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.

LMG1020EVM-006: Driving multiple SPL TL90AT08 withTIDA-01573

Part Number: LMG1020EVM-006
Other Parts Discussed in Thread: TIDA-01573, LMG1020

I want to drive LD  SPL TL90AT08 (IF = 40 A; tp = 100 ns; D = 0.01 %, Vop=11V)  with an avalanche transistor. I need to produce  trains of 100 nc  square pulses with 1000 ns delay after each pulse, n= 100 pulses  followed by 10 ms delay after the train is delivered. Osram recommends to use BSP318S FET but it looks like it can deliver max 30A so I want to try EPC2019. Considering  operating frequency of  1 Mz  and 100nc pulse I suppose I should not expect any significant RF issues and I’d like to attempt to drive 3 LDs connected in series. As I have no experience working with nc pulses and Mz frequencies I'd like to test the idea with TIDA-01573 board. I’ve got a number of questions:

 1.The spec for the board frequency range: 0.1Mz-1Mz. I understand the upper limit but I can’t comprehend  why there is a low  limit .  As my trains are run with 100Hz frequency the board should not work. Is it correct?

 2.Will the buffer shorten any pulse to 1 nc duration? If it does can I bypass it?

 3. Why 4 X 0.1 uf caps are used instead of 1 0.4uf? I would have to replace caps on the board to get longer pulses. How do I calculate caps value for 100nc pulse?

  4. The only issue that I can see with 3 LDs is that the pulses get longer and more triangle like. I don’t see any other  problems in the 3X impedance increase (from my experience LDs  considered a resistive load so additional capacitance will come from leads) if anything it should help with ringing which seems to me has a high probability to occur in this circuit. Are there others?

  • Hi Alex,

    1. The 0.1 MHz is a limitation of the resistor load  that you can install on the EVM. The resistor load is there so you can make sure the pulses are correct before you install the laser diode. You can get around the resistor limitation by reducing VBUS to zero.

    2. Section 8.2 of the EVM User's guide explains how to bypass the pulse shortener buffer.

    3. Are you referring to capacitors C6 through C9? Those capacitors in parallel provide a lower ESR than one big cap. For calculating the pulse length, how are you driving the pulses? To adjust the pulse length from the buffer, you would have to change R9 and C11.

    4. I don't quite understand you are asking for here. To prevent ringing and oscillation, be sure to follow the layout guidelines on the datasheet. Our E2E FAQ home page also provides several guides on reducing ringing and oscillations.

    Best Regards,
    Ethan Galloway

  • Hi Ethan,

    Thank you for the answers. 

    1. I still don’t get why 1 ohm resistor can;t be pulsed with lower than 0.1 Mhz frequency. Lower frequency means lower duty cycle which means lower thermal stress

    3.  I'm referring to c6-c9. According to paragraph   of  TI Designs: TIDA-01573  the capacitance depends of pulse duration and frequency but the formula for cap value calculation does not include frequency. 

    Unless I miss something it is impossible to get 100nc pulse from SN74LVC1G08DCKR as its longest pulse is 3.2nc. Do you make buffers that can be configured up to 100nc?



  • Hi Alex,

    1. I misread the specifications and the thermal limitation applies to the 50 MHz limit and not the 0.1 MHz limit. This limit may apply to the pulse shortening buffer. I will look more into this.

    The LMG1020 driver itself is capable of delivering 100 ns pulses with a 1 ms delay.

    3. Capacitors C6-C9 serve to provide enough power for a high current laser diode to function. I think the equation for the output capacitance required is Capacitance = Current * pulse duration / Allowable voltage droop. I don't know your allowable voltage droop, but I get 4 uF for a 40 A, 100 ns pulse with a allowable voltage droop of 2. Because you have a 1 ms delay between pulses, I do not think frequency will be an issue for the capacitors. The capacitors will have adequate time to recharge between pulses.

    I can't find the longest pulse from a SN74LVC1G08DCKR. Can you show me where you see this spec?

    To get a 100 ns pulse, I recommend that you bypass the SN74LVC1G08DCKR and use a function generator to generate the 100 ns pulse. You can also size R9 and C11 to get the SN74LVC1G08DCKR to get the right pulse length, but using a function generator or a capable microcontroller would be simpler.

    Best Regards,
    Ethan Galloway