[thedoctor]

Quote:

Originally Posted by ATAT *grin* that grinning aside, Class AB is quite a nice way to get long battery life and that design looks nice.. Id probably say its better than a MINT, worse than PIMETA, or any of the buffered ground designs (i mean 3 channel). But BUF634 running in low current mode will draw less than a class-AB IIRC... Also thedoctor, how is sijosae's a diamond buffer? its just a push pull current mirror .. I fail to see how this is a diamond buffer...

ah, sorry for the mistake. it's a push pull config. though from testing (with my ears), i kindda prefer that more to this :

 

[amb]

Quote:

Originally Posted by barkas What if I drove the ground like the L/R channels using a uA741 and 2x220kOhm as rail divider as input for the uA741? That would probably get a lot lower ground impedance.

Yes, if you build the virtual ground to look just like the L & R channels, except wire the input to a resistor divider, then it would work well. However, the 741 is a nasty opamp for audio. If you're going to go through all that trouble, why not use something a little better?

 

[barkas]

Because I have 5 uA741 here

.

However you are right. I should use a better opamp and I will have to mail order new photo boards for the new circuit so I may as well invest in another opa2132, or whatever good opamp I find that is not expensive.

EDIT: I would have to wire the feedback loop not through a 10kOhm and a 5k to earth, but directly back into the opamp without any resistors, right?

 

[Andrea]

The OPA132 or OPA134 should be good for this purpose since they have good output current, and especially a very low output impedance. I'd look at the OPA227 too, and certainly the formerly suggested OPA551 that is powerful beyond necessity, if with unknown output impedance (I couldn't find it in the datasheet

)

 

[ATAT]

Whoops. My bad.. definately not a current mirror.

Swapping the 3k resistor for a CRD will give better results IIRC.

 

[barkas]

If I use a push pull current amplifier behind the opamp, it does not matter one bit, what current the opamp can output, so I may as well use an opa2132.

 

[Pars]

Quote:

Originally Posted by ATAT Whoops. My bad.. definately not a current mirror. Swapping the 3k resistor for a CRD will give better results IIRC.

Do a search here for a BUF634 thread about 6 months ago where this was discussed quite a bit. Mono did a laout of a modified Sijosae buffer with FETs configured as current sources (IIRC). His layout was pin compatible with a BUF634... I've built some of these for a Pimeta and like them so far... better than the BUF634 with 221 ohm BW resistors to my ears.

 

[amb]

Quote:

Originally Posted by ATAT Swapping the 3k resistor for a CRD will give better results IIRC.

Yes. CRDs will make the bias current (mostly) independent of supply voltage, and increase the PSRR.

 

[ozshadow]

But is the sound quality very clean ?

What I want is something portable, with tight, punchy and solid bass. Rich mids to go with my Grado's, and the highs just have to be accurate.

Again, I'd like something very clean, but with the bass very present but not boomy, just with some tight impact.

I've tried the 2227 and like the clean sound, but could be richer and the bass is a bit light. The 2132 was ok, but I felt the bass was more on the boomy side, but it was there, like a ported system in car audio, not a sealed box. It hurt the overall sound to me.

I am trying a 2107 as soon as it gets here. I also have some single 627's on the way, but these will not work in anything I have to this point.

So knowing that I want something clean, accurate, rich, with tight, punchy bass to use with Grados, I'd welcome any suggestions for the amp base and opamps to use.

I really do not care about battery life, and size can be up to SR-71 size. If I had the money, I'd have me one of those !

Thanks all !

 

[ATAT]

Quote:

Originally Posted by Pars Do a search here for a BUF634 thread about 6 months ago where this was discussed quite a bit. Mono did a laout of a modified Sijosae buffer with FETs configured as current sources (IIRC). His layout was pin compatible with a BUF634... I've built some of these for a Pimeta and like them so far... better than the BUF634 with 221 ohm BW resistors to my ears.

Thats good, so the CRD biased buffer sounds good? I've been working on a discrete buffer using that design so that's good news =)

 

[amb]

ozshadow, before you get carried away with subjective descriptives, I believe sijosae conceived this circuit to be ultra-compact as a primary goal, not ultimate high-resolution hi-fi...

 

[ozshadow]

I do not have any ultra hi-fi recordings. I am just hoping for tips on a project. I described what I hoped to build as best I can. I am juggling something like this vs. a pimeta in a portable case vs. scrounging around for a pint board. I enjoy the building, even if a little at a time, but I do not have the cash to build lots of different amps and combos to find what I am looking for. I would prefer to enjoy the building, but have a result I am quite pleased to use as well.

Thanks for the input.

 

[mono]

I could be mistaken but my general impression is that you have to consider the cans. You can choose lower current fast output transistors and lose some output current potential, or parallel them and lose a slight bit of treble definition with the current sharing resistors, or use larger die transitors that are slower. Then again I ended up putting output resistors on my reinvention of the Sijosae buffer after a defective 1/4-1/8" jack caused me to blow out one.

Maybe this is all voodoo superstition, I can only tell you for sure that I have no desire to use BUF634 anymore, after coming up with even the simplistic discrete replacement. Kudos to Amb for the patience and help with it, people like him MAKE head-fi the great place it is!

Anyway, I like Sijosae's AB and only wish he'd fit more capacitors into the designs so I don't have to alter them, as I like to throw in some 0.1uF ceramic and 1uF film where possible. BUT, you don't "HAVE" to built it like that, when i'd suggested it in a previous thread it was just an example of a step-up from a CMOY or A47, a very inexpensive way to get the next step up in performance. I'd think of the Sijosae AB as an unveiled MINT. Personally I would 3 channel this, do a ground channel too and then you have a simplified discrete Pimeta. It is a very high bang for buck, but will it synergize with your favorite cans and music? I can't know.

I can tell you that I would give up a few percent overall accuracy for more treble definition and tight bass punch, but most of all I can't stand the "veiled" sound of BUF634, I can listen to a portable MP3 player or power amp directly if I wanted that.

 

[ozshadow]

Any guesses how it would affect the output if you doubled the amplification stage like this into an A47-type layout ?

 

[ATAT]

Oz, that's pointless since this already has a gain stage and buffer stage.. the A47 just uses an opamp for buffer stage instead of discrete transistors.