Pulse Stretcher This circuit is designed to take a short pulse from a PMT (photomultiplier tube) and make it longer such that an older instrument can record the pulses. The input pulse is about 7-10ns wide and produces an output pulse that's adjustable from 100ns-10us. The heart of the circuit is a high speed comparator (U1) with a output latch function. U1 compares a threshold voltage (~0.25V) set with a potentiometer to the input pulse. Note: The input comes in on a 50 ohm coax cable so I placed a 50 ohm terminating resistor under the BNC connector for impedance matching. Since I was using small, 805 size SMT resistors, I used four larger resistors in parallel where the combined resistance is 49.5 ohms. The count rate is very low in this application but I don't want the resistors getting hot if this is used with higher count rates. The terminating resistor drops the 5-6V input pulse to 2.5-3V but it's clean (no reflections). The comparators latch functions is active low so I plan to feed back the output to latch the output in a low state for a bit making the comparators output stay low long enough to trigger the one-shots (U2, a pulse generator IC). The output of the comparator is normally high when no pulse is present. When a pulse comes in that's greater than the threshold voltage the output goes low. Diode D2 (a very fast schottkey diode) becomes forward biased dropping the voltage on the right side of C5 (see schematic) to about 0.3V. Since C5 was initially uncharged and can't charge instantly the left side of C5 also drops to about 0.3V. This latches the output of the comparator low. The Left side of C5 starts rising back towards 5V (R4 and C5 set the time constant). After 60-100ns the voltage on the latch pin has risen high enough that it's a legal high and the latch is turned off. By this time the input pulse has long since gone low and the comparator output goes high again. This reverse biases D2 effectively removing it from the circuit. C5 is now charged to 2.5V or so. D1 clamps the max voltage at the latch pin to about 5.3V and C5 discharges through R4 & R5. The latch time and discharge time set the maximum count rate. U2 is a pulse generator IC. The /A (inverting) input triggers on the falling edge of the comparator. The Q output goes high on the falling edge and stays high for a time determined by C6 & (R6 + R7). R6 is there because the IC doesn't produce an output pulse if the resistance gets to close to zero (internally the IC can't pull pin 15 low if it's effectively tied high with a small enough resistor). U2's output goes to U3, a line driver IC. By paralleling the outputs this IC can provide enough current to drive a 50 ohm terminated cable with close to 5V. Again, we want a clean pulse at the instrument receiving the pulse. Since U2 & U3 both have two identical circuits I wired both up. This way two different pulse widths can go to two different instruments. Also, if a terminator isn't placed at the end of one output cable its reflection won't effect the other output. The circuit is powered by a 5V wall wart. The input power and every IC is decoupled with a ferrite bead and multiple caps to smooth the power supply and keep the current drawn by one IC from causing the power supply at other IC's to droop. These are low pass LC filters. The three trim pots for setting the threshold voltage and pulse widths are accessible via small holes on the front panel. There are test points on the PC board for measuring the threshold voltage which is not accessible outside the enclosure.