LME49710 -- new passive RIAA Circuit

[jackinnj]

National hasn't yet released this as an application note (AN-1651) -- but it is on their website -- an LME49710 Passive RIAA amplifier for MC and MM transducers -- note that in their advert "tip in" in the July 5, 2007 issue of EDN they had some part numbers screwed up -- we can forgive Joe Curcio for this editorial oversight -- they use a pair of LME49710 as filters for the power rails.

http://www.national.com/nationaledge...l07.pdf#page=2

[jackinnj]

The circuit as shown sims out acurate to -0.04, +0.01 dB using standard value parts. The difference between the schematic shown and the text relates to C3 --- the schematic shows 10pF, the text uses 100pF.

[glennb]

Interesting app note. Joe writes in a nice style which recognises the unique art of audio amplifier design with respect to both objective and subjective performance.

He is right that an amplifier's essential task is to modulate the power supply to provide current into the load in direct response to the input signal. The power supply must therefore present a low and flat impedance across the audio bandwidth for the best result.

Its a pity he doesn't elaborate on the use of a pair of LME49710 as power supply regulators, such as the PSRR, noise and impedance spectra.

The ESL amp is a bit of a worry because the global feedback is only taken from Q2 of the push-pull output mosfets, the non-linear output of Q1 is left to flap in the breeze. He sings the praises of ELSs for 2 columns, and then devotes less than a column to the amplifier design.

[nuvistor]

Originally posted by glennb
The ESL amp is a bit of a worry because the global feedback is only taken from Q2 of the push-pull output mosfets, the non-linear output of Q1 is left to flap in the breeze. He sings the praises of ELSs for 2 columns, and then devotes less than a column to the amplifier design.

Please note the balance pot (R11,200 ohm), this does introduce significant source degeneration which should help linearize the output stage, the output stage transconductance is about 1/210 mho but would be 1/10 mho without R11.

It would have been nice for them to include one of the fully differential ESL driver designs, and measurements for all of the designs. My guess is that the published ESL design was judged "most musical" due to its harmonic distortion spectrum being dominated by 2nd harmonic. Also some construction detail would be very helpful, another of their high performance audio app notes for a power amplifier treats this well.

[Gordy]

Originally posted by nuvistor


another of their high performance audio app notes for a power amplifier treats this well.

Which one please?

[glennb]

Originally posted by Gordy
Which one please?

In LM4702 product folder, http://www.national.com/pf/LM/LM4702.html :

Application Note AN-1490: LM4702 Power Amplifier (PDF, size 1034K, 10 pages) http://www.national.com/an/AN/AN-1490.pdf

Application Note AN-1645: LM4702 Driving a MOSFET Output Stage (PDF, size 656K, 24 pages) http://www.national.com/an/AN/AN-1645.pdf

HTH

[Gordy]

Many thanks Glenn.



Regarding the original thread...
The LM4562 as mentioned in the application note (AN-1651) seems to have issues according to Mr. D. Self: http://www.dself.dsl.pipex.com/ampins/webbop/LM4562.htm .

[glennb]

Originally posted by Gordy
The LM4562 as mentioned in the application note (AN-1651) seems to have issues according to Mr. D. Self: http://www.dself.dsl.pipex.com/ampins/webbop/LM4562.htm .

Yes, I saw that page many months ago.

quote "The LM4562 working in shunt feedback mode (to remove input CM distortion) at a gain of 2.2x with various loads. Test level is 7.9 Vrms output. (approx +20 dBu) +/-15V supply rails. The THD plot is indistinguishable from the analyser output when it has no external load."

Shunt feedback AKA inverting. That's a very credible result. Sub 0.0007% THD across the low audio range.

The issue with common-mode induced distortion in Series feedback mode AKA non-inverting is interesting.

quote "This is not a unique problem with the LM4562. Applying the same test to the 5532 gives almost identical results."

I wonder what NS would have to say.