O2 headamp output booster PCB - Page 4 - diyAudio
Go Back   Home > Forums > Amplifiers > Headphone Systems
Home Forums Rules Articles diyAudio Store Gallery Wiki Blogs Register Donations FAQ Calendar Search Today's Posts Mark Forums Read

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
Thread Tools Search this Thread
Old 2nd November 2013, 01:53 AM   #31
agdr is offline agdr  United States
diyAudio Member
agdr's Avatar
Join Date: Sep 2010
Default OPA140 DC offset lower than OPA827

I was able to get the OPA827+LME49600 end of this project board going with a surprising result: the DC offset of the OPA140+LME49600 is a lot lower than the more expensive OPA827. Both sound terrific though, subjectively.

The photos below:

* The first two photos are the test setup, running on batteries this time rather than the AC adapter. The inputs are grounded. The O2 amp's other NJM4556A chip is gone, with a header running the wires to the OPA827 section. The meter on the left is the OPA827+LME49600 end of the board. The meter on the right is the OPA140+LME49600 end. 24uV for the OPA827 loop, just 1uV for the OPA140 loop. lol - the OPA140's DC was so close to zero that I had to keep occationally injecting noise with my finger to verify a test lead hadn't popped off somewhere.
* closeup of the displays a minute later. The OPA827 loop had wandered a bit to 29uV.
* AC test with music playing and sensitive headphones. Instantaneous values of around 22mVac (rms) on both channels, about mid-volume for the test headphones, with 143uV of DC offset for the OPA827 loop and 15mV for the OPA140 loop.
* Stability and oscillation test setup with an axial 1000pf MLCC capacitor placed across each headphone output, with the headphones connected.
* scope shot of the OPA140 loop channel with amplifier inputs grounded. No signs of oscillation at all, at least to the 5mV/div limit of the scope. Combined probe + scope bandwidth is probably around 80mHz at the 3dB point, so it could miss any higher frequency oscillation. Both OPAs and both LME49600s are running cool though, so most likely no higher oscillation occuring.
* for comparison, a scope shot with the O2 amp turned off (just random background noise).
* scope shot of a 1kHz signal playing through the O2.
* scope shot with music playing.

The scope shots of the OPA827 loop channel were similar. No signs of oscillation. Sergey888's feedback loop is still a really good idea though. I'll add it to the next version. Certainly can't hurt.

And... a typo on my previous post. The DC offset with the 1Vac (rms) into the 8 ohm speaker is 1.6mV, as can be seen in the photo, not 16mV.
Attached Images
File Type: jpg IMG_2112.JPG (205.6 KB, 403 views)
File Type: jpg IMG_2113.JPG (187.8 KB, 401 views)
File Type: jpg IMG_2114.JPG (71.9 KB, 391 views)
File Type: jpg IMG_2119.JPG (63.7 KB, 376 views)
File Type: jpg IMG_2121.JPG (172.6 KB, 375 views)
File Type: jpg IMG_2122.JPG (143.9 KB, 53 views)
File Type: jpg IMG_2123.JPG (129.7 KB, 49 views)
File Type: jpg IMG_2124.JPG (136.2 KB, 48 views)
File Type: jpg IMG_2127.JPG (196.0 KB, 73 views)

Last edited by agdr; 2nd November 2013 at 02:08 AM.
  Reply With Quote
Old 2nd November 2013, 12:48 PM   #32
Shaq888 is offline Shaq888  Ukraine
diyAudio Member
Shaq888's Avatar
Join Date: Feb 2009
Glad there was no smoke on the first power up
  Reply With Quote
Old 2nd November 2013, 07:31 PM   #33
agdr is offline agdr  United States
diyAudio Member
agdr's Avatar
Join Date: Sep 2010
Sergey888 - it actually did smoke on first power-up. I had the tantalums, C13 and C14, in backwards. C13 lasted about 3 seconds then went up in smoke. I doubled checked voltage and polarity - all OK - and replaced it with a second one. That went up in smoke too. I'm new to the SMD stuff. Why on earth would they put a band on the positive end of a part? Lol. Luckily I forgot to order the 47uF tantalums in my first order from Mouser, or they would have gone up in smoke too.

Hey good find with those OPA140's! I think that is the "standard" op amp for this project. Low quiescent current, low DC offset, low cost, and sounds great.

I have to use the 5cm x 10cm standard fab size from Seeed studio for the 28mm x 100mm board. I just realized the board doesn't have to be rectangular. I could extend a tab right over the O2 amp's NJM4556's to reduce the amount of wiring going to the dip headers and still stay within 5cm x 100cm. I was even considering using solid pins between the boards, but too great a chance of a registration error in lining them up. The wiring allows for some flex.
Attached Images
File Type: png O2 output PCB OPA827 LME49600 layout lyr 1_4.png (158.1 KB, 88 views)

Last edited by agdr; 2nd November 2013 at 07:37 PM.
  Reply With Quote
Old 6th November 2013, 04:19 AM   #34
agdr is offline agdr  United States
diyAudio Member
agdr's Avatar
Join Date: Sep 2010
Sergey888 - it looks like I'm going to have room for your second order feedback loops after all. Does this circuit look like what you had in mind? Two 33pF in series with the 3.3K to ground, and the 1K back from the LME49600 output.

I think that I am going to try solid DIP headers after all to eliminate a pile of connecting wires. The board will have the tab with 12 pins that just plugs right into the U3 socket and the top 4 pins of U4 on the O2 amp board. The change makes this truely a "plug-in" upgrade to NwAvGuy's O2 amp. Those are not chips out on the tab. I just created a modified DIP8 socket part in the Eagle library. A 4 pin straight header would actually solder in each side, like #649-68024-413HLF at Mouser, with 4 in a row snapped off for U3 and 3 snapped off for the top 3 pins of U3 (two on board an one with the connecting wire).

That connects everything but ground and one OPA140 input, which will be the only remaining connections needing wires. That input is extended out to the tab, though, so it is just a few millimeters from pin 3 on U4 where it goes with a single pin header. I ran out of board area there with the 5cm fab restriction (staying within the 10cm x 5cm size at Seeed Studio for 4 layer), otherwise the whole U4 header could have been out on the tab too.

I've added another 47uF tantalum rail-to-rail in addition to the two going rail-to-ground.

The tab will mean the board couldn't be panelized 3 to a 10cm x 10 cm fab area, but still could get two in.
Attached Images
File Type: png O2 output PCB OPA140 LME49600 circuit.png (44.8 KB, 80 views)
File Type: png O2 output PCB OPA140 LME49600 layout lyr 1_4.png (162.9 KB, 79 views)

Last edited by agdr; 6th November 2013 at 04:39 AM.
  Reply With Quote
Old 7th November 2013, 12:15 AM   #35
Shaq888 is offline Shaq888  Ukraine
diyAudio Member
Shaq888's Avatar
Join Date: Feb 2009

Yep, it looks correct.
Beside, in this particular case with unity gain you can connect 3.3k resistor to output of LME with very similar result.
Can I ask you what is the purpose of R1 R2?

I still think having two LME on different sides of PCB is not ideal solution. It can be a bit awkward mechanically, but I would place them side by side, and use dual OPA2140. Dual opaamp pinout is actually more layout friendly.
  Reply With Quote
Old 7th November 2013, 04:12 AM   #36
agdr is offline agdr  United States
diyAudio Member
agdr's Avatar
Join Date: Sep 2010
Sergey888 - R1 and R2 are a modification for the O2 headamp that I figured I would add to this board since I had some space left over. They help prevent turn-off thumps. NwAvGuy only had chips as the load on the power rails after the power management mosfets. The resistors help with a linear drain to zero volts below the chips' minimum power supply voltages. So R1 and R2 are optional.

Good thoughts about putting the LME49600s side by side! That actually hadn't occured to me all. I guess that I was focused on maximum heat dissipation by putting them on the ends of the board. The thermal vias work a lot better than I thought they would.

I just whipped up a quick version with the side-by-side LNE49600s and the OPA2140. That does reduce the parts count quite a bit. It looks like DIY Audio's picture upload function is broken right now, so I have the layout, schematic, and BOM here on Google Drive:


I've updated the BOM with the 10uF X7R MLCC's replacing the 4.7uF tantalums for decoupling on the LME49600's. No need for the 0.1uF bypassing MLCCs anymore on those, just on the OPA2140. That rectangular gap in the bottom metal is where the board sits right over two of the O2's electrolytic caps with metal tops. I'm leaving pads for the LME49600's bandwidth resistor pads out entirely. There is just no earthly need I'm aware of to use 180mHz bandwidth for audio, with the associated double quiescent current draw and high(er) frequency oscillation potential.

I'm down to only one connecting wire now, just ground! I realized last night that since the board now plugs into the O2's U3 and U4 dip sockets it won't rattle around and doesn't exactly need to be flush with the front panel anymore. I had it wedged between the front panel and battery connectors, in the top case slot. I moved it 2mm back from the front panel, which added 2mm to the end of the tab (staying within the 5cm fab limit), enough to add a 3rd row of pins on U4 and connect that input.

Last edited by agdr; 7th November 2013 at 04:42 AM.
  Reply With Quote
Old 9th November 2013, 02:41 AM   #37
agdr is offline agdr  United States
diyAudio Member
agdr's Avatar
Join Date: Sep 2010
I'm adding "upgrade board" for the O2 headphone amp to this thread title since I've also added a few modifications to the O2 that I previously posted in the O2 modification thread. I had some extra board space so might as well.

The board now optionally includes:
  • After-mosfet power LEDs on each rail. The O2's let says when the power switch is on but says nothing about whether the mosfets have switch (power management circuit) and if both power rails are up. The LEDs are in series with the anti-thump turn-off resistors so just make use of that existing current flow.
  • Anti-turn-off-thump resistors to bleed the post-mosfet power rails linearly to ground.
  • Power management circuit latch circuit. Prevents the O2 from "motorboating" or "oscillating" when the batteries run down, due to the batteries recharging a bit when the mosfets cut off. Requires two additional wires from the new board to the O2 board. Optional to populate.
  • 1/4" neutrik output jack. If the taller B3-080 case is used there is enough space above the O2' 1/8" output jack and on/off switch to mount a 1/4" slimline Neutrik jack upside down on the bottom of the new board. The board hooks it into the new outputs automatically, so it is plug-and-play ready to go. Adding 1/4" jacks to the O2 using the B3-080 case have seemed to be a popular modification.
In the layout the top layer (red solid) is V- when under the LME49600 and just general routing otherwise. Layer 2 (blue crosshatch) is ground, except when carved out for routing. Layer 3 (red hatch) is V+. The bottom layer (blue solid) is also V-, largely heatsink area connected to the top V- with thermal vias, except for a few routing cutouts.

The files are also on Google Drive at:


Once the PC board is back from fabrication and tested I'll post the Gerber files out at that link.

Last edited by agdr; 9th November 2013 at 02:56 AM.
  Reply With Quote
Old 14th November 2013, 04:48 AM   #38
agdr is offline agdr  United States
diyAudio Member
agdr's Avatar
Join Date: Sep 2010
Sergey888 - while updating the build instructions I remembered why I was hanging onto the single op-amp layout, so that the board could also optionally be used with OPA827, OPA627, OPA1641, LME49710 and LME49990 chips along with the OPA140. So even though it is more space-inefficient I've gone back to two single OPA140s instead of the one dual OPA2140. I've kept the two LME49600s side by side though. Thanks again for all the design input! I've made note of that in the updated build instructions and on the board.

Some other things I've tinkered with:

* The decoupling caps on the OPA140 chips are now the same 2.2uF 50V X7Rs MLCC used in the power management latch circuit, up from 0.1uF COG previously - another good suggestion from Sergey888.
* The signal lines feeding the upside down 1/4" jack now have zero ohm resistors in series to disconnect the jack from the LME49600's on the O2 upgrade board. Allows the jack to use used independently from the rest of the board, and/or the board used just for the jack (wired into the O2's P2 connector instead).
* All the layers directly under the two op amps are cut out, helping to prevent any input to output stray capacitance.
* All the V+ power and output signal traces are "doubled nailed" with two vias to lower via impedance and increase current handling ability.
* The LME49600 output lines now go to both O2 NJM4556A socket "paralleled" output pins on U3 and U4 to run the signal through both of the O2's 1R resistors on each channel, keeping the output impedance at 0.5R. The resistors are really not needed through with the LME49600, of course. One of the O2's 1R resistors on each channel can simply be shorted across under the board to get nearly zero output impedance. But this way the board remains plug and play without having to do that.

Here is a Google Drive link to all the versions of the board. The folder with the latest date is the latest version.


The Google Drive link also has PDF versions of the circuit layout that can be blown up as big as needed, plus it has png's of each layer individually.

The BOM and build instructions now include all the O2 modification sections, plus how to (optionally) upgrade the O2's power rail voltage to +/-15Vdc after the O2 upgrade board is completed.

Last edited by agdr; 14th November 2013 at 04:55 AM.
  Reply With Quote
Old 3rd December 2013, 03:26 AM   #39
agdr is offline agdr  United States
diyAudio Member
agdr's Avatar
Join Date: Sep 2010
Default It fits and it works!

The PC boards for this project arrived today. I just built one up and much to my surprise it not only fits mechanically in the O2, but it works! Didn't screw anything up.

If all goes well with some additional testing I'll post the PC board Gerbers at the Google Drive link above and will sell off the extras in the vendor thread at cost.

The first photo below is the boards being CAT scanned.

The next two photos show the front and back of the PC board. This is one of the stranger shaped PC boards out there. I went with a blue solder mask this time for something different.

The next 3 photos show one of the linear SIP headers being snipped with wire cutters into two 4 pin sections and two 3 pin sections. Those are then inserted into the bottom of the PCB board in the two IC holes in the tab and soldered (on the top of the PC board). They don't have to be exactly straight since they are going to be bent in the next step anyway, but I used a second one of the PC boards slipped over the free ends to align them while soldering.

The next 3 photos show those SIP pins being cut to 2.4mm long with wire cutters, bent outward slightly so they fit on the outside of the coined contacts on the DIP8 headers, then soldered to the header contacts. The net result is the two 8 pin DIP headers ready to plug into the O2 amp's U3 and U4 socket.

I wanted to build up this header contact portion first to make sure I hadn't screwed up the mechanical measurements. The final photo shows the header contacts plugged into the O2 and the whole thing slid into the B2-080 case bottom (O2 PCB) and top (O2 boost board) slots. It fits!

I'm out of picture posting slots on this post, so part 2 will be posted next...
Attached Images
File Type: jpg IMG_2151.JPG (227.4 KB, 77 views)
File Type: jpg IMG_2152.JPG (155.7 KB, 94 views)
File Type: jpg IMG_2155.JPG (204.7 KB, 99 views)
File Type: jpg IMG_2156.JPG (210.4 KB, 76 views)
File Type: jpg IMG_2159.JPG (141.4 KB, 73 views)
File Type: jpg IMG_2161.JPG (133.0 KB, 67 views)
File Type: jpg IMG_2162.JPG (122.2 KB, 74 views)
File Type: jpg IMG_2163.JPG (127.6 KB, 72 views)
File Type: jpg IMG_2166.JPG (112.8 KB, 68 views)
File Type: jpg IMG_2172.JPG (118.9 KB, 89 views)

Last edited by agdr; 3rd December 2013 at 03:40 AM.
  Reply With Quote
Old 3rd December 2013, 04:08 AM   #40
agdr is offline agdr  United States
diyAudio Member
agdr's Avatar
Join Date: Sep 2010
Default Build and test photos, part 2...

The first three photos are more shots of the O2 booster PCB mated with the O2 PCB, since I ran out of picture posting slots above.

Next photo shows soldering on the LM49600s with a wide tip on the iron.

The mext photo is all the amp parts soldered on. I haven't soldered on any of the O2 modification parts yet, just wanted to test the amp at this point. There is no real order to soldering on the rest of the parts, with two exceptions.

The first is that I did manage to fit in pads for the LME49600 bandwith resistors R7 and R10. Normally these are left unconnected (no resistors soldered on) since the 110mHz "low" bandwidth is more than enough for audio and is the low(er) 7mA idle current mode. But if someone really wants to put it in high bandwidth mode, there are the pads. If those two resistors are used they should be soldered on before the LME49600s given how close they are to the chip, although I believe I could have still soldered them on after if needed.

Then the other soldering-order thing is to leave the two OPA140s (or OPA827s, whatever you are using) for last given the static sensitive jfet input pins. There is no resistor to ground on the jfet inputs on the board since it relies on the O2's 40.2K resistors to ground once plugged into the O2. Best to get the rest of the passive parts on there first. There is plenty of room for drag soldering those two chips even after all the other SMD parts are on. Then double check (triple check!) all the solder connections with a magnifying glass for bridges or opens. Note that I'm using an anti-static mat at this point with the iron grounded and a grounded wrist strap.

The green wire is the ground going from JP3 ground hole to the nearby ground of the batteries (middle two batt connectors). The wire doen't have to be nearly that long, of course, I just quickly cut a section of wire and soldered it on to get the measurement show on the road. Another good ground to use is a wire under the board from the JP4 ground hole up front to the ground hole in the O2's P2 jack. You only have to use one or the other ground - your choice - but don't use both. This one ground wire is the only part of the amp that doesn't go through the DIP headers.

The next photo shows the DC test setup since I want to measure output DC offset on both channels first. Two DMMs and some test headphones plugged in.

The next photo shows the O2 amp's new DC offset with the boost board: 10uV and 36uV on the two channels! That is MICRO volts, a 98%+ reduction over the O2's usual 3mV = 3000uV DC offset with the NJM4556A chips.

The next photo shows instantaneous AC and DC captures with music playing. Very sensitive headphones being used, so the 30mV (rms) represents normal listening levels. The instantaneous DC offsets on the two channels read 124uV and 47uV. Instantaneous DC offset jumps around a little bit with the music, of course, since the music's instantateous DC level fluctuates.

The final photo shows fit of the board with all the amp components soldered on, especially the LME49600s and the two 47uF tantalum capacitors which are the tallest. Fits perfectly. I had to pull the metal top of the B2 case up just very slightly to clear the LME49600s to compensate for it not being flat to start with (slight downward bend).

And... quite subjectively, it sounds great! These sensitive headphones are useless for significantly loading the LME49600s. I'm going to attach it to 8 ohm desk speakers again for an abuse test and see how it fares, along with using some less sensitive headphones.

Thanks again to Sergey888 for the great suggestions of the OPA140, the side-by-side LM49600s, and that 2nd order feedback loop! I'm going to throw it on the scope to check for osciallation, but at this point the chips are completely cool. No thermal signs anyway of bad stuff happening.
Attached Images
File Type: jpg IMG_2168.JPG (234.1 KB, 113 views)
File Type: jpg IMG_2169.JPG (161.1 KB, 99 views)
File Type: jpg IMG_2176.JPG (190.2 KB, 84 views)
File Type: jpg IMG_2177.JPG (145.2 KB, 83 views)
File Type: jpg IMG_2178.JPG (182.3 KB, 111 views)
File Type: jpg IMG_2182.JPG (203.2 KB, 105 views)
File Type: jpg IMG_2187.JPG (166.4 KB, 46 views)
File Type: jpg IMG_2184.JPG (162.2 KB, 46 views)
File Type: jpg IMG_2189.JPG (147.9 KB, 77 views)

Last edited by agdr; 3rd December 2013 at 04:30 AM.
  Reply With Quote


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
Zen headamp. PCB, my ver. Stream Pass Labs 43 21st May 2014 08:28 AM
FS - Fully populated & working O2 headamp ThuyTN Swap Meet 0 16th January 2013 02:30 AM
Grommes S35 booster amp output? Dill Pickle Tubes / Valves 1 9th October 2012 07:23 PM
WTT O2 headamp PCB for CCS PCBs regal Swap Meet 0 13th October 2011 03:42 AM
FS: PCB's for the Morgan Jones tube headamp Nisbeth Swap Meet 0 19th October 2004 07:43 PM

New To Site? Need Help?

All times are GMT. The time now is 12:44 PM.

vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2017 DragonByte Technologies Ltd.
Copyright 1999-2017 diyAudio

Content Relevant URLs by vBSEO 3.3.2