Modifying the SMSL SAP VI for low impedance headphones

[AndyBird123]

Hi guys, I'm a complete noob to DIY audio (and electronics in general), but after searching for info on a headphone amp I stumbled across this board... I've done a spot of studying over the past few days and I *think* I now know what I'm doing (famous last words),

everything I'm about to write is based of stuff wot I read on the internets, so if any of it is wrong then please tell me

basically, I have a soundblaster Z soundcard on my PC... virtual surround only works via the headphone socket, which has an output impedance of 22ohms, I've just ordered a set of soundmagic HP200 headphones which have an input impedance of 20ohms, so this presents a problem

after reading reviews of the HP200's I've come to the conclusion that I need a headphone amp with lower output impedance to better match my headphones, however this then means that I'll be double amping, which is also less than ideal

to get best SNR from my soundcard I should run it at about 80%

now, the SMSL has a gain of 6, which is total overkill considering it is an already amped input, the SMSL also has a 10ohm resistor just after the opamp right before the output - and this is where I get a bit sketchy - I am guessing this means it has an output impedance of 10ohms? (I plan on testing this when it comes to be sure)

using the rather handy schematic by rjr; (http://www.diyaudio.com/forums/blogs/rjm/950-modification-s-m-s-l-sap-vi-headphone-amplifier.html)

the way I think I see it, I have 2 options;

option 1, bridge the output resistors entirely (R7 and R8) and cut the return loop on the opamp (R5 and R6 - and R3 and R4?)

this would be the quickest dirtiest route, leaving me with a gain of 1 and giving a minimal output impedance

option 2, replace the output resistors with <2ohm and replace R5 and R6 with 10kohm resistors - leaving some output impedance in place and making the gain 2

have I got my assumptions correct, and which of these would anyone recommend (or something else)

thanks in advance

Andy

 

[Fast Eddie D]

So move the one leg of R5 to the other leg of R7. This solves the output impedance problem.

Eliminate R3 and change R5 to 10K. This gives a gain of 1. Or like you said, make both R3 and R5 10K, for a gain of 2. This solves the gain problem.

It's a very simple circuit, and flexible too. It doesn't have to be complex to work well.

Don't forget bypass capacitors for the chip. 0.1 F ceramics will do just fine.

 

[AndyBird123]

Thanks

The ceramic cap goes across the v+/v- pins of the opamp as per rjm's mods?
What does that do exactly?

 

[Fast Eddie D]

Thanks

The ceramic cap goes across the v+/v- pins of the opamp as per rjm's mods?

Yes. Put it as close as possible to the pins.

What does that do exactly?

Local high frequency bypass. It prevents oscillation. Always use it.

 

[AndyBird123]

Brilliant, thank you.

 

[Fast Eddie D]

You're welcome.

When driving headphones, it's a good idea to set the gain as low as possible. While the 4556 is rated for 80 mA, any op amp's linearity will be compromised when driving a 32 ohm load. So more feedback = less distortion. Unity gain is ideal. Unity gain will provide the lowest DC offset too. It's best to have a separate voltage gain stage if necessary.

 

[AndyBird123]

Yes, i was going to bridge or replace the 50k resistor to give me a gain of 1 for using with my PC, if i end up using the amp with something else then it will be a 10k resistor for a max gain of 2

The amp is just acting as a conditioner to reduce the output impedance, rather than actually as an amplifier as such

 

[Fast Eddie D]

Yes, i was going to bridge or replace the 50k resistor to give me a gain of 1 for using with my PC, if i end up using the amp with something else then it will be a 10k resistor for a max gain of 2

For a gain of 1, you want to eliminate R3 and R4. The 50K resistor can be bridged, or replace with 10K for lower offset (ideally you want R6 = R2). For higher gains, you want a capacitor in series with R3 and R4 for unity DC voltage gain and lowest offset.

At this time I would also like to recommend using an input capacitor.

The amp is just acting as a conditioner to reduce the output impedance, rather than actually as an amplifier as such

Yes but remember the amp must also deliver enough current to drive the load, while maintaining good linearity. That's the primary reason I gave you some extra tips like keep the gain low.

 

[AndyBird123]

It should do, the powersupply is twin 15v / 250mA and the amp is rated 200mw @ 33ohm
My headphones are 95db sensitivity and max 100mw, so i cant see me needing more than 50mw peak

I have everything ordered, so all I can do now is wait and try it and go from there

 

[sgrossklass]

The designer of this amp must have been super cautious. A 10 ohm series resistor, 330 ohm loading resistor and an output inductor? It's not like the buffer caps would be a million miles away either.

I haven't seen new equipment with a SIP package OP in quite a while. Amazingly enough, the part (AL) still is marked as "Active" in this day and age. SIPs were super common in the mid-'80s (and you still find them in things like FM frontend chips like the venerable TA7358AP), but would have been considered unpopular and hence cheap in the mid/late-'90s already.

While DC balance is reasonable, AC impedance never matches at any volume pot setting. At the kind of levels you normally run into headphones, it probably isn't much a an issue though.

Speaking of which, a 22 ohm output impedance is not necessarily problematic. It all depends on how variable headphone impedance is. Some models have such a ruler flat impedance response that even hundreds of ohms do not bother them very much. I have, unfortunately, not been able to find any such measurements for the HP200. You could use a short, high quality splitter cable (plus mini to mini) to run an RMAA loopback test with the headphones connected though. This would directly show response deviation in the FR graph.