Latest Thread Images

Featured Sponsor Listings

A nice(?), simple chip amplifier

Nov 23, 2009 at 8:25 PM Post #61 of 123
OK. I've tweaked the circuit a little and performed some initial analysis with TINA-TI. Find this below...

4128336949_60aa3bff42_o.jpg

One Channel with Analysis - 100K Square Wave

With C2 at 75pF, the leading edge of the wave was distinctly curved, as predicted/shown by diditmyself. 8.2pF provides the squarest wave.

So. What do you think? Does that wave look pretty good or is there more to do?
 
Nov 23, 2009 at 9:11 PM Post #64 of 123
Quote:

Originally Posted by cobaltmute /img/forum/go_quote.gif
I'd say that looks good enough for a sim - remember reality can be a different beast.

I don't know if you'd need a load sharing resistor before the buffers (like the PPAv1)?


Having read Tangent's comments in his parts selection pages for the PPAV1, I think you're right. Actually, I'm a bit surprised by the PPAV1. I thought it used a discrete buffer....obviously that was a change introduced for V2.
 
Nov 23, 2009 at 9:17 PM Post #65 of 123
Do you already have the parts? If you're planning on stacking BUF634 why not go for a better one instead like LME49600? One LME is cheaper than two BUFs. I'm afraid I haven't heard any of them but people report that LMH6321 sounds better than BUF634 and LME49600 should be an improvement over LMH6321. I have experience from LMH6321, high current opamps and discrete buffers. The advantage of discretes is that you can run them at the bias you want. I prefer class A to class AB. It's warmer, softer and more natural to my ears. Some seem to prefer class AB.
 
Nov 23, 2009 at 9:21 PM Post #66 of 123
Quote:

Originally Posted by diditmyself /img/forum/go_quote.gif
That looks like the square wave you want. You're a fast learner. Isn't simulating fantastic? Noobs like us can design amps!


Yeah it's great playing with the software....although I did get a little annoyed at first when I couldn't get anything to work. I took a look at LTSpice as well but felt that TINA-TI was the best option for me to get up to speed quickly.

As far as the design of amps go, I'm definately standing on the shoulders of giants. Without the work by Tangent, AMB, CMoy and the posters on Headfi, on Headwise and on diyaudio I doubt whether I could have come up with the design myself......although the BUF634 datasheet provided a pretty good start point.

What the sim software does allow is for someone like me to tweak designs, combine ideas and come up with something that shouldn't be too shabby. However, there's a long way to go and plenty of opportunities along the way for me to make a mess of it. With a lot of luck and a bit of help it might turn out alright.
 
Nov 23, 2009 at 9:52 PM Post #67 of 123
Quote:

Originally Posted by diditmyself /img/forum/go_quote.gif
Do you already have the parts? If you're planning on stacking BUF634 why not go for a better one instead like LME49600? One LME is cheaper than two BUFs. I'm afraid I haven't heard any of them but people report that LMH6321 sounds better than BUF634 and LME49600 should be an improvement over LMH6321. I have experience from LMH6321, high current opamps and discrete buffers. The advantage of discretes is that you can run them at the bias you want. I prefer class A to class AB. It's warmer, softer and more natural to my ears. Some seem to prefer class AB.


No, I haven't got the parts yet. I've taken a look over at National and couldn't spot a model for the LME49600. Do you have one? Also, I'm not sure I'm keen on the package.....I wanted DIP8
 
Nov 23, 2009 at 11:12 PM Post #69 of 123
Quote:

Having read Tangent's comments in his parts selection pages for the PPAV1, I think you're right


You're looking in the wrong place.
wink.gif



Check figure 1 of the datasheet

http://focus.ti.com/lit/ds/symlink/buf634.pdf

Or consult the Pimeta parts selection guide.

Part Selection Guide

Quote:

R7

We reserved this part number in the PIMETA for those familiar with the META42 or PPA. In the PIMETA, there is no series resistor between the op-amp and the buffer since the BUF634 has 200 ohms of series resistance on its input already. Hence, there is no need for an “R7”.


 
Nov 23, 2009 at 11:53 PM Post #70 of 123
Quote:

Originally Posted by diditmyself /img/forum/go_quote.gif
That looks like the square wave you want. You're a fast learner. Isn't simulating fantastic? Noobs like us can design amps!


Speaking of noobs...
biggrin.gif


I've also been doing some simulations to adjust the compensation cap value, but I haven't been looking at square waves, I just plotted the frequency response.
Right now, with 33pF, the cutoff frequency (gain-3dB) is around 350kHz and it's just enough to keep the gain equal on all audible frequencies (up to 20kHz).

Should I be aiming for something else?

EDIT: I think I answered my own question, or better, AMB did. On the M3 page the caps values are all aimed in order to provide a 1~2Mhz cutoff frequency, so I need to lower the value of my cap.
 
Nov 24, 2009 at 7:25 PM Post #71 of 123
Quote:

Originally Posted by MisterX /img/forum/go_quote.gif
You're looking in the wrong place.
wink.gif



Check figure 1 of the datasheet

http://focus.ti.com/lit/ds/symlink/buf634.pdf

Or consult the Pimeta parts selection guide.

Part Selection Guide


Hmmm. I was looking at R11, R12, R13 & R14 on the PPAV1 and the associated comments....

Quote:

R11, R12, R13, R14

These resistors balance the current from the op-amp among the buffers, and reduce electrical ringing in the amp. R11 goes with BUF1, R12 with BUF2, etc.

There is some wiggle room on the value of this resistor, but for most purposes 1 KΩ is a good value.

Optional? You may be able to get away with jumpering these, but the amp will not perform as well as it should if you do.


Given the 200R on the input of the BUF634, should I look to place a 800R resistor in front or is it really not needed?
 
Nov 24, 2009 at 9:10 PM Post #73 of 123
Quote:

Originally Posted by Ant1Xr1st /img/forum/go_quote.gif
PJPro, its not a good idea to parallel buf634


Why's that?

Quote:

Originally Posted by Ant1Xr1st /img/forum/go_quote.gif
try the discrete Diamond Buffer - Design


Yeah. I've seen that. Thanks. But I was hoping to stick with the simplicity of IC.

Quote:

Originally Posted by Ant1Xr1st /img/forum/go_quote.gif
no, you just need 100-200ohm on input of any opamp or buffer to prevent self oscillating


Hmm. OK. Thanks.
 
Nov 24, 2009 at 9:28 PM Post #74 of 123
Quote:

Why's that?


Two buf634 cant be 100% indentical, so they will produce a bit more distortion, also you will might need to add approx 10ohm to the output of each buffer
wink.gif

I have compared both of them, diamond buffer has a bit more soft and realistic sound
What are purpuse of using 2 buffers? more class A bias?
If you want a simplicity, is's possible to make a diamond buffer just with 4 transistors.
 
Nov 25, 2009 at 10:42 PM Post #75 of 123
I've not made any progress with the modelling tonight. I've been having trouble getting the rail splitter to work properly. Not only that, I wanted to add a non TI component (LM317) and can't work out how to add it to the tool.

Can anyone give me some tips which will allow me to do this?
 
You must log in or register to reply here.

Users who are viewing this thread

Total: 1 (members: 0, guests: 1)
Back
Top