[enemigo]

I am planning on building some amps with buffered power supply and buffered outputs.

The TLE2426CLP will not be used for the power rails due to it's temperatur sensitivity. The TLE2426ID will not be used since it's a SO14 package, and I have no plans to solder that. Could use the BUF634T, but along with 2 output buffers, using the BUF 634 will be far too expensive.

So, I guess I will make 3 voltage followers using cheap opamps instead. I had a look at the LMH6642 and the AD817, but I can only get these in SO8, and I definately need DIP. I want an opamp that will work at lower voltages than that opa2132/2277. I could just use the opa2132 for buffers, my dealer (rs) don't have the opa132. But I figure this could be done cheaper?

Thanks
Knut

 

[Syzygies]

I remain intrigued by Figure C11 on p. 91 reprising Figure 101 in

http://www.stanford.edu/class/ee122/...ets/an47fa.pdf

(link fixed)

...an application note highly recommended by PPL in Diamond Buffer Abstracts. This is a fewer parts precursor to the diamond buffer that Jung describes and many people build.

The focus of the DIY activity I've found online has been to build discrete buffers that would improve the sound of a PPA, rather than to build discrete buffers that approach the sound of a Pimeta and price out closer to a CMoy.

This buffer is basically six $0.60 transistors and a few resistors. The true cost would be the time and effort it takes to match the transistors. I'd bet that a CMoy-grade amp with Jung multiloop topology and these buffers would sound better than a bare CMoy?

If you draw the PIMETA and MINT schematics from memory a bunch of times, then start swapping in discrete buffers, do you see anything you want to build? I'm still trying to understand the buffers better.

 

[enemigo]

Thanks for your reply! Unfortunately your link does not seem to work...

K

 

[morsel]

The TLE2426 is a fine choice for a virtual ground reference. Intersil HA3-5002 buffers are somewhat cheaper and better than TI/BB BUF634 buffers, if you don't want to go discrete. Opamps make bad buffers for audio purposes.

 

[enemigo]

Intersil HA3-5002 might be what I'm looking for, RS don't have it though. They have the HA3-5033-5 at 17.5mhz, but it seems to need +-5v. The HA3-5033 is also called "HA3-5003-5" on the RS pages, don't know if that's an error.

I will test the TLE2426, but I fear 0 degrees Celcius (273K, whatever Fahrenheit) might be a problem for outdoorwear during the winter.

K

 

[Syzygies]

Quote:

Originally Posted by enemigo Unfortunately your link does not seem to work...

http://www.stanford.edu/class/ee122/...ets/an47fa.pdf

(link fixed)

 

[amb]

Quote:

Originally Posted by Syzygies I remain intrigued by Figure C11 on p. 91 reprising Figure 101 in http://www.stanford.edu/class/ee122/...ets/an47fa.pdf

This is basically just the generic diamond buffer, except it lacks the constant current sources at the emitters of Q1 and Q2 (only 1K resistors used), and it also has an output current limiter (Q3 and Q4).

 

[__redruM]

Quote:

Originally Posted by enemigo Intersil HA3-5002 might be what I'm looking for, RS don't have it though.

Newark has it:
http://www.newark.com/NewarkWebComme...KU=06F5241&N=4

I don't know if they ship to Norway, but they do have it...

 

[enemigo]

Syzygies (and everyone else):

Figure C11 in the document looked promising, it might take up too much space for a portable amp, but it should be interesting for my homeamp! Some questions in this regard.

1. This circuit shows one channel of an amplifier with a buffered output, only lacking the powersupply? Or is the LT1220 part of the buffer meaning the circuit only shows the buffer?

2. I can't find the LT1220 OPA, so I guess I will be using a different one, possibly the OPA627 if that works?

3. What voltages will this circuit work with? I see +-15v is specified, but that would depend on the opamp used I assume? Will this circuit work with an OPA2132 and a 9v source?

4. There will be a voltage drop in this circuit due to the constellation of the 6 transistors. Will this voltage drop be to big for a 9v (+-4.5v) source?

Thanks
Knut

 

[Syzygies]

I didn't intend my suggestion as "ready to build", though I too want a home-etched through-hole design, ground channel and Jung multiloop topology, basically a modified PIMETA made as simple as possible, wrapping three discrete through-hole buffers. I'd etch two-sided, with the component side mostly ground-plane, following the suggestions in the datasheets I've been reading.

Walt Jung's Realizing High Performance: Buffers (Part II) gives part numbers for different applications. peranders' SMD Diamond Buffer is the most complete description I know of for a buffer designed especially for headphone amps; perhaps Glassman will document his modifications so we can understand and reuse them. Sijosae has a beautiful, small Jung buffer, though he confesses to not understanding its oscillating properties in a Jung multiloop topology. Hey, I don't either!

I admire CMoy's amp for what he doesn't do as much as what he does do. I'm not sure that simply rolling back the clock to earlier designs is the best way to "subtract" from Diamond buffers.

The SDS Labs Headphone Amp deserves close study, it wraps a very simple buffer into the feedback loop, yet the M^3 team speak of this amp as a source of inspiration.

In any case, I'm not building anything I don't understand. In particular, no solder fumes until I can reproduce ppl's remarks via simulation and analysis. I'm learning to simulate buffers in SPICE to take advantage of the sophisticated datasheet transistor models for different part numbers, though I'd much prefer to be working in Mathematica, where I was able to obtain closed-form formulas for parameters for charger designs. (These were piecewise-linear models, so resistor values for given charging curves satisfied systems of polynomial equations, which Mathematica could solve for me. In contrast, I don't see how SPICE can solve for design parameters for me.)

 

[enemigo]

Thanks for the links, I'm sure I'll find some interesting reading.

I posed the questions to get a better understanding of how the buffer works, and will not necessarily be adding this circuit to my cMoy

I drew the buffer to fit the horizontally leading protoboard, as I find that easiest to understand, and suddenly this circuit wasn't that complicated anymore. I am learing a lot from this!

Knut