I think the standard advice is to make the angles 45 degrees. It will clean up the layout quite a bit. For an example, follow my own learning progression in this thread: http://www6.head-fi.org/forums/showthread.php?t=130517

-d

Maybe better placement of I/O holes that doesnt have traces going from one end of the board to the other for no real reason, hookup wire will do the job better than a pcb trace and it will help on getting the other traces sorted too

I'd recommend putting some small PSU decoupling caps near each of the opamps and buffers.. This is one of the main deficiencies on the MINT and PIMETA IMO.

New board layout, please leave comments/suggestions. I will be adding in decoupling/bypassing, but not yet.

Quote:



I wouldn't bother using 4 electrolytic capacitors. They'll just take up valuable space that could be used for other components.

Quote:



I doubt that it's possible to find a battery holder that will fit that small a space. I threw together a few drawings in eagle depicting a side view of the 1455 hammond case and it barely fits 8 AAA batteries as it is. Finding a way to physically (and reliably) connect the all the batteries end-to-end will probably be a challenge.



Check out the Xin Supermacro3 if you want to see how he was able to fit 8 batteries into the small hammond case.

The pcb layout is going to be entirely dependent on how much pcb area you have leftover after the I/O jacks and pot. If I were you, I would search through the Digikey and Mouser catalogs to find whatever I/O jacks that you want to use, download their respective datasheets, and figure out if they fit into your current design.

Correct me if im wrong, but a dual opamp and 2 caps would save a whole lotta space... wouldnt it?

Rob.

Quote:



the target size was half the size of an unhalfed mint board and we reached that target with the original layout. the opamps we want to use are the OPA627/637. most highend opamps dont come in dual forms, anyways.

Bypassed:

A couple slight changes have been made, added buffer bandwidth resistors. I am not sure what caps to use for opamp bandwidth, as even the Wima MKS-02 is too big to fit. If there are any small smd parts that work, as it seems ceramics shouldnt be used, please post a link. Same for the crds, a smd package would be ideal. Also, if anyone has experience with Eagle, I have about 88 errors in the DRC test, all clearance, all right in the middle of the 1206 resistors. I am not sure why I get these errors, I have been able to sort all of the other ones out, but not this.



Labled up version, I tried to stick to Tangent's numbering scheme

Hey Ctoole -- I like this a lot; the space saving would make possible some of the things I'm discussing in other threads (auto power-on and -off from audio signal; rechargers).

But I wonder if you could improve the size considerably by adding pads for SMD electrolytics. You can get low-esr smd electrolytics whose figures are comparable to Panasonic FC, but are only 10mm high. Modifying your PCB would be a simple matter of placing SMD pads on top of your electrolytic holes, so people could use either type.

Look at these caps:

http://www.rapidelectronics.co.uk/rk...1100&XPAGENO=1

or these:


http://www.rapidelectronics.co.uk/rk...2534&XPAGENO=1

You guys are still thinking in 2D. Why not have resistors on the underside of the PCB? Otherwise, there really isn't any real advantage to using SMD resistors.

-Ed

Quote:



I like the sound of that!

-Ed

Quote:



Um, the PCB cetoole posted does have resistors on both sides of the board. Look two posts above yours.

Also: One of the real strengths of the MINT is the CRDs to buffer the op amps into class A. It's one of the things that makes it a really high-end amp despite its small space. I regard these as a decisive component. (I'd also prefer dual opamps, for the record, but I may be in a minority here -- I have one MINT with an 8620 and one with an OPA2227, and those are about as good as you get for this application).

Quote:




Oops!

that's what I get for skimming.

PiMS. An acronym one letter away from an unfortunate one.

-Ed

To remove the DRC clearance errors on the pads, use the SIGNAL command (you can just type it, or it's beneath the polygon button for mouse users) and connect the relevant pads. Then run a RATSNEST, and now DRC. All sorted You can get CRDs in the MELF package; Eagle doesn't have them in the DIODE library, you'll have to look in the RCL library for them. Use 0805 or 1206 ceramics for decoupling. SMD electrolytics tend to have larger footprints than equivalent through hole ones, so you won't save much (any) space there, and would require yet more vias