You need separate output buffers for each output. As long as the current is enough to run any set of phones you are fine.
Ask your self, what happens when someone plugs in two sets of 32-Ohm cans? They will each need separate volume controls as well. Because, what happens when someone plugs in a set of 600-Ohm phones and then a set of 60-Ohm?
Gain stage needs to be rail to rail. Trust me. You don't want anything normal to be able to be able to clip the gain stage. You can attenuate before the output buffer with a shunt. Which is exactly what a pot is. You will probably have to use fixed resistors as well for additional attenuation control. Or you can use a 3 position switch where the values are closely matched. Then there are no problems with someone not having enough pot travel and someone having too little.
Ask your self, what happens when someone plugs in two sets of 32-Ohm cans? They will each need separate volume controls as well. Because, what happens when someone plugs in a set of 600-Ohm phones and then a set of 60-Ohm?
Gain stage needs to be rail to rail. Trust me. You don't want anything normal to be able to be able to clip the gain stage. You can attenuate before the output buffer with a shunt. Which is exactly what a pot is. You will probably have to use fixed resistors as well for additional attenuation control. Or you can use a 3 position switch where the values are closely matched. Then there are no problems with someone not having enough pot travel and someone having too little.
So you are saying that any desing based on the NJM4558 or the NE5532 would work just fine?
I also need to put a connection for a talk-back mic in the circuit.
I want to use the whole device after a small M-Audio digital console witch has no TB-mic in and convert it from one headphone out into six outputs.
I was thinking of taking the balanced receiver designed by ESP and go right into a NE5532 based circuit before each buffer opamp with a momentary switch that when activated converts the stereo out into mono and when released puts it back to stereo.
I also need to put a connection for a talk-back mic in the circuit.
I want to use the whole device after a small M-Audio digital console witch has no TB-mic in and convert it from one headphone out into six outputs.
I was thinking of taking the balanced receiver designed by ESP and go right into a NE5532 based circuit before each buffer opamp with a momentary switch that when activated converts the stereo out into mono and when released puts it back to stereo.
Sure it does but it says it with a graph. It's called maximum output Voltage vs load resistance. The 4558 won't get any prizes for output current. With dual 15 Volt supplies it will do 28V into 10K, 25V into 1K, 22V into 500 but it's falling off a cliff now. 16V into 300 Ohms and the graph stops at 12V into 220 Ohms. Not a good choice for driving low loads. It's also slower than mud at 1 V/uS slew rate. It _would_ make a fine 8 pin thumbtack though.
The graph is on page 3 of the JRC NJM4558 data sheet.
G²
Last edited:
Not the lowest distortion but certainly easy to get, inexpensive and obviously capable of the headphone load. You can probably use your 4558 as a mic preamp. For intercom use (that's what it is?) the limited performance of the 4558 should be fine. With 2 stages, 25-35dB of gain shouldn't be a problem for a dynamic mic but it will depend on what mic you're actually using as to how much gain you'll require. How long a cable are you thinking of? Will it have multiple stations? Will you want it to interface with any other manufacturers units?
G²
NJM4556AL (SIP-8 version) is 67 cents each at mouser. Put a whole raft of them in your headphone distribution amplifier. Just make sure you stick a few uf of capacitance near the power pins of each one.
Did i mention they are unity-gain stable? Use one for gain and feed 10 of them for less than $10. It won't even be hard to wire. You could do it on stripboard. You could do 10 pair out of a mint tin.
if you already have the 4558, well, it will power one pair of headphones ok. most of the time.
Edit: fwiw NJM2068 is near the same power output as the 4556 with lower noise. for the same 67 cents.
Did i mention they are unity-gain stable? Use one for gain and feed 10 of them for less than $10. It won't even be hard to wire. You could do it on stripboard. You could do 10 pair out of a mint tin.
if you already have the 4558, well, it will power one pair of headphones ok. most of the time.
Edit: fwiw NJM2068 is near the same power output as the 4556 with lower noise. for the same 67 cents.
Last edited:
I don't know where you got that from. Even according to the datasheets, the 4556 delivers nearly twice the output current and seems to have much lower output resistance.
Otherwise: agreed.
Does not look much like a 4556 to me based on the datasheet.
I'm still confused about your project - are you proposing to drive several headphones from a single chip or are you open to using several chips?
Also confused about why you are fretting over jellybeans. The njm4556dd is less than a dollar pretty much anywhere.
What are your project goals and constraints?
(also, wrt the 2086 confusion - I must have been looking at the dissipation column, sorry.)
I'm still confused about your project - are you proposing to drive several headphones from a single chip or are you open to using several chips?
Also confused about why you are fretting over jellybeans. The njm4556dd is less than a dollar pretty much anywhere.
What are your project goals and constraints?
(also, wrt the 2086 confusion - I must have been looking at the dissipation column, sorry.)
You must have gotten something wrong. I never said I want to use one chip to drive several headphones. I am and always have been open to using several chips. It is - AFAIK - the only "right" way.
By the way, I made a multiple headphone amp (1x in - 6x outputs) with 12x LM386's and swapped them for 6x NE5532's: these ones may have a little less volume but they sure produce much less distortion and sound much cleaner.
Sooo, are you saying that the NTE788S is not the same as the 4556? On the datasheet it says that "it is also suitable as a headphone amplifier because of its high load current"
By the way, I made a multiple headphone amp (1x in - 6x outputs) with 12x LM386's and swapped them for 6x NE5532's: these ones may have a little less volume but they sure produce much less distortion and sound much cleaner.
Sooo, are you saying that the NTE788S is not the same as the 4556? On the datasheet it says that "it is also suitable as a headphone amplifier because of its high load current"
The NTE788S has more than twice the slew rate, 2 volts less maximum operating voltage, and drives 50mA vs. the 4556's 70mA. They are not remotely the same chip. There are a lot of chips it could be, and i'm not really interested in finding out.
It would certainly work. If you're getting them for free, go for it.
Beefer options include the opa551, or a single opamp driving a raft of buf634, but those are expensive options.
There are numerous TO220-5 power opamps out there, but headphones don't need much power.
You can also drive a power BJT or darlington within the feedback loop of an opamp. Like so:
Above is the "Tori" headphone amplifier, as designed by he who calls himself PRR on various fora. A quick google search will reveal discussion of this version, and the power mosfet version. You could reduce the power requirements by reducing the bias load on the transistors.
It shows a TIP41 but you can use any npn power BJT used in audio - bd139, mje243, etc. Also darlington parts like TIP120. This amp will drive small speakers if it has to. If you plan to use low-Z headphones you may want a 10ohm or so resistor padding the output.
Most people who built this amp used an lm317 configured as a current regulator for the current sink. at 200ma it will need a heatsink, and the output resistor should be maybe 1W or bigger. 200ma out of an lm317t = 6.2ohms, fwiw. 100ma would be 12.4, etc.
It would certainly work. If you're getting them for free, go for it.
Beefer options include the opa551, or a single opamp driving a raft of buf634, but those are expensive options.
There are numerous TO220-5 power opamps out there, but headphones don't need much power.
You can also drive a power BJT or darlington within the feedback loop of an opamp. Like so:
An externally hosted image should be here but it was not working when we last tested it.
Above is the "Tori" headphone amplifier, as designed by he who calls himself PRR on various fora. A quick google search will reveal discussion of this version, and the power mosfet version. You could reduce the power requirements by reducing the bias load on the transistors.
It shows a TIP41 but you can use any npn power BJT used in audio - bd139, mje243, etc. Also darlington parts like TIP120. This amp will drive small speakers if it has to. If you plan to use low-Z headphones you may want a 10ohm or so resistor padding the output.
Most people who built this amp used an lm317 configured as a current regulator for the current sink. at 200ma it will need a heatsink, and the output resistor should be maybe 1W or bigger. 200ma out of an lm317t = 6.2ohms, fwiw. 100ma would be 12.4, etc.
Last edited:
Copy the circuit in RocketScientist's O2 and make multiple sections, one for each output. I don't think you'd be infringing his intellectual property rights, it's hardly a revolutionary design, besides, as he says much is in the implementation (layout). Drive them all from a single unity gain buffer. You could include a master volume.
Make a good job of it, e.g keep the power rail tracks close together and follow a good star grounding scheme, use plenty of bulk capacitance suitably distributed with 0u1 ceramic caps at each chip power pin, and you'll end up with a result that won't leave much room for complaint, and it'll be about as cheap as it can be done.
Make a good job of it, e.g keep the power rail tracks close together and follow a good star grounding scheme, use plenty of bulk capacitance suitably distributed with 0u1 ceramic caps at each chip power pin, and you'll end up with a result that won't leave much room for complaint, and it'll be about as cheap as it can be done.
I should add that i have used the 4556 in a headphone amp and i find it "punchy" and totally adequate but "unrefined". Grado uses it in their outrageously expensive RA-1 headphone amplifier, but i don't care for their headphones.
I like the 4580 sound signature better but it doesn't have the current ability, needs buffered. for neutral on a budget you want njm2068 and it will drive most headphones 60ohms on up.
really i prefer chips like the opa2107 and lm4562 aka lme49720 (or lme49710 for single-channel) but they don't have much in the way of power output, though the 4562 is good for oldschool 600ohm studio monitor cans.
I like the 4580 sound signature better but it doesn't have the current ability, needs buffered. for neutral on a budget you want njm2068 and it will drive most headphones 60ohms on up.
really i prefer chips like the opa2107 and lm4562 aka lme49720 (or lme49710 for single-channel) but they don't have much in the way of power output, though the 4562 is good for oldschool 600ohm studio monitor cans.
Jrc4558/2068 or NE5532 is useful for line driver applications or unity to some moderate gain (say max 10X gain like in h/p amp gain stage)application. Jrc4556 is unity gain stable & can be used to drive h/p directly(lowest upto 16 ohms). You can use Jrc4558/2068 for gain stage - "x" no. of 10K log pot for vol. followed by "x" no. 1uf polyester/polypropelene series cap. to block out any DC & "x" no. Jrc4556 in unity gain mode to drive "x" no. h/p.
http://nwavguy.blogspot.in/2011/08/op-amps-myths-facts.html
http://nwavguy.blogspot.in/2011/08/op-amps-myths-facts.html
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.