[stellarelephant]

  Silmic 100uF/16V and Wima 1uF here at C2 for high Z 'phones.
 

This is my NHB unit for low Z 'phones. It uses 470uF/16V Silmic's with 4.7uF Wima's at C2. Since the Silmics are so large .......
 

 
FWIW, I much prefer the sound of the Silmic's to the UKA's. The UKA's sound thin to me, with a forward midrange that can sound a bit 'shouty'. The Silmic's are much smoother, but still give plenty of detail.

 
Oh, wow.  That one (with the huge 470uF Silmic) just looks awesome!!!  Wish I could hear that setup....especially since you say the Silmics sound better than the UKAs, which isn't surprising, considering the Silmic reputation.  I am using the slightly bigger 25V version of the UKA, which theoretically means a lot less distortion from the cap...I can't say I've noticed a "shouty" sound from mine, but I would agree that it isn't the smoothest I've ever heard in my life either, at least without a bypass.  Still looking into bypass options...
 
What is the purpose of the little Silmics on the input?  I was under the impression that 2.2uF already gets the low corner frequency down to optimal range(2Hz) coming into the amp...essentially perfect, complete bass extension.
 
Nice work funch.  I'm feeling some amp envy. 

 

 

[funch]

You will have to check with 'X' on the Silmic's at C1. I just copied his lead, and I'm liking what I'm hearing. 
 
ATM, I'm listening to my DIY 'phones using the ubiquitous Fostex T50RP drivers out of the above amp, and this may the best I've heard these 'phones sound.

 

[stellarelephant]

Funch,
Whoa, you are listening to music through headphones you built, fed by the amp you built.  That must be one heck of a feeling of accomplishment!  Do you have a link to your headphone build?
 
So with the big Silmic version, how would you say the amp's treble extension and clarity compares to your other builds?  Any difference?  My goal is to get a bit more "air" out of mine.  More than likely the perfect bypass will do this for me eventually, but I am interested in the Silmics also.
 
X, can you explain the benefits of a Silmic on the input?

 

[xrk971]

StellarElephant,
More capacitance in the form of the venerable silk paper matrix is always better than less. But in fact, I have a more technical reason on my NHB amps - the entire resistor network is re-tuned for more dynamics and resolution than you can get with just plain old cap rolling. One of the resistors is now a smaller value so it necessitates increasing C1 to keep bass extension low. So why not do it by adding 4.7uF of premier silk paper caps?

 

[xrk971]

I'm working on a big SE Class A desktop amp now. Predicted FFT for 1watt into an 8ohm load looks excellent. Similar profile for 50ohm, 16ohms, etc loads. Capable of delivering 22w of pure Class A power Into 8ohm load and 3.2w into 50ohm loads. Uses three actives in the signal path and four actives total (1 for CCS). Draws 1.13amp bias current and dissipates 40w of heat per channel. So quite the little room heater.

Schematic is sent out to pro layout artist for developing PCB at present and I should have a PCB in hand in maybe 2 weeks.

My listening tests with my single channel hand etched prototype have been mesmerizing. The depth and dynamics of the bass is incredible. Works great powering my full range TL speakers that have Fostex FF105WK drivers as well.

 

[funch]

  Funch,
Whoa, you are listening to music through headphones you built, fed by the amp you built.  That must be one heck of a feeling of accomplishment!  Do you have a link to your headphone build?
 

 
Scroll down to post #1945. The white gel pads have been long since replaced by Audeze LCD pads. 
 
http://www.head-fi.org/t/501773/thunderpants/1935

 

[stellarelephant]

StellarElephant,
More capacitance in the form of the venerable silk paper matrix is always better than less. But in fact, I have a more technical reason on my NHB amps - the entire resistor network is re-tuned for more dynamics and resolution than you can get with just plain old cap rolling. One of the resistors is now a smaller value so it necessitates increasing C1 to keep bass extension low. So why not do it by adding 4.7uF of premier silk paper caps?

 
Ah, I see.  Now I remember that on DIYA you said that you needed to use higher capacitance input caps, because with the NHB you changed the local feedback resistor(R2) from 47K to 22K.  Sorry to be dense, but can you explain the benefit of this?  Are you also putting a 68R at R4 in the NHB for higher bias and more slam with low impedance cans?  Any other changes?
 
If you could give rough estimates of how these resistor substitutions (perhaps individually and combined) affect battery life, that would certainly help me, and possibly others who are planning this build, to weigh the benefits and choose how to customize our builds. 
 
Fellow Head-Fi member Macky112 just PM'd me yesterday asking about how to tweak the amp for two builds--one for 50 ohm cans and one for 300 ohm.  I gave him the best answer I could based on my capacitor research, but you're in a better position than me to give recommendations, X. 

 

[xrk971]

The NHB amp tuning is dependent on amp-to-amp basis so I can't give general values used. 

 

[funch]

FYI, I just tried replacing the 4.7 Wima's with 2.2's, but the bass was a tad overpowering, so went back to the 4.7's. 

 

[stellarelephant]

FYI, I just tried replacing the 4.7 Wima's with 2.2's, but the bass was a tad overpowering, so went back to the 4.7's. 

Good to know. I'm going to try the 4.7 too. What impedance are your DIY cans?

Btw, those things look epic. Why did you replace the earpads? Those gel things look crazy cool. Like sandwiching your head between two sticky cream cheese bagels...I bet the seal was incredible.

 

[stellarelephant]

I'm working on a big SE Class A desktop amp now. Predicted FFT for 1watt into an 8ohm load looks excellent. Similar profile for 50ohm, 16ohms, etc loads. Capable of delivering 22w of pure Class A power Into 8ohm load and 3.2w into 50ohm loads. Uses three actives in the signal path and four actives total (1 for CCS). Draws 1.13amp bias current and dissipates 40w of heat per channel. So quite the little room heater.

Schematic is sent out to pro layout artist for developing PCB at present and I should have a PCB in hand in maybe 2 weeks.

My listening tests with my single channel hand etched prototype have been mesmerizing. The depth and dynamics of the bass is incredible. Works great powering my full range TL speakers that have Fostex FF105WK drivers as well.

This sounds like a great project. I have been using my Pocket Class A as a preamp between my xDuoo X3 DAP and my DIY Aurasound full-range boombox, which has a Tripath amp board inside. The Pocket Class A really warms things up and brings the vocals to life. (Pic was taken before I had the PCA).

I'm sure the whole thing would sound even better if I tore out the Tripath board and installed one of your 22W desktop amps. Man, I bet it'd get hot in there though! I'd probably have to figure out some sort of ventilation for the center cavity where the amp would go.

 

[funch]

Good to know. I'm going to try the 4.7 too. What impedance are your DIY cans?

Btw, those things look epic. Why did you replace the earpads? Those gel things look crazy cool. Like sandwiching your head between two sticky cream cheese bagels...I bet the seal was incredible.

The impedance is 50 ohms. I replaced the earpads with the Audeze's as they give much better sound.
The gels are Beyer's and are made for communication use, like in private planes. When you first put them on,
they feel cold and take a while to warm up, which is another reason I sold them.

 

[stellarelephant]

50 ohms--same as my Sennheiser 598. We are in the same boat in terms of output capacitance needs for bass. With 470uF we are doing very well and getting about a 6Hz low corner frequency. I did a bode plot centered on this value, and it turns out the low end rolloff is almost non-existent...better than -1dB at 20Hz and essentially flat from 30 on up. I'll post a pic when I get access to my other computer.

 

[stellarelephant]

A 470uF cap gives excellent results for 50 ohm cans. Bass roll-off is a minuscule 0.5dB at 20Hz.

 

[stellarelephant]

Well, I finally ordered a wide variety of capacitors to try in this amp. Everything from Silmic and Muse electrolytics (even one 47uF bi-polar) to various values of polyester, polypropylene, and polyphenylene sulfide bypasses. I'm even going to experiment with a few C0G ceramics. Eagerly awaiting my Mouser box...