I suggest including a Thiele network (L||R in series with the output) as this is needed to ensure stability with capacitive loads (think long speaker cables and the like).
You may want to put down a couple of vias where Cz terminates to the ground plane. You want a good low-impedance ground for the Zobel network. I suggest rotating Cs4 and Cs6 to minimize the inductance in the ground lead, i.e. ensure that the ground connection in the decoupling network is as short as possible, while keeping the VCC/VEE connections as short as possible. You'll probably find this means the caps are all oriented at 0 deg. rotation. Using smaller capacitors for Cs4 and Cs6 would provide better decoupling and allow for a tighter layout, so that could be something to consider. You can read my thoughts on decoupling the LM3886 on my Taming the LM3886 website. Cs3 could get a beefier connection to the power connector. You're paying for the copper anyway. Might as well use it! 🙂
Otherwise a pretty tight layout. Plus points for the ground plane! I'm confident you'll achieve the data sheet specs for the LM3886 with this layout, provided that you use a clean supply.
~Tom
You may want to put down a couple of vias where Cz terminates to the ground plane. You want a good low-impedance ground for the Zobel network. I suggest rotating Cs4 and Cs6 to minimize the inductance in the ground lead, i.e. ensure that the ground connection in the decoupling network is as short as possible, while keeping the VCC/VEE connections as short as possible. You'll probably find this means the caps are all oriented at 0 deg. rotation. Using smaller capacitors for Cs4 and Cs6 would provide better decoupling and allow for a tighter layout, so that could be something to consider. You can read my thoughts on decoupling the LM3886 on my Taming the LM3886 website. Cs3 could get a beefier connection to the power connector. You're paying for the copper anyway. Might as well use it! 🙂
Otherwise a pretty tight layout. Plus points for the ground plane! I'm confident you'll achieve the data sheet specs for the LM3886 with this layout, provided that you use a clean supply.
~Tom
Last edited:
Thanks Tom!
Will definitely consider your suggestions.
Cs4,Cs6; Big caps long leads low-ESR vs. small caps short leads hi-ESR... I need to do some math there...
Will definitely consider your suggestions.
Cs4,Cs6; Big caps long leads low-ESR vs. small caps short leads hi-ESR... I need to do some math there...
Just a minor update to the board. I gave the Zobel its own ground connection.
A question about the muting cap however: Do I let myself understand correctly, if I let me understand that the voltage rating of the cap could be as low as only 3V? V- rail is -30V.
A question about the muting cap however: Do I let myself understand correctly, if I let me understand that the voltage rating of the cap could be as low as only 3V? V- rail is -30V.
Attachments
A question about the muting cap however: Do I let myself understand correctly, if I let me understand that the voltage rating of the cap could be as low as only 3V? V- rail is -30V.
3 V is cutting it a bit close. If you look at the equivalent circuit schematic for the LM3886 in the data sheet, you'll notice that the voltage on the MUTE pin is limited to 3*Vbe+1000*Imute, where Vbe is around 0.7 V and Imute is the mute current in A. Normally, the voltage on the mute pin sits around 3~3.5 V for 0.5 mA < Imute < 1 mA. I'd use a 6 V cap. 4 V if you're pressed for space.
I explain how the mute circuit works on my Taming the LM3886 page.
~Tom
- Status
- Not open for further replies.