Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › the Pimeta-V2 thread
New Posts  All Forums:Forum Nav:

the Pimeta-V2 thread - Page 31

post #451 of 570
Quote:
Originally Posted by tangent View Post

 

It's nice to see that op-amp getting a little publicity. Its one serious weakness — low output current — totally doesn't matter in a PIMETA v2.


Yes, I have 6 or so various op-amps that I ordered based on your op-amp page. When I started building this amp I sat down with my little Cmoy and did some testing. the Cmoy is easy to roll op-amps in, and its just the op-amps in the circuit so it makes for good A-B comparisons.I really liked the 823, to me it really expanded the "sound stage" and "depth" compared to the others i had (sorry, don't have them in front of me so i cant give the PN's). All the others sounded kind of flat by comparison. Plus the 823 just sounds lively or fun if you will. I actually preferred the lowly little Cmoy with the 823 over my mini3 (the HP build). I did find that it did not have the grunt my mini3 has, but i did prefer the sound. As you say though with the buffers its a non issue. In the Pimeta running 12 volts, even with the standard gain (4.75 R4 if memory serves me, gain of ~5) its plenty loud!.

 

Tangent, just got the parts I ordered from you last week, so i can finish this one up. Thanks.

post #452 of 570

Alternative (better, IMO) to AD823 is AD8397. Not too $$, and HIGH current. It seems to have gotten mildly popular in some audio circles. Anyone try it in Pimeta?

post #453 of 570

The 8397 is a bipolar input chip, and so may cause a DC offset at the output of the PIMETA without tweaking the values of the resistors attached to its input pins.

post #454 of 570
Quote:
Originally Posted by tangent View Post

The 8397 is a bipolar input chip, and so may cause a DC offset at the output of the PIMETA without tweaking the values of the resistors attached to its input pins.

I've found that modern bipolar op-amps like the 8397 are okay to use "naked". In fact, that opa is used as an output for the QLS-350 DAP and several other commercial jobs like the Xin SM3 headphone amp

post #455 of 570

Anyone selling these already built?

post #456 of 570
Quote:
Originally Posted by alphaman View Post

I've found that modern bipolar op-amps like the 8397 are okay to use "naked". In fact, that opa is used as an output for the QLS-350 DAP and several other commercial jobs like the Xin SM3 headphone amp

That doesn't matter.  I think if you look into the history of the AD8397, you'd be a bit more careful in recommending its use.  Tangent himself, along with AMB, both came out with brand new designs when the AD8397 came out (a few years ago).  Tangent went with the PINT - an amp so small, it was intended for use in a mint tin with a couple of 9V batteries.  If I recall correctly, the PINT PCB was 3/4" x 2-1/4" (or same size as the PPAS).  Unfortunately, there were enough problems that Tangent removed the PINT design completely from the market.  He didn't feel that it was stable enough to support as a product from tangentsoft.com.

 

AMB did the same in working with Head-Fi/Headwize user Morsel - who helped on both the M3 and the PPA.  AMB's version used a larger PCB based on 2 - 9V batteries in a large, PIMETA-styled Serpac case.  It was not stable in all regimes.  After more than a year of on-and-off testing, etc., AMB finally came to the conclusion that the AD8397 would only operate with stability with one 9V battery, not two.  That's how the current Mini3 came about.

 

Anyway, all that stuff means is that you shouldn't trifle with the AD8397.  It's not something that you can simply "roll-in" as an opamp choice.  Tangent is being kind in his reply - he knows this more than anyone else.  


Edited by tomb - 7/27/12 at 6:17pm
post #457 of 570

wow...that was an awesome post.


Edited by vixr - 7/27/12 at 7:57pm
post #458 of 570
Quote:
Originally Posted by tomb View Post

Unfortunately, there were enough problems that Tangent removed the PINT design completely from the market.

 

I'm pretty sure the problem with the AD8397 in the PINT was due to the ground channel. If you find the gain vs. frequency chart in the chip's datasheet, you see that the GBW is only flat at gains of 2 and above. The peak you see at ~40 MHz at g=1 causes positive feedback, which in electronics is a Bad Thing if you weren't intending on building an oscillator.

 

I think if someone takes a PINT and replaces the ground channel with a true dual supply scheme with a passive ground, it'll be quite stable.

 

(As for why I didn't just start recommending that with the PINT, it's because I didn't want to stop recommending the AD8397 for OPALR, and there's a segment of the community that won't listen when you tell them it's fine — better even, sometimes — to use a different ground channel chip than for the L&R channels. I've toyed with creating a passive-ground version of the PINT, basically a glorified CMoy pocket amp, which if it ever gets off the ground will be called the PINT/2. Yes, the half pint. Laugh, it's funny.)

 

As for the matter of the AD8397 being "modern," that has nothing to do with anything. Input offset current is all that matters. The physics have no respect for modernity of design.

post #459 of 570

I have not experimented with Pimeta + 8397. This opa has become a favorite in other apps so I suggest it as a possibility for a DEFAULT Pimeta.

 

The DEFAULT Pimeta has/uses:

 

- single voltage source (not a separate 9v for each rail)

 

- multiple "layers" of purpose-built protection. E.g., R11 limits the max. output current on the buffers. 

 

So given the above, experimenting with the 8397 + DEFAULT Pimeta should, at worst, lead to harmless untweakable distortion. 

 

As much as I like the Pimeta, I may like the 8397 more. Specifically, when traveling with my portable rig., a Colorfly CK4 DAP (with both LO and HO opa's modded with 8397's) + Pimeta v2 , I no longer feel I'm missing all that much with an external amp. The only thing I REALLY miss is input-based MLC.

 

Hmmm .... since I like the 8397 so much, maybe I should design a novel/ad hoc amp around it, rather than adapt a "generic" pre-existing job like Pimeta.  

post #460 of 570
Quote:

I've toyed with creating a passive-ground version of the PINT, basically a glorified CMoy pocket amp, which if it ever gets off the ground will be called the PINT/2.

Since you're toying with ideas...

How about using a charge pump configured as an inverter to create a - voltage "rail"?

 

http://www.linear.com/product/LT1054

 

Of course being as there is no such thing as a free lunch,  you would need deal with the switching noise and the uneven "split" from the voltage loss but should be pretty easy. :D  

post #461 of 570

Because if I'm going to create the negative rail with a DC-DC converter, I might as well create both of them so I can run the thing off a single NiMH cell.

 

And the reason I don't do that is that in order to fix the noise created, you need big Cs and Ls, which eats into that space savings budget.

post #462 of 570
Quote:
Originally Posted by tangent View Post

Because if I'm going to create the negative rail with a DC-DC converter, I might as well create both of them so I can run the thing off a single NiMH cell.

 

And the reason I don't do that is that in order to fix the noise created, you need big Cs and Ls, which eats into that space savings budget.

No, Tangent .... MODERN (I know you love that word) i-[devices], DAPs, etc. effectively use MODERN Buck/DC-DC chips with MODERN Li-ion batts in extremely compact packages. No "big" C or L really needed; for C, high-qual. SM tants should be fine, add a few SM ferrite beads for even better SNR. I've 'scoped some of them and even most stock-jobs are clean enough.

post #463 of 570

Okay. Go make that.

post #464 of 570
1) I have question about R11. Tangent says that its 667X power suppy. So if my supply is 24V, the R11 should be 7.3K. But using 7.3K the current limit for 24V is 24/7.3k = 3.3mA. If I want to limit current to 300mA using 24V the appropriate value should be 24/0.3 = 80R. Can someone please help me.
2) I have difficulties sourcing 6.8uF for C4 with the right size. Can I use tantalum here. i have 10uF, 47uF and 150 uF. Which one should I use. Thanks.
post #465 of 570
Quote:
Originally Posted by niksh View Post

But using 7.3K the current limit for 24V is 24/7.3k = 3.3mA.

 

You can't use Ohm's Law that way. The resistor is not simply sitting there across the power supply. It is controlling the amount of current that goes into the buffer's current limiting circuit. We say the current limiter is "programmed" by this tiny controlling current.

 

Read the LMH6321 datasheet for more info on this.

 

 

All the rest of your questions are answered by the C4 section in the docs.

New Posts  All Forums:Forum Nav:
  Return Home
Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › the Pimeta-V2 thread