Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › Amp design question
New Posts  All Forums:Forum Nav:

Amp design question

post #1 of 84
9/17/10 at 5:34pm
Thread Starter 

So I'm thinking about making another cmoy amp, but this time I want to buffer the ground channel using tangent's design:

gnd-channel.png

For this configuration, would this VFB opamp work with this buffer?  I know that the buffer is supposed to be faster than the opamp, but am I missing another spec?  And would this buffered ground work with the standard amplification section for the cmoy? (The amplifier section in the schematic)

Also, what type of caps belong at C1 and Cc?  I'm assuming that C1 is electrolytic, but I dont know what to get for Cc.

I also know that in the PIMETA design, all channels are buffered.  What are the benefits of doing that?

 

All input will be immensely appreciated.

 

EDIT: I will be running this from a 24v wallwart.


Edited by shake - 9/17/10 at 5:40pm
Reply
post #2 of 84
9/17/10 at 6:10pm

All three of your channels should roughly be equal.

There is not much point in using a buffer in your ground channel if your not using

them in your other channels.

To answer your other question, C1 is a low ESR Electrolytic and Cc is a Film.

 

Lastly, the buffer both provide more output current and isolate the op amp from the

load to improve their performance.

 

You will need to learn how to solder power pad surface mount i.c. before you can use

that buffer. If not soldered correctly to a heat sink it will self destruct in a few seconds.

Reply
post #3 of 84
9/17/10 at 6:24pm
Thread Starter 


 

Quote:
Originally Posted by Avro_Arrow View Post

All three of your channels should roughly be equal.

There is not much point in using a buffer in your ground channel if your not using

them in your other channels.

To answer your other question, C1 is a low ESR Electrolytic and Cc is a Film.

 

Lastly, the buffer both provide more output current and isolate the op amp from the

load to improve their performance.

 

You will need to learn how to solder power pad surface mount i.c. before you can use

that buffer. If not soldered correctly to a heat sink it will self destruct in a few seconds.


Thanks for your reply.

What benefits would I get for just buffering the ground channel, using tangent's design?

Would the LMH6321 have to be soldered to a heatsink?  Wouldn't it be fine if I just used some thermalgrease and a heatsink?

And I guess I could use the BUF634 instead.  Would they be stackable in this design?

Reply
post #4 of 84
9/17/10 at 7:27pm

PSOP I.C.'s like the LMH6321 have an exposed pad under the

 I.C. that has to be soldered to the circuit board to for the heat sink.

See page 17 in the data sheet. The TO-263 version would be easier.

It's not like the kind of heat sink you see attached to transistors or MOSFETs.

 

You can use a buffer if you wish, but the benefit it would provide could be had

just as well with the same op amp you used for the L/R channels.

 

BUF634 are stackable. Most people have found that four is the most that

are need, and more than four does not seem to help further. For the PDIP

version, you just stack them one on top of the other.

 

You might want to look at OPA633 as well. It's a buffer like BUF634, it is

not as powerful but it is a heck of a lot faster.

Reply
post #5 of 84
9/17/10 at 7:34pm
Thread Starter 
Quote:
Originally Posted by Avro_Arrow View Post

PSOP I.C.'s like the LMH6321 have an exposed pad under the

 I.C. that has to be soldered to the circuit board to for the heat sink.

See page 17 in the data sheet. The TO-263 version would be easier.

It's not like the kind of heat sink you see attached to transistors or MOSFETs.

 

You can use a buffer if you wish, but the benefit it would provide could be had

just as well with the same op amp you used for the L/R channels.

 

BUF634 are stackable. Most people have found that four is the most that

are need, and more than four does not seem to help further. For the PDIP

version, you just stack them one on top of the other.

 

You might want to look at OPA633 as well. It's a buffer like BUF634, it is

not as powerful but it is a heck of a lot faster.


Yeah I saw the price of the buf634 and decided not to stack XD.  What wold a faster buffer do?

I've decided to buffer the other channels though, so I might be getting 3 buf634's for L/R/G.

 

Do you have any comments on the VFB opamp I've selected?


Edited by shake - 9/17/10 at 7:36pm
Reply
post #6 of 84
9/17/10 at 7:46pm

OPA633, not BUF633.

 

The LT1360CN8 is the little brother to the LT1364CN8 that I use.

Sounds excellent. You could use LT1363CN8 for the ground channel.

Reply
post #7 of 84
9/17/10 at 7:58pm

there is no real reason for active gnd to be identical to R,L channels - the active gnd has to sink both R+L channel op amp current at the same time so in general no even harmonic cancellation will happen since its output is operating at different current than either R or L channel and with constant V drop has higher heat dissipation requirement

Reply
post #8 of 84
9/17/10 at 8:12pm
Thread Starter 
Quote:
Originally Posted by Avro_Arrow View Post

OPA633, not BUF633.

 

The LT1360CN8 is the little brother to the LT1364CN8 that I use.

Sounds excellent. You could use LT1363CN8 for the ground channel.


And I should use it in both the VFB part in the ground channel and as the main amp?  It might get a little complicated 0.o

Reply
post #9 of 84
9/17/10 at 8:13pm


Yes, quite right.

 

Ground channel equal to or better than L/R, never less than....

 

Quote:
Originally Posted by jcx View Post

there is no real reason for active gnd to be identical to R,L channels - the active gnd has to sink both R+L channel op amp current at the same time so in general no even harmonic cancellation will happen since its output is operating at different current than either R or L channel and with constant V drop has higher heat dissipation requirement

Reply
post #10 of 84
9/17/10 at 8:17pm

You should read through the PIMETA.

It's kind of what you are trying to build...

 

Waren probably explains it much better than I can.

Reply
post #11 of 84
9/17/10 at 8:36pm
Thread Starter 

I'm not savvy enough to look at those schematics and know what I need and what I can take out :/

But back to the above ground channel, would the amp benefit if I place bypass caps between the railsplitter and the VFB opamp?  It doesn't show in the diagram, but I think the amp needs them...

Reply
post #12 of 84
9/17/10 at 8:57pm

Bypass caps on the power pins are just a given...

 

Cc shorts out high frequency to keep the op amp/buffer combo

from oscillating.

 

The TLE2426 DIP version has a bypass capacitor input (pin 8).

You should not put caps on the output of the TLE unless it is

the only component in the ground channel.

 

In the above schematic, the TLE, op amp and buffer all act

together. The TLE provides the 1/2 power supply reference,

the BUF provides the current  and the op amp keeps the output

of the BUF stable with feedback.

 

To understand better, the op amp tries to keep the - input at the same

potential as the + input. It does this by varying the output. The - input

of the op amp is connected to the output of the buffer. Since the + input

is being held at 1/2 power supply, the op amp varies it's output until the

- input (and hence the output of the buffer) is also at 1/2 supply.

If the load on the BUF tries to move the output away from 1/2 power supply,

the feedback from the op amp corrects it.

Reply
post #13 of 84
9/17/10 at 9:03pm
Thread Starter 


 

Quote:
Originally Posted by Avro_Arrow View Post

Bypass caps on the power pins are just a given...

 

Cc shorts out high frequency to keep the op amp/buffer combo

from oscillating.

 

The TLE2426 DIP version has a bypass capacitor input (pin 8).

You should not put caps on the output of the TLE unless it is

the only component in the ground channel.

 

In the above schematic, the TLE, op amp and buffer all act

together. The TLE provides the 1/2 power supply reference,

the BUF provides the current  and the op amp keeps the output

of the BUF stable with feedback.

 

To understand better, the op amp tries to keep the - input at the same

potential as the + input. It does this by varying the output. The - input

of the op amp is connected to the output of the buffer. Since the + input

is being held at 1/2 power supply, the op amp varies it's output until the

- input (and hence the output of the buffer) is also at 1/2 supply.

If the load on the BUF tries to move the output away from 1/2 power supply,

the feedback from the op amp corrects it.


Maybe bypass caps aren't what I was trying to get at...  I'm talking about the caps that would be right after the rail splitter:

Which power pins are you talking about for the bypass caps?

 

EDIT: Also, should I worry about input caps?  I've been told that I shouldn't have to worry about these because modern sources already take care of DC offset.  My source is currently the VALAB NOS DAC.


Edited by shake - 9/17/10 at 9:17pm
Reply
post #14 of 84
9/18/10 at 7:44am

In your simple rail splitter above, the resistor divider charges the caps

to 1/2 supply voltage. Without the caps, the resistor divider would have

very poor response to changing loads. The purpose of the caps is to

try and hold the voltage steady. You don't need them in an active

virtual ground.

 

You should also read my CMoy design thread here.

 

Power supply bypass caps are small value (1 to .1uF) film caps that are

placed not more than 1/4 inch(the closer, the better) from the supply pins on op amps and buffers.

Depending on the design, they either go across + and - pins or from + to gnd

and - to gnd. On slow op amps like OPA2134 you might get away with not using them

but on a fast op amp like LT1630 they would be manditory.

If you design is nice and compact, you may get away with "sharing" the bypass caps and

not having to have a set for each I.C.

Reply
post #15 of 84
9/18/10 at 4:04pm
Thread Starter 
Quote:
Originally Posted by Avro_Arrow View Post
Power supply bypass caps are small value (1 to .1uF) film caps that are

placed not more than 1/4 inch(the closer, the better) from the supply pins on op amps and buffers.

Depending on the design, they either go across + and - pins or from + to gnd

and - to gnd.

If you design is nice and compact, you may get away with "sharing" the bypass caps and

not having to have a set for each I.C.

I really hate to ask, but could you explain this more?  I don't know where you mean I should put them.  These power bypass caps aren't connected to the signal at all? Do they just keep a ready charge for the opamps/buffers?  So I place the cap before the + pin? Or is it supposed to be before the + pin and connect to the ground channel?

Also, digikey doesn't have film caps below 68pF...  Can I use ceramics instead?  Something like this?

 

EDIT:

Where in this schematic would they be?

ampschematic1.jpg
I forgot the 100K resistor between the pot and the opamp.

 

EDIT: Ok I think I know what you mean.  It's like the cap used in this pimeta schematic:

pimeta lr.jpg

Since it's a power cap, I'm assuming that a larger value would be better (1µF)?


Edited by shake - 9/18/10 at 5:11pm
Reply
New Posts  All Forums:Forum Nav:
  Return Home
Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › Amp design question