[adydula]

The O2 is not totally perfect, even though I am a ardent user of 2 of them plus an improved ODA of AGDR's design.
 
Recently a person reported frying a NJM4556A chip using a set of HE400's.
 
The cans used were 35 ohms and have a sensitivity of 92.5 db/mW. (HE400's).
 
You can easily go way over the maximum output dissipation for the NJM4556A chips in the upper volume levels of musical peaks. Headphones like that need an ODA type of amp. So it is quite possible to go 100% or more of the maximum allowed package dissipation for certain combinations of low impedance low sensitivity headphones AND higher volume levels AND specific music that has high or sustained peaks.  
 
The maximum power dissipation for the NJM4556A chip's DIP8 package is listed as 700mW on the datasheet here:

http://www.cn.njr.com/PDF/NJM4556A_E.pdf

But that is at room temperature and the insides of the O2 will get warm with those output chips dissipating and the voltage regulator heat. So de-rate that to 600mW.

Then the chip quiescent current is listed as 8mA to 12mA. The higher current end will usually happen at higher output levels. So be conservative and say 11mA, with the O2 on AC power that gives (12 * 2 * 0.011A) = 264mW. Just like that the quiescent current has used up about half the available chip power dissipation just sitting there at idle: 600mW - 264mW = 336mW left for headphones on each NJM4556A chip.

Using 120dB as the maximum level (musical peaks) in the headphone spreadsheet. But at just 112dB the HE400 needs 1.76Vrms and 50mA. So that is 12V - 1.76V across the chip, or 10.24V. The power dissipation for both chips halves then (in the same NJM455A package) is 10.24V * 0.05A = 512mW, which is 35% over the 336mW maximum chip power dissipation from above. At 115dB the power dissipation is 100% over limit at 660mW. That is more than enough to cause the chip to explode in half like that.

Working with the numbers the maximum loudness level to hit the 336mW of available package dissipation would be 107.5dB.

So this person would need to listen to his headphones at lower levels and/or get a higher current amp like the ODA to keep that from happening.
 
The O2 is not safe from internal chip damage for any 16 ohm and up headphone.
 
This recent event occurred with a commercially purchased O2.
 
FYI
 
Alex
 
Note the technical details come from a very highly qualified EE.

 

[Medaud]

The O2 is not totally perfect, even though I am a ardent user of 2 of them plus an improved ODA of AGDR's design.

Recently a person reported frying a NJM4556A chip using a set of HE400's.

The cans used were 35 ohms and have a sensitivity of 92.5 db/mW. (HE400's).

You can easily go way over the maximum output dissipation for the NJM4556A chips in the upper volume levels of musical peaks. Headphones like that need an ODA type of amp. So it is quite possible to go 100% or more of the maximum allowed package dissipation for certain combinations of low impedance low sensitivity headphones AND higher volume levels AND specific music that has high or sustained peaks.  

The maximum power dissipation for the NJM4556A chip's DIP8 package is listed as 700mW on the datasheet here:

[COLOR=0066CC]http://www.cn.njr.com/PDF/NJM4556A_E.pdf[/COLOR]


But that is at room temperature and the insides of the O2 will get warm with those output chips dissipating and the voltage regulator heat. So de-rate that to 600mW.


Then the chip quiescent current is listed as 8mA to 12mA. The higher current end will usually happen at higher output levels. So be conservative and say 11mA, with the O2 on AC power that gives (12 * 2 * 0.011A) = 264mW. Just like that the quiescent current has used up about half the available chip power dissipation just sitting there at idle: 600mW - 264mW = 336mW left for headphones on each NJM4556A chip.


Using 120dB as the maximum level (musical peaks) in the headphone spreadsheet. But at just 112dB the HE400 needs 1.76Vrms and 50mA. So that is 12V - 1.76V across the chip, or 10.24V. The power dissipation for both chips halves then (in the same NJM455A package) is 10.24V * 0.05A = 512mW, which is 35% over the 336mW maximum chip power dissipation from above. At 115dB the power dissipation is 100% over limit at 660mW. That is more than enough to cause the chip to explode in half like that.

Working with the numbers the maximum loudness level to hit the 336mW of available package dissipation would be 107.5dB.


So this person would need to listen to his headphones at lower levels and/or get a higher current amp like the ODA to keep that from happening.

The O2 is not safe from internal chip damage for any 16 ohm and up headphone.

This recent event occurred with a commercially purchased ODA.

FYI

Alex

Note the technical details come from a very highly qualified EE.

Having recently bought an O2/ODAC this is not very comforting...it has handled my 42 ohm ATH-A900x without a problem so far though. The problem with the ODA is that people seem to be taking it in a different direction than what the O2/ODAC was originally all about, e.g. implementing a "bass boost" feature and such.

 

[MrMateoHead]

Anyone that expects at Amp to run at 100% volume sustained for long periods into low impedance loads without frying itself is being a little unreasonable, IMHO. 108 dB is basically hearing damage - totally unnecessary.
 
I wouldn't be TOO worried that the O2 lacks enough thermal dissipation to provide maximum sustained power without damage.
 
I guess the common-sense solution would be to choose more efficient headphones that keep the O2 within a safer power range!

 

[adydula]

The ODA has an option for a bass boost, being a purist I did not implement that. The way the ODA is laid out you can easily leave things out or in as you please.
 
Having three times the current capacity is a plus for me.
 
That said, Mr Mateohead....I agree with your comments that listening to headphones at reasonable levels makes good sense and protects your hearing..... but knowing eactly your amp and headphones are doing is a plus as well.
 
The ODA designer has had many conversations with the O2 designer in the past. This was one topic they debated.
 
Also the ODA designer has tried and has met most of the O2 measurements and capabilities take a good read over at DIYAudio.com.
 
All the best
Alex

 

[potkettleblack]

Hi lads,
 
anyone know how to get the silver odac + o2 with the custom options? it seems it will only let you customise the black one

 

[jring]

...But at just 112dB the HE400 needs 1.76Vrms and 50mA. So that is 12V - 1.76V across the chip, or 10.24V. The power dissipation for both chips halves then (in the same NJM455A package) is 10.24V * 0.05A = 512mW, which is 35% over the 336mW maximum chip power dissipation from above...

 
Hi,
 
care to explain this part, specifically why you subtract the 1.76Vrms from 12V to multiply by 50 mA? I would expect 1.76Vrms times 50mA or 0.09VA (this is apparent power usually expressed in VA instead of W). Probably take this times 2 for stereo.
 
Also the quiescent current can be taken from the supply current vs temperature diagram in the data sheet and is 9mA at Vcc=+-15V and room temperature and will fall with rising temperature.
 
Joachim

 

[jring]

My O2 has stereo RCA outputs. The lower gain is 1x (unity) can I output the signal from my O2 to my Marantz integrated stereo amp?

 
I do this all the time since I had not drill until lately and so couldn't populate the ODAC 1/8" Line-Out socket yet. Of course you are double amping here but I have not heard a problem.
 
Joachim

 

[Solrighal]

I do this all the time since I had not drill until lately and so couldn't populate the ODAC 1/8" Line-Out socket yet. Of course you are double amping here but I have not heard a problem.

Joachim

Thanks for the input Joachim. I have it connected now & it's sounding fantastic. Is it double-amping though. I'm using unity gain with the volume set to max. Surely that's just the same signal my ODAC would be putting out?

 

[potkettleblack]

Thanks mate.
 
I see you commenting on ALL the hd800 youtube videos

 

[adydula]

Joachim, here is your answer from AGDR:
 
The 1.6Vrms * 50mA is the power that gets delivered to the load (the headphones) and not the power dissipated by the output chips.   The voltage that winds up across the output chips is the power supply voltage going into the chip minus the headphone voltage at the chip output, 12Vdc - 1.6Vrms.  That is wiith the O2 on AC, things are better on batteries with the lower voltages.

That is the voltage that winds up across the output transistor in one half of the push-pull output stage in the chip, whichever of the two is "on" at the moment.  The power dissipated by that transistor then is that voltage difference across it, 12Vdc - 1.6Vdc, times the current through the headphones (which has to go through that output transistor to get there from the chip power supply lead), for a total of (12Vdc - 1.6Vrms) * 50mA = 52mW for each amplifier, or 2x that for both halves of the chip 104mW.

But here is the tricky part.   That is just one output transistor.  Using rms voltages takes care of the other automatically.  rms voltages are always the thing to use in power calculations since it is essentially the "heating" value of a sine wave.  The rms voltage for a pure sine wave is 0.707 times the peak votlage, say just the positive half, and takes the negative wave half into account automatically in the math.  In terms of the chip's internal power dissipation that just means the positive push-pull transistor and the negative alternately get the 50mA going through them to the positive and negative power supply rails, respectively, as the sine wave does its thing.

He is quite right about the total quiescent current though!  Good find.  I took the 11mA off the table in the datasheet and didn't notice that graph. So the actual quiescent current at the typical O2 inside temp, say 50C maximum, would be around 8.5mA from that graph, 2.5mA less that what I used.  So the new chip power dissipation due to the quiescent current would be slightly less:  (24 * 0.0085A) = 200mW. But at the higher voltage swings (higher volume levels) the output chip(s) would still be nearly 100% over their maximum power dissipation and would pop.

So to sum it up, the designer's blog and posts leave the impression that any headphone from 16 ohms and up are "safe" with the O2, in terms of not causing internal damage to the chips. But that just isn't the case.  There are combinations of some low-sensitivity low-impedance headphones + higher volume levels + music with higher peaks and/or sustained peaks that will overheat the O2's output chips.  Luckily getting all of those things together at once is relatively rare, which is why you don't see more O2's blowing their output chips.   If anyone does have their output chips pop like that be sure to replace them as a set.  If one cracked open the other was nearly there too and is damaged.
***************
 
Hope this helps...
 
Alex
 

 

[mikeaj]

I'm pretty surprised anybody could stand to listen that loudly, to be honest. For HE-6, I'd believe it. For HE-400? That seems nuts to me. Pretty much if some music is pegged at an average level of -10 dBFS or higher, nobody's listening on 110+ dB SPL peaks. At least, I hope. The time when those levels get actually used may be for more brief peaks on music with some 25+ dB dynamic range.
 
I mean, how many people have even fried their O2s with test signals?
 
But yeah, using planar magnetics, I'd maybe stick chip heatsinks on just out of principle, even without trying to calculate how much that'd help with thermal dissipation:
http://www.ebay.com/itm/New-High-Quality-5pcs-11x11x5mm-adhesive-Aluminum-Heat-Sink-For-Memory-Chip-IC-/131014795047
 
 

I have it connected now & it's sounding fantastic. Is it double-amping though. I'm using unity gain with the volume set to max. Surely that's just the same signal my ODAC would be putting out?

Yes, double amping, while outputting the same level and behaving practically the same, yes. I'm sure whatever noise level the O2 adds is insignificant compared to everything else.

 

[Solrighal]

As long as it doesn't fry anything it's cool I suppose. I've no idea why the RCA outputs on my O2 are connected to the amp though. Unless it's maybe for active speakers.

 

[mcandmar]

Thanks for the input Joachim. I have it connected now & it's sounding fantastic. Is it double-amping though. I'm using unity gain with the volume set to max. Surely that's just the same signal my ODAC would be putting out?

 
Depending on how it is implemented you should be able to tap into the ODAC output using the line in socket on the front of the amp, i.e. bypassing the O2 amplifier.

 

[Solrighal]

I'm thinking about buying a pair of these for the ODAC...

http://www.futureshop.co.uk/product_info.php?products_id=6183#.VDK4fIFwaBY

What do you guys think? It will avoid Double-amping and keep the front of my O2 tidy.

 

[jring]

  Joachim, here is your answer from AGDR:

 
Thanks for the explanation... I would do the math for a Class B output stage like this... Total power consumption is power supply voltage times current.
 
Pps = Vps * Ips
 
assuming a sine wave we get
 
Pps = Vps * 0.45 * Vrms / Rl
 
so with Vps=24V (both rails), Vrms=1.76V and Rl=35 ohms we get Pps=540mW
 
For the power dissipation in the amp we have to subtract the power delivered to the phones from this:
 
Pdiss = Pps - Pout
 
Pout = Vrms * Iout = 88mW
 
Pdiss = 452mW
 
Now we add quiescent current times supply voltage Vps * Iq =  204mW giving 656mW which is a close call with the 700mW max dissipation and in a cabinet but it should not go up in flames. I first thought we had to multiply Pdiss by two for two amps in the package but since they're switched in parallel they shrare the load and each sees only half of the current.
 
Joachim