Submitting to Fabrication for the 1st time

Hi,

This board is designed to do the following:
Relay switching
Phase
Phantom
Pad
PCB mount XLRs
Screw terminal headers for going to the transformers, or you can just solder into those spots obviously.

It takes its power from an external supply and also gives me a mult section to help distribute power to other places.  I was wondering how you more experienced guys felt about the look of it. I'm sure all of my footprints are good so my only concern I guess would be if I made any design mistakes with the traces. What do you guys think? What would you do to improve it before submitting for fabrication?

The buss connector is intended to go to a front panel PCB with the relay switches on it.

Thanks regardless,

Schematic would help to see design mistakes. About the layout, I'd probably lay it differently, still hard to follow the path, but looks some traces are going around too much. Input of pin 1 should be tied to the chasis with a path as short as possible, usually a problem on PCB but I don't get exactly what you are trying to do here... scheme would help!

JS

Are you missing the parallel resistor from the PAD circuit? I always use a 169R switched in across pin 2-3 like the "Jensen transformers" white papers describes...

Also, spell BUS right! ;-)

In general it is better to avoid right angles in PCB traces.

john12ax7 said:

In general it is better to avoid right angles in PCB traces.

Click to expand...

No.

As long as you're below microwave frequencies, right angles in traces are no worse than 45 degree angles.

JD 'myth has been busted over and over again' B.

How are you mounting this in a chassis? Are there screw holes?

jdbakker said:

john12ax7 said:

In general it is better to avoid right angles in PCB traces.

Click to expand...

No.

As long as you're below microwave frequencies, right angles in traces are no worse than 45 degree angles.

JD 'myth has been busted over and over again' B.

Click to expand...

For self etched boards, or when it doesn't have solder mask, 45º corners are less prone to lift, that's the reason I usually put 45º instead of 90º, but as he mentioned going into production I guessed solder mask would be over it which protects the corners from lifting... So, for me, for that reason, 45º are better even in DC.

JS

jdbakker said:

john12ax7 said:

In general it is better to avoid right angles in PCB traces.

Click to expand...

No.

As long as you're below microwave frequencies, right angles in traces are no worse than 45 degree angles.

JD 'myth has been busted over and over again' B.

Click to expand...

Performance wise at low frequencies it won't matter.

But it may become an issue with fabrication especially when lines get thinner.  Certainly fab houses have gotten better over the years, but why chance it? It's just another thing that may go wrong which you can just avoid from the outset.

john12ax7 said:

jdbakker said:

john12ax7 said:

In general it is better to avoid right angles in PCB traces.

Click to expand...

No.

As long as you're below microwave frequencies, right angles in traces are no worse than 45 degree angles.

JD 'myth has been busted over and over again' B.

Click to expand...

Performance wise at low frequencies it won't matter.

But it may become an issue with fabrication especially when lines get thinner.  Certainly fab houses have gotten better over the years, but why chance it? It's just another thing that may go wrong which you can just avoid from the outset.

Click to expand...

Can you link to any professional PCB manufacturer who recommends against (or outright disallows) right angles in traces because it might cause etching/delamination problems? I've never found that, not even with some of the fly-by-night operations that have popped up over the years. Slivers, sure, and they may limit the acuteness of trace angles, but the few board houses that I have seen to discourage right angles did so over the same old "electrons rounding a corner"-myth, not manufacturing issues.

JD 'non-issue' B.
[while why chance it can be a prudent approach, it is unnecessarily limiting if there is no actual risk]

EXPRESS PCB

"When placing narrow traces, 0.015" or less, avoid sharp right angle turns. The problem here is that in the board manufacturing process, the outside corner can be etched a little more narrow. The solution is to use two 45 degree bends with a short leg in between."

ExpressPCB Design Tips

jwhmca said:

"When placing narrow traces, 0.015" or less, avoid sharp right angle turns. The problem here is that in the board manufacturing process, the outside corner can be etched a little more narrow. The solution is to use two 45 degree bends with a short leg in between."

Click to expand...

Well look at that.

I must admit that's a new one to me. I've worked with quite a few board houses over the years, in Europe, the US and more recently China, and have never run into anything like that. It surprises me that they qualify 15 mils as 'narrow'; even cheaper board houses run down to 5mil/5mil trace/space that I've seen. I've asked our CM (as I was talking with them today anyway), and all they'd ever run into was the insides of sharp corners risking to be acid traps. I'm seeing some people from the IPC next month; will report back on what they have to say.

In the meantime: the best course of action is always to check with your board house to see what they support. If the one you want to work with doesn't do <x>, then by all means, avoid <x>.

JD 'or find one that does' B.
[who went as far as putting a few boards under the microscope today, to see if any right angle turns looked dodgy]

Hi,

Sorry for getting back so late! I already submitted to fab for some proto boards. If they don't work, it's time for rev two. that's why I prototype.

Okay lets see -

Schematic would help to see design mistakes. About the layout, I'd probably lay it differently, still hard to follow the path, but looks some traces are going around too much. Input of pin 1 should be tied to the chasis with a path as short as possible, usually a problem on PCB but I don't get exactly what you are trying to do here... scheme would help!

I'll draw one up as soon as I get time 

Are you missing the parallel resistor from the PAD circuit? I always use a 169R switched in across pin 2-3 like the "Jensen transformers" white papers describes...

Also, spell BUS right! ;-)

I'll have to scope that paper out. Got a link? I just used two resistors pre transformer which i'll play with the values of. Does it need one going in between?

about "buss"  .. Dangit.

As long as you're below microwave frequencies, right angles in traces are no worse than 45 degree angles.

JD 'myth has been busted over and over again' B.

Yeah I mean I'll try to do it less just for good measure but I've never had any problems with it. If I ever get contracted to make and X-Ray or Microwave machine I'll remember it's a must.

How are you mounting this in a chassis? Are there screw holes?

It will have PCB mount and chassis mount XLR jacks that will create a right angle support off of the chassis. This is the not exact part, but by using a very similar one which I'm too lazy to look up right now.

http://www.parts-express.com/neutrik-nc3mk-h-k-series-pc-mount-xlr-connector-male-horizontal--092-038?utm_source=google&utm_medium=cpc&utm_campaign=pla

http://ethanwiner.com/gadgets.html

Top Left corner has a pad circuit...

Jensen's webpage is all membership (free) now, so you can't really link to the whitepages.... IIRC, it was in the white paper about "Standard Microphone Input" or something to that effect..

jwhmca said:

http://ethanwiner.com/gadgets.html

Top Left corner has a pad circuit...

Jensen's webpage is all membership (free) now, so you can't really link to the whitepages.... IIRC, it was in the white paper about "Standard Microphone Input" or something to that effect..

Click to expand...

GOOD CATCH! 

I'll have to jump that for now. Here's a quick road map excluding the bus connector and the the screw terminal blocks. It explains my concepts at least.. not enough time right now to draw a full schemo on software.

Now we are talking, for the pad you should look at the jensen app note, that's a better approach than yours keeping impedances right and actually attenuating more or less the same amount independently of the source impedance. The phase reversal is better to be done in the input ass well, since the output not always have transformer but an impedance balanced stage and if you feed that to an unbalanced input you will have no signal...

JS

seems like its gonna work.

to help you guys see what the footprints are for