Construction of ABY switcher box

[Keith Hilton]

Recently I decided I wanted to use two amps. With that thought in mind, I remembered a two amp horror story that happened to me. I simply "Y" connected the signal from my guitar to the two amps. Bad idea! The first problem was a terrible ground loop issue. Different amplifiers have a different ground potential, and that is what creates ground loops and noise. After I solved the ground loop noise, the next issue became speaker phase. The speakers in the two amps were 180 degrees out of phase. The out of phase speakers did not sound good--actually awful. The next issue I faced was input impedance of each amp. One amp sucked all the input signal, requiring me to have to turn the other amp up extremely loud to get enough volume.
I decided on building myself a (ABY) box to solve all of the above problems. I looked on the Internet and saw many (ABY) boxes priced from $50.00 to almost $500.00. When I started looking at the features of the ABY boxes on the Internet, I became somewhat confused. Meaning what features should a ABY box have, and what features did I really need. Of course each feature adds more cost and more work.
I wanted to be able to switch from output A to output B, or visa versa, or run both A and B at the same time. Channel A and channel B would have LED lights. I wanted an active unit instead of a passive unit. I wanted channel A and channel B to have individual gain controls, where I could turn up, or turn down a signal to each amp. None of the Internet ABY boxes had this channel gain feature! Of course I had to have a phase reverse switch on one of the channels. The A channel would have a buffered isolation transformer. A ground lift switch would be installed for the buffered isolation transformer signal. Also transformer output jack would be isolated from the metal enclosure. Channel B would be arranged in differential signaling mode. Differential signaling mode is like an XLR system. Meaning two signals 180 degrees out of phase. The differential signaling system can be run as a balanced system or unbalanced system. Meaning I could run it as a 1/4 inch guitar jack out, or XLR out. If I ran it as a 1/4 guitar jack I would have to install a fully differential op amp to convert the balanced signal to a unbalanced signal. I need the 1/4 inch output, but I am not certain if I need the XLR output, but it could be used as a active direct box with the XLR output. If I do install an XLR output it will have protection against 48 volt phantom power.
I could install a tuner output, but wonder if I need that. Also wondering if I would need more than the two LED lights for channel A and channel B. A power light might be handy of both channel lights were off. Also wondered if I would need two inputs instead of just one. Two inputs would require a summing op-amp. Two inputs should also have input signal controls.
As you can tell, some options are essential, and some options are not essential. The absolute essentials are ground loop isolation, and phase shift option. I could totally eliminate the differential signaling in the B channel, and just run the guitar signal through with the op amp buffer. Foot switches for channels A and B would be essential. I also wondered what other switches would be needed, of course a phase reverse switch would be needed.
What would you eliminate? What would you add that I have missed? At first thought a ABY box seems so simple, but as you can see ABY boxes can be quite different.

[Larry Hobson]

How do you sleep at night! ! ?

[Michael Butler]

How do you sleep at night! ! ?

zackly!

[Steve Lipsey]

I always just order it from these guys, they seem to be deep into this sort of thing, maybe they would be willing to chat with you...
https://saturnworkspedals.com/product/a ... -switches/

Edit:
Never mind. they are too busy to chat...from their website:
"Please note: Saturnworks does not have a phone number. We do not offer phone support or consultations. We can only be reached via email. "

But maybe their site would give you some ideas...

[Mike Auman]

What would you eliminate? What would you add that I have missed?

Hi Keith, if I were doing this, I'd say build it to your current need. If you bump into something later that it's missing, either add it in, or build an upgrade. From what you said, it sounds like you would want two foot switches (A, B) with two LEDs, two buffered outputs, one of them with transformer isolation and a phase switch that flips tip & sleeve at the output, using a jack isolated from the enclosure. If you isolate one of the amp outputs, you won't have to isolate the other one. You can make this as an active buffer/splitter with up to 18dB of variable gain per channel using one dual and one single op amp. With two dual op amps, you can add a tuner-out jack. Runs on external +9V, and should fit in a 125B, including a charge pump to give you +/-9V for lots of headroom.

[Keith Hilton]

Thanks to everyone who responded to my post. Mike, I like your idea that said: "a phase switch that flips tip & sleeve at the output, using a jack isolated from the enclosure." I have never done it that way. I will have to think about that method. Also you brought up something I had not considered: " a charge pump to give you +/-9V for lots of headroom." More headroom is good in most cases.
One issue that I have not brought up is audio transformers. Lots of fancy claims about sound quality with particular transformers. Some of these small audio transformers are selling for $160.00 and up. I think the key to using lower priced transformers is having the input buffered, and the output going to a high impedance op amp. Does anyone have any suggestions on transformers they like. I am going to use a 1:1 transformer, but I have been wondering about the transformer resistances on the input and output. Of course the old telephone transformers were 600:600 ohms, but I have seen some with 10,000 ohms that look interesting.
Larry, I normally sleep pretty good. But---there are times I think about stuff before I go to sleep. I have actually solved some complex problems in my sleep. In other words, waking up and having solved the complex problem in my sleep. Yes, strange but it has happened.

[Mike Auman]

I think the key to using lower priced transformers is having the input buffered, and the output going to a high impedance op amp. Does anyone have any suggestions on transformers they like. I am going to use a 1:1 transformer, but I have been wondering about the transformer resistances on the input and output. Of course the old telephone transformers were 600:600 ohms, but I have seen some with 10,000 ohms that look interesting.

Keith, I've had success with 10k/10k transformers, and as you say, driving the input is key to good frequency response. Without the op amp (passive circuit) you'd need a $50 Jensen transformer like Lehle uses in their pedals. I did a bass buffer/splitter with Triad Magnetics TY-141P transformers for isolation, 10k/10k. The data sheet says their frequency response is +/-2dB from 200Hz to 15kHz, which doesn't sound so great. However, on the bench and driven by an op amp at unity gain, I consistently got -1 dB from 20Hz to 20 kHz, and -3 dB from below 10 Hz to 38 kHz. And they're about $6 each. I've also had good results from Xicon 42TM018-RC 10k/10k transformers which are about $5. Again they perform much better than their data sheet which says +/- 3dB from 300Hz to 3.4kHz, on the bench they're similar to the Triads, their low end rolls off a little quicker but they would still be fine for steel.

[Keith Hilton]

Mike, thanks for the heads up on the transformers. Getting back to switching phase; One would think running the signal trough a inverting op-amp would do the trick of inverting phase. The real problem with hooking up an op-amp as a inverter is two fold. #1 You double the noise. #2 The input impedance is the first resistor going to the op-amp. This means if you wanted a unity gain buffer using a inverting op-amp, the input resistor would need to be around 500K. To get unity gain that means the feedback resistor would have to be 500K. Big resistors adds noise.
Mike, just curious, how would you reverse phase if you didn't isolate the jack and reversed tip and sleeve? What method would be your first choice?

[Mike Auman]

Here's a slick way to invert the polarity that I've borrowed (full description at https://www.runoffgroove.com/splitter-blend.html.) Looking at the upper right quadrant, the signal coming in on Green/Return goes thru a buffer (Q1) then into U2A. If switch S1 is open as shown, the polarity of the signal coming out of U2A is normal. But close switch S1 and the polarity is inverted, without changing anything else - same amplitude and frequency response.

[Keith Hilton]

Mike, that circuit is way too complex. Plus I hate working with individual FETs, because even with the same numbers the parts vary greatly.
I am going to suggest two much easier ways to do it. Look up the "data sheet" for part 1646P08-U. On page 7 of the data sheet, figure 5, you will see exactly what is needed. Very simple and accurate way of splitting the signal. Out + and out - will be two sine waves 180 degrees out of phase. All that is needed then is a switch to select which one you want.
You can also create two sine waves 180 degrees out of phase with a simple 2N2222 transistor circuit. Coming off of the emitter is a sine wave in phase with the input signal, and coming off of the collector is a sine wave 180 degrees out of phase with the input signal.

[Mike Auman]

Yup, more than one way to skin a cat!