How on earth does this power supply work?

[TheMG]

This is a Coutant/TDK Lambda MML400 series switching power supply configuration 24G (24VDC nominal 15A 400W max).

What I simply can not quite understand is how the feedback from the secondary circuits get to the primary switching circuits.

Typically, a switchmode power supply would use either a small transformer or opto-isolator to provide an electrically isolated feedback signal, necessary to obtain voltage regulation.

However, this power supply has neither. The ONLY connection between the secondary circuits to the rest of the power supply is through the main transformer's secondary windings, represented by A, B, and C on the picture. B is the center tap of the transformer secondary.

I can not find any service manuals or schematics for this power supply online.

The only major power-handling components on the secondary side are the usual schottky diode and filter inductor and capacitors. So there's no further voltage regulation done on the secondary side.

There are no surface mount components on the backside of the PCB on the secondary side. The only components on the other side of the board, that can't be seen in the picture, are the schottky diode, capacitors, and a few flameproof resistors. The LM393D ICs that can be seen, are dual differential comparators according to the datasheet. Without reverse-engineering the whole circuit, what sort of voodoo magic is going on here?

Clearly, there has to be some form of feedback across the isolated gap between primary and secondary (otherwise there would be no way of obtaining precise voltage regulation), but how are they doing it? Any ideas?

[TheMG]

Primary-Side Sensing Takes
Complexity out of Isolated
Flyback Converter Design

Unfortunately I don't think that's what's going on here. The voltage on this power supply is adjustable from the secondary side (see the blue 10-turn pot in the picture).

If the flyback primary sensing topology was used, voltage adjustment would have to be done on the primary side if I'm not mistaking. The power supply is also equipped with remote voltage sensing capability.

PS: the main controller IC on the primary side appears to be a UC3844 which to me looks like a bog standard SMPS controller.

[DanielS]

My guess: primary-side sensing for rough regulation, magnetic regulation on the secondary side for trimming.

[wraper]

If it is UC3844 based, there should be optocoupler hidden somewhere.

[wraper]

Post a photo of the PCB where UC3844 is located.

[Andy Watson]

The only major power-handling components on the secondary side are the usual schottky diode and filter inductor and capacitors. So there's no further voltage regulation done on the secondary side.

The regulation is probably performed on the secondary side - so that you can have one main power transformer supplying several secondary modules. Can you see the connections to the "filter inductor". I bet there's more than input and output. The regulation is done by controlling how near to saturation the filter inductor operates. Essentially the filer inductor is being used as a magnetic amplifier.  The primary side being operated continuously at full bananas.

[ajb]

The primary side being operated continuously at full bananas.

  That's either a hilarious autocorrect, or my new favorite bit of jargon.  I'm okay with either one. 

[atferrari]

  The primary side being operated continuously at full bananas.

Sorry but that translates to what?

[Len]

[NiHaoMike]

Is that Freelee the Banana Girl's car by any chance?

Back on topic, I'm going to guess that they just have a postregulator.

[T3sl4co1l]

Yeah, postreg is a possibility.

Nearly impossible to say without a breakdown of all boards.  That thing is packed.  I count six, maybe eight boards, and that's just from this angle?

Tim

[Psi]

It's been like 10 years since i opened my MML400 up, but from memory its a dual stage powersupply design.
One big psu that generates unregulated high frequency AC from the rectified mains. And each module which takes this high frequency AC and generates its own output voltage.
So if i'm remembering correctly the feedback will be internal to each module.

With the high frequency AC available to each module they only need a tiny high freq transformer inside. (possibly a magnetic amp as others have said)

I do remember being confused also when i took mine apart

[Whales]

  The primary side being operated continuously at full bananas.

Sorry but that translates to what?

Full bananas = very full or very high

ie the primary side is being operated at a constant high level, rather than it reducing when less power is needed.  "It's being run at full bananas".

[TheMG]

The regulation is probably performed on the secondary side - so that you can have one main power transformer supplying several secondary modules. Can you see the connections to the "filter inductor". I bet there's more than input and output. The regulation is done by controlling how near to saturation the filter inductor operates. Essentially the filer inductor is being used as a magnetic amplifier.  The primary side being operated continuously at full bananas.

Well, I scoped it (as well as I could with my crappy scope anyways) and can see no change in the secondary waveform while adjusting the output voltage. The waveform amplitude is also quite a bit higher that the expected output voltage.

So there has to be something going on in the output module. The thing is, the ONLY power device is a single TO-247 device (unfortunately I can not see the part number as it is sandwiched between a board and the heatsink) labelled on the PCB as "D30".

I suspect despite being labelled as "D30" it is probably not just one of those dual-diode things commonly found in switchmode supplies, but possibly a MOSFET being used for both regulation and rectification.

This is made even more likely as I had another closer look at how the output module is connected to the main transformer's secondary. There is in fact just one winding without a center tap. The contacts I labelled "A" and "C" in the picture are in fact joined together under the top PCB.

I wish I could tell what "D30" is, but it would take some serious disassembly to get to it.

There's 3 inductors in there, all with only one winding.

I've attached a quick sketch of what lies beneath the top board that you can see in the picture. Please excuse the crudity of the paint drawing.

[DanielS]

I wish I could tell what "D30" is, but it would take some serious disassembly to get to it.

From your diagram, I have a hard time imagining that D30 being anything other than a dual schottky diode. A synchronous or otherwise controlled rectifier would require some form of control/timing circuit attached to its gate, which would usually be pin #1.

[TheMG]

I really wish I could find something somewhere on how this works. It's definitely got my curiosity as it's the first I've seen designed this way.

I have a few other configurable/multi-output power supplies from other manufacturers and while the output modules share a common PFC pre-regulator, each output module is a self-contained switching power supply operating from a common high voltage DC bus.

Can't even find any mention of any patent numbers or anything anywhere.

[mikeselectricstuff]

Look up magnetic amplifier regulator - commonly used for auxilliary supplies on PSUs with multiple outputs.

[TheMG]

I think that's exactly it. Magnetic amplifier. I read up on it a bit and it does seem to match what's happening in this PSU.

Now it makes perfect sense but since I had previously never heard of a magnetic amplifier I couldn't make any sense of what was going on in the circuit.

Interesting concept for sure. I had heard of and seen power supplies using core saturation to provide voltage regulation - but those were big heavy 60Hz units. Never heard of a similar principle being used in switchmode power supplies until now.

As Dave said "don't turn it on, take it apart!". Lots to be learned by doing so! I initially took it apart to check out the build quality (and I'll add that it's a top-notch power supply in that regard, can't fault it), that's when I noticed there was no apparent feedback path between primary and secondary.

[DanielS]

Never heard of a similar principle being used in switchmode power supplies until now.

It has been known and documented for 20+ years but the extra cost and complexity have kept it from becoming common. With today's highly integrated PWMs and drivers, it is often cheaper and simpler to power auxiliary outputs from auxiliary DC-DC converters than magnetic regulators.

[T3sl4co1l]

Saturable reactors have been used for decades on the +3.3V rail in ATX supplies.  Usually a peculiar plastic-cased inductor near the output side (what the heck is that?).  This is also why the 3.3V rail has wonderful regulation (usually much better than spec), while 5 and 12 (which are derived from common windings) are more mediocre (near max spec).

Tim

[Andy Watson]

I thought the banana scale was a widely recognised engineering frame of reference I didn't realise it was so controversial!

The attached pdf is the schematic of a Coutant 5V power unit that employs a magnetic regulator. The input from main power transformer is between the terminals F ans S. L1M is the saturable core inductor - aside from some filtering there is no other regulation between it and the output terminals. The action of the D2 will cause the core to saturate and pass maximum current. Regulation is achieved by driving the other winding on the core, L1R, to take the core away from saturation. That's the oddity with this configuration - to reduce the current through the core it has to be driven (via D3 and TR1).

Apologies for the cruddy layout. This schematic has been recovered from a 20 year old Easytrax file - it originally printed on fan-fold paper. I can't remember why I reverse engineered it - must have had some big plans at the time!

[TheMG]

Saturable reactors have been used for decades on the +3.3V rail in ATX supplies.  Usually a peculiar plastic-cased inductor near the output side (what the heck is that?).

Can't say I've ever seen that in any ATX power supply I've ever taken apart. Maybe this is the case in some higher end power supplies but most run-of-the-mill ATX supplies don't employ such extra regulation. Most of them simply use the 3.3V rail with a remote voltage sensing wire as the reference for the entire power supply. The 5V and 12V rails are usually just along for the ride, as evidenced by the fact that loading the 3.3V output of a typical ATX power supply will cause the remaining unloaded rails to climb in voltage.

More modern ATX power supplies only produce 12V out of the main switcher, and the 5V and 3.3V rails have their own DC-DC converters powered from the main 12V output, therefore as each rail is individually regulated there is no need to have any load on either the 5V or 3.3V rails.

[TheMG]

The input from main power transformer is between the terminals F ans S. L1M is the saturable core inductor - aside from some filtering there is no other regulation between it and the output terminals. The action of the D2 will cause the core to saturate and pass maximum current. Regulation is achieved by driving the other winding on the core, L1R, to take the core away from saturation. That's the oddity with this configuration - to reduce the current through the core it has to be driven (via D3 and TR1).

Schematic appears to be not too far off.

Odd configuration indeed, but seems to work quite well.

I'm guessing the main objective by the manufacturer was to be able to use the exact same primary switcher and transformer for a number of different power supply configurations, and simply bolt in the necessary output modules according to customer's request. I'm also assuming that with the magnetic amplifier configuration, only a limited number of component changes are required for a wide variety of different voltages and currents.

So it's all about minimizing manufacturing and design costs while producing custom power supply configurations to suit the customer's needs. Not having to custom-wind different transformers is probably a big money and time saver - ie, the same transformer is probably used in all 400W power supplies regardless of output configuration.

[DanielS]

Can't say I've ever seen that in any ATX power supply I've ever taken apart.

Most affordable ATX power supplies (as in sub-$80 or so) are group-regulated affairs in either flyback or half-bridge configurations. The oldest ones regulate the 5V output while the not so old ones regulate based on the 3.3V sense line. In modern Haswell-ready models though, 12V is the new king.

I do not remember seeing magnetic regulation in any of the PSUs I have opened either. The only reason I knew about them is because I skimmed a chapter about them in a switching supply design book many years ago.

[TheMG]

12V is the new king.

Yeah. It's been a long time coming but were almost to the point of having PCs that run on all-12V (actually, HP has started doing this in most of their recent business PC lineup).

With the ever decreasing cost of DC-DC converters it just makes sense to put the lower voltage regulation on the boards themselves instead of in the power supply.

Most modern "high-end" ATX supplies use DC-DC converters to produce the 3.3V and 5V outputs. Eventually I think we'll get to a point where even this will be dropped and PC PSUs will be 12V-only.

Personally, I can't wait. This sure will take some of the rat's nest out of computer cases!

They also need to come up with a higher current standard connector. It's getting a little out of hand with some graphics cards having 12 or more pin 12V power connectors to supply the necessary current, each with pin having its own 18AWG wire. IMHO they should pick a new standard connector with a much greater current rating for the motherboard, graphics card, and any other high-current hardware devices.

I have to say as an amateur radio operator I have a slight bias toward Anderson Powerpole connectors. Reliable, configurable, and the connector size is fairly small despite 30A per pin rating! Other nice thing is the gender-less design. I'd like to see those used in PCs too.