Hi all,
As we know:
1 MHz = 1000000 Hz. 1 Cycle per Nanosecond
A period of 1 Nanosecond is equal to 1 000 000 000 Hertz frequency.
1 000 000 000 Herts = 1 000 MegaHerts = 1 Ghz
I'm thinking to start a project and I need a delay no more than 1 nanosecond.
Is that even possible with arduino products/processors.
If you have any kind of info or a such past project please let me know.
It's been a while since I've been here 
3 years or so.
Best Regards,
Domino60
Go for FPGA e.g. Spartan etc. Spartan-6 FPGA Family
Hi,
What is the application that requires less than 1ns delay?
Tom...
I hate to stifle your creativity, the delay that you are expecting is a bit much. The Arduinos run in the MHz (16MHz for Uno and Mega). That translates to about 65ns per clock cycle.
I would recommend taking another look at your requirement for a 1ns delay.
IF you find that it is necessary, I would explore a raspberry pi using python. But again, I have a hard time imagining anything that needs that kind of fidelity.
darrellrread:
But again, I have a hard time imagining anything that needs that kind of fidelity.
Just a silly guess, but perhaps the OP is trying to measure distances using light.
One light-nanosecond is about a foot.
zwieblum
Thanks for the link I'll take a more detailed look on it but a bit too much for my budget 600$+ kit that's a lot.
vinceherman
Just a silly guess, but perhaps the OP is trying to measure distances using light.
One light-nanosecond is about a foot.
Not bad guess
TomGeorge
What is the application that requires less than 1ns delay?
I want to have a a processor, component, circuit that will be possible to send signal to a laser to
be ON for 1ns sending about a foot long light of beam to a receiver sensor which will measure
the time of the light which should be as well 1ns all that in a fully dark room no other lights.
The experiment consist of other parts as well but that's the start of it.
So if anyone got an idea (looking on the internet myself as well) how I could get that ns delay
that would be really helpful.
Note: I'll need to have a receiver which is able to measure ~picosecond/nanosecond time as well.
Best Regards,
Domino60
Well, there are cheeper boards around, e.g. iCE40
Hi,
Have you looked at what frequencies you can FREQUENCY MODULATE a LED of any sort?
Have you looked at what frequency response you get from sensors you will need to use?
Have you considered phase shift in the Tx and Rx devices?
I would be looking for your main output and input elements before looking for driver hardware.
Tom..
TomGeorge
I'm in the process of all that, I've thought about that because having a
processor which can throw signals up to ns delays that's not enough.
I need as well the time the signal is traveling, the time the end
device is triggered by the signal, the time of the trigger to activate
the laser and how much time the laser to light up.
One thing I could say is that I thought using optic fiber to transmit the signal
to activate the end device.
So my current diagram is
I need a beam of laser no bigger than 1 meter long so I have a window of
3.335641 nanoseconds to include all those actions.
This idea is still fresh but it has a purpose so that's why I'm looking
for components/methods but on a budget.
Best Regards,
Domino60
There exist shift registers that can be clocked at GHz frequencies, so a parallel write (transmit side) and a parallel read (receive side) could be done at a clock frequency plausibly within the capabilities of a fast microcontroller.
There are also laser drivers/receivers that can go this fast as used in fiber optic communications applications.
Synchronizing the transmitter and receiver is going to be an issue. Speed of light measurements are done with essentially co-located transmitter and receiver with the measurement path going to a mirror and back.
How about a high-speed two input XOR gate, fed from a single input signal, but with a 1nS delay line on one of the inputs. When you switch the input signal, the inputs to the XOR gate will differ for a period of 1nS, until the signal has time to pass through the delay line, producing a 1nS pulse on the output.
I need as well the time the signal is traveling, the time the end
device is triggered by the signal, the time of the trigger to activate
the laser and how much time the laser to light up.One thing I could say is that I thought using optic fiber to transmit the signal
to activate the end device.
You are aware that light will travel slower in fiber optic cable, so the time of travel will increase for a given distance.
How about a full explanation of what the purpose of this is, rather than just the proposed solutions?
Otherwise, to me, this is looking like an XY problem issue.
--- bill
There are bound to be better ways to do this than trying to switch on and off a laser on command by a MCU, considering time of flight sensors measuring far shorter times are really cheap.
Very short voltage pulses can be produced by an Arduino - pulses orders of magnitude shorter than you expect based on the clock pulse - using a RC circuit. Such circuits may be a way to get your laser to produce a very short pulse as well.
I was just reading up on the ESP32 Remote Control Peripheral. As I understand it, the hardware is a programmable sequencer that is capable of transmitting and receiving pulse streams with a resolution of up to 1/80 MHz = 12.5 nanoseconds resolution.
This is an order of magnitude lower resolution than the original poster desired, but, ignoring the significant issues in transducer speeds, it might be a plausible basis for light-speed round trip time of flight experiments on inexpensive microcontroller hardware.