I have a load which is an external trigger input of the Highland Technology T130 single-channel picosecond EOM driver (technical manual - Google Drive link) which is 50 ohm and it requires a pulse with voltage between 1V to 5V and ideally with a few ns rise time.

As a source to the external trigger input, I want to use the 3.3V PWM output of an STM32F302 microcontroller. But I'm a bit worried about the very slow rise time of the microcontroller. For better performance, I was told to use an external trigger source which has few ns rise time and also the PWM output I guess cannot source a 50 ohm load.

So instead of directly coupling the microcontroller's PWM output to the device's trigger input, I want to first buffer the microcontroller's 3.3V PWM output by using this driver(SN75123N) and then couple the driver's output to the device's external trigger input.

It seems to me the SN75123N is designed for 50 ohm loads so it can source enough current in my case. And in my case the threshold voltage is 0.5V. But I cannot infer the estimate rise time of the SN75123N output. In my case the distance's between the driver and the device can be a few centimeters.

Would this driver be better to use than just directly coupling the PWM? And how can the rise time of the driver be estimated?

  • \$\begingroup\$ It can source the current you want. But it's not a Schmitt. Look at the section on switching characteristics and look at what they use to drive the test circuit. \$\endgroup\$ Sep 10, 2023 at 23:25
  • \$\begingroup\$ If you want to take a signal with a slower edge and make it sharper you need a Schmitt to achieve that. You may only require one edge to be fast -- I would not know because you didn't say and I don't want to read that link to find it. But to get a fast edge (one or both) you need a Schmitt. That's one of the services those things perform for you. And I gather that is what you need. It is not only about sourcing into a 50 Ohm load. You are right to worry about the rise time, in short. \$\endgroup\$ Sep 10, 2023 at 23:43
  • \$\begingroup\$ Perhaps the use of all channels paralleled with this part ti.com/lit/ds/symlink/… Allow also the voltage translation 3.3 -> 5 V. \$\endgroup\$
    – Antonio51
    Sep 11, 2023 at 10:00
  • \$\begingroup\$ The input of the device is a comparator, so while your microcontroller rise time might be a few nanoseconds, you'd see any glitching on a much smaller time scale, probably picoseconds. Does that even matter to your application if you're using a (relatively inaccurate) cheap MCU PWM? You may be cleaning up an optical transition that itself bounces around by 1-2 orders of magnitude more (due to the source), which for a lot of things isn't going to help much. \$\endgroup\$ Sep 11, 2023 at 13:50
  • \$\begingroup\$ @user1850479 You mean input of the external trigger input of the device is comparator? Im sorry I didnt get what you mean by cheap MCU PWM. What do I need as an external trigger source? You mean better to use a slow rise time? \$\endgroup\$
    – user1245
    Sep 11, 2023 at 16:45

1 Answer 1


According to your micro's datasheet, if you set the IO speed register to maximum for this pin you can expect 5-10ns rise time.

enter image description here

This is already quite fast. Looking at the pinout, I'd use a pin close to the VCC/GND pin pairs for that output, for lower inductance.

enter image description here

If you want lower rise time, you can use a buffer to drive your cable and enhance the slew rate. For example a single logic gate like 74LVC1G14/NC7SV14 (inverting Schmitt) 74AUP1G17/NC7SV17 (non inverting Schmitt).

Usually they don't spec rise time, but they do spec propagation time which includes rise time.

For a clean signal without ringing, you need a source termination resistor so the impedance of your driver matches the cable (50 ohms). This impedance is the sum of the actual resistor in series with the output and the internal resistance of the FETs in the output of the chip.

Check the driver's (or micro) datasheet for Voh and Vol depending on current. For example on NC7SV14, at VCC=3V3, Vol is 0.55V for 24mA output current, which indicates an output resistance of about 22 ohms. In this case adding a 27-33R resistor in series will give the proper 50 ohms source impedance to drive the cable.

Remember that when the output is high, it is connected to the chip's power supply. So if you want a nice waveform with a clean edge and no ringing, then there must be no ringing on that power supply when it switches, otherwise it will go straight into the output. Since the driver needs a decoupling cap right next to it (the usual 100nF will be fine) adding a few ohms resistor in the VCC line between the main VCC and the driver VCC should remove any ringing due to decoupling caps and trace inductance.

  • \$\begingroup\$ Thank you for the answer. I really appreciate the help. But output impedance does not necessarily be 50 Ohm in a transmission line as far as I remember. If the characteristic impedance matches the load it should be fine. In my case the device's trigger input is already 50Ohm (drive.google.com/file/d/1pT3gjOOeFLpZ8bzGCFQudrD3BUM79YIe/view) I dont care about the jitter much. I just want to know what should be the rise time of the rising edge of the PWM source which I will use to external trigger the optical device. \$\endgroup\$
    – user1245
    Sep 11, 2023 at 16:58
  • \$\begingroup\$ Also could you elaborate a bit on what is the benefit of using a non inverting Schmitt after the MCU? Basically I have two issues: 1-)What should be a reasonable rise time or deos it even matter? 2-) What part/device to use after the MCU to trigger the 50 0hm input? And why would I need that (a comparator or a Schmitt)? \$\endgroup\$
    – user1245
    Sep 11, 2023 at 16:59
  • \$\begingroup\$ I suggested a driver because you asked for one in your question, but I think it will work find without it, just the micro's output to the cable with a resistor. The rise time your device requires should be in its documentation. \$\endgroup\$
    – bobflux
    Sep 12, 2023 at 7:28

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.