Active XO Bi-amping with Trends TA10.1

[Pano]

Well Mike, I've never heard of someone doing an active 4th order cross to a super tweeter. But you certainly can't be the first.

Why not beg, borrow, buy or hire a Behringer DCX2496 or similar digital crossover? It will give you the flexibility to try almost infinite variations of slopes, filter types, crossover points, phase and time alignment.

Once you find what works, imitate it in an analog circuit. (Easier said than done, of course)

[OzMikeH]

Unfortunately no chance of borrowing a Behringer digital XO. Many people use those, they appear to be fairly highly regarded yet they have A-D then D-A then a fairly dodgy looking op-amp output. I don't see how that could be better than a sensible op-amp based filter. I had a good look at Behringer's two analog XOs but they don't appear to be very accurate. Good for the price but not what I'm looking for.
I tried to cancel my parts order for the XOs but didn't make it in time. The parts are in transit.

Since the money is already spent I may as well experiment a bit. It would be nice to run a pair of shadow subs later on so a three way XO is a likely, this is why I bought two XO boards. It looks pretty easy to alter to 2nd or 3rd order. I'll look carefully at the input circuit to the T-amp - I may be able to delete the output buffers. Likewise with my preamp, I may be able to remove the input buffers too.
Any suggestions for making my half-cocked bazooka more sensible?

[OzMikeH]

Here is a possible layout. I should have ordered a bigger box.
It's a 1RU 19 inch rack case with the "ears" missing, I'm using it the long way so the sides are the front and back. It's well ventilated and I have placed the switched mode supplies and the trends amp boards over the ventilated sections.

One thing that concerns me is running two class D amplifiers in close proximity. They can intermodulate if the clocks aren't synchronised. According to the datasheet for the chip they don;t have a master clock feature. If they start interacting I'll have to shield them from each other. Running two of these from a single power supply would not be a good idea.

After testing theTA10s will be stripped, the boards mounted in the corners and hardwired as power amps with no connectors.

There's easy access to the crossover boards to try different slopes and op-amps. I'll need to swap over some of the TA10 input caps, the amps I got have different caps in them, are the Bennics or the yellow ones with the little triangle logo better?

[OzMikeH]

Wow, took a while to find my old thread, I've been putting this together for a LONG time.

The first version is about 50% complete.
I'm running Burr Brown OPA2134 opamps instead of the stocker TL072.
I'll give this one a run like it is. then I'll remove the output buffer on the XO.

Where to start, I found a nice clean 10A continuous 12V external switchmode supply. So that connects through the banana sockets, then to the right hand side of the capacitor board. those caps are all ultra low ESR 1000uF clinch soldered across two tracks of the vero board with a bit of wire laid on top. There are 2 separate banks, one for each amp. Then out the left hand end of the capacitor board to the underside of the trends boards soldered directly onto the standard filter caps. the wire is only 3 inches long. That's going to be a steadier supply than any battery on the floor can ever be.
The unused tracks in the centre of the capacitor board are bussed together and connected to protective earth.

Trends amps have had the input caps taken off the board, header pins have been put in their place on the Tripath chip side of the capacitor positions. the normal input tracks have been grounded.

The little vero-boards have reed relays to mute the XO board and the input caps off the Trends amps. I chose reed relays because glass encapsulated contacts in an inert atmosphere will be as close to a piece of hookup wire as you can get. All unused tracks and parts of tracks on the vero board have been connected to signal ground.

The ESP boards are pretty self explanatory, I used two because I intend to upgrade with some proper subs in future to take some of the load off the fullranger. I havent installed the op-amps yet, or the decoupling caps. I know I've got some axial ceramic caps somewhere, I prefer to solder them directly to the power supply pins of the chip on the underside of the board. I have no idea how much RF the Trends amps will radiate inside the box.

Output connectors are Neutrik Speakon connectors, these are ideal because they are a 4 pole connector so I only have one cable going to each speaker cabinet. Hopefully I'll have this ready to test in a week. I'll take it into work to measure & align the output, 2 amps in the same box might interact and produce some RF funnies on the outputs.
Once I know it works I'll start reducing the number of op-amps to the bare minimum.

Almost time to start hunting Wabbits. <Kaboom!> hahaha