5
\$\begingroup\$

The circuit steps the voltage down from POE (class 4) input to 5V 6A The circuits I have found so far use opto-isolation to regulate the feedback voltage (for 5V 6A). I'm wondering is it feasible to use the Auxiliary to regulate the feedback circuit when the output 5V/6A? The manufacturers I have found all use opto-isolation for 5V and 6A. Is this because it is better to have opto-isolation for lower voltage and high current? Is there any other reason?

The following uses opto-isolation for 5V 6A:

\$\endgroup\$
9
  • 1
    \$\begingroup\$ Certainly. What level of output regulation do you need? \$\endgroup\$
    – winny
    Commented Apr 26, 2023 at 12:56
  • \$\begingroup\$ Voltage ripple of maximum 20mV \$\endgroup\$
    – Sam Alamad
    Commented Apr 26, 2023 at 13:00
  • 2
    \$\begingroup\$ Opto-isolation is there for electrical safety and EMC reasons both. \$\endgroup\$
    – Lundin
    Commented Apr 26, 2023 at 13:06
  • \$\begingroup\$ Not ripple, regulation. Load and line regulation. \$\endgroup\$
    – winny
    Commented Apr 26, 2023 at 13:48
  • 1
    \$\begingroup\$ Still not answering my question. What level of line and load regulation do you require? \$\endgroup\$
    – winny
    Commented Apr 27, 2023 at 13:34

5 Answers 5

6
\$\begingroup\$

The manufacturers I have found all use opto-isolation for 5V and 6A. Is this because it is better to have opto-isolation for lower voltage and high current? Is there any other reason?

Opto-isolation is used to feedback to the flyback controller when the output voltage is in regulation hence, it's regarded as the most satisfactory method of ensuring output voltage accuracy. It's got nothing to do with the output voltage level or the output current level.

Clearly, if you need galvanic isolation between output and input, an opto-isolator provides this with massive and obvious abundance.

Is it possible to produce 30W (5V 6A) using a flyback topology without opto-isolation?

Other methods can be used that provide galvanic isolation AND accuracy but, why bother if an opto is good enough?

There are also less accurate methods such as using an auxiliary winding or measuring the back-emf on the primary side.

\$\endgroup\$
9
  • \$\begingroup\$ What is "galvanic" isolation ? Also, does the opto isolation really have a linear response for input voltage i.e the light generated and converted into voltage on the other side mechanism. \$\endgroup\$
    – quantum231
    Commented Apr 27, 2023 at 20:12
  • \$\begingroup\$ @quantum231 galvanic means physical electrical connection and, of course, galvanic isolation means no physical electrical connection but, it doesn't mean there won't be some small amount of non-galvanic connection due to isolation capacitance. The opto is driven from zero light to full light for a change of a few millivolts in the output voltage so, it's hardly linear at all but, due to it correcting the duty cycle (inside a strong feedback loop), it sits nominally "in the middle" producing partial light to the detector circuits. \$\endgroup\$
    – Andy aka
    Commented Apr 27, 2023 at 20:41
  • \$\begingroup\$ I see, so the feedback signal is either fully off or fully on, it won't get linear even if the LED and the phototransistor are on the same die and have the same non-linear relationship of light-voltage on both ends? \$\endgroup\$
    – quantum231
    Commented Apr 27, 2023 at 22:45
  • \$\begingroup\$ Also, since the GND is actually the same throughout, isn't there still connection between the two sides inspite of using a transformer? Or have I misunderstood this? \$\endgroup\$
    – quantum231
    Commented Apr 27, 2023 at 22:54
  • \$\begingroup\$ There are two problems with opto-isolation , from what I understand, cost and size. Opto-isolation tends to be higher in cost than auxiliary sensing in every circuit I compared. Also the extra circuitry needed to drive the optocoupler might negate some of the advantaged efficiency that it brings. That's why I'm seeking auxiliary sensing as a replacement to Opto. \$\endgroup\$
    – Sam Alamad
    Commented Apr 28, 2023 at 8:21
4
\$\begingroup\$

Yes. The voltage regulation is not as good. You probably will need to have a minimum load.
There are some off the power line supplies that use the aux winding and some use the primary to see the output voltage. Some of the trick is how the aux winding is made.
Some PWM ICs have a trick where they do not look at the power from the aux because of the diode, but they look at the AC signal right on the winding and sample the voltage at a certain time.

\$\endgroup\$
3
\$\begingroup\$

The optocoupler in these converters is used to create an isolated power supply. This way, the ground of two power lines does not have to be connected to each other. This helps a lot when you are connecting long cables to the power line. When the ground is separated, the voltage difference between the main supply and target does not create any issues.

This page explains some benefits of using the isolated supply:

https://www.cui.com/blog/isolated-vs-non-isolated-power-converters

In this document, they explained when it makes sense to go with non-isolated ones:

https://www.emea.lambda.tdk.com/at/KB/When-can-an-isolated-DCDC-converter-be-replaced-by-a-nonisolated-device.pdf

The voltage, current and power are not a factor to which one to go with. Many applications benefit from isolation on the power supply; for example, a 5V 10mA dc-dc converter is used as a power source (Not in POE, but in some small low-power applications). Also, you can see this in AC-DC converters because of the high risk of un-isolated power supply to humans.

\$\endgroup\$
2
  • \$\begingroup\$ I'm not planning to build one without isolation. I'm looking for a design which utilises the Auxiliary side of the flyback transformer to the feedback circuit and therefore controlling the 5V output. It seems all the manufacturers i have looked at, opt for opto-isolation when it comes to 5V and uses the auxiliary for the 12V. I'm wondering why? \$\endgroup\$
    – Sam Alamad
    Commented Apr 26, 2023 at 13:11
  • 2
    \$\begingroup\$ because with windings, you can only detect current (you actually detect flux, but that's the result of current). You cannot know whether that current is what you need on the output side to achieve the right output voltage. \$\endgroup\$ Commented Apr 26, 2023 at 13:14
3
\$\begingroup\$

Is there any other reason [to have opto-isolation?]

The aux winding, especially when tightly coupled to another winding, often has far inferior insulation resistance. The aux winding may only be rated for 500VAC for 1 minute, where a garden-variety opto could be kilovolts.

\$\endgroup\$
1
\$\begingroup\$

"Why opto on 5V and not on 12V outputs" For an aux winding to regulate well it is best to have the aux and secondary to have the same number of turns. Also tightly coupled.
Also, there are diodes on both windings and the ratio of output voltage and diode drop should be 1:1 or close. In the case of 12V aux and 5V secondary the 5V has more than 2x the voltage loss. The voltage loss on the diode is not compensated for. Another way to say that.... 12.7V winding makes 12.0V after the diode, while a 5.7V winding makes 5.0V. The 0.7V is a smaller percentage of the total voltage on the 12V. As temperatures and load change the diode looks.

\$\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.