1
\$\begingroup\$

For education purposes, I plan to build the following short-pulse generating circuit:

pulse circuit

(picture taken from https://www.radiolocman.com/shem/schematics.html?di=473681, as far as I can tell the original source is from here: https://www.analog.com/en/resources/app-notes/an-98.html)

1.) The 8pF capacitors seem rather small in value so they could be affected by environmental effects Should I remove the ground plane at least on the bottom of the 2-layer board around those caps, maybe even avoiding the use of a GND plane alltogether?

2.) The goal would be to be able to scale the pulse width by making the resistor element of the "variable delay" RC stage variable, i.e. a trimmer/potentiometer. Which type of trimmer could I potentially use? I would think that any "standard" part would have parasitic capacitances way beyond 8pF, essentially building some capacitive voltage divider with the capacitor or even "swamping" that value? Is it even feasible to use a trimmer here?

//EDIT because of comment: The waveforms of the circuit are shown from Figure 20 to Figure 24 of the following article: https://www.analog.com/en/resources/app-notes/an-98.html

\$\endgroup\$
4
  • \$\begingroup\$ Please post a timing diagram showing the input and output waveforms. \$\endgroup\$ Commented Jan 23 at 20:33
  • \$\begingroup\$ What kind of frequencies does this need to run at? \$\endgroup\$ Commented Jan 23 at 20:35
  • \$\begingroup\$ @AnalogKid and Voltage Spike: Thanks for the comments, I edited the text to refer to the Figures of the article from which the circuit originates from \$\endgroup\$ Commented Jan 23 at 20:38
  • 1
    \$\begingroup\$ define Trise minimum and Vout, Iout max or R load \$\endgroup\$ Commented Jan 23 at 21:18

2 Answers 2

Reset to default
1
\$\begingroup\$

Trace and component capacitance are not as bad as you imagine: 8pF is a small capacitor by itself, it's true, but a 1cm long trace on a typical 2-layer board is likely under 1pF total.

Likewise, trimmer resistors are available in quite small sizes (notwithstanding whether you want to use a tiny one!), the capacitance of which can be estimated from the overall dimensions, the length and width of pins and internal components and the distances between them.

Capacitance across the resistor is also not hugely critical, as it creates a partial feed-forward or speed-up effect; it can simply be swamped by using a larger capacitor after the resistor. That is to say, it has the effect of shifting the exact R*C value required to create a given delay / pulse width.

Variable capacitors of this size are also economical, which may be preferable as touching a metal screwdriver to a trimmer resistor (for certain types) will distort the timing while in contact. The relevant metal parts can be avoided (certain types) or connected to ground where the trimming tool has no effect.

I gather from the component types and values, this is for a fairly fast application (some nanoseconds, and perhaps 10s of MHz repetitive). You may desire a 4-layer board to maintain signal quality and matched impedance traces from connectors (e.g. edge-launch SMA) to relevant devices (and termination resistors where applicable).

Not that the circuit as shown cannot be done on a 2-layer board; more that, it will require more mindful design, expertise, to realize with comparable quality (signal quality, EM emissions and immunity) as a multilayer design using internal ground/power planes.

\$\endgroup\$
2
  • \$\begingroup\$ Are you aware of some trimmable caps that have "high" durability? I had a look at e.g. vishay.com/docs/28527/bfc2808-75mm.pdf but they only specify a maximmum amount of 10 trimming cycles as life span =) For Voltronics (mouser.at/datasheet/2/218/…) I don´t even see any specification about "cycles" \$\endgroup\$ Commented Jan 23 at 21:12
  • \$\begingroup\$ "Trimmer" seems contradictory here -- if you need many actuation cycles, a different adjustable type would be desirable (old fashioned air variables come to mind, though I don't know offhand where they're available new these days), or perhaps even electronic methods (e.g. transconductance amplifier, slew rate limiter, varactor diodes, etc.). \$\endgroup\$ Commented Jan 23 at 22:34
0
\$\begingroup\$

What you have works. I have always used XOR gates for narrow pulses with pots or fixed R's. Using an XOR gate with Vol/Io= 33 Ohms as the clock source in my simulation with 100 pF to define the max delay load.

It depends on how narrow and wide you want for a simple pulse generator by choosing the fixed cap value or no cap for the smallest pulse using the x pF Cin and variable R to create a difference in RC delays.

Decoupling caps are assumed near IC and your choice of 5.5V logic (74AHC) or 3.6V logic (74LVC) for faster response with lower Rout and will accept 5Vin.

If the AND-OR inputs of the XOR gate cause issues with a slow-rising input, you can delay to a simple inverter for the delay path.

enter image description here

\$\endgroup\$

Your Answer

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

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.