[surajspai]

Hi,
i had joined a bit back,been lurking the the forums for a while.i owned a pair of sennheiser hd202's and now they're beat-up,thought i'd get myself a new set of cans(?) like the audiotechnica AT-H M50 or AD700,depending on the offer i get or the budget i can manage.my main audio source is my pc and it has a asus xonar-dx for a sound card.i know it aint audiophile,but with my budget limits me.
i spend a lot of time off my pc,and i depend on a sansa clip+ or a ipod nano(2nd gen) for music.
so,i thought of taking my headphones with me(weight and size not a issue),but i cant run it off my portable players so i thought of making an amp for it.
 
1.is this the correct way to go?can i really use a small pmp for running such headphones using a amp?any risks?

so,i got down to it and after a wait of like 3 months(parts mail order was stuck because of red-tape :mad: )
i followed the guide at tangent's site and have used the recommended components.i thought of making it on protoboard but it turned out to be too messy and made my own pcb and all
 

 
 

and it worked on the first go!
but its not perfect.i can hear a hiss and sometimes it noise.the noise has reduced after i switched to battery power but there still is mild noise.
if i cover the case using a metal piece or shield it,there is no issue.i have still not mounted it in a case as its only a prototype and i want to make a smaller and better model.also,the ground plane is not perfect as i didnt do the pcb properly.am yet to add the pot.
http://imageshack.us/photo/my-images/543/dsc03324h.jpg/
the ground plane is kinda perforated,as the board wasnt properly made.all the traces are flawless though
no hum,only mild noise.also,when i plug in/out a interconnect,it crackles,which stays till i switch off and again switch on the circuit.why is that
 
2.how can i totally eliminate the noise?

 Fred_fred2004  sent me his pcb layout but,since i dont have a tle2426 rail splitter,and dont want to mail order again,am not able to use it.
 
3.can i get a power circuit which can be easily implemented using more commonly available components?i see some cmoy's having a lot of caps in their pcb like http://s460.photobucket.com/albums/qq329/Fred_fred_2004/?action=view&current=CIMG2978.jpg  .what are they for?
 
and,i thought of adding a bass boost circuit to the amp.i found a thread http://www.head-fi.org/t/518085/bass-boost-question-for-a-cmoy

4.can i implement it directly?0.47uf for the cap and 10k resistor in parlell?i wont be using a pot there,instead ill use a swit-ch to turn boost on and off.what values for the bass boost?

5.i am using tangent's default components,do i stick to the values of resistors and caps or are there better(?) values i can use?

i thought of trying my hand at other amp designs,like the A47
http://www.head-fi.org/t/544829/i-need-an-updated-a47-amp-parts-list
http://www.head-fi.org/t/531454/mullet-a47-thread/30
also what is the cmoy2?which do i try next?the A47?i am limited to ic based amps as sourcing is an issue and budget constraints.

please advise.
thanks and regards
 
 
(why is s.w.i.t.c.h censored? and is my method of uploading images correct?or do i put everything on imageshack?)

 

[drtturnip]

Have you measured the output for dc voltage? If so then you have too much dc offset caused by a bad opamp or unbalanced circuit. If you aren't going to drive inefficient headphones I would lower the gain. Here is something I did that lowered the dc offset and gain. The gain you have now would probably pick up noise until it is shielded in a metal enclosure.
 
http://www.head-fi.org/t/591941/some-cmoy-mods
 
Along the lines of headphone efficiency, I don't know if there is any advantage to a more complicated circuit unless you go with tubes when using easy to drive headphones. I love the way tubes sound and will probably build a little tube headphone amp since I have a bunch of parts laying around. The more components the more color added to the sound is my opinion. Check out this link comparing a basic single opamp design to a couple of more complicated units.

 

[mono]

Is that a TL072 opamp?  If so, [EEK!] using something else with higher current or paralleling two of them like in an A47, or paralleling two of something else especially, is the next upgrade.  None of the old popular opamps for CMOYs do very good driving a load (cans), it's just sort of the simplest circuit that can be thrown together with an opamp and work without much development. 
 
Put the POT in the circuit or temporarily put a few dozen KOhm resistor between each channel and ground to see if some of the noise goes away (then remove when you add the POT). 
 
CMOY2, if it's this design linked, is just using an opamp as ground channel to do away with the TLE railsplitter, or the buffer chip in an A47, or a dual rail power supply.  There are many ways to more or less get the same job done, with minor audible differences sometimes.
 
http://www.sgheadphones.net/index.php?showtopic=3220

 

[drtturnip]

 

None of the old popular opamps for CMOYs do very good driving a load (cans)

I think this is true for high impedance, super low impedance, or low efficient cans. A good op amp ( NOT TL072! ) can drive modern high efficiency 30-50 ohm cans fine unless you want your ears to bleed. I know, I am doing it. It sounds great and the measurements at the website I referenced prove it. A lot of the tests run on that site were into 15 ohms and it still held up well.  Plus with the small part count there are very few components in the signal chain. You get the benefit of a low impedance capacitor-less output. Electrolytic capacitors suck unless you pay.

 

[mono]

^  depends on how you define "good", if by good you mean an unusually high output current for a DIP8 opamp then yes it's somewhat true.  On the other hand, does one really build an amp with a restriction that it's only adequate for 30-50 Ohm cans?
 
The website you referenced is, to put it politely, deliberately misleading.  Hand picked tests and ignoring all the significant factors that were overlooked while proceeding to still feign a scientific method.
 
You can't just measure output @ load, you have to measure the sound OF the reactive load.  30-50ohm cans also sound MUCH better with a buffer after the opamp.  One example where I auditioned the difference was using Sennheiser HD555, 50 Ohms /112dB SPL.  Different opamps in a CMOY sounded different, different opamps in A47 did too, but once you add the buffer on more advanced amps the difference becomes less and less which opamp you roll in and even swapping which buffer you use makes less difference than the difference between a CMOY and something with any buffer stage.
 
It is easy to hear the sound is far superior with a buffer.  Disagreement is fine, hearing and preference IS subjective, but having built 'em both ways, there is no contest to me.  Some people can't hear well enough to tell the difference though, for them certainly a simpler and cheaper amp makes the most sense.  For others, there is way too much experience from other members here and elsewhere that different amps and amp topologies do sound different when fairly implemented, enough to not jump to a quick conclusion until hearing several amps and deciding for oneself.

 

[drtturnip]

At high listening levels I have no doubt there would be a noticeable difference. I don't listen loud. But I must admit I haven't bothered with any other design because for my needs I don't need it. If I am going to mess with something more complex then it's gonna be tube time. Right now my only other amp is an old Harmon Kardon hk505 which was very highly rated in the day. I have modified the output so the headphones see 10ohms. It is "warm" compared to the cmoy, as it should be. My bass weak  595's sound better with it probably because of their huge impedance peak at 100hz.

 

[mono]

I'm not referring only to high listening levels, at any level loud enough you can hear the details of the music (more than background music) I hear a difference... though each person's idea of a high-volume threshold can differ too. 

 

[drtturnip]

I am sure there is a difference. I also may think that a buffer improves the sound, but is that because said buffer is more accurate or is it perhaps introducing a "color" that is more pleasing?
I just find it hard to get past these facts. The cmoy can drive a range of impedance at a reasonable level with linearity and with low measurable distortion. It has a minimum of components and thus should introduce a minimum of artifacts. In my limited way of thinking it should be more accurate left as is UNLESS you need to surpass what the op amp can do with a demanding load or need massive volume.

 

[mono]

I believe the buffered version is more accurate, because when you swap similar current capacity opamps on a CMOY, it changes the sound more than if you swap the same opamps when they're behind a buffer, and more than if you keep the same opamp and swap in different buffers, UNLESS you were trying to build (or did unintentionally) a buffer or whole amp that colorizes the sound a certain way.
 
The real irony is some of the more (exotic?, maybe high-spec is a better terminology) expensive opamps still sound worse than a humble OPA2134 ($2 or $3 opamp) with $3 worth of buffer components after it... though to be fair, buying very small lots of components at a time, let alone suffering shipping fees, can add up in cost more than that, AND just because I think it sounds worse it doesn't mean much, everyone has to decide for themselves.
 
I should clarify what I wrote previously, that I can hear differences above a background volume listening level.  I do not know if there are differences below this volume threshold, if there are it is not loud enough to hear them so I must state I don't hear any, but it is seldom I would wear headphones for that low a volume level, maybe out of courtesy when sitting next to someone on an airplane or other public areas.

 

[drtturnip]

So let me bounce this. Most, probably all, op amps were not designed to drive a reactive lowish impedance load, however some can and can do it well (within strict limits). When I built my cmoyish thing I went through a bunch of op amps until I found one that worked well and sounded good. 
What the buffer really does is make up for the inability of most op amps in driving headphones. Driving headphones with an op amp is "technically" wrong, but with the right combination works really well. Less is more.  I would also like to say that if a buffer has an electrolytic capacitor in the signal path any gains in it's driving ability would be suspect. I hate how electrolytic capacitors sound.

 

[mono]

I like the less is more philosophy myself, though a couple things happen.  One is that if your buffer is in the global (opamp) feedback loop you mitigate sonic effects of the buffer.  Another is that while we could wish for the perfect opamp to drive cans, manufacturers tend to pimp the low spec stuff for cheap audio apps and view the better performing parts as a modular portion of a larger circuit.  This conversation reminds me of a fellow who posted elsewhere the other day insisting he was going to build a chipamp with JUST the chip, he didn't want any resistors or capacitors, not even a bridge rectifier external -  they should be built in he insisted because less is more.  Obviously at some point it gets unreasonable to put everything in a chip, especially one the size of a DIP8 or smaller.  One issue is heat, especially if you'd like some class A bias you don't want the whole opamp heating up, with external transistors in a buffer you have more surface area for cooling and have isolated the heat.
 
Let me ask this.  What difference is there if you put matched, higher current transistors after the opamp as a buffer instead of the opamp itself having them integrated into itself for the higher current capability?  In the former you tend to reduce costs because you can use more general purpose/popular opamps and transistors, while the latter is a niche market, everything is now moving to SOIC for portable devices with the amp stage built into that IC unless it's fairly expensive gear.
 
Why put a cap in the buffer signal path?  Acting as a current source your DC offset would be coming from the opamp gain stage or already there on the signal input.  Either way if your opamp has no adjustment pin (or the feature isn't implemented) and you're stuck with too high DC offset, a cap has to go somewhere or ideally the problematic portion is replaced.
 
We may be putting far too much thought into surajspai's topic, on my second read it seemed like the goal is not to upgrade a CMOY amp but to build a different amp as the next project, A47 or CMOY2.  If member support is important if/when things go wrong, I'd search the forum to see if there are any current topics with lots of posts, other members having the same type of amp out and fiddling with it makes for easier/faster peer support during a build.

 

[drtturnip]

 
 

We may be putting far too much thought into surajspai's topic

 Oh absolutely!
 
 

What difference is there if you put matched, higher current transistors after the opamp as a buffer instead of the opamp itself having them integrated into itself for the higher current capability?

Cause that's a PITA. MATCHED? Do you buy in bulk? My experience is tolerance is all over the place.
 
 
 

Either way if your opamp has no adjustment pin

By adjusting the overall  gain resistance vs input resistance , not ratio ( [R feedback to ground] /[ R  feedback] =[R  input ] ) you can eliminate dc offset or come real close. If a trim pot were used it could be 0.
 
My whole point was if you have found a really good opamp that drives your particular cans why bother? Now if your talking about glowing glass that's a whole different story but throwing more silicon at something seems counter productive unless necessary.
 

 

[nikongod]

Less is more is very dangerous when dealing with little black boxes. Nothing with an opamp in it should ever say less is more. Unless the opamp is removed and replaced with fewer parts. 
 
Anyways, back on track:
If you have access to an oscilloscope make sure that your amp isn't oscillating. 
 
The hiss could be coming from your computer (computers are neutoriously noisy), the amp its self, or oscillation. 
 
If the amp gets quieter when you cover it, it is probably picking up RF somewhere. Cover it  
 
What gain did you use? What headphones do you have?
 
For a second project, a CHA-47 is fun to build and not much more expensive than a normal Cmoy. 
If you want to branch out into discrete circuits a bit a simple mosfet source follower like the Szekeres isnt really that expensive to build. 

 

[drtturnip]

 
 

Less is more is very dangerous when dealing with little black boxes.

Ok. School me.

 

[drtturnip]

Hey surajspai
 
Master the cmoy before you move on. You will learn a lot.