Hum Free ABY build problems

[Barcode80]

Okay, just finished the active A/B/Y from GGG, and i've got some issues. The a/b switch does in fact switch the A and B signals. The combine also works. HOwever, the LEDs only light up in combine mode. I used 4tM013 ass the transformer instead of 4TM018, which i think may be the problem. Also, the signal is weakened significantly. Any thoughts? I'll try to get a build pic up soon for more help.

[R.G.]

"What to do when it doesn't work"

[Barcode80]

And normally that would be step one. But I guess I'm looking for more understanding of the circuit to know what to look for. I'm just not really clear on how it operates. Could you maybe explain what is actually happening in the circuit to make the switching happen, and to make the circuit function in general? Transformers are sometimes like a magic box to me.

[R.G.]

Sure. I'm referring to this: http://geofex.com/FX_images/humfree2.gif just in case that's not the one you used.

The two output sections are identical, of course. They consist of an opamp set up as a gain-of-one follower. The follower output voltage drives the primary of a 10K to 10K isolation transformer. The secondary provides the isolated output. The 10K/0.001 R/C network on each secondary loads down the secondary of the transformer at very high frequencies and controls extraneous resonances outside the audio band, as well as providing a minimum load at high frequencies.

Transformers ARE magic boxes.    I was always accused of loving them too much back when I was designing power supplies.

In this setup, you can think of them as providing an output on the secondary that's equal to the input on the primary over the range of 60Hz to 22kHz. The secondary is electrically disconnected from the primary - an ohmmeter would show an open circuit. That's really all you need to know for this one. All the A/B/Y/etc goes on back before the two opamps which drive the transformers.

An so let's trace that path. The input opamp just buffers the input signal at a gain of one. Same voltage, but lots of current available if needed. That goes through two paths, one for the + input of each output opamp. The path from input buffer to output opamp is two series 10K resistors.

Mentally, set the top switch to the 'not-Both' setting for the moment. That connects ground to the pole of the A/B switch section that connects between the two 10k/10K signal paths. The A/B switch then shorts either A or B to ground. It has no effect on the other side because of the 10K isolating resistors. So one or the other is shorted out and can't operate.  So both outputs are normally on; the A/B switch shorts the unwanted one out.

Both LEDs are set up to be on all the time; they each go to +9V through an independent resistor. Just like with the signal, the LED side of the A/B switch shorts out one or the other LED. The LED is shorted, no voltage across it, so no light comes out. Current still flows on both sides of the LED paths, but the switch "eats" the current and keeps one or the other LED from lighting.

You can see what's next, right? When you flip the "both" switch on, what it does is open up the ground connection for both the LED side and signal side of the A/B switch. The A/B switch still selects one side not to work, but because the ground path is open, it can't stop either side from working, so you get both.

[Barcode80]

Boy, that does make a lot more sense now! Hmm, it does befuddle me though, if the paths are switching then the LEDs should too if I'm understanding you correctly. If an output is working, since it is turned on/off by a short to ground, then when the path is operational the LED is also connected to ground. so it should work 

[R.G.]

Notice that there is no link between the signal turning on/off and the LEDs turning on/off. They're completely independent.

Here's a way to debug. Remove the switch wires which would short either of the signals and the other ones which would short either of the LEDs. Now both LEDs shine all the time, and both outputs are full on all the time. Right? If not, your signal path is hosed up.

If the signal path works as expected, clip one end of a wire to ground and ground first one, then the other signal-kill points. Do the signals go off as expected? If so, now short one then the other LED. Do they go off as expected? If both of these work, the problem is in your switch wiring.

[Barcode80]

Notice that there is no link between the signal turning on/off and the LEDs turning on/off. They're completely independent.

Here's a way to debug. Remove the switch wires which would short either of the signals and the other ones which would short either of the LEDs. Now both LEDs shine all the time, and both outputs are full on all the time. Right? If not, your signal path is hosed up.

If the signal path works as expected, clip one end of a wire to ground and ground first one, then the other signal-kill points. Do the signals go off as expected? If so, now short one then the other LED. Do they go off as expected? If both of these work, the problem is in your switch wiring.

i see what you are saying. I'm also a little suspicious of the switch. since both of the LED connections are on one side of the switch, that side could be hosed up. I'm using some DPDT switches i got from effectsconnection a few months back and out of 10 i ordered, i think i've successfully used 2 that didn't have mechanical or continuity problems. perhaps this one is in the same boat.

[Barcode80]

Okay, so i swapped out the switch and it looks like that was one of my problems. Everything functions perfectly... except that my guitar signal level is about half what it is when i don't run it through the splitter. I'm guessing it is the transformer, what do you think RG?

[R.G.]

I think it's the transformers, too.

I was lazy at reading your first post and didn't go track down the transformer specs. The 42TM013 is a 1K to 8 ohm transformer, which is an 11 to one voltage step down; yep, sounds a little less than half as loud.

In this case you really do need the 42TM018s for a couple of reasons.

[Barcode80]

excellent! thanks for all your help RG, it is invaluable. Will you be at NAMM this week?

[R.G.]

Yes. I should be in the Visual Sound booth most of the time. Check the catalog for us. I should know the number, but I'm a bit frantic right now getting packed.

We're right next to George L's cables.

[Barcode80]

Yes. I should be in the Visual Sound booth most of the time. Check the catalog for us. I should know the number, but I'm a bit frantic right now getting packed.

We're right next to George L's cables.

can i buy you a beer?