PWM controller for Fixed Frequency Driver for fiber coupled optics

[Fávero Santos]

Hello, guys, how are you all?

I am currently trying to find a PWM controller to interface it to a Fixed Frequency Driver (FFD) from Brinrose - the FFD then drives an AO modulator (at 1550 nm). Such FFD has TTL inputs (0-5V) and 330 ohms impedance. The issue is that I am unable to find a PWM controller that can drive the TTL input of such FFD because we have a requirement of a minimum "ON time" of 10 ns (which may vary up to 500 ns in 10 ns steps) at operation frequency of 100 kHz. The PWMC ICs that I came across are mostly used for power management ICs (such as those used in switched DC-DC converters) and I am insecure if they are adequate for our application.

Could you guys provide me any hint or lead if there is any part or system that is able to drive this FFD? Could you guys provide me any other idea of how to proceed if no PWM controller with such requirement is available? (I am considering the usage of a microcontroller with GPIO drivers).

Many thanks, guys.
I really appreciate the time and effort you spent reading this request.

Fávero

 

[danadakk]

There are logic families that take you into the Ghz area (you need a few hundred
Mhz I would posit) and you could create a PWM using a comparator, latch to hold
comparison value which sets duty cycle) and a counter.

Or FPGA .....


Regards, Dana.

 

[Easy peasy]

nearest would be GaN drivers intended for GaN mosfets

 

[biff44]

you are making the mistake of assuming we know what a fixed frequency driver actually is, and why the heck you would be trying to drive the digital input with a switching power supply circuit, instead of a digital control signal.

are you trying to make a voltage regulator that supplies the electro-optic driver with a dc voltage, say 50 volts or something like that?

 

[KlausST]

Hi,

a (hand drawn) sketch of your situation and what you need maybe is worth a thousand words.

Klaus

 

[FvM]

The specification (100 kHz, 10 ns - 500 ns pulse width, 10 ns step size) can be perfectly met by a small programmable logic device with 100 Mhz clock frequency.

 

[Fávero Santos]

Hi,

a (hand drawn) sketch of your situation and what you need maybe is worth a thousand words.

Klaus

@KlausST, I am trully sorry If I was not clear enough. Please find attached a diagram.
Many thanks!

--- Updated Aug 2, 2021 ---

There are logic families that take you into the Ghz area (you need a few hundred
Mhz I would posit) and you could create a PWM using a comparator, latch to hold
comparison value which sets duty cycle) and a counter.

Or FPGA .....


Regards, Dana.

Thank you, @danadakk for your kind suggestion. For now, I am looking for something OEM, although I am convinced such OEM solution is not available ... I am keeping your suggestion noted for the future custom build.
Thank you so much!
Fávero

--- Updated Aug 2, 2021 ---

nearest would be GaN drivers intended for GaN mosfets

@Easy peasy thank you for answers. Ok, noted your point. You're talking about the driver transistors that will drive such PWM output, right? I am looking for something OEM, but eventually, I will have to make a custom built and I will keep on mind to check GaN MOSFETS.

Thank you so much for considering and contributing to help me solve my problem.

Fávero

--- Updated Aug 2, 2021 ---

you are making the mistake of assuming we know what a fixed frequency driver actually is, and why the heck you would be trying to drive the digital input with a switching power supply circuit, instead of a digital control signal.

are you trying to make a voltage regulator that supplies the electro-optic driver with a dc voltage, say 50 volts or something like that?

@biff44 I am sorry if I was not clear enough. The FFD (Fixed Frequency Driver) is basically a quartz based PLL that oscillates at 200 MHz. When a PWM signal is applied to its input, its output will provide a "copy" of the PWM signal but tuned to the ressonating quartz frequency. In essence, it is a single-tone FM modulator.

"why the heck you would be trying to drive the digital input with a switching power supply circuit, instead of a digital control signal.". Sorry if my text gave you the impression I am trying to do that. I will actually drive the FFD with a digital trigger signal. I commented that I such type of PWM controllers in coomonly used in switched supply ICs (although their requirements are differents), not that I am using a switched supply to drive my FFD.

"are you trying to make a voltage regulator that supplies the electro-optic driver with a dc voltage, say 50 volts or something like that?" No, I am trying to find a solution that meets the "on time", "frequency", and "output impedance" requirements in a OEM solution. If I am not able to find such OEM solution, I will have do design something (but right now, I am just trying to find an out-of-box part).

Finally, just to further clarity, I added a diagram in the answer for mr. KlausST.

Many thanks, have a great day!!!

Fávero

--- Updated Aug 2, 2021 ---

The specification (100 kHz, 10 ns - 500 ns pulse width, 10 ns step size) can be perfectly met by a small programmable logic device with 100 Mhz clock frequency.

Thank you for your repply, @FvM. I am trying to find an OEM solution at first, but designing a custom build will be the next step if such OEM solution is not available. I appreciate your help and I noted your suggestion.

Fávero

 

[KlausST]

Hi,

I see.
You want some digitally controlled PWM.

I don´t know of any IC that meets your specifications.

But I agree with FvM... a simple PLD could do the job. Maybe you find someone to program it for you.
Indeed no rocket science. A 10 bit counter and one (two) binary comparators, a FF.
6 bit parallel IF, or SPI or I2C...

Klaus

 

[Fávero Santos]

Hi,

I see.
You want some digitally controlled PWM.

I don´t know of any IC that meets your specifications.

But I agree with FvM... a simple PLD could do the job. Maybe you find someone to program it for you.
Indeed no rocket science. A 10 bit counter and one (two) binary comparators, a FF.
6 bit parallel IF, or SPI or I2C...

Klaus

Hi, @KlausST thanks anyway for your kind help and effort on understanding my issue. If I do not find any "out-of-the-box" solution I'll certainly have to design such controller.

Fávero

 

[Easy peasy]

Fro 10nS ON time, you need about 500pS rise time & fall time - while such electronics do exist - they may be hard to source.