• State Resolved
  • Locked Locked
  • Replies 8 replies
  • Answers 1 answer
  • Subscribers 38 subscribers
  • Views 998 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.

TIDA-01573: usage with diodes, that have cathode connected with metal case and compare between GaN vs. avalanche transistors.

Nils Thernstrom84
Prodigy 50 points
Part Number: TIDA-01573
Other Parts Discussed in Thread: LMG1020, LMG1210

Hi, 

I'm designing a pulse light source for my master graduation project. I'm using TIDA-01573 as reference design. I'm also basing on application notes delivered by LASER diodes producers - like:  (page 3 fig. 2) or  (page 1 fig. 2).  These designs are little different, probably to allow using avalanche transistors I think. What are advantages of using GaN transistors over avalanche transistors, if any are? In one of TI documents was comparison between MOSFET and GaN FET, but I don't found comparison with avalanche transistor. However in both designs LASER diode case (connected with LASER diode cathode) can't be grounded. I want use LASER diode in TO18 5.6mm case ( leads inductance around 5nH) due to cheap and easy lens holding. So that's mean there will be parasitic capacitance between diode heat sink + lens holder and device ground. Should I care about it as potential noise source (big metal element with high speed potential changes) and as factor that can cause longer rising times (in compare with same case/ inductance, but without anode or cathode connected to case)? 

Best regards, 
Nils 

  • 0
    TI__Genius 12990 points
    Hi Nils,

    thanks for reaching out and welcome to e2e! Im an apps engineer on GaN drivers and hope to help you. Thanks for sharing your studies.
    I found a paper that talks about using BJT vs FET for nanosecond laser pulses.
    jultika.oulu.fi/.../nbnfi-fe201801051097.pdf

    The paper talks about how the BJT has a slower current pulse waveform at its edges (The slowly rising current pulse shape leads to jitter). But in order to get the same peak current pulse from a similar layout is to almost double the bus voltage to allow the current to reach a peak equal to whats capable when using a FET at half the bus voltage.
    The lower bus voltage when using FETs also lead to less power loss, smaller FETs.
    The faster rising of the edges makes for higher frequency and less time walking (or variation in timing of the received echo)
    GaN FETs have much better FOM over MOSFETs like lower Rdson and gate charge for the same package size FET.
    More differences can be found in the TIDA table2 and fig3.
    www.ti.com/.../tidue52.pdf

    Whats your application/EE?
    since the edges are essentially becoming faster the noise is worse,
    are you worried about EMI from your LD heatsink?
    what peak power or current do you need?

    thanks,
  • 0 in reply to Jeffrey Mueller
    Prodigy 50 points
    Hi Jeff,
    My application is rangefinder with direct time measurement. Range it's not specified. In projects I want to compare different light sources ( LASER vs. LED) and different detectors (PIN PD vs. ADP PD vs. Silicon photomultiplier). In this case I want drive a pulse laser diodes 905nm (Ifpeak = 35A, Vf @If = 12.8V, Optical power 75W) and 1550nm (Ifpeak = 65A, Vf @ If = 8V, Optical power 20W) in TO18 5.6mm package. Project construction will be modular to easy replace light emiter and detector circuits. However I want keep a possible compact design.

    >since the edges are essentially becoming faster the noise is worse,
    >are you worried about EMI from your LD heatsink?

    Yes, I worried about 2 things. First about mentioned parasitic capacitance between heat sink/lens holder and ground. In TI design this capacitance will be parallel to switching FET, in reference designs from LASER manufactures this capacitance will be parallel to LASER diode. In both cases it can increase rising times. Second worry is about EMI emission form heat sink/lens holder. Is it a real problem or I should focus in EMI emission from current loop? Do TI have some shielding recommendations for LIDAR applications?

    Thanks for attached documents, I will check it.

    Best regards,
    Nils
  • 0 in reply to Jeffrey Mueller
    Prodigy 50 points
    Hi Jeff,
    My application is rangefinder with direct time measurement. Range it's not specified. In projects I want to compare different light sources ( LASER vs. LED) and different detectors (PIN PD vs. ADP PD vs. Silicon photomultiplier). In this case I want drive a pulse laser diodes 905nm (Ifpeak = 35A, Vf @If = 12.8V, Optical power 75W) and 1550nm (Ifpeak = 65A, Vf @ If = 8V, Optical power 20W) in TO18 5.6mm package. Project construction will be modular to easy replace light emiter and detector circuits. However I want keep a possible compact design.

    >since the edges are essentially becoming faster the noise is worse,
    >are you worried about EMI from your LD heatsink?

    Yes, I worried about 2 things. First about mentioned parasitic capacitance between heat sink/lens holder and ground. In TI design this capacitance will be parallel to switching FET, in reference designs from LASER manufactures this capacitance will be parallel to LASER diode. In both cases it can increase rising times. Second worry is about EMI emission form heat sink/lens holder. Is it a real problem or I should focus in EMI emission from current loop? Do TI have some shielding recommendations for LIDAR applications?

    Thanks for attached documents, I will check it.

    Best regards,
    Nils
  • 0 in reply to Nils Thernstrom84
    TI__Genius 12990 points

    Hi Nils,

    what TI app note are you referring to and what laser manufacture app note are you referring to? The parasitic capactiance of a LD heatsink is between the LD node and whatever other node you decide to attach the heatsink,

    Here is a great note on GaN design considerations:

    It talks about the greatest source of EMI being CSI and to make this parasitic inductance as small as possible with the help of small packaging and tight layout. Noise can manifest as false triggering the GAN or slower switching times. Attaching the heatsink in the right place can reduce common mode EMI.

    does this help? let me know if you have any more questions,

    thanks,

  • 0 in reply to Jeffrey Mueller
    Prodigy 50 points

    Hi, 

    So I decided to modify driver circuit to half bridge configuration, which allow me to connect diode case - cathode directly to device ground, as in picture below. 

    In off state (between triggers)  M1 FET  will be open and shorting LASER diode, also bootstrap capacitor will be charged. Before trigger S1 will be turned off to avoid current flow across bootstrap capacitor/high side of driver and LASER diode. After trigger M2 will start conduct and close LASER diode current loop. What do you think about this design for LIDAR application? I select LMG1210 driver. LMG1020 have better performance for this application, but I have no Idea how to made level shifter using discreet components. High side virtual ground at point A have no reference to device ground and it's only coupled with device ground by parasitic capacitance of M1 FET and capacitance of LASER diode junction. I also need to avoid current flow across laser diode. I'm also not sure about isolated interfaces. Optical signal insulators are mostly slow, and I don't know how magnetic or capacitance coupled insulators will work near LASER diode current loop. Could you confirm that LMG1210 will work in this circuit? 

    Best regards, 
    Nils 

  • 0 in reply to Nils Thernstrom84
    TI__Genius 12990 points

    Hi Nils,

    thanks for this update!

    this should work, S1 is not needed since 1210 has internal bootstrap switch that turns on 10ns after LO has made the path.

    Can you daisy chain 1020 to the HO-HB-HS output of 1210? the 1020 can ride on HS and use HB as VDD with a closely placed cap.

    This will get you the split outputs and high lidar current and the level shifter from 1210.

    what frequency are you using? what pulse width?

    what are you most concerned about right now? do you still need signal isolation with 1210 CMTI capability?

    thanks,

  • 0 in reply to Jeffrey Mueller
    TI__Genius 12990 points
    Hi Nils!

    any update on the above?
    why have you chosen to drive lidar LD highside rather than lowisde?

    thanks,
  • 0 in reply to Jeffrey Mueller
    Prodigy 50 points

    Hi, 

    Sorry for long reply.  I want achieve 5 to 50ns pulse with 1 to 50kHz repeating. High side will allow to connect LASER and lens holder to device ground. I will check it on prototype. Selected TO-56/ TO-18 package have pin inductance around 5nH, so I don't expect big difference in rise time between low and high side driver. For final device I want use low side driver, but with LASER in ceramic SMD package. Now I don't have more questions. I think we can close this case. 

    Thanks and best regards, 
    Nils