Interesting DC-DC converter topology

[pplaninsky]

Hello,

I'd very much appreciate if someone explains what is going on with this power supply.
That is old portable CD-Player from Technics. Runs on DC 4.5V or two AA batteries. The power supply takes either voltage as input.

Questions:
Most interesting:
1. On L12 - seemingly the coils are shorted. Looks like a transformer with shorter primary and secondary.
2. Q3 is designated as regulator. How does it regulate.
3. In case of supply from batteries, the supply voltage could easily below 3V. So, on baterries it must be boosting the voltage to 3.4V and on AC adapter lowering the voltage from 4.5 to 3.4V.

Less interesting:
4. What is the purpose of L11
5. One of the voltage rails is marked as 5.1V. That rail is connected directly to the DC in jack, which is designated as 4.5V. In the manual they say they have used AC adaptor for power supply. Maybe it was outputing 5.1V? Or there is another explanation for the boost from 4.5 to 5.1V
6. Finally, on the block diagram the voltage after the regulator and after the L11 are both marked as +B. Aren't these different voltages? One is regulate the other is coming from the AC adapter? Same are marked on the schematic as 3.4V and 5.1V.


Thank you!

[pplaninsky]

My best guess for point 1 and 2 is:
1. That is not a transformer with primary and secondary coil, even though in the reference application of the DC-DC converter chip, the default schematic uses transformer.
2. It is just an inductor, coupled to another inductor via ferite core and somehow (magically) that is used to regulate the base of Q3. Some kind of feedback mechanism via the couple inductors. Still the whole thing is a mystery to me.

[Doctorandus_P]

The circuit reminds me of the "joule thief", which is a quite ubiquitous schematic
It's even got it's own wikipedia page:

https://en.wikipedia.org/wiki/Joule_thief

But I have not looked closely at your schematic. Redrawing it "properly" first may help in understanding how it works.

[T3sl4co1l]

Seems like Q3 is being used as a synchronous rectifier.  Not a very good way to use an NPN but Veb is probably enough not to mind. Low supply voltage and all.

Tim

[Whales]

1. On L12 - seemingly the coils are shorted. Looks like a transformer with shorter primary and secondary.

None of the coils on L12 are shorted.  Where the thick, red wires on the diagram cross each other: they are NOT joined unless they have a circle drawn at the intersection.  Look above R6 for an example of that.



Looks like a self-oscillating power supply using one transistor and a transformer ("inductor") with a second winding.  The second winding is weak/thin, it only needs to provide a little bit of current back to the transistor base to keep things going.  EDIT: wait there is no diode... T3slacoil you're confusing me, I didn't think it possible to both switch and rectify with a single device?

L11 and the two capacitors either side of it are a "pi filter".  In this case a lowpass filter to try and get rid of the noise made by the (likely very noisy) self-oscillating power supply.  EDIT: wait what, it's on the input not the output.
on a variety.

EDIT2: This thing uses a DC-DC controller chip!  Ignore all of the crossed out section above.  You really need to redraw this schematic in a more sensible (less tangled) way to make sense of it, it looks like they drew it compact just to fit it on the minimum of paper.

[Doctorandus_P]

EDIT2: This thing uses a DC-DC controller chip!  Ignore all of the crossed out section above.  You really need to redraw this schematic in a more sensible (less tangled) way to make sense of it, it looks like they drew it compact just to fit it on the minimum of paper.

Yah!

If you look at the AN8086 chip pin 5 is labeled as "out" and it goes directly to transistor Q5.
All the text pointing to that wire is cut off though 
but even so, I can still see the end of a "MHz" remark.

L11 apparently has polarized capacitors on each side (C13 and C14) both probably 10V 47uF, so that would make it a passive low pass filter.

Maybe Q3 is used as an active diode that switches synchronous with the SMPS circuit?
It looks like that when Q5 switches to GND the base of Q3 is closed too via C11.
But that does not clarify the voltage drop from 5V1 to 3V4 over Q3.

CN8 is the DC input jack (4V5) and it's built in switch opens up the GND lead of the rechargeable battery, so that definitely is where the power comes from. Charging batteries with only some added series resistance (R19 and R20) does not look very good though...

[pplaninsky]

I will redraw it  - but here is the simplified block diagram from the original documentation.
Clearly, L12 has the two coils tied together at one end. No isolation, like in usual primary-secondary transformer.

[Terry Bites]

That looks like a magnetic amplifier. Changing the dc flux in the primary changes the saturation point in the indcutor. Hmmmm

[GumpyGus]

This is a switching converter. Probably runs at a highish frequency, maybe 100 to 500 KHz.   So the transformer is going to look as a dead short with a DC-powered ohmmeter.  Only with like an old Heathkit Q-meter will you see any impedance there.

The voltages shown may be theoretical and actually measured values, or on battery versus external power.  Too lazy to read the footnotes!   Voltages can be off by 25% and still be perfectly fine,  don't sweat the small stuff. 

There are often mistakes on schematics,  that can be ultra confusing.  There also are many unusual circuit topologies with like grounded plates and collectors, see any Tektronix scope schematic.   Not every circuit is as straightforward as in the textbooks!

[pplaninsky]

I have repaired the board and it is working now, but I was just curious to understand how exactly works.
The most confusing thing is the one marked with red on the picture below.

[T3sl4co1l]

The one side goes to Q3, it's just a feedback winding.

Tim

[Whales]

The transformer is isolated as it comes from the factory.  People using it are free to throw away this feature when they wire it up   It's not being used for a safety isolation role, just a measurement/response/control role, so joining its terminals like that is fine.

[Doctorandus_P]

There is indeed nothing special about that connection.

In a transformer each separate winding just gets it's own voltage dictated by the turns ratio between the windings. When you make an external connection between different windings, that just means you have put them in series.

For example, if you've got a normal low frequency mains transformer for 230Vac -> 12Vac, then you can either use it to generate an isolated 12Vac, or you can use the secondary winding to either generate 242Vac or 218Vac.

You only create a short when you make an outside connection between different sections of the same transformer coil.