HD580s + Meta42
Jun 26, 2002 at 5:17 AM Thread Starter Post #1 of 15

andrzejpw

May one day invent Bose-cancelling headphones.
Joined
Jun 25, 2001
Posts
6,636
Likes
10
Any tips for optimizing the meta42 for HD580s? Let's have some fun.
smily_headphones1.gif


I'm talking about resistor values, caps, etc.
 
Jun 26, 2002 at 6:46 AM Post #4 of 15

tangent

Top Mall-Fi poster. The T in META42.
Formerly with Tangentsoft Parts Store
Joined
Sep 27, 2001
Posts
5,969
Likes
56
Quote:

Any tips for optimizing the meta42 for HD580s?


1. Use a gain of about 4 or 5 for full-size players, or 5-6 for portables. If you use lower gain values, you risk running out of volume. If you use higher values, that will work, but you won't be able to turn the knob as much before it gets to deafening volumes, so the knob will be touchy.

2. If you're going to use Chu Moy's modified Linkwitz crossfeed (the only one I recommend with the META42 at this time -- start a new thread if you want details) you need a higher gain because this circuit attenuates the volume.

3. Although paralleling output buffers is said to help with all headphones, you won't get as much bang for your buck with high-impedance headphones like those in the Senn 500/600 series.

4. You don't need R8 or R9 with these headphones. Just jumper them.

5. There are no op-amps I would specifically recommend against. (This is not true for Grados, which is why I bring the issue up.)

Quote:

I'm not even exactly sure what gain really is.


When you're talking about an amplifier, you usually mean voltage gain. (There's also current gain.) This simply means magnifying the op-amp's input voltage by a specific amount.

Concrete example: let's say your headphone wants 1.5V to reach blasting volume. You have a portable source that puts out 400 mV. In between the source and the actual amp circuit, you have a potentiometer, which reduces the voltage further, to, say, 50 to 300 mV in the normal listening range. Result? You need a gain value of 5 to reach full blasting volume at the high end of that volume range.
 
Jun 26, 2002 at 7:02 AM Post #5 of 15

fyleow

500+ Head-Fier
Joined
Mar 15, 2002
Posts
516
Likes
10
Sorry andrzejpw for hijacking this thread but....

Ahh ok, I get it now thanks Tangent. Sorry, you must get dumb questions like this all the time.

About the Linkwitz crossfeed I don't really plan to add that but I am curious. Will it require alot of modding of the META42 or is it designed with the crossfeed in mind?

I was originally planning to use the OPA637 with my META42 but since you recommend about a gain of 5 the 627s will work better then? (Since you said the OPA637 needs about a gain of 10 to be stable in another thread) Or am I confusing voltage gain with some other gain? Haha I'm confused again. Sorry.

I want to use a wall wart this time instead of batteries, what should I be looking for? What type of voltage, A etc?

Sorry for all the questions, any links you can point me to so I can actually get a clue and avoid cluttering the board?

Thanks.
 
Jun 26, 2002 at 7:55 AM Post #6 of 15

tangent

Top Mall-Fi poster. The T in META42.
Formerly with Tangentsoft Parts Store
Joined
Sep 27, 2001
Posts
5,969
Likes
56
You put the crossfeed off on a separate board and plug its outputs into the META42's input pads. Then, you connect the amplifier's input jacks to the crossfeed circuit's inputs. Pretty nonintrusive.

Will the 627s work better? They'll be easier to use, if that's what you're asking. The 637s might sound better -- I don't know yet. (I have a set, thanks to Nezer, but I haven't gotten around to testing them.) If you're set on the 637s, just use a higher gain -- you just won't be able to turn the volume knob up as high. If you go from a gain of 5 to 10, that's only a 3dB gain....

On the power supply, that's been covered here in the past day -- read the thread. Also read my META42 assembly article -- plenty of pontification there, too. For HD580s, 12V-24V is a good range, depending on the op-amp.

"any links you can point me to so I can actually get a clue and avoid cluttering the board? "

http://www.amazon.com/exec/obidos/ASIN/0521370957/
 
Jun 26, 2002 at 1:43 PM Post #7 of 15

andrzejpw

May one day invent Bose-cancelling headphones.
Joined
Jun 25, 2001
Posts
6,636
Likes
10
Thanks a ton for your tips!

I don't think I'll be adding crossfeed at this time.

What would be involved in adding a power supply? I mean, could I just get a wallwart or something, and plug it into the circuit?

Or is there something more involved?
 
Jun 26, 2002 at 2:27 PM Post #8 of 15

Nezer

Antibacterial soap... kills bacteria... bad karma?
Joined
Dec 6, 2001
Posts
1,273
Likes
13
Who said the 637s are stable only over 10?

The data sheet claims stability of G >= 5.
 
Jun 26, 2002 at 6:00 PM Post #9 of 15

kheldar

100+ Head-Fier
Joined
Jun 18, 2002
Posts
156
Likes
0
If you are using the multiloop setup of the meta42, can you configure the opamp to be running a higher gain (for chips like the 637) and lower the second gain so that the total gain is lower (5 or 6) and still make the amp happy?
 
Jun 26, 2002 at 6:05 PM Post #10 of 15

Nezer

Antibacterial soap... kills bacteria... bad karma?
Joined
Dec 6, 2001
Posts
1,273
Likes
13
The purpose of the second gain setting on the Meta42 is to throttle the bandwidth of the opamp.

This is basically the same question I asked above, will limiting the bandwidth on the 637 allow it to run stable at lower gains? I'm guessing someone with a scope would have to sit-down and test this. I also doubt it to be the case because if it were then it would likly be documented in the datasheet.

I still think it best to be safe and use the 627 were G < 5.
 
Jun 26, 2002 at 10:24 PM Post #11 of 15

tangent

Top Mall-Fi poster. The T in META42.
Formerly with Tangentsoft Parts Store
Joined
Sep 27, 2001
Posts
5,969
Likes
56
Quote:

could I just get a wallwart or something, and plug it into the circuit?


Yes.

Quote:

Who said the 637s are stable only over 10?


I did, sort of. When the datasheet says the 637 is "stable at G=5", it's saying that the chip can be made to work without compensation at that gain level, not that you have some automatic guarantee of stability in all conditions. The higher the gain you use, the more stable it becomes, which is why I suggest using a somewhat higher gain -- like 10 -- with this chip than you might normally use.

Quote:

can you configure the opamp to be running a higher gain (for chips like the 637) and lower the second gain so that the total gain is lower (5 or 6) and still make the amp happy?


Yes. All this talk about gains of 5 and 10 and such goes out the window when you use multiloop, because you generally set the op-amp for a gain of 80+, so even the 637 should be nice and stable under all conditions.

Quote:

The purpose of the second gain setting on the Meta42 is to throttle the bandwidth of the opamp.


That's one of the effects, yes, because as gain goes higher, bandwidth drops. What also happens is that phase margin is increased, which means the op-amp becomes more stable.

Quote:

will limiting the bandwidth on the 637 allow it to run stable at lower gains?


The problem with that question is that there's more than one way to limit the bandwidth of an op-amp. If you put a small capacitor (10-47 pF) in the feedback loop, bandwidth is reduced, but not in a fashion that will help the OPA637. This fact is discussed at the beginning of the application note section of the datasheet. With other op-amps, this "phase lead" trick doe help stability at low gains by trading bandwidth for phase margin.
 
Jun 26, 2002 at 11:13 PM Post #12 of 15

Nezer

Antibacterial soap... kills bacteria... bad karma?
Joined
Dec 6, 2001
Posts
1,273
Likes
13
Tangent,

As far as I'm concerned you are da man. You've cleared a LOT of things up for me lately and for that I'm eternally grateful!

So, if we use a Jung multiloop topology we can set the bandwidth-limiting gain such that we are *well* above the required gain of 5 to make the amp stable but still maintain a reasonable gain fo the whole circuit. And as you stated in a PM, stability is a boolean thing, you either have it or you don't.

This makes more and more sense.

Maybe one day I'll be where your at.
 
Jun 27, 2002 at 5:57 AM Post #13 of 15

tangent

Top Mall-Fi poster. The T in META42.
Formerly with Tangentsoft Parts Store
Joined
Sep 27, 2001
Posts
5,969
Likes
56
I don't know if I'd say stability is boolean -- it's more like a black area -- no stability -- and a white area -- unconditional stability even in extremely difficult conditions -- separated by a narrow gray gradient band. In that band, you may have stability or not, depending on conditions. Roughly speaking, with the OPA637, that band is between gains of 5 and 10. This is very vague, and probably wrong in actual fact, but I hope you see the main point: using a gain of 5 with the OPA637 is extremely risky, and risk decreases as gain goes up.

Quote:

Maybe one day I'll be where your at.


Two steps beyond clueless, you mean?
smily_headphones1.gif
 
Jun 27, 2002 at 12:31 PM Post #14 of 15

andrzejpw

May one day invent Bose-cancelling headphones.
Joined
Jun 25, 2001
Posts
6,636
Likes
10
Quote:

Originally posted by tangent
I don't know if I'd say stability is boolean -- it's more like a black area -- no stability -- and a white area -- unconditional stability even in extremely difficult conditions -- separated by a narrow gray gradient band. In that band, you may have stability or not, depending on conditions. Roughly speaking, with the OPA637, that band is between gains of 5 and 10. This is very vague, and probably wrong in actual fact, but I hope you see the main point: using a gain of 5 with the OPA637 is extremely risky, and risk decreases as gain goes up.

Two steps beyond clueless, you mean?
smily_headphones1.gif


heck, if you're two steps beyond, I must be about 7 steps behind!
eek.gif
 
Jun 28, 2002 at 6:44 PM Post #15 of 15

jmedin

New Head-Fier
Joined
Apr 27, 2002
Posts
20
Likes
0
Hey Tangent,

in this thread above you say;

" 5........ (This is not true for Grados, which is why I bring the issue up.) "

What opamp(s) WOULD you recommend for the 1st opamp in the META42 (using EL2001 as buffer) for use with the following headphones:

1) GRADO SR-80 cans ?

and if you have an opion,

2) ETYMOTIC ER-4S plugs ?


Tangent I must also say:

Heartfelt thanks for all your help, your hard work, and truly deep dedication, to this and to the "other" forums!



Thanx much,

Jeff
smily_headphones1.gif
 

Users who are viewing this thread

Top