Designing a 12Volt 10Amp Flyback SMPS with UC3842

[kunal88del]

Hello,
I would like to design a flyback SMPS based on UC3842 with following specs:
Input AC Voltage range: 170 to 265V
Constant Output DC Voltage: 12V
Max. Output current: 10Ampere
Max Power: 120Watt

This SMPS is meant to be used to power LED Lighting strips that operate at 12Volt. In many cases, multiple LED strips may be used with the output that could result in voltage drop at output as the length of each strip is 5 metre. Therefore, I would also want to have some form of voltage preset function in the circuit as well to compensate for the voltage drop & it could be adjusted with some preset.

Now, I have gone thru the datasheets & application notes but I am unable to get a comprehensive list of components to use for this requirement.
If anyone would be kind enough to guide me with selecting components along with calculations then it would be of tremendous help.
We manufacture our ferrite transformers in house hence making prototypes according to design is not an issue, provided I have the winding data. For this design, I was hoping to use ETD39 ferrite core but I would require help with the winding data for the same.
Thanks in advance

Sent from my ONEPLUS A3003 using Tapatalk

[kunal88del]

Hello,
I would like to design a flyback SMPS based on UC3842 with following specs:
Input AC Voltage range: 170 to 265V
Constant Output DC Voltage: 12V
Max. Output current: 10Ampere
Max Power: 120Watt

This SMPS is meant to be used to power LED Lighting strips that operate at 12Volt. In many cases, multiple LED strips may be used with the output that could result in voltage drop at output as the length of each strip is 5 metre. Therefore, I would also want to have some form of voltage preset function in the circuit as well to compensate for the voltage drop & it could be adjusted with some preset.

Now, I have gone thru the datasheets & application notes but I am unable to get a comprehensive list of components to use for this requirement.
If anyone would be kind enough to guide me with selecting components along with calculations then it would be of tremendous help.
We manufacture our ferrite transformers in house hence making prototypes according to design is not an issue, provided I have the winding data. For this design, I was hoping to use ETD39 ferrite core but I would require help with the winding data for the same.
Thanks in advance

Sent from my ONEPLUS A3003 using Tapatalk

One more specific function that I want in this design is that the circuit should go into hiccup mode if maximum load goes beyond, let's say, 150Watt.

Sent from my ONEPLUS A3003 using Tapatalk

[NiHaoMike]

There are so many existing PSU modules available that you'll be better off just buying some.

[kunal88del]

There are so many existing PSU modules available that you'll be better off just buying some.

I would still like to design one from scratch; it will also help me in learning more about power supply circuits.


Sent from my ONEPLUS A3003 using Tapatalk

[BravoV]

When it comes to switching power supply, I don't believe if there is anyone willing to spoon feed you from the scratch, starting from zero up to you able to complete building it.

You have to start yourself some where, start means building the circuit, fail and ask for help, and "pray" someone will help you during the process ... and ... "pray" again that someone is not abandoning you while in the middle of progress.

Your expectation is way too high.

As NiHaoMike pointed out, google for Meanwell or even fake one, buy it, its darn cheap and start from there if you want to learn.

[T3sl4co1l]

I've made them from scratch before (for custom voltages), but not before many years of experience.

I "cut my teeth" on DC-DC converters, 12V in 5V out, that sort of thing.  Learn the basics on low voltage.  Once you have confidence (mind, not that teenaged inexperienced "I can do anything" stage, I mean real experience ), start attempting higher voltages, input and output.  Understand how component performance and parasitics change with scale: both voltage, current and overall dimension of the circuit.

Tim

[floobydust]

At that power level 120W flyback SMPS is not so great, because it only utilizes 1/2 of the B-H curve of the transformer towards energy transfer. I usually go push-pull.

[kunal88del]

A bit of clarification from my end in this regard. We are already manufacturing LED Drivers & SMPS for the past several years. However, most of the time, the procedure has been to get a sample & make it as it is. Now, I have picked up a plethora of knowledge & information while doing it all these years so it isn't going to be from scratch per say. As it turns out, we do have a 12Volt 10Amp SMPS circuit that we have been manufacturing, based on UC3842 that we got from market. However, there are few issues with it, such as it doesn't work at full load. Therefore, we do have starting point & intermediate knowledge but now I was looking for someone pointing me in the right direction & provide help with designing a circuit with UC3842.

Sent from my ONEPLUS A3003 using Tapatalk

[kunal88del]

At that power level 120W flyback SMPS is not so great, because it only utilizes 1/2 of the B-H curve of the transformer towards energy transfer. I usually go push-pull.

I do understand however using flyback topology is something that I am more comfortable with as almost all of my other existing designs are flyback only.

Sent from my ONEPLUS A3003 using Tapatalk

[Richard Head]

A flyback topology is about the poorest choice you can make at this power level. You will find the peak currents in the output diode and primary power switch are huge compared to for example a forward or half bridge. You will also find that the efficiency and power density of a flyback will be poor compared to an alternative topology. A forward converter will require an inductor on the output but you will benefit from much lower peak currents and far less output ripple. A half bridge will allow you to use slightly smaller magnetics compared to the forward but at increased complexity due to the high side power switch. I will omit resonant converters here.
Successful switch mode design is very much an issue of predicting and dealing with non-ideal components such as transformer leakage inductance, reverse recovery of diodes, self resonant frequencies of EMI capacitors etc. Very often these issues can make or break a design. These issues get considerably easier at lower power levels so a first design should be a low power unit. This will allow you to build confidence from a successful design. Once you have a couple of successes a failure is easier to come to terms with. Incidentally, you always learn more from failures than successes so it does have a silver lining.

[xavier60]

A bit of clarification from my end in this regard. We are already manufacturing LED Drivers & SMPS for the past several years. However, most of the time, the procedure has been to get a sample & make it as it is. Now, I have picked up a plethora of knowledge & information while doing it all these years so it isn't going to be from scratch per say. As it turns out, we do have a 12Volt 10Amp SMPS circuit that we have been manufacturing, based on UC3842 that we got from market. However, there are few issues with it, such as it doesn't work at full load. Therefore, we do have starting point & intermediate knowledge but now I was looking for someone pointing me in the right direction & provide help with designing a circuit with UC3842.

Sent from my ONEPLUS A3003 using Tapatalk

Will you show us that design?

[T3sl4co1l]

A bit of clarification from my end in this regard. We are already manufacturing LED Drivers & SMPS for the past several years. However, most of the time, the procedure has been to get a sample & make it as it is. Now, I have picked up a plethora of knowledge & information while doing it all these years so it isn't going to be from scratch per say. As it turns out, we do have a 12Volt 10Amp SMPS circuit that we have been manufacturing, based on UC3842 that we got from market. However, there are few issues with it, such as it doesn't work at full load.

Hmm, I responded based on assumed beginner knowledge; as it turns out you have some experience, so this might go somewhere after all.

Naturally, the challenge then follows: why do you think the "10A" supply can't deliver 10A?  (It's not really "10A" then, is it? )

Will you show us that design?

Seconded -- if it is possible for you to show us, of course!

Tim