[amc]

now on V1.3 links and thumbnails updated...

Hi Everyone,

Requesting a sanity check of the attached cmoy schematic and board layout before I try and etch my first board. Features:

- TLE2426 rail splitter
- Wall and battery power
- decoupling caps on OPA
- ground plane

Is anything missing in the schematic? Is the board layout good? Can you see any problems with it? Suggestions for improvement? Thumbnails attached and large size files are stored at photobucket at the following links:

CMOY-Schematic-1.png
CMOY-Board1_3.png
CMOY-Silkscreen-1.png

Thanks for looking and thanks to all the people who contribute here. I am amazed at the amount of open advise you all provide. And thanks to cmoy and tangent for all the effort they put in to documenting and explaining headphone amps. Without them I am sure I would not have made my first CMOY or taken an interest in electronics DIY.

 

[kvant]

I would consider putting power-rail capacitors after the TLE2426 (from GND to V+ and from GND to V-). With your approach, the ground sourcing/sinking capability is limited to some 20mA, thus some 10mA per channel. This represents a rather unfortunate bottleneck, since the opamps in L&R channels will almost certainly be capable of delivering more current. You will find more details on this in tangent's article on virtual grounds.

 

[Earwax]

no volume pot?

 

[amc]

Quote:

Originally Posted by kvant /img/forum/go_quote.gif I would consider putting power-rail capacitors after the TLE2426 (from GND to V+ and from GND to V-). With your approach, the ground sourcing/sinking capability is limited to some 20mA, thus some 10mA per channel. This represents a rather unfortunate bottleneck, since the opamps in L&R channels will almost certainly be capable of delivering more current. You will find more details on this in tangent's article on virtual grounds.

Thanks kvant. I found tangents article on virtual grounds a bit confusing and had to read it a few times to understand the benefits and drawbacks to adding capacitance before or after the railsplitter. I have made the suggested change.

Quote:

Originally Posted by Earwax /img/forum/go_quote.gif no volume pot?

Not yet Earwax. Have some of the alps pots on order from manelco - more than a million parts but they are on the "slow boat from china". Its been more than a month. I am gonna break down and order some from tangent when the final pimeta v2 boards are availiable.

I have regenerated the images and updated them in the original post. Any other pointers?

 

[Fred_fred2004]

That ground plane is going to give problems you will find its too close to the traces try and get some more space or bridging will be an issue.
Just noticed the diodes are you using 2 9volts in parallel?? you would be better with them in series

 

[amc]

Quote:

Originally Posted by Fred_fred2004 /img/forum/go_quote.gif That ground plane is going to give problems you will find its too close to the traces try and get some more space or bridging will be an issue.

Thanks Fred_fred2004! I decided I wanted to try and etch my own boards after looking at some the fine work you have done in the post pictures of your builds threads.

I have changed the clearances as much as I could for traces and pads, and then used restrict polygons where I could to maximize the distance between the groundplane and the signals. I have updated the board image (now v1.2) in the original post. Do you think the ground plane is a waste of time and should I just use point to point traces for ground connections?

Quote:

Originally Posted by Fred_fred2004 /img/forum/go_quote.gif Just noticed the diodes are you using 2 9volts in parallel?? you would be better with them in series

Not two batteries but one battery and one 24 volt wall power brick. The two diodes approach is to keep the wall power from trying to charge the 9v battery. Tangent describes it as a Diode OR bridge.

Once again I have updated the board to reflect the suggested changes in the Original Post. Keep the suggestions comming guys - its appreciated...

 

[kvant]

Quote:

Originally Posted by amc /img/forum/go_quote.gif Do you think the ground plane is a waste of time...?

I don't think that the ground trace you have now can be called a "plane". You managed to draw a nice ground loop, which is something to be avoided. Break it somewhere, try to make the headphone ground return such that the return current does not flow around input and feedback ground connection.

 

[Fred_fred2004]

I must admit I use ground planes to minimise the amount of copper to be etched.

With the Wall Power why not just use a DC socket/jack with a switch so the battery is out of circuit when you plug in the wallwart. the diode will be wasting 0.6V

 

[amc]

Quote:

Originally Posted by kvant /img/forum/go_quote.gif I don't think that the ground trace you have now can be called a "plane". You managed to draw a nice ground loop, which is something to be avoided. Break it somewhere, try to make the headphone ground return such that the return current does not flow around input and feedback ground connection.

I see what you mean re the ground loop. I will break it. But now I have another question. Are there any rules of thumb aside from avoiding headphone ground current flowing around the input and feedback connections I should be aware of? I was able to get the trace clearance to 24 mil without too much trouble when it was only 10 mil before fred_fred2004 pointed out the problem - would that be sufficient? Is there any guidance you can point me to - examples are nice - of what a good groundplane vs a bad groundplane look like? Here is one example I found in fred_fred2004's photobucket of a ground plane - is this a good example?

CIMG3550Medium.jpg picture by Fred_fred_2004 - Photobucket

 

[Fred_fred2004]

Do you mean this

Your link was'nt working for me,
That board was for a starving student made for another member, I don't know if its a good ground plane or not I was just trying to save etching time and chemicals, but the amp sounded great so I may have jagged it

cheers
FRED

 

[amc]

Quote:

Originally Posted by Fred_fred2004 /img/forum/go_quote.gif Do you mean this...just trying to save etching time and chemicals...

Yes thats the one, and my first goal is to minimize the copper to eat through as well. I will give it another go to see if increasing the trace clearance looks similar to your board tonight.

Thanks,
Ashley

 

[amc]

After a week of messing about with the layout I have got the ground loop addressed and have tried to get as much space between the ground plane and signal traces. Signal traces now 32mil. Clearance now 24 mil. I think it looks better. I had to switch out the caps to get a 7.5mm lead spacing so I could fit two traces between one of the power caps. Any more suggestions?

As usual I have updated the original post so look here for updated images and links. Thanks.