• State Resolved
  • Locked Locked
  • Replies 13 replies
  • Answers 1 answer
  • Subscribers 48 subscribers
  • Views 806 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

DAC39RF10EVM: 300ps short pulse application/design

Siyuan Zhang
Prodigy 10 points
Part Number: DAC39RF10EVM
Other Parts Discussed in Thread: LMG1020EVM-006, DAC39RF10, DAC39RF12, LMG1020

Hi, 

Could I utilize the combination of LMG1020EVM-006 + DAC39RF10EVM to design something that can output minimum 300ps short pulse? the main reason of the DAC with high sample rates here is that I like to program the rise time slope. For example, 1ns short pulse is generated,  and I like to change the rise time but keep the total of rise time, fall time and pulse width constant. Thank you. 

Regards,

Simon

  • 0
    TI__Expert 3400 points

    Hey Simon, 

    The DAC39RF10 can be configured to use all 16 lanes of the SERDES for a single DAC channel. In this mode you can use the DAC with 1X interpolation and have 10.24GSPS of real time bandwidth (12GSPS with the DAC39RF12). From a digital standpoint, this would mean pulse widths just under 200ps. (Imagine a simple square wave pattern with only 2 samples per period).

    For something like a programmable slew rate, the pulse width will be longer as you need more samples per period.

    The DAC has around 12GHz of 3dB bandwidth as well, though after 12GHz you still have some workable bandwidth. 

    Regards, 

    Matt

  • 0 in reply to Matthew Kramer
    Prodigy 10 points

    Hi Matt. 

    Thank you very much for your reply. 

    What's the output level of the DAC? 0~reference voltage, which is 0~0.9V?

    Regards,

    Siyuan

  • 0 in reply to Siyuan Zhang
    TI__Expert 3400 points

    Hey Siyuan, 

    With 2X current mode you get about 7dBm into 50ohm and about 1dBm with 1x current mode (at low frequencies before any rolloff). 

    7dBm into 50ohm is ~500mV RMS. 

    1dBm into 50ohm is ~250mV RMS

    For most applications we recommend ac coupling the output through a bias-T. 

    Regards, 

    Matt

  • 0 in reply to Matthew Kramer
    Prodigy 10 points

    Hi Matt

    Thanks for the clear explanation.

    My next question would be: how I can convert the 500RMS voltage to to 5V peak voltage in such a fast speed? The low threshold voltage to trigger the LMG1020 is  1.1V.

    Regards,

    Siyuan. 

  • 0 in reply to Siyuan Zhang
    TI__Mastermind 38690 points

    Hi Siyuan,

    We are reaching out to the groups internally on who can help support you with the LMG1020.

    I will respond back tomorrow with an update.

    Regards,

    Rob

  • 0 in reply to Rob Reeder
    TI__Mastermind 38690 points

    Hi Siyuan,

    I am still working to find a contact for you.

    I will respond back early next week.

    Regards,

    Rob

  • 0 in reply to Rob Reeder
    TI__Mastermind 38690 points

    Forwarding this to the HVP-HPD team.

    Thx,

    Rob

  • 0 in reply to Siyuan Zhang
    TI__Expert 8660 points

    Hi Siyuan,

    The LMG1020 device is a simple analog to analog driver, so as long as the input signals for the device meet the threshold and the VDD = 5, then the device will output. The input signal cannot be modified with our device or EVM, so it is dependent on if the input signal is being generated properly (whether it's from a function generator or the output of another device). If you would like to adjust rise time with affecting fall time and pulse width, then you can just adjust R1 on the LMG1020EVM (higher value = slower rise time) and just use a function generator for the input signal:

    Thanks,
    Rubas

  • 0 in reply to Rubas Khalid
    Prodigy 10 points

    Hi Rubas

    Thank you for your reply. 

    The limit of LMG1020 to my application is that it doesn't have the ability to output a stable pulse less than 1ns. Say 300ps. 

    Regards,

    Siyuan

  • 0 in reply to Siyuan Zhang
    TI__Expert 8660 points

    Hi Siyuan,

    Unfortunately, we do not have another device that can output a pulse that low. The LMG1020 is our fastest device with minimum input to output pulse of 1ns.

    Best,
    Rubas

  • 0 in reply to Rubas Khalid
    TI__Expert 8660 points

    Hi Siyuan,

    I noticed that you rejected my last response. Is there anything else I can help with?

    Thanks,
    Rubas

  • 0 in reply to Rubas Khalid
    Prodigy 10 points

    Hi Rubas

    Yes If you have any ideas to help me to get 500ps pulse. Otherwise, no. Thanks for your reply. 

    Regards,

    Siyuan

  • +1 in reply to Siyuan Zhang
    TI__Expert 8660 points

    Hi Siyuan,

    There may be a way to generate short pulses by inputting two pulses into the LMG1020, one on IN+ and another identical pulse with a 500ps delay (or whatever your desired pulse is) on IN-. Please refer to section 8.2.2.2 and figure 10 on the LMG1020 device datasheet to visualize this behavior (figure attached for your convenience):

    EDIT: Please note that the datasheet only guarantees the LMG1020 operational performance to 1ns pulse. While this method may provide a shorter pulse, we do not guarantee the LMG1020's performance to be as if it were following the recommended guidelines in the datasheet.

    Thanks,
    Rubas