1
\$\begingroup\$

I'm working on a waveform generator for which I'm using the DDS AD9833 chip to generate sine signals and triangular signals. I'm trying to amplify these waves since the AD9833 output is 600mVpp and I need waves up to 10 Vpp. The problem is to amplify these waves without the insertion of harmonics; to do this, I wanted to use the INA129 instrumentation amplifier but, by using a common potentiometer, everything went well, but the gain has to be controlled by a microcontroller, so I wanted to use the TPL0102 digital potentiometer.

The potentiometer works well on its own, but when I connect it as the resistance gain of the INA129, the output is always saturated, even when I set the minimum gain, which is 1.

Details:

  • INA129 powered with +/- 11V
  • INA 129 V- input connected to AGND
  • AD9833's output connected to high pass filter to remoce DC voltage, then, it is connected to the V+ pin of the INA129
  • TPL0102 Vss pin connected to DGND
  • AGND and DGND connected through a 1MΩ resistor
  • TPL0102 controlled through I2C with Atmega 328 microcontroller

How can I correct this problem?

\$\endgroup\$
7
  • \$\begingroup\$ A circuit diagram would help - how can your INA129 be powered from +/- 11V if its V- is connected to GND? \$\endgroup\$
    – Icy
    Nov 5 '15 at 13:00
  • \$\begingroup\$ It could be the self-capacitance of the digipot's pins being an ever present impedance or your signal could in fact be digital cross talk from the digipots control lines injecting into the very sensitive gain control lines of the INAMP. What frequency range are you using from the DDS chip? \$\endgroup\$
    – Andy aka
    Nov 5 '15 at 13:00
  • 1
    \$\begingroup\$ Why are you using an instrumentation amplifier for this? AD9833's output is single-ended; what you want is a regular opamp. AD9833 isn't ideally suited for amplitude control, either; there are other DDSes in AD's range that have reference current inputs that make it much easier to control amplitude. \$\endgroup\$ Nov 5 '15 at 13:42
  • 1
    \$\begingroup\$ @JsMartinez An instrumentation amplifier has no better noise rejection than the individual amps it's constructed from, and likely has lower bandwidth. You can certainly find a suitable single opamp for your purposes. \$\endgroup\$ Nov 5 '15 at 13:49
  • 1
    \$\begingroup\$ If you want to amplify signals up to 1MHz then the INA 129 is a poor choice. To avoid parasitic capacitance is tricky - I use digipots in gain control scenarios built around an op-amp because one pin connects to the op-amp output and the far end pin connects to ground - wiper produces the op-amp feedback signal. I get reasonably flat gains up to 100kHz and gains from 1 to 16 but my digipot is 10k and not 100k. \$\endgroup\$
    – Andy aka
    Nov 5 '15 at 14:06
2
\$\begingroup\$

You don't specify how TPL0102's supply pin is connected (a circuit diagram would be helpful here), but given its absolute max voltage of 7V, it's a safe bet it's not to the 11V supply.

Digital potentiometers only act as such as long as the signals on them are between their supply rails. It's entirely possible that the voltage on the INA129's adjustment pins is outside this range, which will result in the digipot clipping the signal.

Your approach here is a little strange: using an instrumentation amplifier to amplify a single-ended signal by grounding one terminal is kind of perverse. A regular operational amplifier would work fine here; for digital gain control you can either feed the signal through a digipot beforehand as a variable attenuator, use a programmable gain amplifier, or use a multiplying DAC with enough bandwidth (the DDS output becomes the 'reference voltage' for the DAC, and the DAC value controls the output amplitude).

\$\endgroup\$

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.