THAT1646 Operation From A Single Supply Voltage

[mediatechnology]

Here's one way to operate and bias a THAT1646 for single supply operation:


Image courtesy of THAT Corporation

If the input is at or near the same DC potential as Vref (the one-half Vcc potential at the output of U2A) then C2 and R1 may be elimanted and the input direct coupled.

R3 and C4 minimize the series inductance "seen" by pin 3 of the THAT 1646 to prevent oscillation.

The MC33178 is an ideal low-cost, relatively low power, and low noise rail splitter. C1 and C3 may be sized based upon the expected load impedance.

For multiple 1646s using a shared rail splitter providing 1/2 Vcc Vref:


Image courtesy of THAT Corporation

[nereis]

Is there perhaps an error in the first image?

The first image shows R2 & C4 in series acting as a snubber, while U2A feeds pin 3 directly.
I suppose if output from U2A is close enough and the opamp output impedance at HF is reasonably low this should work, but I am not so sure what R2 + C4 does here.

The second image shows the virtual ground being fed to pin 3 via a 47R isolation resistor, while C2 decouples pin 3 to the negative supply.

From my limited electronics knowledge I am guessing that R2 & C4 values are rather empirical, and I should be fine substituting other small caps and resistors in there place, or am I totally off? I am thinking of 120R, 100nF ceramic disc or 220nF MKTs I have laying around.

[JR.]

I'm sure Wayne will drop in and tell you from the horse's mouth, but it is not unusual to shunt across an opamp output like that to insure a low impedance at very high frequency where cheap opamps run out of gain bandwidth.

The 1646 is a wide band part, and I wouldn't assume their app notes to be wrong. While there is nothing wrong with asking.

JR

[nereis]

Actually I have only DRV134 to play around with, but I remember reading here that they are principally the same.
Anyhow, I'd love to hear more about these things!

[mediatechnology]

nereis - Sorry been AWOL.

Both figures are correct though different in the location of the series resistor.

It's been about 5 years since I did both. I need to go back and look at my notes.
I seem to recall that in the first figure connecting the op amp output directly to pin 3 allowed <<47R impedance at LF to realize maximum output balance in the OutSmarts bridge.

With multiple 1646s sharing a bus, in the second figure, this wasn't practical.

Bottom line is that it doesn't take much capacitance at pin 3 of the 1646 to reduce the output "inductance" (rising impedance vs frequency) of the op amp.

The DRV134 may not require the capacitance.

[nereis]

Thanks for the info, I think I'll still stick a small cap in there just to make sure it works.

I contemplated over using TLE2426 as rail splitter until I saw the Output Impedance vs Frequency plot in the datasheet, the rise starting at 200hz looks quite bad compared to most opamps.

[mediatechnology]

I think it may be a little noisy too. IIRC that's why I didn't use the TLE2426.

The MC33178, considering it's low power, had pretty decent performer as a rail splitter and LP audio op amp.
A 5534 works well but has considerable Icc compared to the 33178.