advertisement


SMPS schematic or similar

stevec67

pfm Member
I recently had an Amazon 5V/1A device charger pack up, these things happen. I didn't intend to repair it, it's not worthwhile, but I cut it open to inspect the innards. As suspected, there was a small capacitor leaking and dead. As always I am impressed with how few components they use. There are 2 transformers, one smaller than the other, a few capacitors, one transistor and 2 diodes I could see. That's it, for a decent quality 5V/1A PSU. It's a "how do they do it so cheaply?" product.

I'm intrigued as to a typical schematic. I know that the general principle is that the incoming mains is switched on and off by a circuit controlled by a high frequency oscillator, this then generates a very rough on-off wave akin to a square wave, and this is then smoothed to some approximation to DC. The mark-space ratio dictates the final voltage. That's as much as I know. Does anyone have a more comprehensive explanation? I'm particularly intrigued by the need for 2 trafos, one of which is tiny, like a bucking trafo might be.
 
Interesting things. I’ll follow the above links and try and get my head round how they work! As with a lot of electronics I kind of almost understand the real high-level broad brush stuff, but the detail is far beyond me.

I was replacing a blown Rifa X2 cap in what must have been one of the very first of this breed yesterday:

52751494793_4341b38387_b.jpg



The PSU from an Apple II computer (1977, though this particular one is from about 1981 I think).

PS I had a Sony TA-F6B amp for a while that I think was around the first hi-fi product to have a switch mode PSU. A really nice thing, though it exploded quite spectacularly and I could never find anyone to fix it so ended up giving it away. Annoying really, I’d have had a go myself now. Quite a nice high-end Sony amp too.
 
Great stuff, thanks gents. Big Clive's post was a good start, the teardowns didn't tell me how it worked, but there's a chap on YT who talks you through it. The regulation stage is brilliant and even uses an opto coupling to feed back output voltage to input switching frequency to ensure mains isolation. The one he buys and reverse engineers has more technology in it than a Hi-Cap, and it costs $20. Somebody, somewhere, is absolutely and totally taking the mick.
 
Well - yeah, at near-LF: unless you want a quiet supply.

The side-effect of SMPS is a roaring-hot-mess of common-& differential mode HF+ noise; & few easy ways around it.
Surprisingly-enough - that can be a far-nastier issue for linear, analogue approaches (esp amplifiers), than ... simple/stupid/linear supplies - by far.

Pick your poison ;)
 
The problem with the Big Clive video is the PSU he is featuring is so modern it has most of the functionality shrunk down on one chip. The ancient Apple II PSU is way easier to understand to my mind (though I’m not suggesting I do fully).


Here’s a video of someone pulling one apart. Sadly he doesn’t go into real circuit detail, but being so old all the parts are big and identifiable with the help of the Wikipedia SMPS page.

To my understanding and starting from the mains input IEC socket end (at the bottom of my pic upthread) we have the high mains input, some filtering for RFI and to stop it kicking crap back out onto the mains (US RFI legislation was very strong at this time), the big switching transistor thingy in its heatsink, the bucking transformer it controls, and then we are onto the low voltage end where it is chopped and inverted into +12V, -12V, +5V and -5V, then smoothed some more with the caps at the far end and out to the appropriate parts of the computer.

PS A remarkably similar PSU in the BBC Micro around five years later:

52756336794_582ed5302e_b.jpg


They’s simplified it a bit, and I don’t think it was anything like as bothered about RFI, but the basic bits are all there to a remarkable degree. It looks so like the Apple design to my eyes. Here the high voltage and low voltage ends have a nice neat physical wall between them!
 
If you have a dead SMPS which does not appear to have bulging caps or burnt out components the likely faults are small electroytics around the chip and/or the high value resistor which drops the voltage from the rectified mains to the chip to start up the regulator. These resistors are often not rated for the voltage even though well within their dissipation spec.

Shorted diodes and faulty electrolytics on the output side are also common, this can cause the supply to simply shut down.
 
I had a BBC master power supply that would not start up. In that case the bootstrap for the startup was taken from the lower reservoir cap that was one of two in series. They needed balancing resistors to keep the voltage similar across both of them and to bleed the charge after switch off. I found out the hard way that one of those resistors was open circuit.
 
I had a BBC master power supply that would not start up. In that case the bootstrap for the startup was taken from the lower reservoir cap that was one of two in series. They needed balancing resistors to keep the voltage similar across both of them and to bleed the charge after switch off. I found out the hard way that one of those resistors was open circuit.

I had an identical problem on a small disc drive supply once. It was sitting on the bench connected but not working. I disconnected it and picked it up to look underneath. Guess how I found that one of the bleeder resistors was O/C
 
  • Like
Reactions: MJS
It's amazing how far you can throw them. I also did it to a Naim SNAPS once where I was attempting a one-handed insertion of the mains lead but forgot the fuse holder at the top of the chassis wasn't well insulated.
 
I recently had an Amazon 5V/1A device charger pack up, these things happen. I didn't intend to repair it, it's not worthwhile, but I cut it open to inspect the innards. As suspected, there was a small capacitor leaking and dead. As always I am impressed with how few components they use. There are 2 transformers, one smaller than the other, a few capacitors, one transistor and 2 diodes I could see. That's it, for a decent quality 5V/1A PSU. It's a "how do they do it so cheaply?" product.

I'm intrigued as to a typical schematic. I know that the general principle is that the incoming mains is switched on and off by a circuit controlled by a high frequency oscillator, this then generates a very rough on-off wave akin to a square wave, and this is then smoothed to some approximation to DC. The mark-space ratio dictates the final voltage. That's as much as I know. Does anyone have a more comprehensive explanation? I'm particularly intrigued by the need for 2 trafos, one of which is tiny, like a bucking trafo might be.

The small transformer is more probably a common mode choke on the mains input. It is there to reduce RFI getting back down the mains. (CE compliance)

Generally speaking there is an optocoupler to provide isolated feedback from the output rail to the control chip to regulate the voltage, by varying the mark space ratio, as you say.

Sometimes each rail has a choke capacitor filter to improve the ripple on the output but often this is not required due to the high switching frequencies involved.
 
Well - yeah, at near-LF: unless you want a quiet supply.

The side-effect of SMPS is a roaring-hot-mess of common-& differential mode HF+ noise; & few easy ways around it.
Surprisingly-enough - that can be a far-nastier issue for linear, analogue approaches (esp amplifiers), than ... simple/stupid/linear supplies - by far.

Pick your poison ;)
Yes, you are right about the noise level. I wasn't suggesting that you would want to use an SMPS on a JLH 1969 or similar, my comment about engineering quality and design input is that a £20 SMPS features all sorts of clever features relating to safety and performance, while a linear power supply consisting of a trafo, BR, some capacitors and (alert the bloody space programme) a couple of LN317 integrated circuit regulators used in accordance with that cutting-edge new science research project "The LM317 datasheet and application notes" sells for BLOODY HOW MUCH? and has technology that wasn't new in the 1970s.
 


advertisement


Back
Top