Seeking advice in balanced -> dual unbalanced -> balanced

[GXAlan]

I want to build an analog balanced differential version of the Bose 901 EQ box using a pair of Bose 901 EQ boxes.

Background:
1) In my room, Dirac Live EQ of the Bose 901 to a Harman +10 dB curve with me sitting and holding the mic (as opposed to using a tripod) is similar enough to the factory EQ that I can go room-correction free and be reasonably happy.

2) I have tested the 901 with a 300B SET/factory EQ box and the Dirac-enabled Arcam SR250 using HDMI. They are both effective solutions, but I am curious about a third option for just curiosities sake.

3) I have two Bose EQ boxes. These are two-prong ungrounded devices.

Goal
I would like to try a Bose 901 with my Marantz SA10/PM10 combo. The DAC and integrated amp keep everything in balanced differential mode from start to finish.

I basically want to build an analog balanced differential version of the Bose 901 EQ box so that I can keep the “fully differential circuitry from start to finish.”

My thought is that I can take a true balanced XLR output and run it through a pair of EQ units and end up with equalized signal that preserves the balanced differential transmission.

I will probably use one of these so it will be easy to tinker.

Questions
1) Does it make more sense to make the EQ boxes “monoblocks” so that the + and - of the same channel are EQ’d by a single unit ?
or
Should one box run L/R+ and one EQ box run L/R-

2) What should I do with the ground pins? Tie everything together? Have separate grounds for the left and right boxes?

 

[GXAlan]

Tagging @Rick Sykora for advice in case the thread title didn't catch his attention. (Sorry, Rick, you've got the reputation of being the local, helpful technical expert who has the patience to answer Electrical Engineering 101 questions).

 

[voodooless]

So if I understand correctly, you want to recreate the EQ box with balanced in and out?

I’d probably just use a balanced line receiver, then single ended EQ, and finally a balanced line driver as output. That would be much simpler than a fully balanced EQ circuit.

 

[pseudoid]

2) What should I do with the ground pins? Tie everything together? Have separate grounds for the left and right boxes?

If you really are going to do a redesign of a board, may I recommend to you to identify all of your returns, and grounds.
*You have 'earth' ground ('chassis' ground from 3-prong AC power supply)
*You have Input L- and R- (returns)
*You also have Output (XLR?) ground potentials (in addition to +/- of L&R channels)
*You have DC ground, sourced from the power supply DC rails.
*I don't know if you will be using digital circuitry in your design...if so, then, you need to treat it as another ground....
Providing earth/ground/return pads (PTH vias) on the PCB will allow you to "play-around" with grounding as necessary, w/o the need for a redesign of the PCB (if any).
Designing a flexible ground system, during build, allows changes in the grounding scheme later.

 

[Rick Sykora]

Tagging @Rick Sykora for advice in case the thread title didn't catch his attention. (Sorry, Rick, you've got the reputation of being the local, helpful technical expert who has the patience to answer Electrical Engineering 101 questions).

Am on the road with limited internet access, so may be slow to respond…

What device is in your pic? Looks like one of the Hypex DSP boards but was not sure.

Recall your main page post said the Bose eq might distort significantly? Are you looking to replace it entirely? If so, what is your budget?

 

[GXAlan]

Am on the road with limited internet access, so may be slow to respond…

Ah, no worries.

What device is in your pic?
Recall your main page post said the Bose eq might distort significantly? Are you looking to replace it entirely? If so, what is your budget?

Sorry for the small pic. Way dumber than that. It's just an XLR to phoenix connector dumb pin-out so I can play with grounding instead of soldering/unsoldering in the case the answer is "just try both options."

Budget is kept low since this is more for fun/curiosity to see how it performs. I already have Dirac as one option to replace the unit which works great.

So if I understand correctly, you want to recreate the EQ box with balanced in and out?

I’d probably just use a balanced line receiver, then single ended EQ, and finally a balanced line driver as output. That would be much simpler than a fully balanced EQ circuit.

In my mind, the options are:

versus

And if the Bose EQ itself introduces a ground loop, maybe I need to do something like this?

 

[voodooless]

Ah, so you want to use the original stereo EQs in a balanced configuration.

I would probably not do that. You’ll need very good matching between the channels to make this work properly. Normally you’d use extra low tolerance caps and resistors for a fully balanced EQ circuit.

I’d still go for my original recommendation. Some actual EE may shed some better light in this though.

 

[Speedskater]

While balanced interconnect systems are always good. An all balanced internal circuit is an engineering challenge and seldom beneficial .

 

[GXAlan]

While balanced interconnect systems are always good. An all balanced internal circuit is an engineering challenge and seldom beneficial .

+1 engineering challenge

+1 seldom beneficial (esp. since I have a Dirac alternative)

But curiosity is hard to ignore (unless I break these units while experimenting).

Still no clear answer on best practices. I think I will just have to experiment on my own…

 

[KSTR]

Ah, no worries.



Sorry for the small pic. Way dumber than that. It's just an XLR to phoenix connector dumb pin-out so I can play with grounding instead of soldering/unsoldering in the case the answer is "just try both options."

Budget is kept low since this is more for fun/curiosity to see how it performs. I already have Dirac as one option to replace the unit which works great.



In my mind, the options are:

View attachment 289424

versus

View attachment 289425

And if the Bose EQ itself introduces a ground loop, maybe I need to do something like this?
View attachment 289430

The whole idea and reasoning of balanced connections is that the receiver subtracts hot (Pin2) and cold (Pin3) wires signals as seen with reference to the local GND of the receiver. By this, the common-mode noise content is subtracted out and not propagated any further. Voltage symmetry from the sender is not required for this and should not be assumed (even though it is often the case) whereas impedance symmetry is actually important. Only with impedance symmetry the subtraction is pretty much perfect but without the balanced advantage still is there for the most part.

What does that mean for your setup above?
The EQs will slightly be different even when adjusted to the same values, from any number of tolerances we may have here... and this spoils the subtraction precision quite a bit. It will still work OK but you don't gain much from using dual EQs and the results will depend on your actual grounding scheme and noise sources and could vary from barely any improvement to perfectly fine, zero hum/buzz.

Therefore I would suggest the following simplification for a start:
- locate one single EQ as close as possible to the source L and R feeds, using shortest and good quality RCA interconnects
- fan out to balanced cable at the EQs output RCA Jacks: Pin1 and Pin3 to shell, Pin2 to core. If you happen know the output impedance of the EQ you can insert the same resistance in the cold (Pin3) wire at the sender for impedance symmetry.
This should really do. At least when the Bose EQ itself doesn't hum like crazy (and I'm assuming the EQ is class-ii, 2-prong mains).

Obviously, for fun and experience I certainly do encourage you to try out several ideas, never blindly rely on anything someone writes on the internet ;-)

 

[GXAlan]

It will still work OK but you don't gain much from using dual EQs and the results will depend on your actual grounding scheme and noise sources and could vary from barely any improvement to perfectly fine, zero hum/buzz.

That’s what I am curious. Improvement? No difference? Worse?

This should really do. At least when the Bose EQ itself doesn't hum like crazy (and I'm assuming the EQ is class-ii, 2-prong mains).

Yes, 2 prong mains.

Obviously, for fun and experience I certainly do encourage you to try out several ideas, never blindly rely on anything someone writes on the internet ;-)

Haha.

I am realizing that no one is really sure what’s going to happen. That’s a good enough as any to try it out.

 

[KSTR]

Oops, brain fart. Your source is balanced, so you would need to use extra balanced to unbalanced modules at the EQ's input for full compliance which complicates things. If we assume a "standard" balanced output from the CD/SACD you might get away with just using hot (Pin2) and GND (1) for the short unbalanced interconnect to the EQs, though.

 

[GXAlan]

Oops, brain fart. Your source is balanced, so you would need to use extra balanced to unbalanced modules at the EQ's input for full compliance which complicates things. If we assume a "standard" balanced output from the CD/SACD you might get away with just using hot (Pin2) and GND (1) for the short unbalanced interconnect to the EQs, though.

I have true balanced differential output from the Marantz SA-10 and I have four channels of unbalanced EQ to process the two channels of balance signals…

That’s the goal.

 

[Rick Sykora]

As has been mentioned, am not sure if any advantage might be achieved here. Am guessing you might end up connecting floating and earth grounds and may do nothing or cause ground loops. If you still want to try, would reduce the wiring to as little as possible to get signals to/from the Bose box.

As I think you already know, any real advantage would be to replace the old electronics with new and upgrade to balanced input and outputs along the way.

 

[GXAlan]

The good is that it reduces noise. The bad is that the EQ curve no longer matches properly, suggesting that the left and right channels are not perfectly matched.

Ah well, not so straightforward. Back to Dirac it seems.

Green is correct, Red is balanced attempt

The distortion obviously drops