Designing a universal diff-in/diff-out Head Amp

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Distortion figures in the image below, for an input signal of 5mVrms or 0dB@10Khz. Left for the current input version and to the right the voltage input version, both as supplied from the 48 Volt Phantom Power with a gain of 20dB
 

Attachments

  • Curl15.jpg
    Curl15.jpg
    403.9 KB · Views: 424
This circuit doesn’t work for several reasons.
1) the Cart has no reference voltage, the Nul in my case, so everything will collapse to Gnd.
2) when you give the Cart a voltage somewhere in between gnd and opamp supply, a DC current will flow through the Cart, that’s exactly what we don’t want.
3) when Cart is playing, opamp tries to Produce a voltage around zero volt, which is not possible with V- at Gnd.

So to let this work, The Cart should get a reference voltage somewhere in between and R2 should be connected to this same voltage. And there you are, that’s what I have done.

Hans
 
Scott, one could even consider to leave the MM impedance convertor and to plug the MM Cart straight into the 2i2's 3KOhm Mic input, without even using Phantom Power in that case.
You will need a different Riaa correction network just like the Aurak design.
With a simple formula like (X*Lcart - Rcart), the missing link in the form of just one resistor can be added to match the digital Riaa filter.


Hans
 
You will need a different Riaa correction network just like the Aurak design.
With a simple formula like (X*Lcart - Rcart), the missing link in the form of just one resistor can be added to match the digital Riaa filter.

You could but every cart and sound card would need a different set of numbers and the noise hit might be audible.
 
Last edited:
You could but every cart and sound card would need a different set of numbers and the noise hit might be audible.
With just one resistor all differences between sound cards and Carts can be solved.
But, it wouldn't be me if I had not simulated the effect on noise, and the effect was a S/N of 72.7 dBA, when using the 2i2 figures for two different cartridges.

What I used was a pole at 71.4usec.
For a 500mH cart with a Rcart of 600Ohm, this meant adding one series resistor of 400 Ohm
For a 580mH Cart with Rcart od 460 Ohm, a series resistor of 1660 Ohm had to be added.

The Math behind, 1) (600+400+6k)/0.5 = (460+1660+6k)/0.58.
For another sound card the 6K (=2x3k) might have another value and can be adjusted with the extra series resistor.

The question is though, is 72.7 dBa good enough or not.


Hans
 
When putting the pole at 75usec, the third Riaa time constant might be completely left without further correction.
The added resistance for a 580mH/460Ohm Cart becomes in that case 1273 Ohm, and for the 500mH/600Ohm cartbecomes 67 Ohm.
It might become a bit narrow for some cartridges and it has no advantage on noise performance.

Anyhow, the calculated 72.7 dBA is with the Cart connected, where most suppliers specify with input short circuited.
And be aware, noise is still quite a bit above the LP's surface noise.
To conclude: it costs no penny to try, and you can't damage anything.


Hans
 
OK, I ran this last night and I hope it confuses others as it did me at first. Pink noise from one of the STR series LP's, I only have one TRS to XLR adaptor so I could only do one channel of the 2i2 direct differential in and the same channel through my pre-amp which measures with a synthesized RIAA pre-emphasis +-0.1dB. This is a Grado MI cartridge.

Plot 1 - Pink is direct in and blue is the RIAA pre-amp. Please bear with my reasoning. What I see here is with direct in we have the pre-emphasis acting like a negative "pinking" (RIAA endpoints are -20dB and +20dB over 4 decades). What's left are the two bumps when you straighten out the RIAA. The blue plot should be straight (the 14K peaking has been in every Grado I have ever tried) and 1/f (pink). All the plots are normalized to be equal at 1kHz.

Plot 2- I used an un-pinking FFT filter on the previous plot. I am disturbed by the slight residual tilt on the RIAA pre-amp, so I need to check a more modern pink noise test record, but it basically makes sense.

The noise floor with reference to a blank groove was within 1 dB on both plots.
 

Attachments

  • riaa1.jpg
    riaa1.jpg
    158.6 KB · Views: 452
  • riaa2.jpg
    riaa2.jpg
    162.2 KB · Views: 463
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.