LTC2757 glitches

I have two LTC2757 DACs. Each has an LT1001 transimpedance amp. I'm generating a square wave with each. The square waves have 0.2V amplitude. I update the DAC values at 100 Ksps, regardless of whether there is a change needed in the output voltage. The square waves are out of phase. Here is a screenshot of the moment when one of the outputs goes high:

Questions:

1. The glitch amplitude here is ~5mV. The datasheet (pg 7 "settling full scale step") shows an amplitude of ~300 uV for a 0 to 10V step. Why are my glitches so much bigger?

2. The glitch amplitude on the green channel gets smaller when the blue channel goes high. Why? The DACs share a data bus, power supplies, and reference.



made more clear
[edited by: justinis at 9:00 PM (GMT 0) on 19 Sep 2019]
  • 0
    •  Analog Employees 
    on Sep 24, 2019 8:24 AM

    Hi,

    Someone is currently looking into your query and will respond to you as soon as possible.

    Cheers,

    Ivan

  • 0
    •  Analog Employees 
    on Sep 24, 2019 4:19 PM

    "settling full scale step" is not the same as glitch energy.  The plot next to the settling time is the plot you want to look at, and it looks like it is 5mV like you are seeing.   

    The reason the glitch changes shape could be any numbers of reasons.  It is likely that when the second channel is turned on the current is split between the two channels lowering the glitch energy in both.  It could also be reference or ground related.  

    If you are concerned with the glitch energy you should only update the DAC when you are changing voltages, not continuously.  This will produce a glitch when the DAC code changes, but not on a 100kHz interval. 

  • When you say the current is split between two channels, what current are you referring to? The current from the reference? The current out of the DAC output? The power supply current? I'm trying to understand why the glitch amplitude of one DAC could be affected by the glitch of another completely separate DAC IC.

  • 0
    •  Analog Employees 
    on Sep 24, 2019 8:30 PM in reply to justinis

    It could be any number of reasons.  The ICs are different but they are running on the same power supply and reference, correct?  So if there is a reservoir of charge that the DAC can draw on and suddenly there are two DACs drawing on it, the peaks will likely be less. Glitch is very much code dependent so when the second DAC changes, it is likely to change the glitch of the first DAC as well. Does the same phenomenon happen if you switch the order of DACs?