I'm mildly surprised that it ever worked at all because if that diagram is correct, there will be spikes coming from the transformer coil (the centre tap is connected to the +12V side) onto the +12V rail and via the 2N3122 into the supply to the CMOS gates. 74xx chips are rated for only 18V and the spikes from that thing could easily exceed 30V or more - even with no load - and more than enough to make the CMOS misbehave or even fail completely.
Another aspect is that the timing Is based on a simple 555 R/C oscillator circuit. With a noisy power supply the timing will be pretty erratic so it was never going to run particularly stably.
Conclusion: it badly needs a few components to give the CMOS a stable supply isolated from the spiky output.
The basic ideas in this circuit are:
- the 555 timer is wired as an R/C oscillator, the timing set by R1, R2 and C1, You can look up a data sheet to see how that works.
- the 7473 flip-flop is wired to divide the frequency by two, and produce two identical square waves exactly out of phase with each other.
- the 7473 doesn't have the capacity to drive low-impedance loads (transistors) directly so a 7406 is used, with the outputs of three gates in parallel to drive a pair of transistors on one side of the transformer, and another three gates drive another pair of transistors on the other side. The transistors are wired as high current-gain darlington pairs, with 4W 2.7 ohm resistors in series to limit the maximum current otherwise you'd exceed the thermal ratings of the transistors or the transformer, and have some rather hot, smelly smoke if you short-circuit anything.
The output circuit is symmetrical, in that the centre tap of the transformer primary is tied to +12V and one side or other of the winding should be energized alternately via the transistors. Since the components are the same and driven by the flip-flop, the timing of the square waves at the output should be quite exact - and symmetrical.
- the 0.27nF capacitor across the output is supposed to round off the top of the square waves, as in the yellow race. That you're setting the reverse in the blue trace suggests to me that one side of that capacitor has been connected to the frame, or to the centre tap.
The oscilloscope traces show several things:
1. The blue trace and yellow trace are quite different. That implies something is quite faulty with the circuit.
2. The yellow trace has a big spike at the leading (rising) edge of each pulse, which I would expect with a strongly inductive load (the transformer winding is an inductor) and the rounded top of the pulse shape is what I would expect. However, after the trailing (falling) edge of the pulse, there is a big spike a few milliseconds after, and this should not be seen and implies here is a fault somewhere in the CMOS circuitry driving the transistor pair.
3. The blue trace likewise has a big spike on the leading (rising) edge which is expected, but the shape of the pulse is wrong and indicates again something is wrong in the circuit as-built. Have you done something with the capacitor 0.27nF capacitor on the output side ?
Changes.
1. An LM7805 is just a 5V regulator and while you can use that to supply the 555 and 7473 logic gates, it won't fix really the above issues, there are other things wrong.
2. It needs decoupling (bypass) capacitors between the +12V and ground, and between the logic supply, at the emitter of the 2N3122 to ground, to stop the spikes reaching the CMOS. These should each include say 10 uF tantalum and 0.01uF ceramic.
3. I would replace the 7406 inverters with a simple pair of op-amps - eg LM 358 - wired as unity-gain voltage followers providing a much better buffer between the CMOS and the analog output.
4. Lastly beware of the output. Transformers being inductive the output spikes will be a lot more than 100 or 120V and quite enough to give you a painful shock if you make contact with bare wiring or terminals. Hazardous, actually.
Which is why ultimately I would put the whole thing in the trash - with the synchronous motor - and build a circuit to drive a 4V stepping motor which can be easily built and diagnosed at home with little more than a multimeter - and safely. Motors, drivers and more available at www.pololu.com. You could copy the 555 timer and the flipflops to drive a stepper driver. No transistors and no transformer required. The hardest part will be to choose a suitable stepping rate for nominal tracking, and tweaking the timer components to produce that frequency.
Edited by luxo II, 27 February 2024 - 03:07 AM.