Lets ask the other way - why would OPA827 + OPA2209 be a bad choice?
I'm currently running this combo wit 24ohm headphones (Sony MDR 7520) and couldn't really distinguish from the stock O2 opamps. Extremely clean, undistorted, tight, punchy sound, although a little bit on the bright side, just as original O2. Apart that OPA827+2209 take lot less battery power.
I'm not sure about LME49990 - uncompensated would most probably oscillate and its perfect specs would be lowered with 4556...
I'm currently running this combo wit 24ohm headphones (Sony MDR 7520) and couldn't really distinguish from the stock O2 opamps. Extremely clean, undistorted, tight, punchy sound, although a little bit on the bright side, just as original O2. Apart that OPA827+2209 take lot less battery power.
I'm not sure about LME49990 - uncompensated would most probably oscillate and its perfect specs would be lowered with 4556...
LME49990 checked with O2 and working OK, thanks to good layout, and proper bypassing of PCB.
O2 is perfect amp, and there is no to much reasons to change here something, except 2 cases: 1. Some phones required more output from amp. 2. O2 will be used as preamp in some demanding system.
But there is more short answer to you question: You don't have audio analyser to check what are you doing. And you going to alter very fine tuned system. So best approach do not to touch, what you can't to check. Just to not to make a harm.
That was objective point, now subjective.
You still have your ears. You can to give it a try. There is a chance (probably not to big) that you will get an improvement, that you looking for.
O2 is perfect amp, and there is no to much reasons to change here something, except 2 cases: 1. Some phones required more output from amp. 2. O2 will be used as preamp in some demanding system.
But there is more short answer to you question: You don't have audio analyser to check what are you doing. And you going to alter very fine tuned system. So best approach do not to touch, what you can't to check. Just to not to make a harm.
That was objective point, now subjective.
You still have your ears. You can to give it a try. There is a chance (probably not to big) that you will get an improvement, that you looking for.
For the gain stage: because it is untested, because NwAvGuy's own opamp comparison showed that few other opamp could even match the NJM2068 on the bench. If I am not mistaken, the OPA2227 was the one that was able to just match the NJM2068 on the bench. Then you have cost considerations. One goal with the O2 was affordability, hence through-hole components. Agdr tested tested LME49720 (i think) and LME49990 for oscillation and it turns out they work fine, but they also cost much more.
For the output: because OPA2209 doesn't have the current capability hte NJM4556 has, that could be an issue with low impedance phones. And it doesn't come in DIP8.
I have tested LME49990 and OPA2227 in the gain stage. I am not sure I hear a difference compared to the NJM2068. They certainly do not sound five or ten times better, if you look at prices.
So that could be an anser to your question: any other opamps than the default ones would be a bad choice, either because they just are worse, or because they are a waste of money!
But as Tolik says, it is up to you ears (and wallet)!
LME49990 exceeds all parameters of NJM2068... On the paper, or in front of analyser system. But the differences are quite below of human's ears capability threshold. It is true for O2 circuit. NwAvGuy vanished before LME49990 was introduced. But NwAvGuy mentioned that LME49600 is quite good. The mine reason that this IC was avoided - O2 was designed as portable amp, and there was an issue to supply these powerful buffers from simple rechargeable batteries. LME49600 will shine in desktop amps, or amps with LI-xx accumulators, where it will be able to benefit from higher rails voltage, delivering a solid power at output.
True... and I do have the combo LME49990/LME49600
(agdr:s booster board) on one my own O2:s, but it also has an odac (no batteries) and I only use it as a sound card/desktop amp. 
Re: gain stage opamp, there is more than one way to skin the proverbial cat.
As evident from NwAvGuy's own measurements, the '2068 struggles with high levels at a gain of 6.5x. That's because output loading approaches 1k5 at that setting, and the part is not exactly a load-driving king. Hence why he noted that the more robust '5532 did better at high gains. That leaves us with two options for improvement:
a) Put in a more powerful opamp. Downside: These are either more power-hungry (like 5532) or considerably more expensive.
b) Ease up on feedback network resistor values a bit (factor of 2 higher or so). Not a major problem as long as you keep an eye on impedance balance. Noise is not an issue anyway, as it'll always be swamped by source noise.
One other thing the '2068 is not is a rail/rail output capable part. If you insist on using the O2 on batteries while feeding high-level sources like a CD player or ODAC (at full digital level) into it, you may have found that the poor thing clips even at a gain of 2.5x. That's because the gain stage can output maybe 4 Vrms on batteries, and 2 Vrms times 2.5 is 5 Vrms. A RRO capable part like the OPA2209 (possibly in conjunction with item b) above) may just buy enough headroom for things to work out again. This amplifier concept is not user-friendly, that's the price for its superb noise performance.
BTW - has anyone ever bothered to measure what the stock O2 gain stage does when having to output like 5..6 Vrms (mains operation, of course)? Published results only include 2 Vrms, and you can't always pick your source levels.
As evident from NwAvGuy's own measurements, the '2068 struggles with high levels at a gain of 6.5x. That's because output loading approaches 1k5 at that setting, and the part is not exactly a load-driving king. Hence why he noted that the more robust '5532 did better at high gains. That leaves us with two options for improvement:
a) Put in a more powerful opamp. Downside: These are either more power-hungry (like 5532) or considerably more expensive.
b) Ease up on feedback network resistor values a bit (factor of 2 higher or so). Not a major problem as long as you keep an eye on impedance balance. Noise is not an issue anyway, as it'll always be swamped by source noise.
One other thing the '2068 is not is a rail/rail output capable part. If you insist on using the O2 on batteries while feeding high-level sources like a CD player or ODAC (at full digital level) into it, you may have found that the poor thing clips even at a gain of 2.5x. That's because the gain stage can output maybe 4 Vrms on batteries, and 2 Vrms times 2.5 is 5 Vrms. A RRO capable part like the OPA2209 (possibly in conjunction with item b) above) may just buy enough headroom for things to work out again. This amplifier concept is not user-friendly, that's the price for its superb noise performance.
BTW - has anyone ever bothered to measure what the stock O2 gain stage does when having to output like 5..6 Vrms (mains operation, of course)? Published results only include 2 Vrms, and you can't always pick your source levels.
Okay let me explain more why I've chosen OPA827 and OPA2209:
- they have both lower offset voltage
- they have both less noise, less distortion
- they are both less power hungry
- they have reasonable gain bandwidth products (i.e. less high-speed as popular LME49xxx)
- both are unity gain stable (even my HD800 can get driven pretty loud at 1.0x gain)
- 2209 can output cca 50-60 mA if I read the datasheet correctly
Have I made something wrong there with my decision?
- they have both lower offset voltage
- they have both less noise, less distortion
- they are both less power hungry
- they have reasonable gain bandwidth products (i.e. less high-speed as popular LME49xxx)
- both are unity gain stable (even my HD800 can get driven pretty loud at 1.0x gain)
- 2209 can output cca 50-60 mA if I read the datasheet correctly
Have I made something wrong there with my decision?
What resistor values for 2.5k gain
Hi All,
I bought an O2/ODAC kit from Headn'hifi. I have now soldered everything together and done the initial testing and all has passed and appears to be working, although I am yet to install the gain resistors. I could just do with double-checking what I am doing.
I am using this with Beyer Dynamic DT990 Pro 250ohm and DT1350 80ohm headphones. I think that gain settings of 1x and 2.5x (or possibly 3.5x) would be suitable. The trouble is I can't work out what resistor values to use for R19 and R23 (the higher gain). The calculation is 1+1500/value, so would I need something around 600ohms in R19 and R23?
Head n'hifi have included a selection of very small resistors which dont seem to fit the resistor sockets so I was going to use the standard size ones I have to hand.
Thanks, Will
Hi All,
I bought an O2/ODAC kit from Headn'hifi. I have now soldered everything together and done the initial testing and all has passed and appears to be working, although I am yet to install the gain resistors. I could just do with double-checking what I am doing.
I am using this with Beyer Dynamic DT990 Pro 250ohm and DT1350 80ohm headphones. I think that gain settings of 1x and 2.5x (or possibly 3.5x) would be suitable. The trouble is I can't work out what resistor values to use for R19 and R23 (the higher gain). The calculation is 1+1500/value, so would I need something around 600ohms in R19 and R23?
Head n'hifi have included a selection of very small resistors which dont seem to fit the resistor sockets so I was going to use the standard size ones I have to hand.
Thanks, Will
Member
Joined 2009
Paid Member
I'm a newbie regarding headphone amps but I would reason that if you don't need more than unity gain then a less complex buffer circuit might be imagined ?
I just bought some Grado's and unity gain is enough. I see a lot of headphone amplifiers with gain, but I assume they are for 600 Ohm 'phones that need more than line-level.
Easy: If Gv = 1 + 1k5/Rg, then
Rg = 1k5 / (Gv - 1)
If you know the maximum output level of your source, you can calculate the maximum sensible gain for mains operation as approximately
Gv,max = 6.5 V / Vout,source
(It is important to keep the gain stage out of clipping at all times, since it has to sustain full maximum output voltage regardless of volume setting.)
So with a CD player specified at 2.3 Vrms at 0 dBFS you're advised to stay below about 2.8, and really it doesn't hurt to account for an extra 1-2 dB of intersample-overs. Hence with a "hot" source like that, you're not likely to need more than 2.5x. Anything above that is merely needed for portable sources (about 800 mV max for standard MP3 players, 1.2 V for more upscale ones).
250 ohm Beyers tend to be quite happy with 2 Vrms max as-is, so 1x / 2.5x (Rg = none / 1k) should work fine in such a setting.
Last edited:
I can second that! My 250 ohms DT-770 are very happy with unity gain. I have my batteryless desktop O2 (with ODAC) set with gains of 1x/3x, and I do most of my listening through the odac at gain 1x. With the volume at about 2 o'clock it is plenty loud for most music.
I can also confirm that my other O2 clips at 2,5x with the ODAC when low on batteries.
Many thanks for this and to the other repliers. So that would be exactly as you say: zero resistors in the one slot for 1x, and 1ks in the other for the.. 1.5x gain I make it?
Would it not need to be a 600ohm resistor instead for a 2.5x? I'm sorry for my ignorance - I previously have used an O2 with 1x and 2.5x and I'd like to get it the same if possible. Thanks and I will shut up. Will
Many thanks for this and to the other repliers. So that would be exactly as you say: zero resistors in the one slot for 1x, and 1ks in the other for the.. 1.5x gain I make it?
Hang on I get it now! My over reliance on windows calculator, I was feeding in the equation wrong as (1+1500)/1000. 1k it is for 2.5x, now I can get on and enjoy it!
Hi everyone, DYI beginner here.
I need help finding replacement parts for Q1 and Q2 but I know almost nothing about electronics. I looked at the parts from the BOM but they don't seem to be the right ones anymore (1 is a surface-mounted piece...). Could someone link me to components on Mouser or any other website that ships to the US?
I'd really apprecieate it
Thanks!
Worldinc
I need help finding replacement parts for Q1 and Q2 but I know almost nothing about electronics. I looked at the parts from the BOM but they don't seem to be the right ones anymore (1 is a surface-mounted piece...). Could someone link me to components on Mouser or any other website that ships to the US?
I'd really apprecieate it
Thanks!
Worldinc
Hi everyone, DYI beginner here.
I need help finding replacement parts for Q1 and Q2 but I know almost nothing about electronics. I looked at the parts from the BOM but they don't seem to be the right ones anymore (1 is a surface-mounted piece...). Could someone link me to components on Mouser or any other website that ships to the US?
I'd really apprecieate it
Thanks!
Worldinc
Good find! The Q1 and Q2 part numbers in NwAvGuy's BOM are still correct, it is Mouser's picture on the 512-FQU11P06TU that is wrong.
I've just sent them a message to update their photo. That part comes in both surface mount ("TM" at the end of the part number) and through-hole ("TU" at the end). Mouser put up the wrong picture. Everything else in their part listing is correct though, so you would be safe in ordering those parts as the listings sit.Good find! The Q1 and Q2 part numbers in NwAvGuy's BOM are still correct, it is Mouser's picture on the 512-FQU11P06TU that is wrong.
Thank you for this! I just learned to look at the datasheets instead of the pictures
Learning something every day...Thanks!Hi,
I need a little help here. I built the O2+ODAC unit. When used as an amp, it works great (audio in, USB unplugged). When I use the DAC though, and I max out volume everywhere (PC, volume knob and gain switch on), I can barely hear the sounds. It's whisper-quiet. The ODAC seems to be detected properly in Windows (8.1 x64). What am I missing here? I used P1 to connect the ODAC to the O2. The O2 input jack is grounded to the case.
Thank you
I need a little help here. I built the O2+ODAC unit. When used as an amp, it works great (audio in, USB unplugged). When I use the DAC though, and I max out volume everywhere (PC, volume knob and gain switch on), I can barely hear the sounds. It's whisper-quiet. The ODAC seems to be detected properly in Windows (8.1 x64). What am I missing here? I used P1 to connect the ODAC to the O2. The O2 input jack is grounded to the case.
Thank you
Did you remember to cut the two ground traces under the O2's input jack? If not that will essentially short out the DAC's output. Here is a picture on NwAvGuy's website:
o2%2520odac%2520input%2520jack%2520wiring%255B5%255D.jpg (image)