O2 amp CRC, diode, cap, and heatsink mods - Page 13 - diyAudio
Go Back   Home > Forums > Amplifiers > Headphone Systems

Headphone Systems Everything to do with Headphones

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 25th January 2013, 11:08 PM   #121
agdr is offline agdr  United States
diyAudio Member
 
agdr's Avatar
 
Join Date: Sep 2010
Default Improved DC Servo for O2 amp output stage

I've posted a schematic and discussion for a high accuracy DC servo for the output stage of Rocket Scientist's O2 headphone amp here as a wiki page:

DC Servo For O2 Headphone Amp Output Stage - diyAudio

Figured that I would give the site's wiki system a try. The circuit is based on a fixed/upgraded version of the DC servo in the National Semiconductor LME49600 evaluation board. The circuit uses a OPA627 for high DC accuracy for input offset voltage and extremely low input bias current (FET input). The circuit makes use of a 3 pole passive input low pass filter to extract DC from the O2's output signal with just 40uV of AC ripple left on it, about half the maximum input offset voltage of the OPA627, effectively pure DC for purposes here.

The goal of this servo is to zero out the 3mV or so of DC offset voltage on the O2 amps outputs caused by the 1.5mV or so input offset voltage of the NJM4556 chips, plus their relatively high input bias current pulled through the 40.2K resistors to ground causing an additional IR drop that is added to the NJM4556 input. The servo allows those resistors to be bumped up to 160k, reducing the low frequency cutoff of the output stage coupling cap filter. The caps can also be increased to 4.7uF, as mentioned in one of the mods above, for even a lower Fc.

The servo likely won't make any audible difference, and as per AKG tech support really won't even matter for most phones since a 3mV deviation of the diaphram out of 130mV or so of excursion range really doesn't matter. However many folks like the idea of the transducers in the headphones sitting at their natural resting positions, not 3mV one way or the other. The small 3mV offset might make a difference in some sensitive IEMs.
  Reply With Quote
Old 14th February 2013, 08:44 AM   #122
diyAudio Member
 
availlyrics's Avatar
 
Join Date: May 2011
Default How/were to solder .01 to .1 uF bypass caps. in O2

Pls read "A question about bypass cap." #3092 & comments by sofspud & gootee on page 310.

The Objective2 (O2) Headphone Amp DIY Project

Since you are good at hacking /moding O2(& many more things) at the same time having measuring instruments to chk. your hacks/mods can you pls test if bypass caps(.01 to .1 uF) at U1 and maybe at U2/U3 can change the performance of O2 w.r.t. transient response or HF stability.
__________________
Thanks,
availlyrics
  Reply With Quote
Old 14th February 2013, 12:13 PM   #123
agdr is offline agdr  United States
diyAudio Member
 
agdr's Avatar
 
Join Date: Sep 2010
Quote:
Originally Posted by availlyrics View Post
Pls read "A question about bypass cap." #3092 & comments by sofspud & gootee on page 310.
In the case of the O2, my two cents is you are OK with what RocketScientist has in place for bypassing. The speed of the chips involved and the speed of the information going through figure into things, since the problem is inductance in the traces. The NJM2068 and NJM4556A are both "audio" chips that just are not all that fast, which is fine (preferable) for audio and less likely to launch into HF oscillations.

By way of comparison, take a look at what OPC has done in the "Wire" for bypassing (schematic link at the bottom):

"The Wire" Headphone Amp Build Wiki - diyAudio

Now there the LME49600 is another animal entirely. That is a "video" chip that is fast enough to pass video and does require what you are saying. OPC has 10uF X7R MLCC ceramics right at the pins of the LME49600, which is a great design imho. In the LME49600 evaluation board data sheet they have 1uF X7Rs in parallel with 10uF tantalums at each rail to ground, as I recall, but those were likely the days before you could just buy a 10uF X7R MLCC. A ceramic cap is good at high(er) frequencies. The 10uf is for energy storage while the MLCC ceramic is to nuke HF oscillations. The one 10uF XLR MLCC gives you both.

Same holds for the LME49990. The gain-bandwidth product on that guy is 110mHz! Compared to a GBW of just 19mHz for the NJM2068. Again OPC has put 1uF X7R MLCCs right at the pins. Less energy storage is needed with the gain chip than with the high current output chip there, hence the lower capacitance for the bypass cap.

Like sofaspud posted in #3094 I had one 10uF MLCC X7R sitting between the two OPA627 chips in the ODA layout I posted yesterday in an earlier layaout where the chips were further away from the power supply caps. In that case I just had one cap going rail to rail, none rail to ground. The power lines fed the cap which then star'ed off to the chips on either side. I may add it back, actually, was thinking about that today.

Bypassing can get into a real religious discussion! Lots of good thoughts on it all out there.

Last edited by agdr; 14th February 2013 at 12:38 PM.
  Reply With Quote
Old 15th February 2013, 05:18 AM   #124
diyAudio Member
 
availlyrics's Avatar
 
Join Date: May 2011
Quote:
Originally Posted by agdr View Post
In the case of the O2, my two cents is you are OK with what RocketScientist has in place for bypassing.....
That cleared my doubts.BTW I'm atheist
__________________
Thanks,
availlyrics
  Reply With Quote
Old 15th February 2013, 07:36 AM   #125
diyAudio Member
 
sofaspud's Avatar
 
Join Date: Nov 2005
Location: San Antonio
AFAIK, Jerald Graeme, Bonnie Baker, et al (Baker has a co-author whose name escapes me) have the most definitive examination of op amp bypassing that I'm aware of. I think some of it can be found at the Electronic Design magazine website.
__________________
It is error only, and not truth, that shrinks from enquiry. - Thomas Paine
  Reply With Quote
Old 19th February 2013, 03:57 PM   #126
diyAudio Member
 
availlyrics's Avatar
 
Join Date: May 2011
Looks like some more reading needs to be done
__________________
Thanks,
availlyrics
  Reply With Quote
Old 21st February 2013, 08:52 PM   #127
aphocus is offline aphocus  New Zealand
diyAudio Member
 
Join Date: Jun 2012
agdr, I'm looking at your Bass-boost mod.

I'm curious as to how one would bring the frequency + bandwidth down, I'm hoping to achieve +10dB at 30hz and about +5dB at 60Hz and maybe about +1dB(or less) at 100hz

I'm really in to my Subbass and considering my Headphones naturally drop to -10dB at 20hz.

I was originally looking at doing this Passive RC low-shelf pass but your method looks more interesting.
  Reply With Quote
Old 22nd February 2013, 02:27 AM   #128
agdr is offline agdr  United States
diyAudio Member
 
agdr's Avatar
 
Join Date: Sep 2010
Quote:
Originally Posted by aphocus View Post
I'm hoping to achieve +10dB at 30hz and about +5dB at 60Hz and maybe about +1dB(or less) at 100hz
The first order filter in the boost circuit can't roll off quite that fast, but can get in the ballpark. Looks like the closest it can get is with the values below, 0.39uF and 2K for the boost part and 7.5K for the regular feedback resistor. The circuit winds up looking a lot like the rightmost diagram in your link for the shelving filters, except their configuration is for an inverting op amp and here it is non-inverting.

In the plots below the first is an O2 with no bass boost. The second is the regular boost parts, and the 3rd is with these values. The net result is:

10Hz 18dB (+10.3dB)
30Hz 16.7dB (+9dB)
60Hz 14.3dB (+6.6dB)
100Hz 11.5dB (+3.8dB)
273Hz 8.7dB (+1dB)
from 1Khz on 7.7dB.

If you need less than 7.7dB of gain (2.5x) after 1kHz use the attenuation mod earlier in this thread to knock the incoming signal down as needed. That modification just substitutes a larger value for the incoming 274R resistor to form a voltage divider with the 10k resistor to ground on the input of the O2. For example, if the 274R is replaced with a 10K that would form a 50/50 voltage divider. The input signal would get cut in half, then 2.5x from 1Khz on, yielding about 1.25x. The boost amount would get cut in half though, too, with the attenuation mod. In that case the amount of boost would need to be doubled to compensate.
Attached Images
File Type: jpg Original O2 No Bass Boost AC plot.jpg (38.9 KB, 170 views)
File Type: jpg Original Bass Boost +4dB.jpg (39.4 KB, 168 views)
File Type: jpg O2 boost 16.8 dB at 30Hz circuit.jpg (71.6 KB, 167 views)
File Type: jpg O2 boost 16.8 dB at 30Hz plot.jpg (40.5 KB, 164 views)

Last edited by agdr; 22nd February 2013 at 02:49 AM.
  Reply With Quote
Old 22nd February 2013, 07:58 PM   #129
aphocus is offline aphocus  New Zealand
diyAudio Member
 
Join Date: Jun 2012
Thanks! I have LTSpice open with the Passive RC design in it and it looks like it suffers from the same bandwidth issues, my current one is peaking around 4hz , and hitting (-9.5dB) at 100hz.

The only way I can see getting it to work my way is to peak it near 1hz at +20dB, so the slope lands at +10dB at 20Hz, and 1dB at 100hz. I would never consider this except the other shelf filter has a -8dB from peak at 1hz which might offset this excessive gain.
  Reply With Quote
Old 23rd February 2013, 01:23 AM   #130
agdr is offline agdr  United States
diyAudio Member
 
agdr's Avatar
 
Join Date: Sep 2010
Quote:
Originally Posted by aphocus View Post
The only way I can see getting it to work my way is to peak it near 1hz at +20dB, so the slope lands at +10dB at 20Hz, and 1dB at 100hz.
Yes, that might do the job! Certainly worth trying. I messed around with that a bit, trying for just the last half or so of the roll-off curve, but it never seemed like the roll-off was steep enough to hit the 30Hz, 60Hz, and 100Hz specs. My guess is at least a second order filter would be needed in the feedback loop to get a steeper slope. I don't see a way to do that offhand with the shelving effect. A 2nd order MFB or salen key wouldn't level off, they would just keep going. Maybe someone out there has worked this out before and will post. Seems like the kind of thing that would come up a lot with subwoofers.
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Reduced diode switching by placing a cap across Goetz Power Supplies 10 19th December 2011 01:38 AM
DC blocker diode/cap orientation Glen B Solid State 4 22nd August 2011 06:46 PM
Mounting a small signal diode on heatsink - need suggestion IG81 Solid State 14 3rd November 2009 05:12 PM
diode bridge - 10000uF cap - diode tube - small cap - HV+ engels Tubes / Valves 5 29th January 2008 11:16 PM
Diode Heatsink dmk Tubes / Valves 8 9th September 2007 05:03 PM


New To Site? Need Help?

All times are GMT. The time now is 01:40 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2