[raoultrifan]

   
That is not the spec of the IC. That is the spec of the sample circuit fed with +-5V.

That's exactly what I said after in my previous post.
 
Thank you!

 

[HotIce]

To figure out max power, if the spec does not clearly spell it, you need to consider that while the IC is providing Vo to a load Zo, hence emitting a Vo^2/Zo output power on the load, it is dissipating (V-Vo)*Vo/Zo power in the drivers.
Where V is the rail voltage.
Once you have that, once you have the maximum IC silicon temperature, and the silicon to air (unless heatsink is provided) thermal resistance, you can figure out if you can afford to dissipate such power while being within the max temperature range.

 

[raoultrifan]

I believe this can be easily tested with a dummy 32ohms resistor and a IR thermometer or some contact thermometer (or just use carefully the fingers).
 
I'll probably purchase a pair of these soon, because AGDR had did some great tests with 1688 on diyaudio. 

 

[adydula]

I spoke to AGDR and here are his comments"
 
  The OPA1688 can do way more than 50mW into 32ohms (before clipping) with higher chip source voltages, like the +/-12Vdc in the O2 if running on AC.
 
An OPA1688 should be pretty much equivalent to a NJM4556A in terms of power dissipation.  A person might ask how can that be, given the NJM4556A is a DIP-8 package, while the OPA1688 is smaller surface mount SOIC-8 package?  The answer is the dramatically reduced idle (quiescent) current of the newer OPA1688 part.  The Iq of the OPA1688 is just 1.8mA max per amplifier, 3.6mA total, vs a whopping 12mA maximum total for the older NJM4556A.  That means the about half of the NJM4556A’s package power dissipation is taken up just dissipating its own idle current!  At +/-12V (12V – (-12V) * 12mA = 288mW.  So although the OPA1688’s package is smaller, it has a lot more dissipation “room” left for the signal.
 
Alex

 

[adydula]

There is a concern with chip oscillation depending on the capacitance of the load...I dont hear anything but its indeed a possible issue. The O2 circuit and TI's recommendation circuit is different...so putting the 1688 into the o2 is taking a chance with oscillations, but the o'scope traces over at DIY by mlackey look good...but again at what capacitance?
 
If I hear anything more I will add to the thread...
 
Aled

 

[HotIce]

  I spoke to AGDR and here are his comments"
 
  The OPA1688 can do way more than 50mW into 32ohms (before clipping) with higher chip source voltages, like the +/-12Vdc in the O2 if running on AC.
 
An OPA1688 should be pretty much equivalent to a NJM4556A in terms of power dissipation.  A person might ask how can that be, given the NJM4556A is a DIP-8 package, while the OPA1688 is smaller surface mount SOIC-8 package?  The answer is the dramatically reduced idle (quiescent) current of the newer OPA1688 part.  The Iq of the OPA1688 is just 1.8mA max per amplifier, 3.6mA total, vs a whopping 12mA maximum total for the older NJM4556A.  That means the about half of the NJM4556A’s package power dissipation is taken up just dissipating its own idle current!  At +/-12V (12V – (-12V) * 12mA = 288mW.  So although the OPA1688’s package is smaller, it has a lot more dissipation “room” left for the signal.
 
Alex

 
If the extra current of the NJMs is drivers bias current, like 99% probability is, the argument above is not quite correct.
Think about a class A amp, where power dissipated by drivers actually lowers with load (if you keep operation under the class A range).
IOW, the output power and the idle one, do not sum up in the thermal output.
In a very crude form (which is incorrect, among other things, because it does not account rail voltages for PO), if PI is idle power, and PO is output power, dissipated power is more like max(PI, PO), more than PI+PO.

 

[adydula]

This am I soldered a pair of 1nf caps across the outputs with the 1688's in the O2 to see if I could get it to oscillate and it did not. Temps were the same...and I was able to place the stock 4556's back in and compare somewhat. The issue when changing out parts and the time it takes from one setup to another may affect my opinion subjectively....I like to do blind AB testing to leave the bias part out.
 
That said, I thought there was a improvement using the 1688's playing several demo tracks with really well recorded bass it seems the 1688s version is indeed better in my humble mind IMO.
 
I guess I could keep increasing the capacitance to see where it oscillates but for me with 35 ohm and 250 headphones there seems to be no issues that I am concerned with now.
 
Alex

 

[thuNDa]

  The most elegant way would be to change R14 and R20 to something lower than 10K IMHO.
The lower this resistor, the more signal gets shunt to ground, while the resistor(R3/R7, i guess standard value was 100 Ohm) which the signal has to pass, which is responsible for noise, can be left untouched.
The filter characteristics still work too:
 

I must correct myself:
 
It's generally not a good idea to attenuate by decreasing R14 / R20 if someone wants less than effectively 1x gain, because those resistors determine the input impedance of the O2.
In my example you would lower the input impedance from 10K to only 100 Ohm for "0.5x gain".

 

[adydula]

I recently built AGDR's O2 Booster board with 827's and you can remove or short out the output resistors and have "0" ohm output imedance!!
 
Alex

 

[Arkady Duntov]

Does agdr still sell his (her?) pre-assembled ODAs? My amp died this week.

 

[adydula]

Its a "him"....send him a pm directly over at diyaudio.com.
I know he is a bit out of pocket due to some personal issues, but hopes to be back soon, dec? 
 
Oh forgot....agdraudio.com
He has contact info there as well.
 
Alex

 

[chimney189]

Very interesting thread here. I wonder what happened to the designer of this little beast?
 
I have a question to anyone who might know: would the 02 + ODAC be able to power ALPHA PRIMES?

 

[raoultrifan]

Definitely yes! More details about the drivers here: http://www.fostexinternational.com/docs/products/RP-Series.shtml#content-3-tab-tab.
Also, feel free to search on HeadFi or Google some reviews related to Alpha Prime and amplifiers compatibility.

 

[gkanai]

I haven't looked over the whole thread recently but does anyone know if there is a 100-240V power source for the O2? I have the 100V one, but now live in a 220V country, and travel sometimes to the US, where mains is 120V. I'd pay for a larger power source if it could work globally.

 

[dukefx]

Why not get one for Europe and one for the US? The wall sockets don't match anyways.