Guitar amp effects send balancer/de-balancer

[Newmarket]

I do actually have some 10k:10k transformers that i intend on using as one final check that i do in fact have an rfi problem, but ultimately i would like to pursue this active solution, even if its only a learning exercise. the parts count is low enough, that i dont mind giving it a try

Fair enough. Great experiential learning exercise. Let us know how you get on.

 

[abbey road d enfer]

Not sure that would be the case. Since the 'Ground' (and I hesitate to use the term !) ie Screen is directly tied between the Pedalboard and distant Amp in the scenario that I think is being outlined.

They may be, or not.

The amplifier has an unbalanced input so using that Screen connection as the 'Signal Reference' by definition. Looks like a baked in 'Pin 1 Problem'

The amplifier "ground would be connected to the interface "ground, with a very short length of cable. This is excatly like teh balancing and debalancing stages would be built in the amp or the pedal board.

The situation "between an outboard and a mixer" seems fundamentally different if we are considering a connection that is 'balanced' at both output and input.

The proposed scheme exactly provides balanced connections.

 

[grid_stopper]

So far there is only one "glaring" mistake. The voltage follower in the debal has no galvanic connection at the input. It should have a resistor to ground.
However, this stage is utterly dispensable, since the THAT line receiver is perfectly capable of driving any of the usual unbalanced inputs in a guitar rig. You may want to move the 100R resistor to the output, though.
The balancer is fine, although the 10uF/10k between the input buffer and the THAT line driver is also dispensable.
Ther are commercial units that do that, I think Radial.
There is one point that should be considered. Adding gain to the balancer, compensated by complementary attenuation on the debal runs the balanced lines at a higher level, which makes them somewhat less sensitive to some forms of interference. Turning teh follower into a non-inverting gain stage just takes 2 resistors.
The THAT line driver already has 6 dB gain. Adding a little additional gain to the input buffer provides some noise reduction. Of course it must be compensated by some passive attenuation after the receiver in the debal. It takes just an additional resistor.
The same circuit would make a perfect -10/+4 interface.

Ok, so here is where i got to, based on your advice. I added some gain at the unbalanced input, and added a simple/equal voltage divider at the debalancers output.

 

[Newmarket]

They may be, or not.

Indeed. I thought this was what ccaudle was positing. But guitar amp screens / grounds / earths get handled in various ways.

The amplifier "ground would be connected to the interface "ground, with a very short length of cable. This is excatly like teh balancing and debalancing stages would be built in the amp or the pedal board.

The screen of the unbalanced cable is also a signal conductor here. But yes - if the cable is very short it likely becomes insignificant.

The proposed scheme exactly provides balanced connections.

I get that. It's the unbalanced sections I am considering. But, as above, if the cable runs are short it likely all good.

Pulling back from the detail of the solution sketched out and looking at the problem itself.
As I understand it, we want to minimise interference noise problems into an unbalanced input stage.
So it occurs to me that there might be a benefit in looking at a Ground Sensing type output or a Cross Coupled type solution (discrete or ic) that gives a degree of CMRR into an unbalanced input (with appropriate wiring).

 

[Newmarket]

Ok, so here is where i got to, based on your advice. I added some gain at the unbalanced input, and added a simple/equal voltage divider at the debalancers output.
View attachment 93755

As Abbey pointed out previously - you don't need unity gain buffer after the THAT balanced receiver. But you could use it to advantage by placing the attenuating divider before the buffer. Thus giving a low impedance output compared to the version drawn with the output coming directly from the divider junction. Also more accurate though that may not be an issue. The 100R resistor would then have some function. As drawn it is superfluous since you already have 5K0 in series with the op amp output.
Small detail - 'GND' connects to the 'Chassis Earth' symbol in the power supply section ?

 

[abbey road d enfer]

So it occurs to me that there might be a benefit in looking at a Ground Sensing type output or a Cross Coupled type solution (discrete or ic) that gives a degree of CMRR into an unbalanced input (with appropriate wiring).

Agreed. That would be another level of noise rejection. Actually both the unbal output and the unbal input could be ground-sensing.

 

[abbey road d enfer]

Small detail - 'GND' connects to the 'Chassis Earth' symbol in the power supply section ?

In the same vein, the 47pF cap is supposed to go the chassis ground.

 

[Bo Deadly]

Agreed. That would be another level of noise rejection. Actually both the unbal output and the unbal input could be ground-sensing.

It's not clear to me how ground sensing would work since the ground of the pedals and guitar must be connected to earth ground of the amp through the shield. Maybe it would help a little if the balanced send at the amp was connected directly at the pedal board using a TR plug with cold and shield connected together. That's usually how I connect balanced to unbalanced stuff. Sometimes you can leave the shield open at the unbalanced end but not for guitar unless it's getting earth from somewhere else.

But again, IMO simple impedance balanced drivers would be more appropriate for a variety of reasons and more generally, I would identify and characterize the noise more carefully before doing anything.

 

[abbey road d enfer]

It's not clear to me how ground sensing would work since the ground of the pedals and guitar must be connected to earth ground of the amp through the shield.

Well, the interface has its own "ground", which is not the same as the amp's. Using ground sensing would cancel whatever noise exists between these two grounds. It don't think it is significant in practice.

But again, IMO simple impedance balanced drivers would be more appropriate for a variety of reasons and more generally, I would identify and characterize the noise more carefully before doing anything.

Agreed. Ground-sensing would be overkill in this configuration.

 

[Newmarket]

Agreed. Ground-sensing would be overkill in this configuration.

I think there may be some misunderstanding about what I intended to propose. Possibly I didn't articulate it properly.
My suggestion was simply to drive the long, currently unbalanced, cable runs with a ground sensing circuit and appropriate screened twisted pair wiring.
That should be able to go straight into the "unbalanced" input jacks (I caveat that with "should" only to allow for something 'odd' about eg a valve (tube) amp input stage that is an issue. I can't think that there is but valves aren't my thing).

I realise that the OP wants to implement the original idea as part of a learning process. So just thinking around the issue.

The Impedance Balanced into a balanced receiver route seems a good option too.

Obviously we don't know the exact nature and severity of the actual problem. Definitely agree with " I would identify and characterize the noise more carefully before doing anything."
Particularly since the electrics seem a bit on the 'dodgy' side. That would concern me from a safety point of view apart from the problems with the audio kit.

 

[abbey road d enfer]

I think there may be some misunderstanding about what I intended to propose. Possibly I didn't articulate it properly.
My suggestion was simply to drive the long, currently unbalanced, cable runs with a ground sensing circuit and appropriate screened twisted pair wiring.

I believe you mean ditching the idea of a bal/debal, thus replacing the balanced connections with ground-sensing connections.
It would actually be as complicated as a bal/debal, because it is hardly possible to modify the amp and pedal board to convert the i/o's from strict unbalanced to ground-sensing, so a box with electronics should be added to the set-up.

 

[Newmarket]

I believe you mean ditching the idea of a bal/debal, thus replacing the balanced connections with ground-sensing connections.

Yes

It would actually be as complicated as a bal/debal, because it is hardly possible to modify the amp and pedal board to convert the i/o's from strict unbalanced to ground-sensing, so a box with electronics should be added to the set-up.

Certainly a 'Box of Electronics' would be envisaged. By saying ""simply to drive..." I didn't mean to suggest I thought it was trivial to physically implement .
I think it might be a little simpler / cheaper than the THAT based bal/debal option depending on the detail. But not much in it. But possibly not as good in technical performance. I do like the THAT stuff.

Looking at the OP's schematic - I'd be wary of using the DC/DC converter there to provide +/-15V rails.
I've used/specified DC/DC converters in that form previously (+/-15V from a single +5V rail) in pro audio product. But they were a type with integrated linear regulation after the DC/DC stage. And we had input filtering as advised by the manufacturer (Newport IIRC) and additional passive L-C filters.

 

[ccaudle]

I'd be wary of using the DC/DC converter

Could be worth a try, although I just noticed the datasheet does recommend two 1500pF caps from outputs back to inputs "to meet BS EN/EN55032 class A/B emission." The caps (2, not just one) are 150pF for class A, but 1500pF for class B. I would always shoot for class B if possible.

 

[Newmarket]

Could be worth a try, although I just noticed the datasheet does recommend two 1500pF caps from outputs back to inputs "to meet BS EN/EN55032 class A/B emission." The caps (2, not just one) are 150pF for class A, but 1500pF for class B. I would always shoot for class B if possible.

Yes. Definitely worth a try. I'd just be wary if I wanted to get good performance with least hassle/delay.
But it does look a potentially useful part. When I've done similar I've had to stack up individual single output supplies. Apart from the component count it does potentially give rise to other issues if the clocks are not sync'd.

 

[abbey road d enfer]

Looking at the OP's schematic - I'd be wary of using the DC/DC converter there to provide +/-15V rails.

I'd be too.

And we had input filtering as advised by the manufacturer (Newport IIRC) and additional passive L-C filters.

Datasheets are notoriously lacking regarding characterization of noise.
"Ripple & noise: 100mV p-p" is not enough.
I would like to see something like a noise spectrum in datasheets.
I know it's veering significantly, but it may be the most interesting point in the whole thread.

 

[grid_stopper]

Sorry for dropping off the map- had a mid-month move to a new apartment, and only recently got settled and what not.

Regarding the DC/DC convertor- I have successfully used that schematic before in a preamp pedal, and although I have never measured its noise-it has worked well to my ear. I would be happy to dig in and measure it though, if people are curious. perhaps some improvements could be made!

As Abbey pointed out previously - you don't need unity gain buffer after the THAT balanced receiver. But you could use it to advantage by placing the attenuating divider before the buffer. Thus giving a low impedance output compared to the version drawn with the output coming directly from the divider junction. Also more accurate though that may not be an issue. The 100R resistor would then have some function. As drawn it is superfluous since you already have 5K0 in series with the op amp output.
Small detail - 'GND' connects to the 'Chassis Earth' symbol in the power supply section ?

With regards to this, I ended up changing the schematic to reflect these sugesstions(attenuating divider before the buffer). My intention is to build it up early next week and see where im at.