# Delaying a ns pulse by a few μs - is it possible with a 555?

#### Turksen

May 18, 2011
8
Hi,

I'm trying to create an electronic delay circuit, to be used as a trigger signal.

I have a train of pulses (~1 ns width with 5 μs spacing (~200 kHz)) and want to be able to delay the pulses by a few microseconds (say, 3 μs). As the resulting pulse train is to be used for rising-edge triggering, it doesn't matter if the pulse width broadens to a few microseconds too, providing the 200 kHz frequency is unchanged.

I considered using a 555 timer for this as a monostable:
The 1 ns pulse would trigger the monostable, with time period = 3 μs, say.
An inverter follows the monostable so that after 3 μs, the output voltage rises.
Since the input into the 555 trigger is periodic (200 kHz), the output would be periodic too, with 200 kHz frequency.

From the 555 datasheet, I think the 555 should be able to produce a 200 kHz output, although I wonder if the 1 ns trigger pulse might be too short?

Any suggestions or comments would be greatly appreciated. Thanks!

#### woodchips

Feb 8, 2013
43
A 1nsec pulse has a frequency bandwidth of something over 3GHz. Quite definitely a 555 will not cope with this.

How do you even know it is 1nsec wide? Sure you don't mean 1us? If so then 74 logic will work with that.

#### Turksen

May 18, 2011
8

The pulse is ~2 ns on an oscilloscope. I want to delay this by ~5 us.

A 555 datasheet suggests the minimum input pulse width for triggering is ~ 50 ns.
I also considered a simple RC delay, but I'm not sure if a capacitor could be charged by a 2ns pulse.

Could the trigger pulse be extended by a latch or some other circuit perhaps?

Or any other ideas warmly welcomed! Thank you.

#### woodchips

Feb 8, 2013
43
Are you really really sure it is 2ns? What oscilloscope are you using? As I said, 2ns is fast, also a 555 just won't work.

Bob

#### Turksen

May 18, 2011
8
Yes, I'm confident the pulses are of ns duration as they're from a 1 GHz bandwidth photodetector, upon which sub-ns optical pulse are incident. Therefore, effectively, the pulse I'm seeing (on a fast 'scope) is characteristic of the impulse response of the detector.

My aim is to delay the pulses by 5us and use them to trigger another device.
I appreciate ns pulses are probably too short for a 555, but I wonder if other delay methods might work? Perhaps an RC circuit?

Last edited:

#### woodchips

Feb 8, 2013
43
Almost any photodiode detector could be described as 1GHz bandwidth, but it is the amplifier after it that really defines the bandwidth. A 2ns pulse, not spike, has a 1/3 time rise, 1/3 time top and a 1/3 time fall.

Exactly what scope are you using? How about a screen photo of the pulses? A 2ns pulse has a bandwidth of something like 5GHz, reciprocal of 1/3 of the rise time of 0.6ns. A 100MHz scope has a rise time of about 3ns, one of those useful rules of thumb.

Delaying a 5GHz signal needs a delay line, so getting 1us requires about 100m of coax cable. The coax is a broadband RC circuit.

Bob

#### Electrobrains

Jan 2, 2012
259
You could probably use this simple, cheap circuit.
It shows adjustable delay and pulse length, which of course you could fix at whatever value you want.

I have not tested the circuit, so there could be a bug somewhere, but I don't think so.
Your task would be to calculate the RC values. For your super fast application, the capacitors would be low value types (pF).

The Pulse Delay is triggered from the falling edge of the input signal. With some changes, it could be triggered on the rising edge.

The suggested IC(s) should be fast enough.
The diodes too, or you might use Schottky Diodes instead.

#### Turksen

May 18, 2011
8
Thanks woodchips - the scope has 50 GHz bandwidth. I did consider a delay line from coax cable, although for the desired 5 us delay, I'd need a few hundred m of coax, which is a bit impractical.

Thanks too electrobrains - that's a really helpful circuit and the ability to vary the delay could be very useful.

I assume R1&P1 with C1 set the delay time constant, and R2&P2 with C2 set the duration constant. I've checked out the spec sheets for the IC but don't fully understand the purpose of the schmitt-trigger inverters, especially the two at the circuit output. Are they there to help sharpen-up the pulses after the RC delay?

Many thanks,

#### Electrobrains

Jan 2, 2012
259
..I assume R1&P1 with C1 set the delay time constant, and R2&P2 with C2 set the duration constant. I've checked out the spec sheets for the IC but don't fully understand the purpose of the schmitt-trigger inverters, especially the two at the circuit output. Are they there to help sharpen-up the pulses after the RC delay?
The Schmitt-Triggers are needed for quite accurate switching points. They give the necessary hysteresis and as you say, "sharpen-up" the pulses.

Yes, R1&P1 with C1 set the delay and R2&P2 with C2 set the pulse duration.

The third inverter just inverts the signal and gives you the same polarity as the input.
It also adds a small delay (ns).
If that disturbs you, it could be removed and the second inverter could be used as output. Even with the right polarity, if the RC components would be referenced towards +5V.
I prefer using ground as reference and because there are 6 inverters in the package I just added the third one.

#### Turksen

May 18, 2011
8
Great, many thanks for your help.

#### (*steve*)

##### ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,505
I would wonder what effect the propagation delay of the 74HC14 would be on this circuit.

At 5V, the typical propagation delay is 12ns. Added to this, the typical transition time is 6ns.

The pulse you get out of this circuit is not going to look very "square"

The input capacitance of the gate (approx 3.5pF) might also be a problem.

Also, is the 1N4148 able to recover inside of 2ns? (The minimum is 4ns, and the capacitance is 4pF)

#### Electrobrains

Jan 2, 2012
259
Steve, the circuit is 74A... not 74HC. The A versions are faster.

About 1N4148, you are right about the recovery time. That's why I mentioned the Schottky Diode. It could well be that another (HF) diode must be used instead.

What also concerns me a bit is the input impedance. The source of the input signal must be very low impedance to be able to charge the input capacitor (C1) within the pulse width (2ns)!
Maybe the rest of the inverters in the package should be used as input buffers.

Replies
5
Views
975
Replies
3
Views
805
A
Replies
8
Views
2K
Jim Wilkins
J
M
Replies
13
Views
2K
Fred Bloggs
F