LM6171 as CMoy buffer?

[Sir Nobax]

My bad, although I am damn proud that i didn't totally messed it up, a week ago i didn't even know what a op-amp was

.

So i put the Resistor values in the 'new' revision, are they okay?

U1 would be the Main op-amp i suppose, and U2 the buffer?
If U1 would be a 2-channel op-amp, which lanes are used to go to U2B, or is that not possible?

A simple guess would be: (edited)
7 -> R5B
6 -> R1b/R2b
5 -> input (10K Ohm Pot)

 

[jcx]

yes I think you've got the dual pinout right - a dual input opamp should work fine

you probably want more V gain, portable sources only put out .5 to 1 Vrms, this amp should swing over 5 Vrms so a gain of 10 would be the min I would use

for a gain of 10 I would use R1 = 9K, R2,4 = 1K, R3 = 2K

 

[Sir Nobax]

Yeah, Awesome !!!

Thank you for your help jcx, i really appreciate you help and comments ! !
So this really could work?

Any other Elite-DIY'ers can confirm this ?

(btw, why aren't there any Cap's used in this design, except the optional Cf)

 

[Sir Nobax]

Euhm, more people can confirm this to work????

(i'll be ordering stuff tomorrow, at least if i am completely sure that this design will work)

 

[jcx]

most sources are already AC coupled on their output

250 Ohm headphones aren't highly sensitive to offset

but if you want to be completely safe you could add some DC blocking caps

The caps you shouldn't leave out are the power supply bypassing caps shown in the Jung schematic: C1-4, you really want 0.1 uF stacked film or ceramic right at each LM6171's power pins

read through a couple of Tangents projects and parts lists to get ideas about a complete amp:
http://tangentsoft.net/audio/

the circuit should work, but it is short of being a "noob proof" or complete "project", some things that might be added: series diodes to prevent reverse battery damage, anti reversal diodes from each power supply to gnd to prevent damage if only one battery is connected at a time

also prototyping technique/layout can be an issue - what are you building on?
gnd plane with clearance hole perf board is good, although 30W may not be enough to solder to the gnd plane
a bare perf board could work but you should understand how to put down a low inductance gnd grid with bare tinned "bus" wire

 

[NelsonVandal]

Quote:

Originally Posted by jcx /img/forum/go_quote.gif yes I think you've got the dual pinout right - a dual input opamp should work fine you probably want more V gain, portable sources only put out .5 to 1 Vrms, this amp should swing over 5 Vrms so a gain of 10 would be the min I would use for a gain of 10 I would use R1 = 9K, R2,4 = 1K, R3 = 2K

A gain of 10 is absurd with IEM's. The only thing you "gain" is noise. I'd say stick to your lower gain. Buy resistors so you can fine tune the amp to your needs.

 

[Sir Nobax]

Quote:

Originally Posted by jcx /img/forum/go_quote.gif most sources are already AC coupled on their output, 250 Ohm headphones aren't highly sensitive to offset but if you want to be completely safe you could add some DC blocking caps.

This amp will 'complete' my 'Big rig' .. in my sig (PC's SBLive! 5.1), so the line out will be totally dependent on my PSU, which I don't see putting AC on one of the voltage rails of my computer. However, if that would be the case, DC offset on the DT-550 will be the least of my worries when that happens.

Quote:

Originally Posted by jcx The caps you shouldn't leave out are the power supply bypassing caps shown in the Jung schematic: C1-4, you really want 0.1 uF stacked film or ceramic right at each LM6171's power pins

Damn, i missed those, i thought that was the Power part of the PCB

.
So C1-C4 will have to go to the LM6171's pins, and also paralel to the main amp?

Quote:

Originally Posted by jcx read through a couple of Tangents projects and parts lists to get ideas about a complete amp: http://tangentsoft.net/audio/ the circuit should work, but it is short of being a "noob proof" or complete "project", some things that might be added: series diodes to prevent reverse battery damage, anti reversal diodes from each power supply to gnd to prevent damage if only one battery is connected at a time also prototyping technique/layout can be an issue - what are you building on? gnd plane with clearance hole perf board is good, although 30W may not be enough to solder to the gnd plane a bare perf board could work but you should understand how to put down a low inductance gnd grid with bare tinned "bus" wire

I will be using another LM6171 as supplier of the virtual ground, i don't know if reverse voltage will still be a problem when using a op-amp based virtual ground. If so, i have to put some diodes in the power supply scheme, directly from the 18V's - and + , is that correct? And what kind of diode should i be using, as Farnell (the shop where i will buy everything) has about 15 catagories.

And i will be using a Perfboard identical to the one Tangent is using in his CMoy guide for a part, but alot bigger. I was thinking of two levels, one for the power and the main opamp, and a second level for the buffers, but i guess the cables connecting those two levels would be a bad thing.

Tangents site is indeed very helpfull, but i think it is a shame that he doesnt make any how-to's without a professional PCB, like his CMoy.

Btw, i replaced the R1 and R2, someone was pretty sure the amp would be useless because of a huge amount of oscillation without the change.

 

[jcx]

Quote:

Originally Posted by Sir Nobax /img/forum/go_quote.gif ...Btw, i replaced the R1 and R2, someone was pretty sure the amp would be useless because of a huge amount of oscillation without the change.

too many cooks! - "someone" should read and understand Walt Jung's article on the composite line driver:
http://waltjung.org/PDFs/Composite_L...Distortion.pdf

the circuit is intended to have the buffer with gain U2, LM6171 inside the feeback loop of U1

the circuit is limited to using unity gain stable op amps of < 20MHz GBW for U1 with the values I suggested
with that limitation the circuit will be stable as shown, faster unity gain stable op amps can be used in U1 position by adding an appropriately sized Cf cap as shown in Walt's drawing

I r a engineer - I have designed and built over a dozen different multiloop circuits (I hate prototyping so I make really sure the theory's right before i build - I've probably simmed hundreds of composite op amp circuit variations), I have several thousand instances of more complicated composite op amp circuits in the field inside industrial/scientific measurement instruments I have designed for real companies dating from the mid 90's - nobody's come back to complain that they are oscillating yet

if you are planning on dual battery operation I would definitely skip the active buffered ground

 

[NelsonVandal]

If you want to use LM6171, why not be inspired by a good implementation. Look at Meier-audio's homepage "Projects/Headamp".
http://www.meier-audio.homepage.t-online.de/

I don't know anything about electronics, but to me this is all theories and measurements. Not a word on how it sounds. IMO AD823 is one of the worst sounding opamps, buffered or not - multiloop or not, and I doubt Walt Jung's amp will sound good. It would be more interesting to read how things actually sound. Multiloop is probaly a good thing, but I haven't found any review of how it sounds vs separate loops or just one global loop.

Using LM6171 instead of TLE2426 for active ground is asking for trouble. At least I have bad experiences from doing that. Adding an active ground channel like in PIMETA, PINT, PPA etc is a good idea though.

 

[Sir Nobax]

... Mmm, I think it will be a good idea for me that I shall study some DIY designs, and more information about the components to use. And use some things from the populair designs.
The current design indeed not excactly n00b-proof like JCX said

.


Ahum... . . . i just found an design in the SGheadphones forum, which would be a alot eastier to re-create without all the troubles.

 

[Heady]

Quote:

Originally Posted by Sir Nobax /img/forum/go_quote.gif ... Mmm, I think it will be a good idea for me that I shall study some DIY designs, and more information about the components to use. And use some things from the populair designs. The current design indeed not excactly n00b-proof like JCX said . Ahum... . . . i just found an design in the SGheadphones forum, which would be a alot eastier to re-create without all the troubles.

I have built a few of those and they sound pretty good too.