Post Go back to editing

LTC2666: Voltage output glitches when outputting negative voltages

Category: Hardware
Product Number: LTC2666

I am observing glitches when updating outputs on a LTC2666 DAC (12-bit, part number LTC2666CUH-12-PBF), specifically for negative output voltages. These glitches are small (~5mV, 1usec) but are specific to writing DAC values below 0V. Details and description of my system are below. My questions are:

  1. 1. Is this behavior expected? I didn't see it documented in the datasheet, and haven't seen this from other DACs, so it comes as a bit of a surprise for me.
  2. 2. Are there obvious steps I can take to ameliorate/eliminate these glitches?

Thanks in advance for your help!!!

Here's an example, in which I'm driving a +-10mV 1kHz sine wave:

Zoom:

Operating conditions:

  • Using command code 0011 (Write Code to n, Update n) to update a single DAC channel
  • Voltage rails are +-15V
  • Softspan range is +-2.5V
  • Using internal reference.
  • have tested this on multiple channels with the same result. I have also tested this with channels with no load attached and verified this result remains (and isnt obviously a result of some weird load behavior)
  • I'm updated the DAC at about 250ksps.

Here's a snippet of how my LTC2666 is configured:

I am not using DAC_TOGGLE or LDAC (they are tied low and high respectively by the microcontroller).

Apologies for the necessarily somewhat incomplete description of my system, and please let me know if there's more info I can provide here.

Thread Notes

Parents
  • Hi  ,

    Your query has been acknowledged and queued with. We'll get back to you after some investigation.

    In the meantime, you can try to connect a pull-down resistor to the MUXOUT pin as it is a high impedance pin if mux is unused.

  • For what it's worth, a 10k pulldown resistor on MUXOUT seems to have no effect (oscilloscope traces below). I don't follow the logic for why this would have helped, so I guessed 10k somewhat randomly, please let me know if another value seems more likely to help.

    Any updates from your end?

    Thanks so much!

    No pulldown, glitches clearly seen when outputting negative values, (on par with mid-scale glitch!):

    With 10k pulldown: seemingly no difference.

Reply
  • For what it's worth, a 10k pulldown resistor on MUXOUT seems to have no effect (oscilloscope traces below). I don't follow the logic for why this would have helped, so I guessed 10k somewhat randomly, please let me know if another value seems more likely to help.

    Any updates from your end?

    Thanks so much!

    No pulldown, glitches clearly seen when outputting negative values, (on par with mid-scale glitch!):

    With 10k pulldown: seemingly no difference.

Children
No Data