[wakibaki]

From the datasheet:-
 
The NJM4562 integrated circuit is a high-gain, wide-bandwidth, low noise, dual operational amplifier capable of driving 20V peak-to-peak into 600Ω loads. The NJM4562 is frequency compensated for closed loop gains greater than 10.
 
The NJM4562 is frequency compensated for closed loop gains greater than 10 - or to put this another way - the NJM4562 may oscillate at closed loop gains less than 10.
 
An amplifier can have audibly degraded performance due to oscillation without the oscillation itself being obvious to the senses. Such differences are sometimes perceived as improvements. (Any difference I can hear is a good difference.)
 
The driver stage in the O2 is a unity-gain buffer. The NJM4556 is unity-gain stable, i.e. it is unlikely to oscillate in this configuration. 
 
20V peak-to-peak into 600Ω loads.
 
The NJM4556 is rated 70mA into 150Ω.
 
The NJM4562 has a dissipation of 500mW in DIP8 as opposed to 700mW for the NJM4556.
 
The 4562 may be OK driving some phones but the 4556 is the preferred option for stability, power and almost certainly distortion and frequency response (gain flatness) and particularly versatility in terms of the phones it will drive well.
 
w

 

[cheapskateaudio]

 
Quote:

From the datasheet:-
 
The NJM4562 integrated circuit is a high-gain, wide-bandwidth, low noise, dual operational amplifier capable of driving 20V peak-to-peak into 600Ω loads. The NJM4562 is frequency compensated for closed loop gains greater than 10.
 
The NJM4562 is frequency compensated for closed loop gains greater than 10 - or to put this another way - the NJM4562 may oscillate at closed loop gains less than 10.
 
An amplifier can have audibly degraded performance due to oscillation without the oscillation itself being obvious to the senses. Such differences are sometimes perceived as improvements. (Any difference I can hear is a good difference.)
 
The driver stage in the O2 is a unity-gain buffer. The NJM4556 is unity-gain stable, i.e. it is unlikely to oscillate in this configuration. 
 
20V peak-to-peak into 600Ω loads.
 
The NJM4556 is rated 70mA into 150Ω.
 
The NJM4562 has a dissipation of 500mW in DIP8 as opposed to 700mW for the NJM4556.
 
The 4562 may be OK driving some phones but the 4556 is the preferred option for stability, power and almost certainly distortion and frequency response (gain flatness) and particularly versatility in terms of the phones it will drive well.
 
w

 
That was very informative and helpful, thank you for the post!
 
Trying to compare data sheets with the different way companies spec and rate their products is a pain, but knowing just a few of the key pieces to look for is helpful.
 
 

 

[cheapskateaudio]

Update: I swapped out the other two elctrolytics, C8 & C9 on the board, and while it results in a perceived reduction in bass, bass quality is improved while the body and richness of the mids was increased. This has a good effect on some types of music like old Kraftwerk which is not quite as robust sounding as modern synth music. It has a somewhat congesting effect on some other types of music, but it could just be an adverse reaction to more musical information as this amp has thin mids (with HD650's) by default. So to test things out in the extreme I went to youtube and searched for "bass", I listened to a hilarious song called "Speakers Going Hammer" that was Bass Boosted, I swear I could hear the license plate on a trunk rattling somewhere in the distance... Bass test your setup: http://youtu.be/nPUJwgZZiYM
 
 
 
 

 

[samsquanch]

what caps did you put it?  brand, series, values (farads, voltage, esr)?

 

[cheapskateaudio]

 
Quote:

what caps did you put it?  brand, series, values (farads, voltage, esr)?

For C8 and C9 I used same value low esr Panasonic caps, http://www.mouser.com/Search/ProductDetail.aspx?R=EEU-FR1H221virtualkey66720000virtualkey667-EEU-FR1H221
 
I also swapped C2-C5 with 470uf Panasonics.
 
And C13 & C14 with Panasonic Polypropylene film caps, http://www.mouser.com/Search/ProductDetail.aspx?R=ECW-F2225JAvirtualkey99990000virtualkey667-ECW-F2225JA
 
Everything fits on the board if you bend the leads and surface mount the Polypropylene caps (the leads are too big for the through holes), but likely the 470uf Panasonics are too tall for the thin case and the Poly caps definitely are. I got the tall case though, yay!
 
I also got the beefier wall wart, WAU16-400.
 
The wall wart's effect was more noticeable at first, after adding everything else the wall wart's effect is diminished, but not entirely. I just did 3 blind tests again, between the two wall warts, and it took me about 5 minutes of comparing to feel confident the first time, but once I found what to listen for the other two goes took less than a minute. The difference was most noticeable on Tabla Beat Science's Tala Matrix CD. (A great recording for testing amps because it really stresses some of the more intangible qualities of rhythm and cohesion.) What you get with the bigger PSU is a fuller and smoother overall sound, more and better controlled bass, better separation between instruments, bigger broader dynamics. It's a subtle effect at this point, but completely 100% detectable.
 
The caps improved those same things to varying degrees. It was kind of weird how the sound came together. I'd describe it like this:
 
No mods: A bit thin and smeared with drum driven music, aggressively revealing. Great for classical.
WAU16-400: Fuller sound, HD650's starting to sound like HD650's, still harsh at times.
WAU16-400 + C13, C14: Smoother fuller sound, more cohesive.
WAU16-400 + C13, C14 + C2 - C5: More bass, more bass control, a more familiar dark tone, and a more full and smooth sound, was great with the uDac 1..
WAU16-400 + C13, C14 + C2 - C5 + C8, C9: More full in the midrange. After this change the uDac 1 dac output sounded more rolled off from the lower treble up. Switching to the PC line out showed the difference easily. 
 
I feel like it might be a little tipped up in the bottom end, or rolled off on the top end, but that is the basic shape of the HD650 FR curve, so based on this not so scientific ear test I'd say the above mods produce improvement across the board. 
 
I'm going to build another one, and build it 100% stock at first, just to compare the two and maybe measure the output before I mod it and give it away.
 
 

 

[cheapskateaudio]

Well, I think something got damaged with all the handling and soldering/unsoldering, and now the left channel is slightly lower in volume. Any ideas on what could have caused that? I have no MM so basically I can reflow the solder joints and thats about it. But I do have a few spare/upgrade parts here so I might try swapping them at random and seeing what happens. Could a bad resistor cause that? I touched a resistor with the soldering iron for a few seconds trying to surface mount the poly caps.

 

[snakeeater69]

How hard is the build this amp? Is it as easy as looking at the instructional, plugging parts into proper place and then soldering it? I want to try this amp but have never built an amp or anything before.

 

[Paul Blythe]

cheapskateaudio: Check between S2 [gain switch] and R21, its been reported that they're a little too close togeter and can cause a short, nothing harmful apparently... that's if you've been working in that area....
 
How about soldering thinner leads onto the thicker ones so you cant still mount through-hole!

 

[Paul Blythe]

It's an easy build, so long as you take your time you should be fine. The bits from Mouser are bagged individually & marked with the componant details which relates to the BOM, cross-refrence this with the pcb layout for part placment and you can't go wrong. If you've never soldered before, spend 20-30 minutes practicing beforehand and you should be ok.
 
Quote:

How hard is the build this amp? Is it as easy as looking at the instructional, plugging parts into proper place and then soldering it? I want to try this amp but have never built an amp or anything before.

 

 

[snakeeater69]

Most you guys building the desktop version with 1/4mm plugs?
 
For portable one, has any been able to fit it with an altoids case? Just wondering because I'm cheap and want this be as cheap as possible.

 

[i_djoel2000]

Quote:

Most you guys building the desktop version with 1/4mm plugs?
 
For portable one, has any been able to fit it with an altoids case? Just wondering because I'm cheap and want this be as cheap as possible.

not possible. even the size of the pcb is already bigger than the outer dimension of altoids case

 

[Paul Blythe]

If you want to use a 1/4" output socket the NRJ3HF-1 socket and NRJ-NUT-B locking nut are the recomended replacement parts... this is mounted off-board and wired to P2.

 

[tomjtx]

Quote:

Well, I think something got damaged with all the handling and soldering/unsoldering, and now the left channel is slightly lower in volume. Any ideas on what could have caused that? I have no MM so basically I can reflow the solder joints and thats about it. But I do have a few spare/upgrade parts here so I might try swapping them at random and seeing what happens. Could a bad resistor cause that? I touched a resistor with the soldering iron for a few seconds trying to surface mount the poly caps.

This sounds like 'bad amp karma"  :=)
 
After all, you have done everything to this amp that Voldemort said was unnecessary  to do. Indeed, he designed the amp to demonstrate that all your mods are a waste of time and money.
 
It has been entertaining though, and we thank you for that :=)

 

[AKG240mkII]

Quote:

Update: I swapped out the other two elctrolytics, C8 & C9 on the board, and while it results in a perceived reduction in bass, bass quality is improved while the body and richness of the mids was increased. This has a good effect on some types of music like old Kraftwerk which is not quite as robust sounding as modern synth music.
 

WOW !!
Swapping caps to compensate for poor digital remastering of old analog tape-recordings, that's dedication
But it isn't HiFi . It is entirely possible that it sounds better to you for some reason but I would like to see some measurements
establishing the deficiency you claim and also of the fix . 
 
Regarding the RMAA debate :
Dedicated audio-analysis hardware costing thousands of dollars combined with math that adds up vs a more or less free PC-program, often using non-calibrated hardware ?
That's a tough one ..
 
Now excuse me, I'm working on a sensational report on the huge sonic differences between cables made from
South-African vs Russian gold .
 

 

[micmacmo]

Quote:

For portable one, has any been able to fit it with an altoids case? Just wondering because I'm cheap and want this be as cheap as possible.

The dimensions of the PCB alone are 8 cm x 10 cm. Too big, me thinks, for an Altoid tin.