TPA6120 Power Requirements

[dfdye]

I am CERTAIN this has been covered before, but I cannot for the life of me find the info I am looking for! I am putting together a TPA6120 based headphone amp for a set of Audio-Technica ATH-A900 Art headphones I just picked up. Yes, I know they work fine without amplification, but I wanted something a little beefier in line to drive them.

Anyway, I have a couple of really nice +/-15V 100mA PSU's laying around that I was planning on using with the chips, but the data sheet indicated that at full load they put out 700 mA. Of course this will be heavily dependent on music and won't come anywhere near 700 mA continuous, AND caps will easily handle quick transients, but despite this it got me wondering if 100mA would be enough juice to power the chips.

Is there a concensus on how large a PSU you need to get these chips to behave? Should I double up and give 100mA per channel? (I do have 2 of these PSU's, and I would be more than happy to throw both into this project) Or should I throw in the towel and build a 2A +/- 12V PSU just to be on the safe side? (I know Bob Ellis's PSU from the group buy would be perfect for this)

Thanks for the input!
David

[lineup]

hi

worst case, if your headphones are 32 Ohm
you will do well with 100mA

for higher impedance phones
you can do with less currents ( 300 Ohm ~ 30-50 mA MAX )

Recommended supply is +-12 volt, but using +-15 volt DC is not bad, more like better!

see a little info in my own class A discrete Headphone amplifier
this post:
http://www.diyaudio.com/forums/showt...900#post999900

regards to any heradphone listener
from lineup

I attach the basic electrical data for TPA6120A2, from the datasheet:

[jcx]

you should learn how to do the math

amps require some quiescent current – even when not hooked up to a load

beyond that they draw only as much additional current as need to drive the load impedance to the output V determined by their gain times the input signal V

headphones have impedances all over the map as well as hugely differing sensitivity

for the ATH-900 google gives 40 Ohms and 101 dB/mW - the last # could be suspect: sometimes headphone sensitivity is dB/V but mistyped in ads - the 2000 mW power rating makes this error more likely

120 dB is a nice target - for peak dynamic headroom - a average loudness of even 100 dB would lead to hearing loss

100 mWrms power (+20 dBre 1mW) into 40 Ohms is ~70mA, 2.8V peak

on average only 1/2 the current comes from each supply polarity so 100 mA average rating is OK if you have enough C to hold up for the 140mA from both channels peaking in phase at low frequencies - and you have to allow for the TPA quiescent current

with +/-15 V you have quite a bit of supply headroom, so as long as the supply doesn't hard current limit @100mA you could tolerate quite a bit of V droop

on the upside however you don't have much margin with +/-16 max supply V allowed for the TPA6120 - you need to be sure you're not using a unreg supply that could easily be 10-20% high at light loads

regulated supplies can't be just wired together you'd need some current sharing resistors -or you go dual mono which would require 2 separate TPAs

if, as seems possible, the sensitivity is really dB/V then you need lots more current - 350mA peak per channel

[dfdye]

Originally posted by lineup
hi

worst case, if your headphones are 32 Ohm
you will do well with 100mA

Well, my back of the envelope math said so too, but it is good to have a second opinion to make sure that I am not on crack. BTW, these cans are ~40 ohm.

for higher impedance phones
you can do with less currents ( 300 Ohm ~ 30-50 mA MAX )

Of course I am trying to design for as wide a range of headphones as possible in case I ever get some of the highe impedence Senheisers or something like that, so this is always good to know.

Recommended supply is +-12 volt, but using +-15 volt DC is not bad, more like better!

Right, and the +/-15 is within the spec of the part, I have the supplies laying around, so why not??

see a little info in my own class A discrete Headphone amplifier
this post:
http://www.diyaudio.com/forums/showt...900#post999900

regards to any heradphone listener
from lineup

Thanks

I attach the basic electrical data for TPA6120A2, from the datasheet:

Of course the first thing I did was read the datasheet, but thanks for the backup in case I was an idiot.

I'll be sure to let everyone know how it turns out.

David

[dfdye]

Originally posted by jcx
you should learn how to do the math

Probably! Actually, the thing that threw me was the datasheet saying it needed 700mA which didn't make any sense to me in light of the load it would be driving. My back of the envelope math said 100mA should be fine. Thanks for the example problem though

with +/-15 V you have quite a bit of supply headroom, so as long as the supply doesn't hard current limit @100mA you could tolerate quite a bit of V droop

on the upside however you don't have much margin with +/-16 max supply V allowed for the TPA6120 - you need to be sure you're not using a unreg supply that could easily be 10-20% high at light loads

These supplies are bullet proof and are extremely well regulated as far as voltages go. I am not sure about current since I have never tried to stress them. The supplies were pulled from some hard core instrumentation that had rather high noise rejection and stability requirements, so I am guessing they will be fine for audio work.

regulated supplies can't be just wired together you'd need some current sharing resistors -or you go dual mono which would require 2 separate TPAs

AH! But here is where the beauty of the TPA comes into play! The chip has inputs for separate power supplies to cut down on crosstalk between channels. As such, as long as the voltages of the two PSU's are the same, I can power each channel off its own supply.

if, as seems possible, the sensitivity is really dB/V then you need lots more current - 350mA peak per channel

I'll double check that number in the packaging (though most of it is in Japanese, which doesn't do me much good), but if indeed it is, I think I can use the ~1A PSU I have and call it a day, but that supply unfortunately doesn't fit nicely into the enclosure I had all ready to go. I guess that means I'll have to suck it up and build another project to go in that enclosure!

Thanks for the feedback!
David

[jcx]

TPA supply pins are not exactly "independent"

V- are shorted internally and "A" V+ must always be on, with power to the "B " side V+ being optional

combining these constraints with ps gnds connected together and with 2 ps you still want current sharing resistors on V-, probably anti-parallel diodes across the 2 ps V+

you might want to ask on http://www.head-fi.org/ or http://www.headwize.com/ about the sensitivity # for the ATH-A900, the 2000mW max power rating only makes sense if sensitivity is really in dB/V - and then you should want more current and need to set your amp up for higher V gain

[dfdye]

Originally posted by jcx
TPA supply pins are not exactly "independent"

V- are shorted internally . . .

That's good to know, and the data sheet doesn't exactly make that clear. Thanks for the heads up.

BTW, these headphones are rather efficient, so I am going out on a limb and saying that the power requirements will be lower rather than higher. I still haven't looked at the box or manual, but I think I have the info I need to get the power requirements squared away.

Thanks again to all who commented!
David

[dfdye]

Originally posted by jcx
for the ATH-900 google gives 40 Ohms and 101 dB/mW - the last # could be suspect: sometimes headphone sensitivity is dB/V but mistyped in ads - the 2000 mW power rating makes this error more likely

I wanted to add that I have the correct number, and it is indeed 101 dB/mW. The HUGE power handling is due to the extremely high end build quality of the drivers in the unit. This makes me very happy since there is little chance of overdriving them!

So, for the time being I am going to run with a single 100mA PSU and see how well that works.

[peranders]

As for power supply, my 2 x 2.3 VA is very well working - for headphones that is. You can find more inspiration here. The power to the headphones is more than enough.