I can't believe that TI engineers write such nonsense. Just mentioning "sound room evaluating" doesn't exactly show engineering skills. You can prove with AP measurements that capacitors have no effect on the sound if you use the right capacitors.
placing a shorting switch over the caps is something i can do, but to do a fair comparison between DC substraction and caps? Would a switch in series with R16 and R17 work? I have a stash of DPDT's somewhere...
Hey hey hey. The most heated debates happen over those issues where the debaters have no personal, firsthand experience. Such debates are also the most pointless, and I feel we are edging into one such debate right now. The remedy is to ground yourself in facts, whether subjective or objective, that you personally observed.
In this particular case, I personally have not been able to measure the effect of (good quality) DC blocking capacitors carefully placed at the input of amplifiers with -140dBc level of distortion - this includes the Omicron and a power amp that has not been published here. Your experience may be different. BTW I am familiar with Cyril Bateman's work on capacitor distortion, although I have not attempted to verify or duplicate his results.
Also, I meant something quite a bit less ethereal than SQ. The LM4652 typically has an input current of 10nA (72nA max according to the datasheet). Two such currents from the two half of the input opamp create a DC voltage drop of several tens of mV on the 2.2 megaohm resistor. I measured between 50 and 100mV on several channels on several boards (I have built quite a few, both Putzeys and my own designs). This DC is common mode, but the CMRR is limited. If you (chase SQ and) build Putzey's MicroPre with 1% MELF resistors, you will get CMRR at about 40dB, give or take. 100mV less 40dB is 1mV DC at the output. It is not a problem whatsoever, but it does produce audible clicks when switching inputs. The DC removal circuit elimates those clicks. Or, resistors with 0.1% tolerance (specified as "precision thin film" in the original orticle) improve the CMRR and make those clicks inaudible. Or, DC blocking capacitors would eliminate them, too - but the need to charge and discharge these capacitors (typically via large-ish resistors from inputs to the ground) would interfere with the role of that 2.2 megaohm resistor. Or, you can just DC couple everything, throw away the DC removal and live with the clicks. It is good to have a choice.
In this particular case, I personally have not been able to measure the effect of (good quality) DC blocking capacitors carefully placed at the input of amplifiers with -140dBc level of distortion - this includes the Omicron and a power amp that has not been published here. Your experience may be different. BTW I am familiar with Cyril Bateman's work on capacitor distortion, although I have not attempted to verify or duplicate his results.
Also, I meant something quite a bit less ethereal than SQ. The LM4652 typically has an input current of 10nA (72nA max according to the datasheet). Two such currents from the two half of the input opamp create a DC voltage drop of several tens of mV on the 2.2 megaohm resistor. I measured between 50 and 100mV on several channels on several boards (I have built quite a few, both Putzeys and my own designs). This DC is common mode, but the CMRR is limited. If you (chase SQ and) build Putzey's MicroPre with 1% MELF resistors, you will get CMRR at about 40dB, give or take. 100mV less 40dB is 1mV DC at the output. It is not a problem whatsoever, but it does produce audible clicks when switching inputs. The DC removal circuit elimates those clicks. Or, resistors with 0.1% tolerance (specified as "precision thin film" in the original orticle) improve the CMRR and make those clicks inaudible. Or, DC blocking capacitors would eliminate them, too - but the need to charge and discharge these capacitors (typically via large-ish resistors from inputs to the ground) would interfere with the role of that 2.2 megaohm resistor. Or, you can just DC couple everything, throw away the DC removal and live with the clicks. It is good to have a choice.
One could activate the muting relay when switching inputs. Those relays, you gotta love them! 🙂
Despite thinking source selectors being necessary I noticed the following when I was cleaning my devices and changing my setup. A recently bought source/digital hub does about all with digital input switching (which I did not use till now) and it has XLR outputs to an also recently acquired XLR equipped test amplifier thereby eliminating the need for a balanced preamplifier completely.
Quite a shocker as I was thinking of building a new balanced line stage with source selection and analog volume control.
Despite thinking source selectors being necessary I noticed the following when I was cleaning my devices and changing my setup. A recently bought source/digital hub does about all with digital input switching (which I did not use till now) and it has XLR outputs to an also recently acquired XLR equipped test amplifier thereby eliminating the need for a balanced preamplifier completely.
Quite a shocker as I was thinking of building a new balanced line stage with source selection and analog volume control.
Last edited:
