2
\$\begingroup\$

Could someone explain how the auxiliary winding controls the output voltage on the secondary winding?

For instance,

  • How do the zener diodes function?
  • How do you calculate the output voltage as a function of zener voltage?

RCC #1: https://www.electroschematics.com/diy-rcc-smps-circuits/
enter image description here

RCC #2: https://www.electroschematics.com/diy-rcc-smps-circuits/
enter image description here

Edit: Here is how I assume the feedback/auxiliary circuitry works.

For RCC #1, the voltage across the auxiliary winding is rectified into a negative DC voltage across C2. I assume the goal of the auxiliary rectifier circuit is create a DC voltage that's as close as possible to the secondary voltage across C3. The auxiliary voltage is then used as a reference voltage. The (negative) reference voltage is applied to a voltage error detection circuit for turning on/off the switching transistor. The error detection circuit consists of zener ZD1. I assume the goal error circuit is to turn the zener diode on and draw current away from the base of the switching transistor when the reference voltage is too high (or low i.e. negative) across C2. As current is drawn away from the base the switching transistor the switching transistor begins to turn off. If the reference voltage across C2 is significantly out of range, the switching transistor is turned off completely. I assume the transistor operates in the linear, saturation, and cutoff regions. Is this all correct?

The question that is bothering me is this. If the output of the secondary apparently is supposed to be 9V. And the zener diode is sized for 8.2 V. How does the secondary get to 9 V?

For RCC #2, the voltage across the auxiliary winding is rectified into a negative DC voltage. The rectified DC voltage across C3 should ideally be the same voltage across C4 of the secondary. I noticed the diodes are slightly different. Wouldn't you want the diodes to be the same, especially since the negative voltage across C3 is used as a reference voltage? I assume then that if the negative voltage is too low the zener diode conducts and draws current away from the base of the switching transistor. When current is drawn away from the base, the switching transistor is either slightly on or completely off. The zener voltage is 6.2 V. The voltage at the output across C4 is apparently 5.3 V, according to the website. How is 5.3 V achieved? Also, D7 is connected from ground to the base of the switching transistor. What is the purpose of D7? Is it to protect the B-E junction from reverse voltage? I assume so the base doesn't go too far below ground. If so, then why is the diode not on RCC #1?

Last question for RCC #2. How is 5.3 V achieved at the output when the zener diode (ZD1) is rated at 6.2 V?

I still don't understand how the voltages of the zener diodes are chosen.

\$\endgroup\$
5
  • \$\begingroup\$ Does this answer your question? Zener diode function in ringing choke converter \$\endgroup\$ Oct 30, 2023 at 5:36
  • 1
    \$\begingroup\$ Have you tried to simulate it? \$\endgroup\$
    – winny
    Oct 30, 2023 at 6:41
  • 1
    \$\begingroup\$ Winding polarities are not shown. I don't see dots. \$\endgroup\$ Oct 30, 2023 at 8:11
  • \$\begingroup\$ That helps. The issue I'm having trouble understanding is whether or not Vbe of the switching transistor needs to be accounted for when sizing the zener diode. \$\endgroup\$
    – mrbean
    Oct 30, 2023 at 16:11
  • \$\begingroup\$ I edited my question to add more detail about things that are unclear. \$\endgroup\$
    – mrbean
    Oct 30, 2023 at 17:03

1 Answer 1

2
\$\begingroup\$

You're close.

A note: while phasing dots aren't given, the circuit only works correctly one way around: with #1 having dots at the top of each winding, and #2 having L1 and L2 with dots at the top and L3's at the bottom.

I assume the transistor operates in the linear, saturation, and cutoff regions.

Linear operation is unstable, due to positive feedback between collector and base windings, and the R2-C1 / R3-C2 feedback network. Since C2-ZD1 / C3-ZD1 steals base current, and linear mode is unstable, the effect is to hasten turn-off, reducing energy stored in the inductor and thus throttling down the output. If T1/Q1 ever turns on, it turns on hard, but how long it turns on for is determined by this, the RC time constant, and hFE.

Selected Details of Operation

It's not clear if this should be a mostly CW (continuous wave) oscillator -- as implied by the "RCC" (Ringing Choke Converter) label -- or if it's more of a burst-mode blocking oscillator. #1 seems the most likely candidate, as its bias current is "robbed" by the zener; C1 is charged up by a forward pulse, then it stays off until it charges to threshold again -- or if the inductor voltage rings to a high enough peak to trigger another cycle. The lack of damping (RC) network on L1 seems to suggest this is intended, and it probably operates in a tone-burst mode ("squegging").

In contrast, #2 is forced into CW by D7: when L2 reverses, C2 is fully recharged through this path, ensuring turn-on on the next swing. This would normally make the circuit uncontrollable -- it's forced to full throttle -- but ZD1 absorbs the difference. This works because R1's voltage drop is proportional to load current, and thus VB rises during the pulse, eventually being clamped by ZD1 and turn-off is forced.

L2's turns ratio is determined by the same mechanism, not just the voltage at C3 (ZD1 in turn) but avalanche of T1/Q1 as well; this limits the output voltage range of the burst-mode blocking oscillator, lest avalanche conduction serve the purposes of D7 and force it to full throttle. (That is, full throttle, not counting the effect of ZD1.)

I think the throttle range of both of these is rather narrow, so that an unloaded output is not exactly recommended. #1 may need a minimum load (basically to mirror R3 on the primary side); #2 shows one explicitly, and a rather generous one at that (16mA at the voltage shown).

Diodes and Turns Ratio; Regulation

The zeners should be good, sharp types; true zeners (the Zener tunneling effect is dominant) are quite "soft", i.e. current increases more gradually with voltage.

A common question on this Stack is novices being confused why their "3.3V clamping" circuit only outputs 2.2V or something like that; simply, the 3.3V zener used, only develops that voltage at some whopping like 10 or 100mA, and its voltage drop is much lower otherwise.

The avalanche effect takes over above 9V or so, but zener and avalanche effects are balanced around 5-7V rating, with 6.2V zeners having the sharpest knee. So, there is some priority to choose parts around this value.

Higher voltages are unsuitable, as E-B breakdown is typically not much above 6V, i.e. the transistor will break down for you. And as mentioned, this is undesirable in #1, and other more conventional blocking oscillators; and would waste power in #2 (hence the use of D7 instead).

Regulation is determined by the relative voltage drops of D2/D3 or D6/D8, and their loading conditions (#1 has a pull-down resistor R3, #2 has the C2-D7 charge pump current). It also depends on leakage inductance between L2 and L3. It will generally be poor; using these for something like a phone charger would be dubious, I think.

Output voltage is simply the ratio of L2 peak voltage to L3, less each diode. For #1, say C2 idles at -7.5V (T1 VBE minus V(ZD1)); L2 peak must be around -8.2V then. L3 can be a few turns higher to make say 10V peak, and D3 drops that to the 9.something V desired. Mind I'm being loose with all these values, because generally, power transistor VBE will be higher than the 0.6-0.7V rule of thumb, and likewise 1N4148 makes a rather poor rectifier and probably drops more like 1-1.1V at maximum load (if it survives at 100s mA at all..). These days, schottky like MBR140 or PMEG4005 would be used, reducing VF to more like 0.4-0.6V.

But these days, proper controllers/regulators with direct primary voltage peak sensing are available, at bargain prices, with much better efficiency, and more reliable operation, completely obviating the need for such circuits. (They're still a neat curiosity, some of the more nuanced things you can do with naught but a single transistor.)

Other Notes

#2 also has a superfluous C1-D5, presumably missing a discharge resistor across C3 to act as peak clamp snubber.

Neither has a bulk capacitor, meaning the output will be pulsating DC. This is especially egregious for #2, where C4 is the largest capacitor on the power path; but #1 is even half-wave rectified, giving 50 rather than 100Hz output ripple.

And needless to say, neither shows a line filter, which is required for commercial use, but we might simply assume their use, with these being mere sample circuits.

And a "Y type" capacitor between primary and secondary common. Inrush limiting and fusing too.

On the upside, power factor might not actually be terrible, if "throttle" is somewhat proportional to mains input voltage. A side effect of the lack of primary bulk capacitance. This will probably be better for #1 than #2 (because the latter is self-powered by the charge pump effect mentioned).

The poor regulation might also not matter for fixed-load applications, like LED lighting. A current-limiting resistor would be needed on the output, then as many LEDs in series to make up the bulk of the output voltage. A current feedback scheme might also be applicable. (But here again, we have ICs available that are better in all respects, so it's not a very interesting commercial proposition today.)

\$\endgroup\$

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.