LM4562 Passive RIAA

[jackinnj]

Looks like the capacitors in the RIAA network should be switched. The 47/33nF combo should be attached to the bottom of R7 as shown below, the 27.2nF combo is attached to the non-inverting input of the second amplifier. (Yeah, I did point this out to National, but haven't heard back from them.)

Looks pretty good with the theoretical values -- there isn't a LM4562 model in the public domain yet so I used the values from the product file. The min/max on the scale is 0.2dB.

An externally hosted image should be here but it was not working when we last tested it.

 

[jackinnj]

This is the response as National originally published -- note that the Y-Axis is -5dB to +10dB.

An externally hosted image should be here but it was not working when we last tested it.

 

[jcx]

It would also be fun to hear what Bob Pease would have to say about the prominent positioning of a topology without technical merit that only panders to the prejudices of a particularly ill informed corner of the “no feedback” crowd – and is even worse technically than the obvious “active” equalization alternatives when one accepts some of the premises these confused “audiophools” might spout as justification for the topology

 

[jackinnj]

It would also be fun to hear what Bob Pease would have to say about the prominent positioning of a topology without technical merit that only panders to the prejudices of a particularly ill informed corner of the “no feedback” crowd – and is even worse technically than the obvious “active” equalization alternatives when one accepts some of the premises these confused “audiophools” might spout as justification for the topology

uuuh, I dunna know but Walt Jung discusses the topology at length in "Op Amp Applications Handbook".

 

[jcx]

Walt is just being complete, assuming the gain blocks are perfect and giving them no further consideration is the logical error in recommending this design

The LM4562 is a high gain, "low bandwidth" device with integrating frequency response with a corner frequency below 10Hz

from 20 Hz to 20 KHz the open loop gain and the feedback of the "flat gain" blocks will fall 1000x

I think some of the usual criticisms of feedback amplifiers might include "falling feedback factor", "rising distortion, output impedance with frequency", and Otala's PIM

all of which are worse with the flat gain/ passive eq topology than if you implement the RIAA roll-off as active feedback

then of course you're feeding the input with a pre-emphasized signal that boosts 20 KHz by 40 dB over the 20 Hz level – to the point that the circuit has to “burn” some S/N at high frequency to keep any overload margin

I suppose you could argue that the LM4562 is so good that this bad topology would make the LM4562 superiority stand out from most other op amps in the flat gain blocks

 

[jackinnj]

-3dB @57kHz ---- +/-15V Rails, 2K load

 

[jcx]

???

"bullet" items, sept 2006 data sheet LM4562:

Gain Bandwidth Product 55MHz (typ)
Gain (RL = 600Ù) 140dB (typ)


55 MHz / 10^7 ~= 5.5 Hz ol gain corner freq



see also "Gain Phase vs Frequency" p22 of 33


this op amp's open loop gain is close to a pure integrator for all audio frequencies

 

[ahaja]

Hello.
One month ago I made a few pcb's for this preamp, based on National pdf (Lm4562). 2 weeks ago I soldered elements and gave him a chance ... played good. But I got distortions, sth like dirt in sound. Then I realized that feedback loop is too big on pcb ...and I soldered smd 3320R beetwen opamps pins - that was that . I found this post on forum and I changed positions caps in riaa - bingo !!!! - jackinnj was right . Now sound is out of any disstortions; no hum; big, deep bas; vocals are very good; stage is also deep and large; is just beautiful music.

PSU: old russian transformer (2x15V); schottky diode (8) with bypass caps on each; chokes; 12x5600uF/25V (Rubycon). LM317/337 as close as it was possible to opamps (on preamp board).

I changed also gain - instead of 150R I placed 100R - with my Ortofon 510mkII (3mV output) is great.

Here are pictures

>jackinnj - thanks for idea.

 

[Johan Potgieter]

It would also be fun to hear what Bob Pease would have to say about the prominent positioning of a topology without technical merit that only panders to the prejudices of a particularly ill informed corner of the “no feedback” crowd – and is even worse technically than the obvious “active” equalization alternatives when one accepts some of the premises these confused “audiophools” might spout as justification for the topology

Excellent! (wish I had said that) .... and if I may be a rebel with a cause. I have always wondered about the exemplary design of RIAAs down to say 30 Hz ... and then just flattenig out further down. Why not make use of the opportunity already there with several time constants within a feedback loop to include useful rumble filtering below whatever frequency one desires?

I can almost sense the lifting of several eyebrowns referred to in Jcx's quote: "One buys a decent turntable, of course ...."

Yea - but what about the occasional bump, etc., attempting to dislodge any half-decent low frequency driver? Apart from the point that I would rather invest a few $'s worth of components to have the same effect as an extra few hundred $ in the price of a turntable. In the process one could also up response a dB or 2 from say 60 Hz down to cut-off frequency, to compensate for most loudspeakers' tendency to begin falling off there. I have yet to not find an advantage in this kind of sub-audio topology. Academically probably to be frowned on; audio sense IMO.

Regards.
_________________________________
Edit: Sorry, did not notice that the original was several months ago - my post hopefully still relevant.