Is making an XO on a PCB a bad thing?
 

I had the idea of making the XO on a PCB and I read that some recommend against that. I also left big areas of copper (I etched only a few millimeters around the traces because I had little etchant left and I was too lazy to go buy some) and I'm wondering if the capacitance made from the big copper areas will hurt the sound. Should I bother to redo the XO's?

I'm a big fan of PCBs for everything. The capacitances from your board will be absolutely negligible in a speaker level crossover.

There's a feeling that PCBs are an unnecessary complication and may even make things worse - by hard wiring you can wire the components directly together (reducing the solder connections and intervening pcb tracks), it's easier to fix them to the board ina more robust fashion and you can space the components farther apart to reduce interactions. I remember Russ Andrews telling me of some experiments he did when working as a consultant - they took the x-over components off the pcb, spaced them farther apart and reconnected them - then fell about laughing at the size of improvement it made.

Colin

My question is kinda stupid and worrying about this is probably overkill, but it will help me sleep better tonight :D

The major thing that affects crossover perfromance is inductor spacing and orentation. I f you want to use a small PCB then mout the inductors off the board and run reasonable sized lead to them.

If the inductors are pretty near each other you have to orient them 90 degrees twisted otherwise you will get inductive coupling and mutual inductance.

I have a 1mH coil of approx 5.5cm diameter and a 0.22mH one of approx 3cm diameter. They are placed maybe 3 cm away from each other. I gues this isn't enough?

A pc board will work just fine. Just make sure your trace width can handle the expected current. Close spacing if inductors is fine if they are in the same branch. Even if they are in different branches, unless you expect large currents, they dont need to be placed very far apart. If you have a scope, look at the output of the mid or high section while feeding a low frequency signal to the bass section. How much energy do you see?

The traces are about 4-5 mm wide and I covered them with solder alloy. I think I'll hardly get into the region where current is a concern.

Besides capacitance, you can worry about eddy currents. Dahlquist tried beautifully made pc boards on the DQ10 in the late 80s and had to recall them because of too many complaints of "distorted" sound. And yes, it could be measured as well as heard. Big wide traces, with lots of conductor surface area, make nice eddy current pickups.

Like everything, it is not a case of pc boards or not pc boards, but how a particular project is designed and constructed.

With all of the free fft based spectrum analyzers available it is easy to measure component interaction when prototyping board layout.


Best of luck to you,

Mark