Another PCB Critique (Qi TX)

[ANTALIFE]

Hi All

I just finished laying out my 3rd PCB. I have tried minimizing the size of current loops where I can and overall it looks good enough to me, but would be good to hear your guys opinion on this.
Also the board is effectively a big DC-AC converter. First stage is Solar -> DC, second stage is DC -> DC, and last stage is DC -> AC (Qi TX).

Cheers

[Rerouter]

Part of me is concerned about 3 AAA's running an AC converter,

Layout wise it appears mostly well done,

On a current loop level, your input PV converter has a pretty big loop for its input capacitors, though it would be pretty well contained for path of least impedance.

C107 and C106 can be shifted left, allowing you to connect the current feedback without via's,

I would probably shift CE200's ground points to its left, so C201 actually does something, currently it has to detour back to those ground vias for its return current.

The test points on your switching nodes are something i would relocate, they will make unintentional antennas, at least place them in the main bulk of the switching nodes copper pour.

Not required, but I have started with a best practice of for screw type external connectors, I run a ring of vias around the pads, It makes it much much harder to tear them off the board, and prevents most of the long term thermal cycling caused by the fiberglass changing thickness with temperature.

And final note, your left 2 mounting holes, I would probably give a little more room around them, I dont know if your mounting to plastic or metal, but metal would risk the tool shorting PV or 5V to ground.

[ANTALIFE]

Part of me is concerned about 3 AAA's running an AC converter,

It's not mains AC that it's producing, it's AC for inductive charging. The circuit should be functional on x3 AA's but want to see how well it works on x2 AA's

On a current loop level, your input PV converter has a pretty big loop for its input capacitors, though it would be pretty well contained for path of least impedance.

Good point, I don't think extending the GND from the IC is going to do much as the pathway through the via's seems sufficient. Will leave as is for now and see how it goes

C107 and C106 can be shifted left, allowing you to connect the current feedback without via's,

Yea shifting C106/107 left will remove one of the via's but I will still need to do a layer jump for the other. As these are sense signals I want to keep them as similar as possible so will make both of them use via's to jump layers. Though one net is 19mm in length while other is 22mm, I think I'll just leave this alone and see how it goes. The datasheet/appnote did not mention anything about length matching and the eval boards look more or less the same

I would probably shift CE200's ground points to its left, so C201 actually does something, currently it has to detour back to those ground vias for its return current.

Another good point, have moved the via's closer to IC

The test points on your switching nodes are something i would relocate, they will make unintentional antennas, at least place them in the main bulk of the switching nodes copper pour.

True, if it becomes a problem will just cut the tracks

Not required, but I have started with a best practice of for screw type external connectors, I run a ring of vias around the pads, It makes it much much harder to tear them off the board, and prevents most of the long term thermal cycling caused by the fiberglass changing thickness with temperature.

Oh yes, have seen those before. Did know they are good for improving mechanical strength of connection, did not realise they help thermal expansion/contraction too. Have added

And final note, your left 2 mounting holes, I would probably give a little more room around them, I dont know if your mounting to plastic or metal, but metal would risk the tool shorting PV or 5V to ground.

Mounting to plastic, just rechecked the clearance and should be OK. Will leave as is but yes good point

Thanks for the help