| Electronics > Projects, Designs, and Technical Stuff |
| CMOS Crystal Oscillator duty cycle |
| (1/4) > >> |
| fabiodl:
Hello, I am trying to get a 10.7386 Mhz oscillator, 50% duty cycle. I used the schematic in section "CMOS Crystal Oscillator" of https://www.electronics-tutorials.ws/oscillator/crystal.html I used a 10.7386 Mhz crystal, 62p for the capacitors, R1 = 10M and R2=2k2. The inverter is a SN74HCT04PWR I measured a duty cycle of about 30% (at the output of the first gate). How can I make it closer to 50? I need to bring it into the range 45 to 55. I suspect the asymmetry is due to the difference between low (0.8-0) and high (5-2) voltage ranges, but I don't know the best/standard/simplest way to adjust the duty cycle |
| Circlotron:
Have a crystal at double the requited frequency then shove it through a flipflop? Exact 50/50 duty cycle guaranteed. |
| RoGeorge:
Use a XOR gate to double the frequency, then a flip-flop to divide by 2. The resulting output from the flip-flop will be 50% duty cycle. |
| Benta:
Use 74HC instead of HCT. The HCT is not symmetrical, as it's laid out for TTL levels. |
| Ian.M:
--- Quote from: RoGeorge on October 31, 2018, 11:59:21 am ---Use a XOR gate to double the frequency, then a flip-flop to divide by 2. The resulting output from the flip-flop will be 50% duty cycle. --- End quote --- No it wont, as the XOR circuit will produce pulses starting at the edges of the original asymmetric signal, so you'll get a pair of pulses closer together, then a longer interval then another close pair and so on. Circlotron's suggestion of staring with a crystal of double the frequency and dividing it down is the best one. If you *must* use the existing crystal, after replacing the 74HCT04 with a 74HCU04 to get near symmetric thresholds, and a cleaner, more sinusoidal waveform at the driven side of the crystal, if the symmetry isn't good enough, any remaining asymmetry can be trimmed out by capacitively coupling the input of the buffer gate and biassing it via a 1Meg resistor fed from a 10K potentiometer across its supply rails. Due to the lower gain of the 74HCU04 gates, running the buffer output through an additional one to further buffer it and sharpen the edges would be advisable. Its also possible to use low pass filtered high resistance negative feedback to bias the first buffer gate so it actively maintains a 50% duty cycle over temperature or supply voltage variation, but that gets more complex, and can produce a very asymmetric duty cycle during startup. |
| Navigation |
| Message Index |
| Next page |