Super fast pulses with inexpensive electronics

[Mat99]

Hi, just a quick question, does anyone know how if it is possible to generate low voltage (~5V) short pulses that are in the 50-500ps regime withouth resorting to expensive electronics?
Thanks

 

[Papabravo]

Probably not, but then again --
What is cheap? What is expensive?

 

[Mat99]

...with cheap I'd probably say <US$100.

 

[DSGarcia]

How much current do you need? I.E., what will the pulses be used for? Will a 555 work for you?

 

[Ubergeek63]

How much current do you need? I.E., what will the pulses be used for? Will a 555 work for you?

LM555 100nS rise/100nS fall. The OP was for 50pS... The LM555 is right out.

 

[Mat99]

In the end I'd like to drive a laser using up to 250mA.

 

[Ubergeek63]

Hi, just a quick question, does anyone know how if it is possible to generate low voltage (~5V) short pulses that are in the 50-500ps regime withouth resorting to expensive electronics?
Thanks

What do you call expensive?

The short answer is no.

The long answer is ECL gate delays, if you know the tricks It is possible to use a single dual input gate and feed in a pulse to both inputs with a 200pS gate delay. If you then feed in an offset into one of the inputs you should be able to tweak that inputs delay resulting in a variable output pulse width. Of course you also need the level shifters...


Dan

 

[Mat99]

Thanks for the tip, however, I know few tricks in electronics. Would an avalanche transistor with a fast capacitor as current source (but with very limited current supply) do the trick?
Are there any common capacitors out there that are fast?

 

[Ubergeek63]

generally the lower values of NPOs are the "fastest" or in other words the have the lowest parasitic inductance. But you still have to be able to supply the laser that fast.

 

[ccurtis]

I did a little research and best I have determined anyway is that the semiconductor avalanche method is the simplest (and likely cheapest) way to go. I saw one article in an application note where Analog Devices describes an avalanche transistor circuit it uses to test the rise time of its own fast comparators.https://www.linear.com/pc/downloadDocument.do?navId=H0,C1,C1010,C1096,P2069,D4183 The adjustability (amplitude and risetime) of the circuit is nice. The base capacitor was several picofarads, and not critical. The pulse width is determined by the length of hardline.

As mentioned, ECL gate delays can be exploited. The trick is to increase the voltage swing to 5V from the ≈600 mV that ECL provides. Maybe a commercial monolithic wide-band microwave amplifier, or a Class C microwave amp can be built.

 

[Dean Huster]

Use a tunnel diode circuit. Finding tunnel diodes these days is difficult, however. But time domain reflectometers often used TDs for their picosecond-area pulse generators, e.g., the Tektronix 1502. However, the higher-power TD used in the 1502 even back in 1979 when the 1502 was in its heyday ran around $150 and went up to around $500 around 1982. But smaller TDs (GE was a major supplier for Tektronix) used in the triggering circuit of Tek scopes such as the 465, 475 and 485 weren't so bad.

Tunnel diodes, the darling of the electronics industry from around 1967 through around 1982 were also used for UHF oscillators.

Dean

 

[ccurtis]

Figure 75 of this app note shows a similar avalanche pulse generator with an 800ps pulse with rise and fall
times inside 250ps. Pulse amplitude is 10V with a 50Ω
source impedance. Experimenting with different transistors, the circuit should be able to generate faster rise and fall times.

See also, figure B1.

https://www.linear.com/pc/downloadDocument.do?navId=H0,C1,C1154,C1009,C1021,P1461,D4161

 

[BTX Sistemas]

If you don't need precision, you could do a short circuit between input and output of a logic inverter, you'll get a signal of about T=2*InverterDelayTime.

 

[Mat99]

Many thanks everone, I shall look closely at all your replies. Electronics in general and fast electronics in particular is a new world for me, and I am experiencing the daunting effects of a steep learning curve!

 

[mneary]

You do know that at these pulse widths and edge rates a 10mm wire or trace is a major transmission line!

 

[Mat99]

I estimated that the pulse duration would equal roughly 100 mm in propagation distance, so I have sort of figured out that I might have to solder the Emitter (..or whichwever pin the output comes from) straight onto the laser with a minimum of wire length. Hopefully the laser does not cook at the same time.

 

[duffy]

Those references from ccurtis look good, they both have sub-nanosecond pulse generators in them and good construction and layout tips, cover issues with probes and all.

Hope you have access to a real kick-ass scope, you are going to need it!

 

[Mikebits]

How was this pico range spec derived? Is this a normal laser thing?

 

[Mat99]

Short pulse durations are very common in laser applications, but normally they are achieved through optical mode-locking techniques etc. You can also buy a very expensive pulse generator and drive diode lasers directly, but I prefer to make my own.

 

[duffy]

That optical mode-locking thing can get down into the femtosecond range. Seems like an electronic technique ought to be able to get down into the picosecond range, though.

I think you are going to need that laser held pretty close to the conduction point. This short pulse isn't going to transfer much charge.