[sci.electronics] Simple RIAA phono filter

MKUHN@auvm.auvm.edu (Marty Kuhn) (11/08/90)

Ok, here's the problem:
   I've hooked up a (ceramic-cartridge) turntable to my boom-box's AUX input.
This works fine, execpt that records sound much 'tinnier' than they should.
I would guess this is due to the RIAA equalization curve that records are
cut with.  What I want is a simple filter to de-equalize the signal to match
a standard AUX-type input.  This should preferably be a simple R-C passive
filter, so that I don't have to build a separate power supply for it, but if
the signal loss would be too low, a simple single-transistor per channel
active filter would be OK.  Incidently, this circuit does not have to be
super-accurate; this is not rec.audio.high-end here!

Sorry if this is sort of a basic question, but I don't know how the RIAA
equalization curve is defined.  What is the RIAA curve anyway?


Anyway, thanks in advance for your help...

                                 -- Marty

(Insert random quotes, jokes and clever ASCII pictures here)

whit@milton.u.washington.edu (John Whitmore) (11/16/90)

In article <90311.114631MKUHN@auvm.auvm.edu> MKUHN@auvm.auvm.edu (Marty Kuhn) writes:
>Ok, here's the problem:
>   I've hooked up a (ceramic-cartridge) turntable to my boom-box's AUX input.
>This works fine, execpt that records sound much 'tinnier' than they should.
>I would guess this is due to the RIAA equalization curve that records are
>cut with.  What I want is a simple filter to de-equalize the signal to match
>a standard AUX-type input.  This should preferably be a simple R-C passive
>filter, so that I don't have to build a separate power supply 


	The RIAA playback curve is a three-pole curve with corners
at 50 Hz, 500 Hz, and 2200 Hz.  It is flat to 50 Hz, -3dB/octave to
500 Hz, flat to 2200 Hz, and -3dB/octave from 2200 Hz on up.  In
normal filter-specification manner, the piecewise linear curve I just
described is the ASYMPTOTE of the actual curve (which is more easily
implemented).
	National Semiconductor application note AN-64 gives an example
of an RIAA preamp; I'll try to show it here.


Input    C1    |\ U1
------+--||----| \
      |        |  >------+-------- output
      >  +-----| /       |
   R1 >  |     |/        |
      >  |               |
      |  +--^^^^-+-^^^^--+
    GND  |   R2  |  R3   |
         |       |       |
       +-+-+    --- C2   |
       |   |    ---      |
       >   >     |       |
       >   >     +--^^^^-+
    R4 >   >R5   |   R6  |
       |   |     +--||---+
       |  ---
       |  --- C3    C4
       |   |
      GND GND

R1: match the transducer (circa 1Mohm)
R2: zero ohms (read the article; it's there for a reason)
R3: 1.2Mohm
R4: 100 kohm
R5: 180 ohm
R6: 100 kohm
C1: coupling (value not critical); 0.1 uF
C2: .003 uF
C3: 15 uF
C4: .001 uF


Their example uses LM381, an audio 'op amp'; if you use a general purpose
op amp, be aware that the (+) input should NOT remain connected only
to the capacitor C1; you'll need a resistor to ground as well (10 Mohm 
should suffice).

	You'll need two channels, for stereo; dual op amps suitable for
this include LM381, TL072, NE5532.  Similar (simpler) circuits exist
for dedicated audio units (if you can find LM382, an RIAA preamp with
it will only require 4 capacitors and one resistor, per channel).
You'll also need 10 to 30V of power supply (this should already
be available inside the unit you're using).

	Dontcha wish you hadn't asked?

		John Whitmore

francis@cs.ua.oz.au (Francis Vaughan) (11/16/90)

|> In article <90311.114631MKUHN@auvm.auvm.edu> MKUHN@auvm.auvm.edu
(Marty Kuhn) writes:
|> >Ok, here's the problem:
|> >   I've hooked up a (ceramic-cartridge) turntable to my boom-box's
AUX input.
|> >This works fine, execpt that records sound much 'tinnier' than they should.
|> >I would guess this is due to the RIAA equalization curve that records are
|> >cut with.  What I want is a simple filter to de-equalize the signal
to match
|> >a standard AUX-type input.  This should preferably be a simple R-C passive
|> >filter, so that I don't have to build a separate power supply 
|> 

Umm, No. Most crystal/ceramic cartriges are built with a natural RIAA curve
in them. Of course it will probably be a pretty lousy match to the legit
curve, and you may have other problems. Might be worth looking at
impeadances. Maybe  put 10k across the input and see what happens.
If your boom box has an attempt at a graphic equaliser you
can play with that.

						Francis Vaughan.

dunphy@col.hp.com (Dan Dunphy) (11/17/90)

Normally you only need an RIAA preamp/filter for a moving magnet
cartrige. This curve is required becauuse the MM cartiige is
a velocity device, in other words, the highest signal occurs
at the zero crossing of the groove wiggle. The output is
usually only a few millivolts.
A ceramic cartrige is an amplitude sensitive device, and
would require a different type of compensation if any.

whit@milton.u.washington.edu (John Whitmore) (11/19/90)

In article <11200@milton.u.washington.edu> whit@milton.u.washington.edu (John Whitmore) writes:
>In article <90311.114631MKUHN@auvm.auvm.edu> MKUHN@auvm.auvm.edu (Marty Kuhn) writes:
>>Ok, here's the problem:
>>   I've hooked up a (ceramic-cartridge) turntable to my boom-box's AUX input.
>>This works fine, execpt that records sound much 'tinnier' than they should.
>>I would guess this is due to the RIAA equalization curve that records are
>>cut with.  What I want is a simple filter to de-equalize the signal 

>	The RIAA playback curve is a three-pole curve with corners
>at 50 Hz, 500 Hz, and 2200 Hz.  It is flat to 50 Hz, -3dB/octave to
>500 Hz, flat to 2200 Hz, and -3dB/octave from 2200 Hz on up.  In
>normal filter-specification manner, the piecewise linear curve I just
>described is the ASYMPTOTE of the actual curve (which is more easily
>implemented).

	The rest of my original posting varied from 'inappropriate' to
'wrong' :-0 .
	As several folks have pointed out, the intrinsic properties
of a piezoelectric cartridge guarantee that there's 3 dB/octave
droop, and relatively good matching to the RIAA curve, without
much filtering.  In fact, it's only the 'flat' region in the RIAA
curve that requires treatment with a filter.  Something like this


Input ------^^^^^^^-----+---||---+------- Output
              R1        |   C1   |
                        +---^^^--+
                            R2   >
                                 >
                              R3 >
                                 |
                               GND

should do it.  
	R1 +R2 + R3 = Rcartridge (I think 1Mohm is a good guess)
        R2 * C1 = 1/(2*pi* 1.1kHz)
	( R1 + R2 + R3) =  4.4* ( R1 + R3)
	R3 = 47 kohm        

Note that R3 is the internal resistance of the AUX input socket; USE NO
RESISTOR, it's already part of the amplifier!
	If this is truly a mid-fi application, this set of equations
should give something listenable.  For hi-fi, give up!  Buy a good
cartridge (they start at $15 or so) and a cartridge preamp (those
used to be available as black boxes for circa $20).  Then put
the record player in a conspicuous spot in your home, and listen
covertly to CDs.


	John Whitmore