Custom active crossover

[Puoskari]

Hi there!
I just wanted to share the crossover I designed for the Protones. The design specifications were created by simulating in vituixcad and the component values were calculated as described at linkwitzlab.com

The response was measured using picoscope's FRA tool. The vituixcad simulations were pretty much spot on with the actual response of the crossover, which I have to admit surprised me since it's not really meant to be used as a circuit simulator. They were designed to be used in junction with the Icepower 50asx2se amplifier modules so they are powered by the amplifiers auxilary power.

Here's the response:

They didn't produce any unintended hiss, which was a relief.
More about the project can be found at the wiki.

If you have any questions about the design or tips for the next iterations (I'm quite new to pcb layout design) please ask and comment!

The total cost for this batch was 45€/pcs, but that will go down to 16€/pcs when ordering ~50pcs.

 

[jamescarter1982]

hi so could you go Into some more detail about how you designed this did you use someone else's design ? I am interested in making something like this

 

[justjason]

If you have any questions about the design or tips for the next iterations (I'm quite new to pcb layout design) please ask and comment!

You are rather sparse on power rail decoupling capacitors and those that you have , are placed on the wrong side of the power line and far away from the device which requires them.
The correct psychical placement of a decoupling cap is power -> cap -> device. In your case you have cap -> power -> device

These caps should be placed as close as possible to the op-amps.

I do not understand your choice of electrolytic caps in the input (C1 + C8).Electrolitic caps are polarized and your input signal is AC - it has both positive and negative peaks relative to ground. . At the voltage and current in your use case, is is unlikely to die immediately, but I am still confused as to your reasoning.

The ground plane is unnecessarily broken and is forcing large ground loops which can be avoided

 

[Puoskari]

You are rather sparse on power rail decoupling capacitors and those that you have , are placed on the wrong side of the power line and far away from the device which requires them.

Good point. The reason is that the power supply is designed as an afterthought as always . This should indeed be changed.

I do not understand your choice of electrolytic caps in the input (C1 + C8).Electrolitic caps are polarized and your input signal is AC - it has both positive and negative peaks relative to ground. . At the voltage and current in your use case, is is unlikely to die immediately, but I am still confused as to your reasoning.

This is the reason why I made the post. I would have never noticed this obvious design mistake... The electrolytics were used since jlcpcb had them in stock in the right size, but they are clearly the wrong choice of caps and will be changed.

Thanks for the recommendations, I will take them into account for the next revision.

hi so could you go Into some more detail about how you designed this did you use someone else's design ? I am interested in making something like this

The overall design is my own and the scripts used are my written by me. The scripts are only a glorified system of equations solver and the equations can be found at linkwitzlab.com.

The design specifications (notch filter parameters, high/lowpass frequencies) are specified by trial and error, programming the crossover on the minidsp and listening how it sounds, then measuring the response. Rinse and repeat.