active crossover unstable

[pjwd]

Hi - I have built a balanced active xover for a woofer and it is quite noisy when input cable is plugged in and very sensitive to cable and equipment movement - it works very well when I can get it quiet but will not necessarily be quiet when I turn it on again.

this is the first time I have built anything like this and I expect it could be the way it is constructed causing capacitance and inductance in circuit as it is a point to point wiring on a tag board with large copper voltage supply rails - here is the circuit and a picture of construction.

I suspect it needs to be on a printed board - if someone could let me know if this is a hopeless way to proceed I would appreciate it

thanks,

Phil

[wintermute]

Hi Phil, what are the opamps, and are you using any local PS bypass on them (ie small caps very close to the power leads of the opamps)? Low bandwidth opamps should be relatively safe, but high bandwidth ones are very sensitive to layout.

Tony.

[pjwd]

Thanks Tony - they are all burson units and there are caps to a centre ground rail at each PS to opamp - I think they are .2Uf with a 10Uf at the end of the rails. slightly embarrased about the layout and soldering

Cheers,

Phil

[wintermute]

oh ok "discreet" opamps... any idea what the bandwidth of them is? Just did a quick search and specs weren't forthcoming.

Have you tried it in a shielded case? Could simply be picking up rf or other airborne noise that a metal (grounded) case might eliminate. Also what are the black leads that aren't attached to anything? could they be picking up noise?

Tony.

[pjwd]

Burson is a bit slim on info but they do get called wide bandwidth - I have made a rough zincalume box and that did not seem to help - I have put the PSU in a metal box - those black legs are the unused pins of the opamp.

My concern is that the basic layout being spread apart like that is inherently a problem compared to the short signal paths you get in a printed circuit board

Cheers,

Phil

[Calvin]

Hi,

maybe the decoupling leads to probs. Instead of film caps of 2µF use ceramics (X7 or NPO) and parallel Electrolytics with a not so low ESR.
Keep the Decoupling Caps as close as possible to the OP-pins

jauu
Calvin

[pjwd]

Thanks Calvin - presume electrolytics allow flow to earth only - that makes sense - what is the ceramic for

cheers,

Phil

[wintermute]

Hi Phil, being spread apart certainly means more area to pick up noise, I'm not an expert on layout, so someone else will be better able to comment.

The suggested use of ceramics is because they still function properly (as a capacitor) at very high frequencies, not all capacitors work as well at high freqencies. The wider the bandwidth of the opamp, the more liklely it is to have high frequency oscillations, so minimizing high frequency noise on your powersuply rails with ceramics (very close to the opamp supply pins) will help stabalise it. Go with the NPO's if you can, over the X7's

Tony.

[gmphadte]

What is that speaker in the pdf. Is it connected to this circuit...
What is op4

Gajanan Phadte

[pjwd]

Thanks Tony

Gajanan - the OP4 is a line driver - the speaker is an exodus tempest 15" - 2 of them in separate dipole H frames wired to give 8 ohm - the base is fantastc when no humm buzz etc. interferes

The filter is input buffer , shelving filter and lowpass cutting off at 250 Hz

Cheers,

Phil