Buck converter with transformer output

[throbscottle]

Suppose I connect a buck converter with a transformer as the inductor so I can get an extra supply.

Is this better than using 2 converters? The main supply would be 5v at maybe 100mA, the tertiary winding would supply -24v at 10-20mA. It would be nice to pull a low AC supply off it as well but that might be asking a bit much.

Is this a good way to go?

 

[Papabravo]

I don't think so, and I've never seen an analysis of such an arrangement. Changing load on either side would be reflected to the other side. Sounds like a bad business. You can certainly do forward, flyback, SEPIC, and Ćuk converters to your heart's content.

 

[Diver300]

https://www.ti.com/lit/gpn/lm25576 On page 18 of that it shows how a transformer can be used to give an additional ouput.

I've used that idea to get a 4.5 V supply where the main regulator runs at 3.9 V. It only needs a few turns to get the additional voltage, which are wound around a standard inductor. To get -24 V you are going to need many turns, around 5 times as many as on the main inductor, and so there may not be room to wind them onto an inductor.

The voltage on the winding is proportional to the output voltage in one direction and to (supply - output) in the other direction, so you should arrange the diode that rectifies the secondary winding so that it conducts at the same time as the main output diode, and that will give you an output voltage that is doesn't vary much as the supply voltage changes.

There will be some change in voltage with load, both the auxiliary load and the main load.

I don't quite understand what you mean by a tertiary winding. Buck converters normally have an inductor with just one winding, so an additional one would be a secondary winding.

 

[Papabravo]

Did you mean Figure 19. on page 23?

 

[throbscottle]

Yes, I meant secondary. But I was thinking more in terms of how they describe it in the TI datasheet - "additional winding".
The load will be pretty stable. Total input to the existing circuit (containing a linear regulator, an inverting MC34063 and a 555 based DC-AC inverter) varies by about 5%
Well I've taken apart the transformer from a dead wall-wart now, primary's on. Now I just have to get my head round that diode agreement on the secondary.

 

[schmitt trigger]

Suppose I connect a buck converter with a transformer as the inductor so I can get an extra supply.

Is this better than using 2 converters? The main supply would be 5v at maybe 100mA, the tertiary winding would supply -24v at 10-20mA. It would be nice to pull a low AC supply off it as well but that might be asking a bit much.

Is this a good way to go?

The correct name for this contraption is called a coupled inductor, and it can be done as long as the energy drawn from the auxiliary winding is ALWAYS LESS THAN that from the primary winding.
After reading about this technique in an old ST Micro databook, I used it successfully to provide a -5V supply to a lone opamp in a circuit where everything else was powered from +5V, including the opamp.

 

[throbscottle]

Using a buck converter with a transformer as the inductor to create an extra supply can be a viable solution if carefully designed. It consolidates the process, reducing complexity and cost compared to using two separate converters. However, it requires precise engineering to ensure proper regulation and isolation. While obtaining a low AC supply from the tertiary winding is possible, it might demand more sophisticated circuitry. Professional consultation and testing are recommended to determine the feasibility and optimize the design for your specific requirements.

If this was a commercial thing then yes that would be appropriate. And thank you for the confidence BTW. But it's just me doing a one-off, so it comes down to tinkering and taking advice from people who know what they are talking about...

 

[Diver300]

The correct name for this contraption is called a coupled inductor, and it can be done as long as the energy drawn from the auxiliary winding is ALWAYS LESS THAN that from the primary winding.

On the circuit that I use, the peak power on the primary winding is considerably larger than the load on the coupled winding. However it does run for considerable periods of time with minimal load on the main output and a significant load on the coupled winding.

The circuit that I used is basically what is shown in the TI datasheet that I linked to before. The coupled winding only provides the boost in voltage from the main output to the additional output. On my circuit that is only a boost of about 0.6 V, so given the drop in the Shockey diode, the voltage boost is about 1 V. The current taken by the additional output has to come from the main output anyhow.

Where the additional output is negative, the current cannot come from the main output, so there could easily be the situation where there is insufficient current taken by the main output for the additional output to work correctly.

 

[throbscottle]

So, as long as there is some load on the primary, it's the total load that's important more than how they are distributed? Anyway, the -24 supply will draw much less current than the main one anyway so there shouldn't be a problem.

 

[BR-549]

It's a terrible idea. Play with it for a week. See for yourself. The only way this could be viable is if the primary and secondary loads were constant. Adding transformer action will change the entire coil dynamic. Unless that dynamic is kept within the primary feedback load.

 

[Diver300]

It's a terrible idea. Play with it for a week. See for yourself. The only way this could be viable is if the primary and secondary loads were constant. Adding transformer action will change the entire coil dynamic. Unless that dynamic is kept within the primary feedback load.

It can work in the right conditions. Many small switch-mode power supplies use a transformer with three winding, and they rely on the voltage ratio between the two secondaries to get the correct output voltage.

https://electronics.stackexchange.com/questions/207738/smps-mobile-charger-design

In a buck converter, when the switch is off, the voltage across the inductor is equal to the supply voltage, usually plus one diode drop. A second winding on the inductor will have a voltage across it, and the ratio of voltages will be just about the turns ratio.

 

[tonigau]

This article should have some helpful info for your project...

https://www.eetimes.com/save-a-buck-with-a-coupled-inductor

 

[throbscottle]

Lost count of my secondary turns, got 34v! Works until I load it... However, construction (SBB) leaves a lot to be desired and I forgot to clamp my core together. Lots of ringing.

 

[ChrisP58]

What are you using for a core? My choice would be a powdered iron toroid.

Texas Inst. calls this topology a Flybuck. They have a calculator for it here:
https://www.ti.com/tool/FLYBUCK-FLY...=&tisearch=search-everything&usecase=software

 

[throbscottle]

It's just the core and bobbin from an old wall-wart adapter - or something a little bigger. Now I've got it assembled on perf board the 5v output is nice and solid. Core is held together with a twist-tie.
I've only got 2 sizes of toroid - too big and too small. I've got a couple of power noise-suppression ferrite rings at the small end of the too big scale. Do you think one of those would do?
With the 5v output loaded with 22R, I can get potentially useful output from the overwinding. If I lighten the load, progressively less current is available.
Anyway, searching for "Flybuck" pulls up a shedload of resources, so that's helpful.
So far it looks like my inductance is too great (about 240uH), and possibly I need a more advanced regulator than the 34063.