch of PIMETA v1s out there with C1s installed and the jumper left uncut. On the other hand, at least then the amp functions from the start.

Another point in favor of the "pre-installed" jumper is that I don't recall getting any complaints about it. I do remember all those threads about people getting no sound.

[Edited by tangent on 02-23-2009 at 06:36 PM.]

how about having that neat DIP/so8 dual style with just pins to get access to the 8 chip pins? it would be pretty cool to be able to have a scratchpad browndog/DIP hybrid, so to speak

I can airwire DIP8 parts if I really want to, but having some scratchpad where I can break-out SMD style chips would be really cool! and if you do your trick of having 8 thru-holes around the 8 SMD pads like you do on the other areas, so much the better!

also, since its 4layer, you could put a heat-sink pad in the center so it could support another buffer, and all the other 3 layers would carry the 8 pins mapped to 8 pads. (well, not sure if using the bottom layer for the heatsink pad is worth it or not, just wondering aloud..)

[Edited by linux-works on 02-24-2009 at 09:06 AM.]

is there any benefit to each of the headphone drivers having its OWN private ground return via a buffer?

I'm not sure I'd call that balanced since the ground buffer isn't swinging equal and opposite signal as the L and R outputs, but it is isolated and you'd think the specs would improve even more with 2 gnd's instead of 1 (?)

there's also good DIY sense in both OPAs being the same part, or at least both being dual-channel OPAs.

---

[Quoting linux-works]

how about having that neat DIP/so8 dual style with just pins to get access to the 8 chip pins?

---

If you want to put an SO-8 in the scratchpad area, well, that's what chip adapters are for.

---

[Quote]

is there any benefit to each of the headphone drivers having its OWN private ground return via a buffer?

---

---

[Quote]

there's also good DIY sense in both OPAs being the same part, or at least both being dual-channel OPAs.

---

---

[Quoting linux-works]

there's also good DIY sense in both OPAs being the same part, or at least both being dual-channel OPAs.

---

So make them all the same part and run all single op-amps:
cimarrontechnol...021001.aspx

---

[Quoting cobaltmute]

---

[Quoting linux-works]

there's also good DIY sense in both OPAs being the same part, or at least both being dual-channel OPAs.

---

So make them all the same part and run all single op-amps:
cimarrontechnol...021001.aspx

---

my point is density, which is important in a small sized enclosure this is meant for.

going with 'stereo' devices makes sense for the L/R op amp. but I'd still think that a stereo op amp on the ground channel would allow you to run in single-ground channel (normal) mode or even add a 2nd if you really wanted.

sure, you need to break out your phones and make it 4wire instead of 3wire but some people seem to like that and if you offer 2 independant ground buffers, isn't that almost free in this design, needing only a few extra parts on that 'unused' amp inside the 2-ch opamp pkg?

(at times it IS useful to swap parts to verify things; or simply to order only part X and keep those on hand instead of X and Y).

not foolish at all!

Also, access to V+, V-, and ground near the breadboard might be useful.

edit:

---

[Quote]

- Split the ground ring in perfboard area so four of the previous group of 9 are separate, leaving only 5 hard-connected to audio ground.

---

I see that ground is already available. V+ and V- would still be useful.

[Edited by dsavitsk on 02-24-2009 at 04:24 PM.]

If you were to put an op-amp in there, you only need two jumpers to connect it to the supply layers: a very short one in the lower right corner, connecting the V- layer-access pad to the DIP-8's pin 4, and a longer one from the V+ layer-access pad in the upper right to either pin 7 or 8, depending on whether the op-amp is a single or a dual.

EDIT: I expect it's obvious to any experienced DIYer why I didn't hard-connect the DIP-8 pins to the supplies: even in the op-amp case, you don't know the connections in advance. If it's a microcontroller, or a 555, or any number of other DIP-8 chips you might want to use, the connections change again. The best I can do is make the needed connections easy to implement.

You can restore the previous ground ring with two jumpers, in the same way I recommend in my CMoy perfboard layout. Or just one, giving a 'u' or 'h' shape. I'd do this on the top before adding a socket to the DIP-8 area.

Keep in mind that adding a DIP-8 here isn't required; it's just the easiest thing to add, if you did want to add something. The average PIMETA won't have anything in the scratchpad area, except perhaps some hookup wires going to the group of 5 ground pads for some panel component that wants to be grounded.

[Edited by tangent on 02-25-2009 at 06:09 AM.]

[Edited by Mister X on 02-25-2009 at 06:16 AM.]

More commentary about why it is the way it is:

The group of 6 in the lower left is for linux-works: it lets you run taut wires from the O pads, say on the bottom side of the board to the "rear" set of 3, then put an 0.1" spaced connector in the front set for the output jack, so it's easy to disconnect the board from the front panel.

The pad pair behind the group of 6 has no special meaning. I have no particular problem with removing the trace between them and connecting them perhaps to the supply layers. The silkscreening for that might be tricky.

Another thing we might think about is adding another row or two of pads extending under C2A, which you can either access on the bottom side of the board, or by just leaving out C2A if expansion is more important than rail capacitance.

[Edited by tangent on 02-25-2009 at 09:28 AM.]