0
\$\begingroup\$

I am new to power electronics. I am designing an isolated DC-DC converter using an LT3002 and a ZA9672-BED (12V, 1A) transformer from Coilcraft. My target is 12V DC output with 10W power. The transformer can provide 12V with 1A.

I want to use an LT3088 linear regulator at the transformer's output to get a stable 12V. To do this, I must get around 14V at the ZA9672-BED (12V, 1A) thanks to the LT3088.

I am confused about if it can provide 14V (10W) or higher voltage on the ZA9672-BED (12V, 1A) rated 12V, 1A transformer. Will that cause any problems?

Circuit:

enter image description here

Result:

enter image description here

\$\endgroup\$
2
  • 1
    \$\begingroup\$ Show your proposed flyback circuit design please. There is some confusion in the words you are saying that should be solved with a schematic. You talk about the transformer's output connecting to the LDO regulator but, I don't seriously think you mean that. Please show your proposed design schematically. Add data sheet links for the two chips you mentioned as well. \$\endgroup\$
    – Andy aka
    Commented Apr 10, 2023 at 12:54
  • \$\begingroup\$ I added all of the related information. \$\endgroup\$ Commented Apr 10, 2023 at 13:09

2 Answers 2

1
\$\begingroup\$

I am designing an isolated DC-DC converter

[snip]

want to use an LT3088 linear regulator at the transformer's output to get a stable 12V.

Any reason that you want to use a linear regulator on the output of the SMPS DC-DC converter?

Linear regulators are used typically on the outputs of a 230V 50Hz transformer with capacitor filter. In this case, the transformer won't provide a stable voltage since most transformers are lightweight and to create a lightweight transformer, you need lots of turns or else the core will saturate merely from the idle magnetizing current, meaning the resistance of the transformer windings is huge, meaning that any current at the output will create huge voltage drop. Also the capacitor filter creates pulsating DC, not true DC, meaning the linear regulator is needed.

This can be observed by an unregulated non-SMPS "DC" power supply that has a transformer and capacitor filter. Typically, a 12V nominal power supply can give you 16V or more when unloaded. The current rating is typically ridiculously low, let's say 0.5 amperes. When you put the 0.5 amperes, you will observe that the average voltage at the output isn't even the advertised 12 V, it's less than that! None of that happens in SMPS DC-DC converter power supplies.

However, SMPS DC-DC converters are typically a different beast. The frequency is very high, so a lightweight transformer will do without huge winding resistance. The ripple at the output is measured in millivolts. Because of the high frequency, the output capacitor can hold the voltage during a switching cycle, meaning no more capacitor ripple (unless you care about a millivolt ripple). No matter how much you load the output, it gives exactly the advertised voltage, until you hit the current limit, in which case it will start to reduce the voltage to maintain the current limit.

I guess someone might want to do an exceptionally low-ripple SMPS DC-DC converter. Typically you would want to do that by bumping up the frequency as high as possible and ensuring the capacitors are large enough, and ensuring the control algorithm is good. However, you might in addition further filter the output with a linear regulator to further reduce the ripple. This is made somewhat harder by the ripple rejection of linear regulators being optimized for mains frequency (50 Hz), and being much lower at typical SMPS operating frequencies. But it might reduce the ripple even lower, to allow using an SMPS in applications that absolutely demand the lowest possible ripple you can get.

This would be extraordinary, however. I don't believe any common lab power supply for example would use SMPS and linear regulators at the output of the SMPS.

Typically, SMPSes do not require linear regulators at the output.

I am confused about if it can provide 14V (10W) or higher voltage on the ZA9672-BED (12V, 1A) rated 12V, 1A transformer. Will that cause any problems?

Transformers have no voltage rating, unless we are talking about isolation (and no transformer would have isolation rated only for 12 volts). Transformers have current rating and inductance. If the transformer is rated for 1 A, you can put 1 A through it. If they say it's rated for 12 volts (whatever that means) and you use it at 14 volts, you're fine. It will work. However, in some circuits bumping up the voltage from 12 volts to 14 volts might require slightly larger inductance for optimal operation. However, the increase from 12 volts to 14 volts is so small that it is extremely unlikely to cause any problems.

Furthermore, any general purpose SMPS IC allows setting an arbitrary voltage. The older answer explains how to set it with resistors.

\$\endgroup\$
5
  • \$\begingroup\$ > Linear regulators are used typically on the outputs of a 230V 50Hz transformer with capacitor filter. | Uh, what? \$\endgroup\$
    – MrGerber
    Commented Jun 9 at 8:00
  • \$\begingroup\$ >This is made somewhat harder by the ripple rejection of linear regulators being optimized for mains frequency (50 Hz), and being much lower at typical SMPS operating frequencies. | [Citation needed]! \$\endgroup\$
    – MrGerber
    Commented Jun 9 at 8:02
  • \$\begingroup\$ > any general purpose SMPS IC allows setting an arbitrary voltage. | Again, if you define "general purpose switchmode controller" as one that allows you to set the voltage, then - yes - but there are definitely a plethora of common switchmode controllers that only provide fixed output voltage. \$\endgroup\$
    – MrGerber
    Commented Jun 9 at 8:03
  • \$\begingroup\$ @MrGerber Here's the citation: ti.com/lit/ds/symlink/lm317.pdf -- figure 8 shows much less ripple rejection at SMPS frequencies than at mains frequencies (like 50 Hz). \$\endgroup\$
    – juhist
    Commented Jun 9 at 8:44
  • \$\begingroup\$ Ripple rejection goes down with increasing frequency, yes - but to generalize that into saying LDO's are designed for use at 50Hz is incorrect. LM317 response is flat into the 10kHz range - 3 orders of magnitude over mains. \$\endgroup\$
    – MrGerber
    Commented Jun 10 at 6:32
0
\$\begingroup\$

First of all, it's not recommended to use components at their maximum rated power or very close to it. You should always leave some "room", at the very least 10-20% to make sure your circuit will not fail or misbehave due to overload.

Second of all, you can set the output voltage from the transformer with resistors connected to RFB, RREF and TC pins of the LT3002, as specified in its datasheet on page 5:

The ratio of the RFB resistor to the RREF resistor, times the internal voltage reference, determines the output voltage (plus the effect of any non-unity transformer turns ratio).

Third of all, do you really need the LDO at the output? Do you know why are you using it?

\$\endgroup\$
2
  • \$\begingroup\$ I want to limit the output power to 10W. LT3002 and the transformer can provide 12W at 12V output. I think that if I can use LDO with a maximum output current of 800mA at 12V, like in the circuit above, I can limit the output power while having a stable 12V. But I must increase the output voltage (out_flyback) to get 12V from LDO. By increasing Rfb, I can get 14V, but the transformer (ZE9672-BED) is rated for 12V. My question is increasing this voltage to 14V in the circuit (still 10W) might be a problem for the transformer? \$\endgroup\$ Commented Apr 10, 2023 at 13:25
  • \$\begingroup\$ @FiratDagkiran No, it wouldn't be a problem, that's a relatively small change. You shouldn't use the LDO's absolute current limit to limit the output power. Depending on your load, the power can be limited via constant voltage, constant current or both. \$\endgroup\$ Commented Apr 10, 2023 at 16:42

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.