Amplifier input's low pass (RFI filter), and other questions

[Legis]

Hi,

one phase of my amp's input after XLR socket goes like this:

1) 10µH inductance in series with the input (RFI filter?)
2) 50k resistor connected between the signal line and ground
3) input swich (rca/xlr)
4) 300R in series after the input swith
5) 660pF cap connected between the signal line and ground (RFI filter?)
6) 4,7µF DC blocking caps in series
7) and after that the signal goes to buffer stage.

What is the purpose of the 300R resistor and 50kR resistor? 50kohm seems quite big value. What components can be bypassed, I presume at least the inductors? I would like to clean up the signal path as much as possible. All components are through hole and could be easily removed or bypassed.

The rca input has an op amp which inverts another phase from rca input's signal. I once tried bypassing the DC blocking caps, but after that I measured over 300mV of DC offset from outputs without anything connected to input. I don't remember if I had the input swith at RCA or XLR, but I don't see how the DC offset could be present if XLR input is on. But the opamp of the rca input could definitly generate the DC offset, am I right?

Regards,

Legis

[DF96]

All those components do something useful, so why do you want to bypass them? Do you prefer poorer sound or less reliable equipment?

You have an RFI filter, to reduce incoming clicks and pops from your fridge and mobile phone. You have a DC ground reference (50k) whch ensures you don't get loud clicks when you switch inputs. You have a stopper resistor (less RFI). You have DC blocking. The designer of this equipment obviously knew what he was doing. 'Cleaning up' the signal path would reduce it to the level of a poor DIY item.

[Legis]

Originally Posted by DF96

All those components do something useful, so why do you want to bypass them? Do you prefer poorer sound or less reliable equipment?

You have an RFI filter, to reduce incoming clicks and pops from your fridge and mobile phone. You have a DC ground reference (50k) whch ensures you don't get loud clicks when you switch inputs. You have a stopper resistor (less RFI). You have DC blocking. The designer of this equipment obviously knew what he was doing. 'Cleaning up' the signal path would reduce it to the level of a poor DIY item.

After posting that I came to think bypassing the inductor and DC blocking caps only. The DC blocking caps are very essential regarding the sound and I would like to take them out of the equation by removing them (and make the amp DC coupled at the same time), rather than buying some expensive Mundorfs. I have heard some good stuff about DC coupled amps' sound.

edit. by taking out the inductors, I would reduce the RFI-filter's slope to 2nd order? There are a 660pF "X-cap" between the phases and 660p "Y-cap" between both phases and ground. Don't the caps create a 2nd order filter by themselves, and the inductor adds one order more?

[DF96]

There is nothing special about a DC-coupled amp. It replaces a few simple capacitors with a DC-servo (unless you want to blow up your speakers). The DC-servo includes caps, so it still won't satisfy those who are afraid of caps. I assume you know enough to re-engineer your amp?

[Legis]

Originally Posted by DF96

There is nothing special about a DC-coupled amp. It replaces a few simple capacitors with a DC-servo (unless you want to blow up your speakers). The DC-servo includes caps, so it still won't satisfy those who are afraid of caps. I assume you know enough to re-engineer your amp?

I wrote here how I would go forward: Emotiva XPA-1 vs Jungson 99D

I also like the idea of bypassing the switch in the signal path.

If my preamp is DC-free, is there any possible hazard in bypassing the DC blocking caps? For example if I clip my preamp, does it generate DC?

[Tekko]

If the amp was designed to have the dc blocking capacitors, bypassing them would likely lead to several volts appearing on the output when the input is pulled to ground by the preamps output impedance. I have done this and i had to modify the input curcuitry of the amp to bring the dc offset back down, absolutely no improvement what so ever in sound quality was obtaioned by doing this.

I have designed a completely capacitorless amplifier with a frequency response down to DC, in simulation it sez 6mVC dc offset with about 1-2 mV difference in input open(input impedance of 33k) or shorted to ground.

In reality it puts out 60mV dc offset which is still pretty good for a dc coupled amp. I made this just for fun to see if i could pull it off. This one uses jfet input stages.

So if i were you, i'd leave the dc blocking caps alone as they have next to no audible impact on sound quality, if anything replace them with higher quality ones if the originals are a cheap questionable brand.

[cbdb]

is there any possible hazard in bypassing the DC blocking caps

Yes your output DC offset could get large enough to damage your speakers.

by taking out the inductors, I would reduce the RFI-filter's slope to 2nd order?

Probably not. The corner freq of the inductor is well bellow the caps. The caps probably take over at a freq that the inductors parasitic capacitance renders it useless as a filter. I get the impression you dont know enough electronics to really know what your mods will do. The designers didnt add these components because they were bored. They are there for a reason.

Why do audiophools (Im not saying your one, Legis, Im generalizing) always think less components are better? Theres probably a few transistors and resistors you can get rid of too!

[Legis]

Originally Posted by Tekko

If the amp was designed to have the dc blocking capacitors, bypassing them would likely lead to several volts appearing on the output when the input is pulled to ground by the preamps output impedance. I have done this and i had to modify the input curcuitry of the amp to bring the dc offset back down, absolutely no improvement what so ever in sound quality was obtaioned by doing this.

Hi, I have now bypasssed the DC caps and the DC off does not rise at least when driven from DCX2496, modded with balanced passive outputs (1k resistor in series with both phases). I quess I will have to try to take a signal from my preamp with XLR-rca adapter to see if the situation chages with lower output impedance. The adapter connects the other phase to ground.

edit. Yep, you were right. After I tried with the other pre amp the DC offset jumped to ~400mV. So just bypassing the caps is not going to do any good for the amp.