Manson 80W Constant Power Switching Mode Power Supply

[caryg]

I'm trying to get the above power supply working. On the outside it's a neat little thing that you can dial up any voltage from 0-36 V and limit the current via two rotary encoders connected to two LED displays, and independently switch the output on and off (link to manufacturer's site and also sold under various brand names). Someone's done an independent stripdown here of the version shown by Maplin in the UK, which reveals it all to be a bit of a tight squeeze with lots of output capacitors sandwiched between two boards and further insulated with a nest of wires.

I got it as not working with no history of what had gone wrong. It lit up, but the two displays showed 0.00 or would occasionally flicker with a hundredth of a volt, but I could still set a couple of limits.

A look inside showed effectively two power supplies on the board, one to run the internal electronics (the smaller, top, inverter in the picture attached) and the other to supply the output voltage. After following a few dead leads I worked out that the PWM chip (just below the lower heat sink) for the output wasn't working. I changed that and now I can get an output voltage, but it's neither adjustable nor stable. It starts at about 3.5 v then gradually wobbles its way upwards at about an average speed of a hundredth of a volt every three or four seconds, although it isn't smooth.

But what is more puzzling for me is the unstable voltages I'm getting on the primary side. At the bridge rectifier (top left) I'm getting a stable 237 V AC input, but on the DC side pins it's anywhere between 288 and 318 (I don't have a scope to get any more detail). I pulled the big red capacitor and measured it with my ESR meter but it seemed OK. Then across the two input capacitors (Capxon 250v 220 μF) I'm getting between 428 and 449 volts. That's only 20 volts of wobble rather than the 30 at the rectifier, so they are doing some smoothing, and again they check out OK on the ESR meter.

So does anyone have any suggestions on what could be causing the unstable voltage on the primary side? Could that yellow transformer-like thing above the smoothing capacitors have anything to do with it? What exactly is that yellow thing? It's labelled F8080000 which throws up too much noise for Google to be any help. And what in the circuit is boosting the voltage on the primary side? I would have thought that ordinary rectified mains would be enough for the Mosfets.

Thanks in advance for any ideas or hints.

[b700029]

Re: Manson 80W Constant Power Switching Mode Power Supply

At the bridge rectifier (top left) I'm getting a stable 237 V AC input, but on the DC side pins it's anywhere between 288 and 318 (I don't have a scope to get any more detail). I pulled the big red capacitor and measured it with my ESR meter but it seemed OK. Then across the two input capacitors (Capxon 250v 220 μF) I'm getting between 428 and 449 volts. That's only 20 volts of wobble rather than the 30 at the rectifier, so they are doing some smoothing, and again they check out OK on the ESR meter.

Are those two caps in series or parallel? If parallel, then 428-449V is well past their voltage rating and they should've exploded a long time ago, or you're not measuring correctly. If series, then the voltage is still too high --- 237VAC should result in 335V across both in series (half that voltage across each one.) I would also check all 4 diodes of the rectifier.

Tracing the primary circuit and posting a schematic would be good, that will help confirm the slight but present possibility that the primary uses a voltage doubler.

[momaka]

Re: Manson 80W Constant Power Switching Mode Power Supply

Are those two caps in series or parallel? If parallel, then 428-449V is well past their voltage rating and they should've exploded a long time ago, or you're not measuring correctly. If series, then the voltage is still too high --- 237VAC should result in 335V across both in series (half that voltage across each one.) I would also check all 4 diodes of the rectifier.

I'm more inclined to think the PSU has active PFC. That would explain the high voltage across the two caps on the primary side. The yellow transformer right above the primary caps is likely the PFC boosting coil.

Judging by the lack of a 3rd transformer between primary and secondary and the presence of the optocouplers on the bottom of the picture, this PSU likely uses a single-transistor forward configuration. Since there's also 3 ICs on the primary, 1 of them is most likely the APFC controller, 1 for the standby circuit, and 1 for the main PSU circuit.

That's about all I can determine, though. Will need the part numbers of the 3 ICs on the primary side as well as some underside pics of the PSU to determine how exactly it functions.

My guess would be that the unstable voltage on the primary side is perhaps caused by a fault in the APFC circuit. Or perhaps the NTC thermistor has gone high in resistance?

[caryg]

Re: Manson 80W Constant Power Switching Mode Power Supply

Thanks for the suggestions so far. It's given me a lot more to think about, even if I now feel I understand a lot less than I did yesterday. Attached are a couple more shots of the board, front and back.

The IC circled in red is (I think) the PWM controller for the current the PSU outputs. It is a UC3843BN, leading to a K2717 Toshiba N-type Mosfet specced at 600V and 5A and then to the large transformer. This initially had nothing on the output pin, so I swapped it out and now I can output power, although I can't yet control how much.

Circled in green is a VIPer 12A-DIP which combines PWM and Mosfet on the one chip. Through its associated transformer it powers the two ATMegas and associated circuitry which are the brains of the PSU. My multimeter shows big fluctuations on the drain leading to the transformer from 0V to over 1000 (I guess as that is its limit). As further upstream I get a fairly stable 7.5 volts going in to a couple of 7805 regulators, I guess that the fluctuations are due to limitations on the way my multimeter works.

Marked in blue is a MC33262P Power factor controller, feeding a F21NM60N 600V 17A N-type Mosfet, then my "yellow transformer-like thing". That then feeds back in to the primary side, so is I guess the APFC. I've been unable to find a useful guide to how these work, there being lots of "this is a good thing and will save the planet" stuff around, and a fair bit of heavy maths but nothing in between, so here are some readings I took with my multimeter:

Pin  description         voltage

1  feedback             2.7
2  input compensator      0.8 - 1.7
3  multiplier input        2.5 - 2.9
4  current sense input       0
5  zero current detect input 0 - 1.3
6  GND
7  Drive Output          0 - 0.5
8  Vcc                8.3 - 10.5

And lastly, just to the right of the Mosfet circled in blue is a space silk screened and drilled for a capacitor, but with no Cx identifier and nothing there. Measuring between the two points gives me up to 450v. Would it help to stick a small ceramic cap there?

[b700029]

Re: Manson 80W Constant Power Switching Mode Power Supply

Look at the MC33262P datasheet, there's a good explanation there along with example circuit.

The extra capacitor is not needed.

You might want to check the APFC diode (looks like D2, unless there's another component next to the red transistor on the heatsink.)

[momaka]

Re: Manson 80W Constant Power Switching Mode Power Supply

My multimeter shows big fluctuations on the drain leading to the transformer from 0V to over 1000 (I guess as that is its limit).

1 thing to note here:
DO NOT measure the voltage on the HV side pins on any of the transformers with your multimeter because you can damage your multimeter.

The IC circled in red is (I think) the PWM controller for the current the PSU outputs. It is a UC3843BN, leading to a K2717 Toshiba N-type Mosfet specced at 600V and 5A and then to the large transformer. This initially had nothing on the output pin, so I swapped it out and now I can output power, although I can't yet control how much.

I don't know if that's a good thing or not. In most PSUs using this design, the output of the PWM controller should be driving the main PSU MOSFET *only* when the power supply is switched on. Therefore, you need to check what tells the PWM controller to turn on - that way we know if the result you got above is desirable or not. The PWM controller is likely enabled/disabled and driven through the two optocouplers below T2 (the transformer you circled in red). One likely enables/disables Vcc to the PWM controller, while the other is for feedback.

Circled in green is a VIPer 12A-DIP which combines PWM and Mosfet on the one chip. Through its associated transformer it powers the two ATMegas and associated circuitry which are the brains of the PSU.

All of that is standby circuit. This circuit should always be on when the PSU is plugged in. Looks like everything is working well here.

Marked in blue is a MC33262P Power factor controller, feeding a F21NM60N 600V 17A N-type Mosfet, then my "yellow transformer-like thing". That then feeds back in to the primary side, so is I guess the APFC.

Yes, that's the APFC circuit.
From the voltages you measured, the voltage on pin 1 (Feedback) seems a bit high. Based on what I looked up in the MC33262P data sheet, I think it should be closer to 2.5v.
The voltage on pin 8 (Vcc) is within the allowed range, but it seems to be a bit unstable as well. So for that, you need to check the supply feeding pin 8. If you look at the sample circuit on page 12 of the MC33262P data sheet, D6 and C4 in that circuit supply voltage to the MC33262P after it starts up. On your PSU, these components correspond to capacitor next to U1 and the diode and resistor next to R2 (next to the large red capacitor). Make sure the resistor and the diode are both fine that that the capacitor next to U1 has not gone high-ESR. If you don't have an ESR meter, replace that capacitor with a known good one of the same capacitance and voltage rating.

Apart from those things, check that thermistor NTC1 hasn't gone high-resistance as this can also cause the unstable voltage on the primary. The easiest way to check it is to measure the voltage across its terminals for a few seconds - it should stay fairly low, 2 to 3V I would guess.

Lastly, if you have an ESR meter, also check the two large 250v, 220uF CapXon caps. While high-voltage caps like these rarely fail, there have been a few cases here on these forums where CapXon high voltage caps like these went high-ESR/open circuit and caused the PFC to malfunction. So those can be suspect as well. If nothing seems to work, change them.

Let me know what your findings are.

[momaka]

Re: Manson 80W Constant Power Switching Mode Power Supply

As for how the APFC circuit works, let's have a look again at the sample circuit on page 12 of the MC33262P data sheet (I will use the circuit and components there as a reference)...

When the PSU is initially plugged in, let's say at a line voltage of 110 VAC, the bridge rectifier (diodes D1 to D4) rectifies this voltage into approximately 155 VDC and charges capacitor C5. Current flows through R6 and also begins to charge C4. While this is happening, the voltage across C3 (the HV bulk capacitor(s) in the PSU) is also ~155 VDC minus a voltage drop from diode D5.
Once the voltage across C4 rises to the threshold voltage of the MC33262P controller, MC33262P begins pulsing MOSFET Q1 ON and OFF rapidly.

When Q1 is ON, one side of the inductor marked with "T" becomes grounded while the other is still at ~155 VDC. This causes a large amount of current to flow through it momentarily. After a short while, MC33262P turns Q1 OFF. But due to the nature of inductors, the current through that inductor cannot instantaneously change back to what it was before Q1 was ON. (This is because the inductor has stored some energy in its magnetic field. Turning Q1 OFF causes the magnetic field to collapse, thus releasing the energy stored in it) Therefore, for a short amount of time, there is a large current flowing in the direction of Q1 and D5. Since Q1 is turned OFF, the only place where the current can go is through D5. Once the current gets rectified through D5, it is stored in C3. If the load across C3 (the PSU main and standby circuits) is small, then the extra current that just flowed through D5 raises the voltage across C3.

As the MC33262P controller keeps pulsing MOSFET Q1 ON and OFF, the voltage across C3 keeps rising. The voltage divider formed by resistors R1 and R2 is used by the MC33262P controller to set the output voltage across C3.

Also, C4 is charged through R6 only once to start up the MC33262P controller. After the MC33262P controller starts, an auxiliary coil on the APFC inductor is used to generate the Vcc voltage needed for the controller. The AC voltage from this auxiliary coil is rectified though D6 and then filtered through C4. So C4 is very important to the function of the PFC controller.

[retiredcaps]

Re: Manson 80W Constant Power Switching Mode Power Supply

At the bridge rectifier (top left) I'm getting a stable 237 V AC input, but on the DC side pins it's anywhere between 288 and 318.

I'm by no means a SMPS guru, but in the lcd monitor world, the DC voltage reading across the large cap is always stable.

As a test, I took my monitor and measured the DC voltage across the bridge rectifier and it is stable (165V DC - I get 82.4 V on each side).

So if you are getting unstable or fluctuating DC voltage, could it be that your bridge rectifier is bad?

This is the same info as post #2, but I was curious and wanted to test this myself on the bridge rectifier.

[caryg]

Re: Manson 80W Constant Power Switching Mode Power Supply

That APFC stuff is interesting. So presumably the efficiencies come about from being able to charge the smoothing capacitors a bit higher, so there's less need for big swings in draw from the mains when the load changes suddenly. Would I be right in concluding that the claimed improvements relate more to places that run on 110V than to those which run on 220-240V?

[caryg]

Re: Manson 80W Constant Power Switching Mode Power Supply

The resistance across NTC1 when cold is 7-8 Ω and the voltage drop across there settles down to 0.05-0.1 when it has been running for a while.

I've pulled all the caps on the primary side to check the ESR and they all check out fine. That's not to say they won't get replaced once I've got it running, but I'm happy for the moment that they are not the trouble.

I tried swapping the bridge rectifier out for another but it made no difference.

I've pulled all the optoisolators and taken some measurements and they all checked out the same, so I think they're fine too. Plus they all went back in at random and the problem remained the same.

Diode D1 is OK. The resistor reads at 12.5Ω which is close enough to the marked value of 12 for me to blame any differences on my cheap meter.

Last night I tried for the first time to work it with a load. Just a big LED lamp that draws around 120mA. As I said, when switched on and with the output enabled, it starts off showing about 3.5V then slowly wobbles its way upwards. My lamp needs around 8V before the LEDs start to flicker, and after a minute or so flicker they did. At this point the display was clearly showing the voltage and current being supplied, but the PSU couldn't manage to supply more than 8 volts. I then switched the output off and found out that although the display went to zeros and the small led on the front of the PSU indicating whether or not it was supplying power switched off, the unit still carried on supplying power. I then switched the supply output back on again to get the display working and switched off my big LED lamp at its own switch. The voltage display jumped immediately up to 14.6v which a multimeter confirmed as right. I then switched the load back on, and the display showed a current draw of c130 mA and the voltage dropped to around 10.

With the load connected I measured across the bridge rectifier and got a stable 337V, and across the big capacitors where it was a rock solid 450.

So now I'm starting to think that the primary side is OK, and that the problem is in the control circuitry, which is a bit scary because I'm moving from nice chunky discretes on a well laid out board to complicated surface mount stuff, hidden in a tangle of random connectors, that probably needs a scope to understand it properly. Time to bury myself in datasheets for a few days.

And here are a couple of shots for your rogues gallery. As I said before, the primary side, indeed the whole main board, is pretty cleanly laid out. But look what you find when you remove the board with the ATMega, opamps and ADAC. They couldn't decide whether to use ribbon cables, headers or directly soldered wires, so went for some of each. And note the large silk screen markings for capacitors C34 and C35, and the extra holes drilled to enable them to stick in something smaller

[PCBONEZ]

Re: Manson 80W Constant Power Switching Mode Power Supply

That APFC stuff is interesting. So presumably the efficiencies come about from being able to charge the smoothing capacitors a bit higher, so there's less need for big swings in draw from the mains when the load changes suddenly. Would I be right in concluding that the claimed improvements relate more to places that run on 110V than to those which run on 220-240V?

The only efficiency PFC improves is the efficiency of the power companies billing department.
.
It doesn't do A THING for YOU. - Nadda.
.
Think about it...
Does adding parts that dissipate -additional- heat make something more efficient?
.
SMPS by their nature tend to 'spoof' power meters a bit.
All PFC does is reduce that so the power company can bill you more for the same amount of power.
.
Newer PSUs do tend to be more efficient but it's -NOT- because of PFC. - Just isn't.
.

[momaka]

Re: Manson 80W Constant Power Switching Mode Power Supply

With the load connected I measured across the bridge rectifier and got a stable 337V, and across the big capacitors where it was a rock solid 450.

So now I'm starting to think that the primary side is OK, and that the problem is in the control circuitry...

Yup, everything (major) on the primary is looking good from your results.

Question: Is there an ON/OFF switch on the front of the PSU?
Reason I'm asking is because I don't think the main power supply should turn on when the PSU is plugged in. Only the standby should. For that, you need to see what exactly on the secondary controlls the two optocouplers under the main transformer as these optocouplers enable/disable the main PSU and are responsible for the voltage adjustment.

Could you also post a picture of the front panel of the PSU?

[caryg]

Re: Manson 80W Constant Power Switching Mode Power Supply

I didn't bother posting any general pictures of the unit because there are lots here including this one of the front of the unit - if anyone would prefer me to upload a couple here to make the archive a bit more complete then just say and I'll do it.

The power switch on the front is a hard one - both live and neutral cables are routed there directly from the plug socket then back again - another nice design decision partly undone by the fact that the insulated conduit they are enclosed in blocks a good number of the already few ventilation slots on the base. So the standby circuit isn't for standby as such, but for the two ATMegas that control the unit. When it is switched on there is a delay of a couple of seconds as the big capacitors charge up and the processors boot, then when it is switched off, the display stays on until those capacitors have been drained.

But now it looks like I've found what's wrong, and sadly it's likely to be terminal. Attached is a close-up of one of the processors with a couple of what seem to be burn marks. I'll poke around a bit more yet, but it looks very likely to be on its way to the great parts box in the sky.

[Behemot]

Re: Manson 80W Constant Power Switching Mode Power Supply

It seems more like burned from solder pistol than from inside. Usually the actual die is in the middle and when it gets shorted, it melts in the middle of package.

[momaka]

Re: Manson 80W Constant Power Switching Mode Power Supply

But now it looks like I've found what's wrong, and sadly it's likely to be terminal. Attached is a close-up of one of the processors with a couple of what seem to be burn marks. I'll poke around a bit more yet, but it looks very likely to be on its way to the great parts box in the sky.

Yup, looks burnt to me as well, unfortunately . Perhaps you can find a data sheet for it and then use the block diagram in that (if given) to see if there are any short circuits between various pins that shouldn't be there.

If bad comes to worse, you may be able to make a simple circuit to directly controll the voltage output of the main PSU (via the 2 optocouplers). I remember seeing a video on YouTube where someone did just that, but with a computer PSU.

[ben7]

Re: Manson 80W Constant Power Switching Mode Power Supply

Yup, looks burnt to me as well, unfortunately . Perhaps you can find a data sheet for it and then use the block diagram in that (if given) to see if there are any short circuits between various pins that shouldn't be there.

If bad comes to worse, you may be able to make a simple circuit to directly controll the voltage output of the main PSU (via the 2 optocouplers). I remember seeing a video on YouTube where someone did just that, but with a computer PSU.

Good idea, cause if the SMPS still works, why waste it!?

-Ben

[b700029]

Re: Manson 80W Constant Power Switching Mode Power Supply

Those "burn marks" look more like traces of glue or flux to me.

[Outtasight]

Re: Manson 80W Constant Power Switching Mode Power Supply

Hi, I'm not posting about this particular PSUs "issues" but generally about this bad boy design. It's just ironic that I found this on a forum called "Badcaps".

I've had a few of these Maplin branded Manson 80W bench SMPSUs and there's definitely some problems with the design.

I've had two explode in use. Seems the issue is that the output has no back-feed blocking diode and it's fatally sensitive to back-feeding. Even when the output is "turned off" with the software button, the output isn't completely disconnected. The output has a capacitor with a probably a 700 Ohm-ish bleed resistor permanently connected to the output.

I've been using the PSU for battery charging. When you connect a 12V battery to the disabled output, you get a big spark as the battery charges the cap.

Seems that at up to about 12V you can get away with this back-feed but then I connected the PSU (with all mains turned off) to a 24V battery and a massive spark ensued. This obviously damaged something in the output by the huge back-feed impulse current to the cap. When I turned the mains on and then the output, the PSU displays showed 0.03V for a bit with cycling 0A to 0.3A output. There was then the unmistakable sound of electrolytic caps boiling, followed by a pop and then electrolytic cap smoke rising from the air vents.

When it happened the first time a few months ago to one unit, I put it down to a bad example and the store changed it for a new one. I'd then only used the replacement with 12V and 3.2V batteries and cells, before using it again today on the big 24V battery and BOOM.

Setting aside that issue with the design, another unit also failed after a few days because I got one where they accidentally soldered across the LED display current limiting resistor or something. The LED display was much brighter than the other units I had and then it burned out some segments, so that one got returned for a swap as well.

Another got returned because one of the plastic spring loaded wire sockets on the back snapped off as I tried to press the release lever to get a remote sense wire out of the block.

I notice Maplin has placed a product label on the side of the units with a unique serial number printed on the base label but then pasted on top a separate little hardware version sticker that reads "R1.0"... Maybe that ought to read "R0.1", as it's clearly not yet a release candidate design.

As the original poster mentioned that these Manson units get rebranded to various names, beware!

Another reviewer on the Maplin web page for the product also noted the back-feed issue and added a Schottky blocking diode to the positive output lead to stop it. But this meant the Volt meter didn't read right as the diode dropped 0.37V after the sense point. I might hack the two surviving PSUs I have to prevent any more fireworks in the lab...

[Outtasight]

Re: Manson 80W Constant Power Switching Mode Power Supply

Ok, so I got the the last PSU that exploded replaced.

Then the next day I had two of these atrocities charging a 12V battery, using one of them with the remote sense wires to set the charge Voltage at the battery terminals.

Upon stopping the charge, I made the fatal mistake of disconnecting the positive power lead before the sense wire with the output still on. The result was an immediate series of explosions from the PSU and total destruction.

These PSUs are a menace.

Admittedly, it does say in the user manual, "make sure to disconnect the wiring to the remote sensor first before disconnecting the main output connection".

What it should say is, "make sure to disconnect the wiring to the remote sensor first before disconnecting the main output connection or else your power supply will detonate and set fire to your house. Don't connect the PSU to any batteries with the output turned off, as this may subsequently cause it to detonate when turned on. On second thoughts, don't connect the PSU to anything and store it in a non-flammable box in the garden."

Made in China

[Behemot]

Re: Manson 80W Constant Power Switching Mode Power Supply