Minimal Headphones Amp

[k brown]

I've been quite pleased with the HP amps in my Scarlett and Saffire interfaces - any idea what circuit/chip(s) they use?

But then I may have easy-to-drive phones (60 ohm, I believe).

 

[ccaudle]

for the protection of the op amps and for the protection of peoples ears a build out resistor can be a good thing.

I guess it depends on the fidelity you need or expect out of your headphone amplifier. Would you make the same argument for putting a 20 or 30 Ohm resistor in the output of a power amplifier driving speakers? That does not seem to be much in fashion these days.

 

[JohnRoberts]

I guess it depends on the fidelity you need or expect out of your headphone amplifier. Would you make the same argument for putting a 20 or 30 Ohm resistor in the output of a power amplifier driving speakers? That does not seem to be much in fashion these days.

I feel like I am repeating myself for perhaps the third time in this thread alone. Indeed the build out resistor is a compromise forgoing some fidelity for the constraints of working inside a master section with "all" possible headphones.

I made a headphone amp that happily drove loudspeakers (HB-1), but it was a stand alone accessory product. I never built anything like that inside a console/mixer master section, for the obvious reasons already expanded upon.

JR

 

[Newmarket]

I guess it depends on the fidelity you need or expect out of your headphone amplifier. Would you make the same argument for putting a 20 or 30 Ohm resistor in the output of a power amplifier driving speakers? That does not seem to be much in fashion these days.

Power amplifier circuits also need to account for load reactance and it's effect on stability, short circuit conditions etc.
Additionally there are clearly hearing safety concerns that should be of paramount importance with headphones.

 

[ccaudle]

Power amplifier circuits also need to account for load reactance and it's effect on stability

Is that not also a concern for a small amplifier driving a small reactive load?

 

[JohnRoberts]

Is that not also a concern for a small amplifier driving a small reactive load?

yes it is... the biggest (only) difference is that headphone amps are modest power devices, typically only hundreds of mW output power....
===
In my HB-1 I had to current limit the power supply because it would open the thermal fuse inside the 1A wallwart, from current draw when playing speakers too loud for too long.

JR

 

[Newmarket]

Is that not also a concern for a small amplifier driving a small reactive load?

Yes of course. But the choice of method of achieving that may differ due to the different load and power requirements.
Especially given the extremes of impedance that can be seen with headphones compared to "normal" loudspeakers (excluding 70V/100V line speakers that are transformer driven).
But above all hearing safety is paramount.
fwiw a Malcolm Toft anecdote about the Beatles comes to mind ...

 

[Tubetec]

I think the best mixer Hp amp I ever heard was in my Amek BC3 ,
not sure exactly what it is ,maybe Matt Syson does .

A badly clipping HP amp ,I suspect does much more hearing damage than one with the headroom to carry the program material without distortion ,even if the levels are louder ,
Theres absolutely no doubt a headphone amp thats running out of steam and clipping even short term transients will cause massive ear fatigue , , the subjective nastiness of an amp distorting correlates closely with the amount of higher order distortions it produces , the unpleasentness tells you its doing damage ,its a physiological reaction .

I get the point that a Hp amp running off the same rails as other audio circuits cant be so power hungry as to cause unwanted interactions ,
But isnt it the case in Ians mixers that all the other audio circuits are tube driven , so that isnt likely to be an issue .

I read Thorz recent point about his IFI ZEn HP amp in another thread , its supplied with a 2A supply , but replacing that with a high quality 5A supply makes the whole thing perform better , even if the extra headroom isnt needed ,it still improves the performance markedly .

Agreed the safety of your customers must take high priority ,
but in very loud rock music your not going to get the take if the drummer cant hear things properly . If you have hits coming off the snare that are as loud as a gun shot your lucky bag IC headphone amp is useless , I cant see any safe way of doing it , but if he needs more level ,he needs more level ,if you cant give it to him it looks like your not doing your job properly ,
so your going to pull out your SPL meter and give the talent a lecture on hearing damage

 

[Migs 31]

so your going to pull out your SPL meter and give the talent a lecture on hearing damage

Sorry, what did you say, can't hear you?
(I'm a sound engineer turned drummer or vice versa)


M

 

[ruffrecords]

If you have the space available (ca. 15 x 15 cm pcb) it might be an easy fix, it wasn't expensive. If you are interested I can try to find it for you.

Cheers,

Michael

Yes please,

Cheers

Ian

 

[gswan]

I am working on a very simple headphones amp for a mixer. The attached pdf shows my thinking so far. There is a level control before the input in each case and appropriate coupling and decoupling components are not shown

I started with a standard hi-fi type phones amp using a pair of 5532 op amps shown in the top figure. The two op amps work in parallel with 47 ohm build out/ current sharing resistors. The NFB is taken from the output so the output Z is very low and the gain is about 10dB.

The next version is one I found in the Mackie 1202 VLZ mixer sown in the middle figure. The negative feedback is now taken from the output of the first op amp and the build out resistors have been increased to 120 ohms. In this configuration the output impedance is essentially the two 120 ohm resistor in parallel which presumably gives it some advantage when working with a wide range of headphone impedances. The gain is still about 10dB.

I only need unity gain as my tube mixer internal level is +4dBu. So the third and I hope final version simply removes the two gain setting resistors and turns the first op amp into another unity gain follower.

I have built and tested the first version and as expected it has too much gain. I plan to modify it to match the other two and retest.

Comments?

Cheers

Ian

Not really a fan of 5532's in this configuration. They really don't like driving much below 1K.
I've a page on my site with a couple of headphone amp designs we use in HP distribution boxes around the studio. The output drivers are the JRC NJM4556 device, which I've found to be far superior in driving headphones of all types.
https://axtsystems.com/axtsystems/proj_hpm.php
- Geoff

 

[JohnRoberts]

Sorry, what did you say, can't hear you?
(I'm a sound engineer turned drummer or vice versa)


M

Drummers are notorious for having hearing damage.... crash cymbals can be loud at HF.

Dudley (RIP) the great drummer who inspired my HB-1 design with his too loud for school headphone mix, surely experienced some hearing damage .

JR

 

[ruffrecords]

Not really a fan of 5532's in this configuration. They really don't like driving much below 1K.
I've a page on my site with a couple of headphone amp designs we use in HP distribution boxes around the studio. The output drivers are the JRC NJM4556 device, which I've found to be far superior in driving headphones of all types.
https://axtsystems.com/axtsystems/proj_hpm.php
- Geoff

Thanks for the tip. I was not aware of that device. It certainly has some serious drive capability. Fortunately I can just plug them into my existing circuit. I think this is the one Mackie used in their headphones amp. I (wrongly) assumed is was just a cheap version of the 5532.

Cheers

Ian

 

[ruffrecords]

I guess it depends on the fidelity you need or expect out of your headphone amplifier. Would you make the same argument for putting a 20 or 30 Ohm resistor in the output of a power amplifier driving speakers? That does not seem to be much in fashion these days.

No but the inverse square law is the saviour for your ears with loudspeakers. But I agree it is a relevant concern. For some reason build out resistors are very common in headphone designs.

Cheers

Ian

 

[ruffrecords]

I quite like the NJM4556AD as a substitute for the NE5532 in my current design but they seem to be on at least a 16 weeks lead time. Ebay and Amazon sources look distinctly dodgy.Anyone have four to spare I could buy?

Cheers

Ian

 

[Newmarket]

I quite like the NJM4556AD as a substitute for the NE5532 in my current design but they seem to be on at least a 16 weeks lead time. Ebay and Amazon sources look distinctly dodgy.Anyone have four to spare I could buy?

Cheers

Ian

Mouser seem to have a couple of thousand in stock

https://www.mouser.co.uk/ProductDetail/Nisshinbo/NJM4556AD?qs=Dx6nUK05NTELd0aHqFd4%2Bw==

 

[ruffrecords]

Mouser seem to have a couple of thousand in stock

https://www.mouser.co.uk/ProductDetail/Nisshinbo/NJM4556AD?qs=Dx6nUK05NTELd0aHqFd4%2Bw==

Well done. I just saw the 22 weeks lead time but missed what they currently have in stock. If I buy 10 from them it will cost £8 plus £12 shipping. A bit expensive but at least they will be genuine.

Cheers

Ian

 

[Newmarket]

Well done. I just saw the 22 weeks lead time but missed what they currently have in stock. If I buy 10 from them it will cost £8 plus £12 shipping. A bit expensive but at least they will be genuine.

Cheers

Ian

Digikey have them in stock if that works out better although without checking their shipping costs I'll guess it's similar ?

https://www.digikey.co.uk/en/products/detail/nisshinbo-micro-devices-inc/NJM4556AD/673767

 

[thor.zmt]

I am working on a very simple headphones amp for a mixer. The attached pdf shows my thinking so far.

You need to be clear what kind of headphone you want to drive.

To give a few extreme examples.

AKG K1000 - 86dB/1V @ 120 Ohm
AKG K240 Monitor - 90dB/1V @ 600 Ohm
Sennheiser HD600 - 105dB/1V @ 300 Ohm
Beyerdynamic DT-770 Pro/32 - 111dB/1V @ 32 Ohm
Sony MDR-7506 - 118dB/1V @ 63 Ohm

As you can see you have a relative SPL difference at 1V out of 32dB.

And you have impedance levels of as low as 16 Ohm or even lower and as high as 600 Ohm.

So for a "universal" HP amp you should be able to drive at the very least full output into 32 Ohm. Also, most modern lower impedance headphones need low (< 10 Ohm) drive impedance.

I would suggest to use a inverting mode Volume control (using a Fet Input Op-Amp may be desirable over 5534) to give a wide range of adjustment of gain, possibly with a switchable max gain.

See here:

Headphone Amplifiers

Further, many headphones need EQ and often some form of crossfeed is desirable. You can use DAW plugins or use suitable analogue circuitry, I tend towards the latter.

Parallel Op-Amp's have serious current limits.

You could try the so-called "O2" Design, with NJM4556, it may be "adequate" from an objectivist viewpoint (it does not sound particularly transparent):

NwAvGuy

The design could be extended with (say) dual NJM4556 as followers with 3R9 buildout resistors and an added "main" Op-Amp using for example OPA1656 which drives the NJM4556 Inputs and the HP output via a 10 Ohm resistor in a variation on the Quad "Current Dumper".


In this case the OPA1656 delivers 30mA with each of the 4 NJM4556 sections delivering 70mA peak for 0.3A Peak and full output of ~ 16V Peak / 11V RMS into 60 Ohm (less into lower impedances).

Noise at unity gain and volume full up would be around 1.5uV and thus ~ 137dB dynamic range re 11V Output. Distortion is mostly determined by the OPA1656 which is very low, with a 32 Ohm load as shown the load "seen" by the OPA1656 is around 328 Ohm.

Personally I would suggest using a discrete output stage using BD139/140 following the topology of the Neve BA640 Line Amp design with 1 Ohm emitter resistors running on 12V...18V rails.



Don't go with the "diamond transistor" output, based on my experience.

With ~ 180mA Iq you get Class A for 12V peak into 32 Ohm. With 45mA Iq you get Class A for 2.8V Peak into 32 Ohm. It is recommended to keep normal operating levels in Class A.

For Drummers the standard 12V...18V rails may be insufficient, you can move up to TO-220 output devices and run the rails up to +/- 20V and make a balanced Amp (with ~ 80V P-P output) and either use a 4-Pin XLR with a balanced drive, or stick to 6.3mm Jack and build a M/S matrix headamp with M driving the "ground" and L-R / R-L driving the Tip & Ring.

Thor

 

[thor.zmt]

I read Thorz recent point about his IFI ZEn HP amp in another thread , its supplied with a 2A supply , but replacing that with a high quality 5A supply makes the whole thing perform better , even if the extra headroom isnt needed ,it still improves the performance markedly .

Note, this design (commercial, not DIY) uses a DC-DC converter to take in 5V DC and create +/-12V. The efficiency of this converter is not that great at lower loads, even at maximum 20VA input power it ends up around 80%.

As it is a balanced design, we have four channels. Each channel runs a discrete output pair with 1R emitter resistors and in effect the SOT-223 SMD versions of BD-139/140.

The 1R emitter resistors mean optimum bias is theoretically at 26mA, in practice more like 30mA+.

This means the output stages idle at 120mA and 24V or 2.88W, in practice with other consumers in the device and the limits on the DC-DC converter efficientcy at lower loads (~70%) we end with well over 1A @ 5V idle.

Thus when used with low impedance AND low efficiency headphones (which are all the rage now - magnetic planars with 90dB/1V and under 20 Ohm impedance, the (external 5V/2A supplied, originally specified by the designer 5V/4A) power supply becomes the limit.

This does not linearly translated to more conventional power supplies.

But, if we want to reach 20V RMS into 18 Ohm to get 115dB peaks with these 90dB/V & 18.Ohm & 879 UKP headphones...

...well then we need to supply 1.1A RMS at least on short peaks, yup, that is 22W RMS for musical peaks to drive a headphone! Per channel!

By comparison, with a 140mA peak limit (NJM4556 in parallel) the same headphone would limit the output to 1.8V which means peak SPL's would be limited to ~96dB.

So as always context is important. NJM4556 in parallel will do fine driving 300 Ohm Sennheiser Headphones with 105db/V sensitivity for example. Merely 3.16V are needed for 115dB and thus 10.5mA RMS current.

Now if NJM4556 actually did not sound like "dead feet" even on 300R headphones they would be a good choice for such headphones.

Thor