[Luvya]

I have no idea what a good design should looks like. So maybe you guys can enlighten me.
http://www.myhpw.com/Others/DL01.htm

Some IC design pictures....I dunno which one is relevant or which is not. So please disregard the ones that are not comparable. Or maybe they are all relevant? @_@

 

[eric343]

That's a funky design

It looks like he's got not one, but two different feedback loops (in the first design). One's a DC-blocking servo, by the looks of it (the one with the capacitor feedback loop), however it's configured in inverting mode and fed into the positive input of the main chip (servos, and any feedback componetns, are usually fed into the negative input for a non-inverting configuration) but he still has a capacitor on the input...

Maybe I'm just up too late...

 

[Luvya]

So the design has flow? wouldn't even work? @_@ I was under the impression that at least the first (LC-icute) and the third (LC-1) are proven design.....

 

[eric343]

No, I'm not saying that. I'm just saying that it's not something I've ever seen before...

 

[rickcr42]

simply a DC servo circuit to eliminate DC from the output without using an output coupling cap or a DC balance pot.

Should work fine

 

[eric343]

Ah. Live and learn, I guess; I forgot about the ability of amps to generate DC offset without a DC offset on the input (there's an input coupling cap)...

 

[rickcr42]

especially with a single ended transistor sitting at the output.

Unless provisions are made for a DC balance control the output would ride at an AMPLIFIED DC level !

Not only that , but say there were a balance control , they are known to stray with age of the part.

Rickmeister

 

[wildmonkeysects]

Really, it is the *implementation* that makes or breaks a design.

It does look like the author is having a blast with all kinds of topoligies...

But, technical points/quibbles:

The U1a DC offset loop is unity gain, while the closed loop signal gain, U1b is 6x...methinks it was not listened to critically.

DC offset loops, while seemingly simple, can wreak subtle havoc on the sonics. The R1/C1 time constant needs to be orders of magnitude shorter then the servo, R8/C2 time constant to reduce servo thermal tails, which contaminate the signal path...also, R8 should be broken into two series resistors, with a cap from the center of the two to ground. The loop will still be stable, and the signal "seen" by the servo opamp is further reduced.

Sometimes, it is just as easy to AC couple the output with Panasonic HQ (with film bypass) or Black Gate electrolytics.

 

[Luvya]

So...is it better or worse than Meta? @_@
It certainly look more complicated though..

 

[wildmonkeysects]

I would not be qualified to pass judgement, as I have not listened to both...but, the 42 is a known quantity, so probably safer for an early DIY.

The impression I have is that Mr. Lin is having fun with servos, probably more as an intellectual exercise than an esthetic one. This is neither good nor bad, happens to lots of us in early stages of circuit discovery...we want to try it all and frequently end up with Rube Goldbergish designs, just because we can.

I really do not mean to step on any design toes, it's all fun, right?

The residual offset from a decent monolithic opamp such as the BB/TI opa627/637 is low enough that it is a non-issue for unity gain stages (627 only!) and probably not an issue for stages with lowish gain such as 5x or 6x, so, again DC offset servos are more for intellectual fun than need in such a case.

Do use good stuff to state the obvious...implementation is as important as the raw design in terms of resultant sonics.

 

[rickcr42]

I agree that the design is more than likely an exercise and not optimized.
in fact it looks more like it is for a speaker not headphone load.

maybe you want to look at this instead

http://www.tnt-audio.com/clinica/dynamicopamp_e.html

 

[Maniac]

Well, about those designs...

I've got two of those varients on my desk now...

First one is LC-iCute, that Class A headphone AMP is tiny. Currently running using a pair of OPA2132P.

Second one is LC-King, a similar design but running in Class AB mode (this one's mode of operation is determined when you are soldering the parts on, you get to pick A or AB). And with a rather interesting power supply to boot.


Both of them EASILY achives SNR ratio of over 100 db... My LC-King is installed in a rather good looking Arcylic case that offers no protection to EMI... But... the circuit doesn't seem to care, absolutely no interference...


As for sound... It sounds really good... REALLY GOOD... Good enough that I really have no big desire to get other head amps, other than just for kicks/comparison...

The circuit is currently designed to make it more fool proof, but some parts can be changed to make it sound even better. Tho the fool-proof-ness factor would not be as high.

David Lin, the designer is really good with what he is doing, and don't like to do things the way that other people had done... I'm currently waiting for his newer design and modifications to current designs...

 

[radrd]

Now this looks like a setup...

Poster asks "Anyone think this will be better than a META42?"

...a few posts later...

New member enters....

New member says: "I have it and it is outstanding...."

10 days later the damn thing will probably go on sale!

Hey, I'm probably wrong, but this looks fishy.

 

[Joe Bloggs]

Well you can buy it off Lin himself if you want:
http://www.myhpw.com/Business/LCiCute/LCiCute.htm

NT$4500, however much that is

This guy is the boardmaster of a big in Taiwan: www.myhpw.com

Hey, you guys have butted heads before, and this circuit has been discussed before too. Have you guys got a short memory or what?

 

[Maniac]

Quote:

Originally posted by radrd Now this looks like a setup... Poster asks "Anyone think this will be better than a META42?" ...a few posts later... New member enters.... New member says: "I have it and it is outstanding...." 10 days later the damn thing will probably go on sale! Hey, I'm probably wrong, but this looks fishy.

I'm *NOT* selling mine. And I'm not going to send one overseas.

I'm going to have fun with it and that's why I got it in the first place. Does that satisify you?

I might build several varients of it and put it into this chassis, and each should give a different flavour or no flavour at all...