2
\$\begingroup\$

I'm pretty inexperienced in EE, but I'm adapting the circuit given in this question to amplify microphone inputs to drive an ADC.

I am going to be driving 4 ADC channels (on a Beaglebone Black) from 4 microphones, so I'm planning to use a quad op-amp.

op-amp driver circuit for 1.8V ADC

The upper left of the circuit is a voltage divider providing bias voltage, with a 1MΩ resistor R3 isolating it.

My naive question is this: How much of the bias voltage circuit do I have to replicate per-channel? Can I drive all four channels from the node at the right of a single copy of R3?

\$\endgroup\$
2
\$\begingroup\$

You can use the same divider for all channels but you need to replicate R3 for each channel.

In other words, there can be a single version of R1, R2, and C2, but there should be an \$R3_A\$, \$R3_B\$, \$R3_C\$, \$R3_D\$, where they all connect to the same node on the left, but each connects to the appropriate input on the right.

\$\endgroup\$
7
\$\begingroup\$

The input signal is present on the node at the right of R3, so R3 must be duplicated for each channel (otherwise you would be joining all the inputs together!). R1, R2 and C2 can be common to all channels.

There are two possible issues with using a common voltage divider - bias current, and crosstalk.

OP295 op-amps have a maximum bias current of 20nA each, so in the worst case you could have 80nA flowing in or out of the voltage divider. The DC impedance at this point is 50k (100k in parallel with 100k) so the maximum voltage change that might be caused by bias currents is 50k*80nA = 4mV, which is probably not significant in your application.

Without C2, signals could leak from one channel to another through R3 (greatly attenuated perhaps, but still might be audible). However C2 should shunt virtually all of this crosstalk to ground.

The only thing I would be concerned about is the 1.8V supply voltage. The OP295 is only rated down to 3V.

\$\endgroup\$
  • \$\begingroup\$ I was planning to use an OP490, which is rated down to 1.6V and maximum bias current of 25nA. Looking at the specs of that part, though, I'm now concerned about the slew rate as I'm after ultrasonic signaling (~18kHz). The slew rating is 12 V/ms on a +/-15V supply, and the gain/frequency plot is, I think, telling me it won't amplify at that speed. \$\endgroup\$ – Russell Borogove Feb 15 '15 at 16:00
  • \$\begingroup\$ Also the OP490 has a bipolar output stage with 0.8V saturation - not much voltage swing on 1.8V! You need an op amp with greater bandwidth and rail-to-rail output, eg. TSV634. \$\endgroup\$ – Bruce Abbott Feb 15 '15 at 16:38
  • \$\begingroup\$ "Bipolar output stage with 0.8V saturation" versus "rail-to-rail output" -- what am I looking for on a datasheet to make that distinction? I'm hoping to find a suitable part in a DIP. \$\endgroup\$ – Russell Borogove Feb 15 '15 at 17:03
  • \$\begingroup\$ You want to look for a Voh (output high voltage) near Vcc, and a Vol (output low voltage) near ground. Alternatively, many op amp data sheets have a figure showing peak-to-peak output voltage swings for different supply voltages. Op amps which swing very near the plus and minus supply voltages are called "rail-to-rail" amps. In your case, you might look into the MC33204 op amp. It seems to fit your requirements with two problems - 1) input offset voltages are pretty high, and 2) it's very hard to find in DIP. eBay does have them, and cheap, if you're willing to take the chance. \$\endgroup\$ – WhatRoughBeast Feb 16 '15 at 0:00
  • \$\begingroup\$ If you can't get what you want in DIP, get the SOIC version and use a SOIC to DIP adapter (eg. futurlec.com/SMD_Adapters.shtml). \$\endgroup\$ – Bruce Abbott Feb 16 '15 at 1:48

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.