Balanced Inputs on Leach Amp

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
m0tion said:
Sorry, but those answers are a little over my head. I'd perfer to construct something that could intergrate into the existing design (so it still allows the unbalaced input to function) and would not require modifications to the leachamp PCB. Anyone have a diagram of a circuit that could accomplish this?


Well, you're going to have to alter SOMETHING. You have to connect the input to the board. If you don't want to change the input resistor on the board, you can add a 20k resistor across the connectors of the XLR so that it will be in parallel to the 20k input resistor on the Leach. You can find a circuit at the link. It ain't cheap, though. You could also wire the unbalanced input in parallel with the XLR, but don't connect to both inputs at the same time.



Jensen input transformer
 
The one and only
Joined 2001
Paid Member
m0tion said:
How would I add the ability to accept a balanced input (via an XLR connector) to the Leach Amp design? Anyone ever done this?

As I recall, the Leach is a simple enough differential input, and
is easily adapted to balanced inputs. Check out the discussion
of differential inputs on the A75 article at www.passdiy.com
and you'll see good enough examples.
 
It does have an input filter cap that Leach says is tied somehow into the compensation of the amp. It may be nothing more than the setting of the proper filter frequency for slew rate considerations when selecting miller comp cap.) This may have to be addressed. In fact, the input filter frequency may be better raised if an input transformer (with inherent bandwidth) is used.
 
Balanced inputs are another BIG LIE in audio.

"Resistors with 1% purchase tolerance yield a worst-case CMRR (common-mode rejection ratio) of 34 dB, and added drift tolerances of 2% limit the CMRR to 24 dB.

"Difference amplifiers normally have equal-value resistors, but this situation leads to unequal input resistance. The approximate input resistance that the source, V1, experiences is Z1, and the input resistance for the source, V2, is (Z3+Z4). The source resistors act with the difference-amp input resistors to form unequal voltage dividers. Thus, the resistor mismatch creates a differential error signal that the difference amplifier cannot reject.

"Author Information
Ron Mancini is staff scientist at Texas Instruments. You can reach him at 1-352-569-9401, rmancini@ti.com. "

When you need it dead silent, only a transformer will do.
 
I'm saying a transformer input works better than any electronic balanced input.

You will get over 90dB hum rejection even with an unbalanced system, and you can just plop one on the input of the Leach and be done with it. You can drive the transformer balanced or un-balanced, it doesn't care.

If you really think you need the CMRR a twisted pair provides, think again. The wire itself can't be made any better than 60dB~70dB. That's fine for voice and data, not music.

Sometimes a balanced transformer system is needed, but make no mistake, the majority of the improvement is from the transformer.
 
Transformers have their own set of drawbacks :

- They pick up all the stray magnetic fields due to their huge leakage inductance and there is little to do to avoid this. Shielding them in a grounded conductive case may result in about 20dB improvement in magnetic field rejection. Of course, there are strong stray magnetic fields near any audio power amplifier, not only 50Hz fields but also fields caused by the amplified audio signal and odd harmonics of 50Hz due to mains rectification current pulses

- Inter-winding capacitance ruins CMRR at high frequencies. Inter-turn capacitance and leakage inductace may cause ringing and ruin the signal at high frequencies. Magnetizing inductance and its hysteresis cycle ruins the signal at low frequencies. Transformers should be driven from a small power amplifier to reduce these effects

Constructing a high performance transformer is a more complex and expensive task than constructing a high performance balanced input with the same CMRR [including manually matched resistors and premium op-amps]. Not to talk about the winding tools and the exotic iron core required for the transformers, and the high current capability [both at LF and HF] required to drive the primary

I think that transformers are much more useful for power or SMPS/pulse applications where galvanic isolation of hundreds or thousands of volts is required, than for audio where usually 1 volt of galvanic isolation is enough

Op-amps also allow to use much higher balanced signal levels than transformers, even up to 48Vpp [achieved with a balanced line whose wires carry +-12V] and this may be used to improve CMRR [not only op-amp CMRR but also twisted pair CMRR] toghether with outputs with built-in gain and inputs with built-in attenuation
 
The one and only
Joined 2001
Paid Member
djk said:
Too bad you can't hear a good transformer system, you wouldn't bother arguing. At the risk of repeating myself, a cheap $8 transformer works better than an electronically balanced system.

Depends on the criterion for quality. I've heard and measured
the best transformers, and if you're on stage, that's what you
want. If your in my living room, they come in second.
 
The one and only
Joined 2001
Paid Member
djk said:
Balanced inputs are another BIG LIE in audio.

"Resistors with 1% purchase tolerance yield a worst-case CMRR (common-mode rejection ratio) of 34 dB, and added drift tolerances of 2% limit the CMRR to 24 dB.

"Difference amplifiers normally have equal-value resistors, but this situation leads to unequal input resistance. The approximate input resistance that the source, V1, experiences is Z1, and the input resistance for the source, V2, is (Z3+Z4). The source resistors act with the difference-amp input resistors to form unequal voltage dividers. Thus, the resistor mismatch creates a differential error signal that the difference amplifier cannot reject.

"Author Information
Ron Mancini is staff scientist at Texas Instruments. You can reach him at 1-352-569-9401, rmancini@ti.com. "

When you need it dead silent, only a transformer will do.

That's B.S. Analysis shows that for common mode noise the
input impedance is equal on both inputs, and with 1% resistors,
I routinely see 60 dB CMRR.
 
The numbers were for worst case 1% resistors, I stated that.

Are you just not being careful in your reading?

To get 60dB the match would have to be closer to 0.1%

But you know that.

I used an electronic balanced system in one of my last systems, the 80dB of CMRR was OK for the bowling alley it went into.

80dB was not OK for the last house of worship I built the crossover/amplifier/subwoofer for.

The main contractor returned the crossover/amplifier and requested that I install an octal socket on the chassis.

With a transformer he was able to get it up to 120dB.

It made a believer out of me.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.