[Lucias_D]

I have some questions about ac/dc op amp based amplifier coupling, with some pertaining to the cmoybb
 
First I would like to discuss the types (physical construction) of capacitors used in coupling, and it's measured effect on sound quality. Does anyone have directed reading or personal findings on types of capacitors (electrolytic, mylar, polystyrene, teflon, etc..) regarding this question? Most notably I'm tackling THD or floor noise, and associated coloration of the waveform. If I had a scope I might try a bunch of different caps and try to take some measurements but I'm quite limited over here. I'm limited to my DMM and ears.

 
Secondly, JDS Lab's on there cmoybb modifications page highly recommend using a DC coupled signal path when using a "Line Out" input source. Does anyone know why this is? What is it specifically about a line level signal that would benefit from DC coupling in contrast to AC coupling?
Maybe this benefit is because of the specifically recommend caps in the bill of materials for this amp. From this thinking I would theorize that if 'high quality' [read: transparent audio] caps were used in this specific ac coupling in place of resistors (DC coupling), there would be no benefit to going the DC route.

 
Lastly, has anyone tried using trim potentiometers to 'equalize' DC offset in op-amp based headphone amplifiers or specifically the cmoy/cmoybb. Outcomes?
 
 
 
This forum has been an invaluable resource for many situations I've encountered. Thanks for all those dedicated to helping others!

 

[Budgie]

The input capacitor is there as a safety device to protect from DC voltage at the input of the amp.That voltage would get amplified and passed on to the output damaging your headphones. That is a rare problem with modern equipment (most source components have capacitors in series with the output to prevent any DC output) so many people chose to leave the input capacitor out of the circuit, which will eliminate one source of distortion and non-linearity in the amplifier. There is tons of information on this site and on the net in general regarding types of capacitors to use for audio coupling chores and my experience has been that different caps (presuming equal rated capacitors are being compared) do have different "sonic signatures". After that it becomes a matter of personnel preference as to which caps to use. Since caps are non-linear devices and cause distortions, I prefer eliminating capacitors from the path whenever possible.
 
http://sound.westhost.com/articles/capacitors.htm  is a good read and I believe reasonably factual. Google "audio capacitors" for tons of opinions and providers.
 
In general, Teflon, polypropylene and polystyrene usually sound the least colored to me. Film dielectric type caps tend to be very large for useful values at audio frequencies. Electrolytic types will be much physically smaller, compared to film types, but don't sound as transparent, to me. A common practice is to use an electrolytic cap paralleled with a smaller value film cap to try to get the best of both worlds. If you must use capacitors, quality types can reduce their sonic impact. For electrolytic types used for signal coupling, I prefer "Cerafine" or "Simic II" over others, but "Blackgate" and Certain Panasonic versions (FC?)  are popular. (I am sure others will post there favorites caps here and to greater specificity.)
 
It would be possible to build a DC coupled cmoy, but without any capacitors to block DC voltages, you could damage your headphones if something goes wrong. A servo circuit could be used to keep DC from showing up at the output, but the added complexity makes a capacitor more attractive, in my opinion.  (I don't know which would sound better. Some people report DC servos affect the sound negatively also)
 
(edit -spelling, typos, and incoherent rambling corrected.)

 

[Lucias_D]

Thanks for the response,
 
Doing some further reading from the site you mentioned as well as others, there are much better sonic signatures to be had with different types of capacitors. I guess the main drawback to using a lot of so called "exotic" capacitors is that most require a large footprint due to the capacitance values needed for audio circuits. Most polystyrene, teflon, silver mica, and other "exotic caps" have low capacitance density compared to electrolytic and various ceramics/tantalum. If you're going to build a full size rig then some of these components might fit, but I'm working in a finite space so "exotics" I've found work for my application.
 
Also, I tore apart my amp today and the results were two fold. I guess I answered my own dc coupling question. the cmoybb uses the Ti rail splitter and it's quite accurate. I just put in some really high tolerance resistors .1% and measured my offset. .1mVDC+ and -.1mVDC-. Looking good now. Once I play around with some different op amps we'll see if I need to come back to this coupling issue.
On the bad side of things, some solder fell on to my pcb and filled a hole accidentally. getting this cleaned out so I could shove a lead through it was the biggest PITA and I actually messed up the through-hole by scraping off too much. Luckily it's for the bass boost circuit which I rarely use :\

 

[MrTom]

Quote:

On the bad side of things, some solder fell on to my pcb and filled a hole accidentally. getting this cleaned out so I could shove a lead through it was the biggest PITA and I actually messed up the through-hole by scraping off too much. Luckily it's for the bass boost circuit which I rarely use :\

You could run a wire jumper from one side of the board to the other after tracing the traces to a solder pad? I have had to do this when repairing plasma TV PSU's...

 

[cobaltmute]

Quote:

On the bad side of things, some solder fell on to my pcb and filled a hole accidentally. getting this cleaned out so I could shove a lead through it was the biggest PITA and I actually messed up the through-hole by scraping off too much. Luckily it's for the bass boost circuit which I rarely use :\

Get a desoldering pump.  They are cheap and make life way easier.
 
 
 

 

[Lucias_D]

Quote:

Get a desoldering pump.  They are cheap and make life way easier.
 
 
 

Yeah, I wish I had one here! I had some desoldering equipment back in Minnesota but I didn't bring all my tools

 

[Soymilk]

one way to deal with that if you dont have any desoldering related things (pump, wick) is if you can heat up the solder, then just shove a wire through.

 

[Budgie]

Wow, that"s excellent results for DC offset!  I hope you post some final thoughts on your build when your finished with it.
 
I think it would be interesting to build several versions of a cmoy, using the same op-amp, but with different topologies, to compare the sonic results. Bipolar power versus active rail dividers versus resistive dividers, and passive and active grounding channels, etc. I am too lazy/busy to undertake it, but it would be a fun experiment.
 

 

[jseaber]

As Budgie has pointed out, DC offset is very rarely an issue with modern equipment. Our recommendation to go for DC coupling if using an LOD is, quite honestly, based on distortion experienced by a small number of cMoy users. I've personally had no trouble listening to a cMoy with AC coupling and an LOD (this is actually the test setup for the cMoyBB), but others have occasionally expressed difficulties as they increase volume. You can find these discussions here at Head-Fi. When this happens, I send out a DC coupled amplifier--problem solved. As I'll explain below, this issue isn't limited to the cMoyBB.
 
Line level signals can generate signal voltages of as little as 0.45 Vpk, or as much as 2.2Vpk. The common-mode input voltage range of the OPA2227PA opamp is |V+/-| minus 2V. With a 9V battery, voltage supplied to the opamp is +/-4.5V. Subtract 2V and we have a range of +/-2.5V left at the opamp's inputs. However, many cMoy users drain their 9V's well below 6V before swapping batteries. A strong line-level signal (2.2V) requires a +/-4.2V power supply, which means the 9V battery can only drop to 8.4V before distortion begins at maximum volume. This is also the approximate peak voltage of a fully charged NiMH battery. The volume control does attenuate the line-level signal voltage, and no one should ever need to turn the volume to max, but as the battery drains further and further past 8.4V, it's easy to understand why strong line-level signals can become problematic. Alternatively, a higher voltage AC/DC adapter can eliminate this trouble, but what's the point of a portable amp which cannot be transported?
 
Driving a 1uF load also isn't easy, but I'll not go into that explanation. Most consumer audio devices are stable enough to do so. By replacing the 1uF input capacitors with resistors, we solve two problems:
 
1) The audio source now drives a resistive load with a comfortable 10k impedance.
 
2) The DC coupling resistors slightly attenuate the strong line-level input signal.
 
 
--JDS
Quote:

I have some questions about ac/dc op amp based amplifier coupling, with some pertaining to the cmoybb
 
First I would like to discuss the types (physical construction) of capacitors used in coupling, and it's measured effect on sound quality. Does anyone have directed reading or personal findings on types of capacitors (electrolytic, mylar, polystyrene, teflon, etc..) regarding this question? Most notably I'm tackling THD or floor noise, and associated coloration of the waveform. If I had a scope I might try a bunch of different caps and try to take some measurements but I'm quite limited over here. I'm limited to my DMM and ears.

 
Secondly, JDS Lab's on there cmoybb modifications page highly recommend using a DC coupled signal path when using a "Line Out" input source. Does anyone know why this is? What is it specifically about a line level signal that would benefit from DC coupling in contrast to AC coupling?
Maybe this benefit is because of the specifically recommend caps in the bill of materials for this amp. From this thinking I would theorize that if 'high quality' [read: transparent audio] caps were used in this specific ac coupling in place of resistors (DC coupling), there would be no benefit to going the DC route.

 
Lastly, has anyone tried using trim potentiometers to 'equalize' DC offset in op-amp based headphone amplifiers or specifically the cmoy/cmoybb. Outcomes?
 
 
 
This forum has been an invaluable resource for many situations I've encountered. Thanks for all those dedicated to helping others!

 

[Lucias_D]

 
Quote:

one way to deal with that if you dont have any desoldering related things (pump, wick) is if you can heat up the solder, then just shove a wire through.

Yeah, I tried that, but only after I tried scraping it a little. Not so effective after scraping. Live and learn I guess

 
Quote:

Wow, that"s excellent results for DC offset!  I hope you post some final thoughts on your build when your finished with it.
 
I think it would be interesting to build several versions of a cmoy, using the same op-amp, but with different topologies, to compare the sonic results. Bipolar power versus active rail dividers versus resistive dividers, and passive and active grounding channels, etc. I am too lazy/busy to undertake it, but it would be a fun experiment.
 

Yeah the offset was quite low from what others have posted as well. I'd say it's because all the resistors to and from the op-amp were all .1% tolerance, but I'm sure there are other factors (board trace layout, etc..)
The different topologies would be nice. Maybe someday I'll get crackin on a bigger box to house those components. My only reservation is that JDS labs seems like they put a lot of time into optimizing the board layout for reduced line capacitance, cross talk, etc.. that I'd have to break out a breadboard or build a revised pcb to do those designs. I think that would negatively affect sound quality, and I'm not sure I could accurately surmise if the revised circuitry actually improved the sonic characteristics because of that. The cmoybb I build sounds better than the regular cmoy I heard previously...
 
Quote:

As Budgie has pointed out, DC offset is very rarely an issue with modern equipment. Our recommendation to go for DC coupling if using an LOD is, quite honestly, based on distortion experienced by a small number of cMoy users. I've personally had no trouble listening to a cMoy with AC coupling and an LOD (this is actually the test setup for the cMoyBB), but others have occasionally expressed difficulties as they increase volume. You can find these discussions here at Head-Fi. When this happens, I send out a DC coupled amplifier--problem solved. As I'll explain below, this issue isn't limited to the cMoyBB.
 
Line level signals can generate signal voltages of as little as 0.45 Vpk, or as much as 2.2Vpk. The common-mode input voltage range of the OPA2227PA opamp is |V+/-| minus 2V. With a 9V battery, voltage supplied to the opamp is +/-4.5V. Subtract 2V and we have a range of +/-2.5V left at the opamp's inputs. However, many cMoy users drain their 9V's well below 6V before swapping batteries. A strong line-level signal (2.2V) requires a +/-4.2V power supply, which means the 9V battery can only drop to 8.4V before distortion begins at maximum volume. This is also the approximate peak voltage of a fully charged NiMH battery. The volume control does attenuate the line-level signal voltage, and no one should ever need to turn the volume to max, but as the battery drains further and further past 8.4V, it's easy to understand why strong line-level signals can become problematic. Alternatively, a higher voltage AC/DC adapter can eliminate this trouble, but what's the point of a portable amp which cannot be transported?
 
Driving a 1uF load also isn't easy, but I'll not go into that explanation. Most consumer audio devices are stable enough to do so. By replacing the 1uF input capacitors with resistors, we solve two problems:
 
1) The audio source now drives a resistive load with a comfortable 10k impedance.
 
2) The DC coupling resistors slightly attenuate the strong line-level input signal.
 
 
--JDS

 

Thanks for that explanation. Didn't even think about the attenuation reduction from the resistors. I rarely turn my pot above half way on mine anyways, but the effects of clipping could definitely be noticed when using a drained battery, especially a rechargeable one with a lower initial voltage