GEOFEX ABY splitter problems.

[Greg Moss]

Hey everybody,
I'm working on the a/b/y splitter from GEOFEX:



I'm having a little problem not so much with the electronic component end of things,  but with the switching.  It looks to me like when the right hand dpdt (that is, the "A+B") is in the "up position" (which I am assuming is puting the pedal in A or B mode), it is grounding the input of the "Output B" stage.

This seems like a good idea when you are routing the signal to "output A" but when routing to output "B" it is sort of counter productive....

I've been messing around on paper, and I think I can get it to work using 2 DPDT's but without the led's.  I think I could get it working with a DPDT and a 3PDT.  Other than that, I'm Stumped.

Help!

Greg

[Greg Moss]

Here's the addendum to the above post:

Specifically, what is working and what isn't is as follows.  

*Output "A"- the one that's hardwired - works fine,  
*the "Both setting works, but the
*"Just B" (the buffered output) won't work unless I lift the ground lead going from the right side DPDT.  

*If I just cut that ground lead, the "Both A+B" setting works fine, but when output is switched to the "A" output, the "B" output buzzes like crazy....

Should I be addressing a problem with the wiring configuration between the switches, or is there maybe something wrong with the buffer stage going to output "B?"

[moosapotamus]

Well, I guess you've confirmed that the B output needs to be grounded when not in use, right?

You're using an isolated jack for output B, so neither the sleeve nor the ground ring connect to ground, right?

Grounding the unused output is nice. But, you need a way to break the ground connection to both outputs when you switch to A+B. To do that, it sort of seems like you would need to use a 3PDT for one of the switches. But I'm not sure exactly how that would work.

It doesn't look, to me, like there is anything wrong with the buffer stage for output B.

Best...
~ Charlie

[Greg Moss]

Well, I guess you've confirmed that the B output needs to be grounded when not in use, right?

You're using an isolated jack for output B, so neither the sleeve nor the ground ring connect to ground, right?

I am using one of those isolating nylon washers on the "B" jack.   I checked for continuity between ground and the sleeve of the jack, and they are indeed isolated.  (Thank you, nylon).

Having the the "B" output being grounded when in use doesn't make much sense to me, But grounding the it when not in use definatly does.

Grounding the unused output is nice. But, you need a way to break the ground connection to both outputs when you switch to A+B. To do that, it sort of seems like you would need to use a 3PDT for one of the switches. But I'm not sure exactly how that would work.

I figured out a way of using the LED swiitvhing portion of the first DPDT (the top pole) to ground one of the outputs  (and at the same itme completing the LED curcuit for the other channel.  The only problem is that this scheme would send 9V to the output not grounded (Bad!)

I wonder what would cause the output buffer to become noisy when not in use?


Anyway, thanks much for your help,

Greg

[moosapotamus]

I'm only guessing, but...
Come to think of it, I'd guess maybe when the B output is not in use the buzzing is actually to be expectd. Ever build an effect and somehow leave off the ground connection from the output (or the input, ftm)? It'll buzz like crazy. In the active splitter, isolation keeps the B output quiet in use, but makes it hum when not in use... maybe??

~ Charlie

[Peter Snowberg]

That schematic makes no sense to me. This is the switching for my A/B/Y box using two 3PDT switches. Pole 1 for each switch is for status LEDs.

Pole #2 of switch #1 takes the audio in and switches that between output A or output B.

Pole #2 of switch #2 shorts both outputs when "both mode" is engaged. There is no ground connection to this pole as in the shown schematic.

Pole #3 of switch #1 switches a wire to the opposite output of switch #1 pole #2.

Pole #3 of switch #2 switches that wire to ground when single output mode is on. When both mode is engaged, that ground gets removed and the wire floats.

I hope that helps.

BTW: I use dual color LEDs. The A/B switch shows the current side with green. The "both" switch turns on both reds. In this way one becomes yellow and one red so you can see which side you will come out on when you disengage "both mode".

-Peter

[R.G.]

It's been a few years since I drew that one up. I'll see if I can retrace the operation.

[Greg Moss]

Thanks, any suggestions are greatly appriciated,

Greg

[Greg Moss]

I wouund up just using the schematic pretty much as posted.  I elimiated the problem of grounding the "B" out put by caving in and using a 3PDT switch on the A/B selector, and using the third pole to break the connection to the pole of the"A+B" switch which goes to ground.

I was hoping to get away with just using DPDT's.  Drat!

BTW: I use dual color LEDs. The A/B switch shows the current side with green. The "both" switch turns on both reds. In this way one becomes yellow and one red so you can see which side you will come out on when you disengage "both mode".

Hey Peter,

That's a cool idea! What kind of dual color LED's are you using?  Pretty much the only ones I could find easily were some Vishay ones through mouser.  - I used them on an earlier project.  I like trying different stuff though.

Thanks again for the help folks,

Greg

[Peter Snowberg]

That's a cool idea! What kind of dual color LED's are you using?

They're the dual color type with three legs. I don't remember if they're common anode or cathode, they were just in the bottom if the box one day. They're not very bright so it's not a very good solution for battery power but since they're switching amps anyway...

-Peter

[Mike Burgundy]

you could get away with DPDT's the way I see it - I'm thinking the ground connection there is a mistake.
Sorry RG, I hacked your drawing for this (so all rights reserved and kudos etc to you )
This is the bit it's all about:

The upper part of both switches is concerned with the LEDs, the lower part with the signal. As shown, the pedal is in A, and in "or" setting. Let's ignore the A+B switch for now, and pretend it's not there. throwing the AorB switch does what's expected:

the LEDs switch to LED and signal now gets routed to opamp B.
If we sneak a peek at the second switch, we see that indeed input B is *always* grounded when that switch is set to "or". This is not right.
So what does that switch do?
The upper part is now not doing anything, the lower is wrongfully grounding B.

Now, the upper part shorts the A and B LED lines together (both on, no matter what), the lower part shorts the signal lines together.
It's a logical step to conclude RG meant to have both switches to do the same thing, so:

Just leave out that ground connection.
No grounding of any input though. On the other hand, if we were to ground B when not in use, we should also ground A when not in use...
hih

[MetalGuy]

Hi,

I'm resurrecting this thread because I didn't want to start a new one and because it has the schematic I'd like to comment on (the lower part):
http://www.geofex.com/FX_images/oaspltr.gif
I needed a quick 1 to 4 splitter (fixed, no switching involved) to run guitar signal to PC and 3 tube amps simultaneously. This splitter from GEO site came first to mind so I did a PCB sketch but went for a perf board. I used a charge pump to get the negative voltage.
In two words this thing doesn't work or to put it mildly it has PROBLEMS. By problems I mean strong audible buzz mixed with some kind of pitch shifting white noise present at all non grounded outputs.
The board is placed in a grounded 1590B box. All jacks are plastic, isolated from the box ground. I'm feeding the splitter from a 9V regulated adapter. Tried batteries without improvement.
In my current test setup I'm running signal to one PC and two tube amps. All devices are connected to the mains AC with three prong cables. Where I live we don't have a real ground but - all AC outlets' grounds and electronic equipment chassis are grounded to the "0" line (via the outlets).
After several hours of experiments and voodoo rituals I got finally this thing to work by connecting the floating isolated output jacks' sleeves to ground via 100nF caps (10uF provided even better results) and after adding 150pF caps across pins 1-2 and 6/7on each opamp section.
I would appreciate if you share your comments, suggestions or experience if you built this splitter.

[R.G.]

By problems I mean strong audible buzz mixed with some kind of pitch shifting white noise present at all non grounded outputs.

That's a reasonably good description of problems with the ICs oscillating.

[MetalGuy]

Thanks for chiming in.
The ICs are not audibly oscillating now. Maybe schematic needs some update.
I forgot to mention I tried NE5532 and LM833 with same results. They are placed in a single DIL-16 socket all power supply legs are bypassed with 100nF caps.

[R.G.]

The ICs are not audibly oscillating now. Maybe schematic needs some update.
I forgot to mention I tried NE5532 and LM833 with same results. They are placed in a single DIL-16 socket all power supply legs are bypassed with 100nF caps.

They may not oscillate audibly. It may be well above audio. "White noise" and "pitch shifting" are some of the things that can come out when an opamp is oscillating and modulating the oscillation frequency on top of any audio going through it.

This is not always the case, but it's the thing that comes to mind first with that description.

Notice that this version of the splitter is running on the idea that an opamp will act as both a mixer and a differential amplifier at the same time, subtracting any ground differences between its output jack and the input signal while adding the input signal. It's trickier than your every day application of opamps.

I'd have to see your schematic to get further than that. This is one set of symptoms that would be much quicker to debug with an oscilloscope to confirm or deny the presence of oscillation.

[MetalGuy]

This is one section of the splitter:



I added C1 and C5 to make things work.

This is roughly how perfboard connections look like: