O2 AMP + ODAC - Page 20

Quote:

Originally Posted by purrin 

Ah I forgot the regulation. Do you what the +/- PS rails are after the regulator? I forgot what the stock op-amp is in the O2. What do the spec-sheets say the op-amp can actually swing with the actual voltage on the rails?

I took the +/- 11.8 V figure from the "O2 Details" (listed under "Popular Posts") page of the designer. The O2 uses 12 V regulators, but the voltage is reduced slightly by the diodes that isolate the two sources of power supply (these are special low voltage diodes that drop less than the typical ~0.6 V), and the MOSFETs in the under-voltage protection circuit. The "Op Amp Measurements" page from August 2011 includes graphs that shows the THD+N vs. voltage performance of the gain stage with a few op amps (the stock one is NJM2068), both on battery and AC power, and at 2.5x and 7x gain. These can be used to find out what the exact clipping level is. Of course, the supply voltage will have a small amount of random variation.

OK. I'm measuring 23.3V peak-to-peak (consistent with +/- 11.65) at the power rails at the chip. (Just thinking out loud now - for my own sake and understanding - correct me if I've made any mistakes.) Lets be very conservative and say the op-amp can output 18V peak-to-peak with those rails. 2.0 Vrms = 2.82 peaks = 5.64 peak-to-peak.

5.64 peak-to-peak * 6.5 gain = 36.66 p-p. (> 18 p-p) Oh boy. Not good.

5.64 peak-to-peak * 2.5 gain = 14.1 p-p. (< 18 p-p) OK. No problem.

Just curious. That explains lot and why I had several friends tell me why their O2s were clipping.

the outputs don't swing full current near the rails - the datasheet shows that they drop ~ 2V from each rail at 150 Ohms load +/-12 Vsupply (the O2 paralleled out could drive ~1/2 the Z but you also have to add the current sharing R drop)

but it isn't unusual to have some more amp gain than that which just reaches the max output V at the fullscale source V but lots more gain does constitute poor system gain structure and you can be wasting potential dynamic range

optimal use of DAC + fixed gain amp requires willingness to use software digital volume control, and knowledge of the target headphone V sensitivity

Wouldn't the load on the NJM2048 be negligible? The input Z for 4556 op-amps in the next stage should be in the k or M Ohm range? At least this was my assumption.

Still, the pot in the middle arrangement is retarded - mainly because the amp's lack of portability has resulted with some people using it as a "desktop amp" fed directly from their PC line outs or desktop CD players - and the 6.5 gain won't work with, among all things, the ODAC (at least without digital volume control). And 2.5/6.5 gain is just barely enough for some headphones, especially since the O2 seems to have plenty of power / current capability to spare.

Quote:

Originally Posted by sling5s 

To fix distortion, lower the pre-amp volume on your computer or source.  Just by a tad.  It works for me when I'm using high gain.  I don't know if it's just my imagination but I too think the LCD-2 sounds better with high gain.

What kind of voodoo is this? We all know subjective impressions are not reliable.

Quote:

Originally Posted by purrin 

 

What kind of voodoo is this? We all know subjective impressions are not reliable.

Yeah...I have very little knowledge when it comes to the science of music.  Sometimes I speak out of ignorance.  Please excuse. :)

But I do love the subjective experience of music  :) 

my numbers were for the output

the NJM2068 looks like it is limited in negative output swing to just past -10 V with +/-12 Vsupply even with negligible load

Quote:

Originally Posted by purrin 

OK. I'm measuring 23.3V peak-to-peak (consistent with +/- 11.65) at the power rails at the chip. (Just thinking out loud now - for my own sake and understanding - correct me if I've made any mistakes.) Lets be very conservative and say the op-amp can output 18V peak-to-peak with those rails. 2.0 Vrms = 2.82 peaks = 5.64 peak-to-peak.

 

5.64 peak-to-peak * 6.5 gain = 36.66 p-p. (> 18 p-p) Oh boy. Not good.

5.64 peak-to-peak * 2.5 gain = 14.1 p-p. (< 18 p-p) OK. No problem.

 

Just curious. That explains lot and why I had several friends tell me why their O2s were clipping.

(for NJM2048)

so looks like the negative side clips first, won't get more than -10V there, so +10V on the top gives you 20V peak-to-peak, around 7.07V rms (sine wave).  Any more, and the gain stage will clip.  18V peak-to-peak was a pretty decent estimate, conservative as you said.  edit: whoops jcx just a couple minutes before I responded

I wonder what's the percentage of users feeding roughly RedBook 2V rms sources, as opposed to portable devices or something that could actually use the 6.5X gain without clipping.  That said, Epiphany configured it as 1X / 2.5X by default, and hopefully any DIY builder knows what's up... it's still a hassle though.

Quote:

Originally Posted by sling5s 

Yeah...I have very little knowledge when it comes to the science of music.  Sometimes I speak out of ignorance.  Please excuse. :)

But I do love the subjective experience of music  :) 

Just a friendly reminder, but sometimes you may need to reboot the sarcasm detector.

Quote:

Originally Posted by sling5s 

Yeah...I have very little knowledge when it comes to the science of music.  Sometimes I speak out of ignorance.  Please excuse. :)

But I do love the subjective experience of music  :) 

I was being sarcastic. Long story if you don't know where the sarcasm comes from.

Quote:

Originally Posted by mikeaj 

That said, Epiphany configured it as 1X / 2.5X by default, and hopefully any DIY builder knows what's up... it's still a hassle though.

You would have figured they all came configured like that or similarly. I just noticed the JDS labs version comes in 2.5x/6.5x by default.

Quote:

Originally Posted by mikeaj 

I wonder what's the percentage of users feeding roughly RedBook 2V rms sources, as opposed to portable devices or something that could actually use the 6.5X gain without clipping.  That said, Epiphany configured it as 1X / 2.5X by default, and hopefully any DIY builder knows what's up... it's still a hassle though.

Keep in mind that you can always attenuate (make smaller) the input signal if 1x gain is even too much to prevent the first O2 stage from clipping.  Just make R3 and R7 bigger so they form voltage dividers with R20 and R14, respectively, such as 5K resistors.  Has the side benefit of increasing the O2 input impedance.

I tried to talk "Voldemort" into including a first-stage clipping indicator circuit last year with no luck (plus no PCB space left on the O2 to add circuitry in the B2-080 sized box).  Without that a person really has no way of knowing when clipping on either rail is occurring, at least until the clipping gets bad enough to be clearly audible.   A lot of faith is put into the user's ability to know exactly what their source voltage level is, then be able to do the math with the O2 first stage gain settings to not exceed the 7V (rms) maximum swing.

Quote:

Originally Posted by BleaK 

 

This is my main rig at the moment. With O2 in low gain. This is personally the best combo I have heard with the HE-500.

 

 

Again, personally I don't get the craze about having alot of power to drive orthoes. I have owned the NFB-10SE which puts out 6W at 50ohm, and I still thought the ODAC/O2 combo sounded better.

 

This my preference of course, and someone or many will surely don't agree with it, which is of course totally ok

What is your low gain setting? 2.5x or 3.0x etc.

Also, how far do you have to turn the volume knob to get it to be too loud?

Thanks.

Quote:

Originally Posted by Smoke Signals 

 

What is your low gain setting? 2.5x or 3.0x etc.

 

Also, how far do you have to turn the volume knob to get it to be too loud?

 

Thanks.

I have the standard 2.5x.

At two/three O'clock my ears can't take it anymore. My normal listenig volume is around 10/11. This is with every volume (windows and foobar) on max.

Great, it seems like the O2 on the standard low gain easily handles them then.  Thanks for the valuable info, probably going to pick up a pair soon now.
 

I hope you guys (who seem a lot more knowledgable than me) could give some advice. I purchased an odac and o2 (separate units) from JDS labs, however between when I ordered and it arrived, I impulse bought an awesome Woo Audio 6 tube amp. So now I am using the odac and wa6 and the o2 is doing nothing, so I want to get a new dac to pair with the o2 for my desktop pc. If I were to get a dac other than the odac, what are some of the best options around the same price range? what about the same performance range?

The odac seems fantastic from my use so far, but I don't want to buy another one unless it is by far the best option as I would like to try new hardware out.