Mean Well PSC-160A-C, 160 W power supply with UPS function

Today I'm reviewing a rather unusual power supply. The key feature that sets it apart from the ones I've reviewed before is its uninterruptible power supply function.


The power supply was sent to me by a regular reader, Vladimir, and I'm very grateful to him. But he didn't send it just for the sake of it: there was a specific reason, namely the idea of adding it to one of my electronic loads for battery testing.

I've come across UPS units from Mean Well before, but from an older series. I can say that I liked them quite a lot, primarily because of their reliability, although their power output wasn't particularly high (55 and 155 W). I've also built various similar UPS units myself, and posted a guide to converting a regular 12–15 V power supply into a UPS.
In fact, all the options I've mentioned work on exactly the same principle. The implementation and service features differ slightly, such as indicators for a discharged battery and a mains failure.

Like other Mean Well power supplies, this one came in a cardboard box with a small window so you could see which model was inside.


This model comes in two versions: in an enclosure, and as a board with a heatsink. It has protection against overheating, overvoltage, overload, and battery over-discharge.


As with the previous models, there are two versions: 13.8 and 27.6 V, for 12 and 24 V lead-acid batteries, respectively. There used to be another version, the AD-155C, rated at 155 W and 53.5 V, but the PSC series doesn't include a model with that voltage.


And here's the “beauty”—at least, it looks neat.


The power supply is built on an L-shaped aluminum chassis covered by an enclosure.


The 12 V version was a better fit for my project, mainly because it lets me build a more compact battery. It's also easier to find a 4S protection board than an 8S one, and I didn't want to combine two boards. So this review will cover the 12 V version.
The specs say the power supply can deliver 13.8 V at 7.6 A to the main output, and 4 A to the battery charging output. That power is too low for my needs, since the device draws around 160–180 W at peak, while the battery charging current is actually too high for me. I'd much rather have a 10.6/1 A configuration than 7.6/4 A.


Compared with regular power supplies, this one may seem quite large, but the previous version was even bigger: 200 × 110 × 50 mm versus 152 × 76 × 37 mm for the one under review. The older model was rated at 155 W, while this one is rated at 160 W.
For example, the AD-155A had a more useful balance of output currents for me: up to 11.5 A on the main output and 0.5 A for charging.


The drawing gives you more information about the dimensions.


The mains input connector is on one side, and the output connectors are on the other: power connectors for the load and battery, and low-current connectors for the relay outputs. There's also a trimmer near the output connectors for adjusting the output voltage.

Unlike many Chinese power supplies for DIY projects, this one doesn't use screw terminals for the input and outputs, which may not be very convenient.


The table shows the connector types and pin assignments.


There's nothing interesting on the outside, so I removed the enclosure. The screws holding it in place were, as they say, tightened with all their might. The enclosure was also glued to the power transformer with silicone sealant.


The input has a nearly complete filter; the only thing missing, perhaps, is a varistor. The bridge rectifier has a heatsink with an insulator glued to it, which also prevents the enclosure from being pushed in.



1. The transistors and PFC diode are pressed against the chassis. The power section is laid out very tightly.
2. Since the power supply has power factor correction, the board has an inductor as well as a transformer.
3, 4. The inductor and transformer appear to be wound with Litz wire, unlike those in cheap power supplies.


The output transistors are mounted on a separate heatsink. There are also three relays and a fuse here.


1. IRFB3607 synchronous rectifier transistors.
2. The output has four 1000 µF, 25 V capacitors, as well as an inductor to reduce ripple.
3. Trimmer for adjusting the output voltage.
4. From the very beginning, I wasn't too happy with the way the fuse was installed. Then I noticed that its insulation was also slightly damaged, so there's a real chance of pushing the cover in far enough to cause a short circuit.


It's also odd that the manufacturer used three different relays. The top one signals that the battery is discharging, the middle one detects mains power, and the bottom one connects and disconnects the battery.


During the initial inspection, I also noticed that an insulator on one of the transistors was crooked. After I removed the clamping bar, it fell into place. It's a minor thing, but it doesn't look great. The clamping bar has an additional insulator on the side facing away from the power components. Although it's glued on, it's best to make sure you don't lose it.


As it turned out, removing the board isn't that easy. There's a spacer glued underneath, so you have to lift the board up rather than slide it out, which seemed more convenient at first. During the "photo shoot", the spacer kept sticking to the table, which made things even more awkward.


With the board removed, it was a little easier to identify the components: the PFC diode and transistor are a BYV29X and a 20N60S1, respectively; the inverter transistor is a 65E6280.
The Y capacitor is 2.2 nF. All the electrolytic capacitors except the input one are made by Rubycon.


As usual, the board is populated on both sides, with most of the components on the underside.


The "hot" side. Unfortunately, I didn't trace the schematic. I may still do that, but drawing schematics this complex takes quite a lot of time, and as always, there's never enough of it.


There are two controllers here: NCP1608 and NCP1380. The first controls the active PFC, and the second controls the inverter.
This isn't the first time I've seen PWM controllers and power transistors placed quite far apart. I've always thought that the closer they are to each other, the better.


There aren't many fewer components on the output side. That's because it includes not only voltage feedback, but current feedback as well, along with protection circuits for overvoltage, battery overdischarge, and so on.


Most of the functions are handled by two chips: the TEA1761, a synchronous rectifier controller with current and voltage feedback, and the AP4310, a dual op-amp combined with a voltage reference, which controls the relay.
There's also a 20 mΩ shunt in the battery circuit.


The datasheet includes a block diagram showing the key functional sections, but there's nothing new here: the primary side is a conventional power supply with active PFC and an additional circuit to protect against excessive output voltage. Its signal comes from the monitoring circuit on the output side through an optocoupler.
The separate mains-voltage detection circuit is simply powered from an auxiliary winding. In my own UPS designs, I implemented this even more simply: I powered it from the same winding that supplied the output, using an extra diode and a small-capacitance capacitor. That way, when the inverter stopped, power to the output of this rectifier disappeared too.



I once designed an adapter board that turns an ordinary 12–15 V power supply into a UPS. It added not only battery backup, but also charging with CC/CV mode, plus protection against battery overdischarge and reverse polarity.
The circuit is very simple and, of course, a little less efficient than commercial solutions, but that trade-off is worth it for the simplicity.


The board itself was compact, too, and only needed a few wires to connect. For higher-power systems, all you need to do is install a higher-rated relay and possibly a diode.


Now we can move on to the tests.
First, I simply connected the power supply to the mains using a cable with the appropriate connector. I had some of these left over from when I modified intercom monitors to work with low-voltage UPS units.

1, 2. As expected, the voltage at the load and battery terminals is the same. More precisely, the voltage at the main output depends directly on the battery voltage because they're connected to each other, and the power supply has a common feedback loop. As a result, depending on the operating mode, the output voltage varies from 10 to 13.8 V.
3, 4. The trimmer potentiometer lets you adjust the voltage over a fairly wide range. This is a major drawback, since you don't need that kind of range in real life and it only makes adjustment more difficult.
5. By turning it very slowly, I set the output to 13.5 V. I needed this voltage because I plan to connect a battery made from LiFePO4 cells. Their float voltage is 3.35–3.40 V per cell, or 13.4–13.6 V for the pack, so I chose 13.5 V as the midpoint.
6. With no battery or load connected, the power supply draws about 2.3–2.4 W.

There is a small difference from the circuit I used. In my version, if you connect the power supply to the mains without connecting a battery, there is no voltage at these terminals. In other words, when connected correctly, the battery itself triggers the relay. In the unit under review, voltage appears almost immediately, and the relay switches on even without a battery, which makes it suitable for battery packs with a protection board.



The seller also sent me the connectors needed to hook up the power supply, but since I didn't know what wire length I needed, I decided not to use them for now and simply soldered the wires to the board. I soldered two pairs of wires to the output: one for the load and one for voltage measurement.


The power supply's load characteristics were a bit surprising. I expected the protection to trip at the combined load and charging current (7.6 + 4 A) when no battery was connected, but in reality it tripped at 8.9–9 A.


Accordingly, I tested efficiency up to 9 A, with 0.5 A between points on the horizontal axis. The specs claim 88% efficiency, and that's roughly what I measured. I think it could have been higher, since efficiency increases with load up to a certain point, but I could only load the main output enough to measure it.


With a charged battery connected, it easily delivered 10 A as well.


With the mains disconnected, it also easily delivers 10 A when running from the battery. I think the current is limited only by the fuse, meaning there is no electronic protection, though it might be worth trying higher currents too.
Using a regular fuse is a very inconvenient solution, especially when it's soldered to the board, since replacing it if it fails will be quite difficult.

Don't worry about the voltage drop: the battery was connected through long wires, plus the multimeter leads.


I didn't test the charging output with a load in CV mode, and instead limited myself to a test showing that the power supply changes operating mode at 4–4.1 A.


The ammeter showed that the charging current was almost exactly 4 A, although it rises slightly as the unit heats up. With no mains power and a load connected, the current drawn from the battery is 35 mA.


While the battery was charging with a 9 A load current, even the load displayed this behavior, and the relay that disconnects the battery kept clicking.
To prevent this, you need to reduce the load current to at least 8 A, although I also found a few quirks there.


For the thermal test, I decided to fully discharge the battery and monitor temperatures throughout the entire cycle: a full charge from zero, followed by the transition to CV mode. While I was at it, I also checked the battery's low-voltage cutoff.
The protection relay disconnects the battery at 10 V and doesn't reconnect it until mains power is restored, at least not during the time I waited. At around 10.5–10.6 V, the "battery discharged" relay trips. But since I'm using LiFePO4 batteries, there is very little time between this relay tripping and the battery being disconnected. At 8 A, it was about 20–30 seconds; I didn't time it precisely.
Yes, using LiFePO4 batteries comes with some caveats because they have a different discharge curve, as I've mentioned several times before.


I put together a "test setup" and started the test.


1. The load current is 7.6 A according to the datasheet, and the battery charging current is 4 A. Initially, the power drawn from the mains was about 180 W.
2. After a very short time, the power draw started to fall, although logically it should have been rising. The reason was simple: at this charging current, the battery voltage quickly reached the set 13.5 V, and the charging current began to drop. I had to raise the voltage threshold to 14 V.
3. The power drawn from the mains gradually increased, which is normal: the load current stays constant while the voltage rises as the battery charges. The output voltage rises accordingly, and the current there is also regulated. The peak was about 188–190 W.
4, 5. A little over an hour and a half later, the battery was fully charged, the output voltage had risen to the set value, and input power had dropped to 123 W. After that, I ran the power supply for another 20 minutes or so to see how it would behave with the battery already charged.
6. With the load disconnected, power consumption was 3–3.2 W. In effect, the battery was being topped up with a small current, indicating that charging was complete.


Since the test took two hours and I took measurements every 20 minutes, I ended up with 6 sets of results.
The thermal image shows that the transformer reached the highest temperature, over 100 degrees. In reality, though, the diode next to it also got hot, reaching roughly the same temperature. As a result, the thermal camera kept switching the maximum-temperature focus between the diode and the transformer.


At 1 hour 20 minutes into the test, the temperature started to drop gradually. This was because the total power consumption fell as the battery charged, and by the end the temperatures were more manageable.


Interestingly, judging by the graph in the datasheet, the unit has slightly stricter temperature limits in its enclosure than in the open-frame version. I would have thought the chassis would dissipate heat more effectively.
By the way, the diagram below shows how to connect the load and battery. The manufacturer recommends connecting like-named connector pins in parallel to reduce the current through each pin.


Temperatures of various key components. Output power is estimated from input power, assuming 88% efficiency.



Oscillograms for these operating conditions: no load, 4 A, 8 A, and 8 A with the battery charging.

Despite the output inductor, the output is pretty "dirty". The manufacturer specifies 150 mV p-p; I measured about 200 mV at 8 A and 250–270 mV while charging the battery at the same time. As you can see from the oscillograms, I measured only the main ripple, excluding the short "spikes". Of course, there may also be some interaction between the load, power supply, and oscilloscope, but my experiments showed that if there is any, it isn't very significant.


In reality, though, things were even worse. Let me explain.
As I mentioned, the power supply turned out to be very "musical": it would sometimes squeal, buzz, and make strange noises. For example, the battery might be charging at 7.2 A with everything working normally. Then I would touch the oscilloscope probe to the output (as you're supposed to, ground first, then the signal), and it would suddenly start squealing. I'd remove the probe, but it would keep squealing. Over the next few minutes, the pitch would change slightly and the sound would gradually fade away.

So I also captured some "alternative" oscillograms while it was behaving strangely. As you can imagine, the behavior was unpredictable, so I captured what I could. Note that the first oscillogram was taken at 100 mV/div, while the others were taken at 200 mV/div.
1. Battery charging starts with no load on the output.
2. Charging starts with a 7.6 A load; squealing begins.
3. Same as above, but the load current is reduced to 5 A.
4. Charging with an 8 A load at a low sweep rate.


And now for the verdict.
To be honest, I'm a bit disappointed. After testing it, I was left with the impression that the product is still rough around the edges, even though my initial inspection left me with nothing but positive impressions. But let's take it in order.

First, the good. The unit has decent efficiency, good build quality, equally good-quality components, a full set of protection features, signal outputs for AC input and battery discharge, and protection against overdischarge.

Now for the bad (in my opinion). Very strange behavior, occasional squealing and beeping, and so on. It also gets quite hot, has a high charging current that's awkward to balance against the output current, high ripple, reverse-polarity protection in the form of a soldered-in 15 A fuse (I'll check whether it also has electronic protection), and coarse adjustment of the output voltage and, consequently, the charge termination voltage.

A separate note on the high charging current. With the advent of lithium batteries, where 4 A is a normal charging current for a 4–8 Ah battery, many people have probably forgotten that this is by no means normal for lead-acid batteries. For example, I recently tested a name-brand 9 Ah battery whose maximum charging current is 3.4 A. For cheaper batteries, it's often around 2–2.5 A, so a 4 A charging current would simply kill it.
The older model had a charging current of 0.5 A, which is arguably too low. In my opinion, around 1 A would be optimal, and ideally the current should be adjustable, even in steps.
I had planned to reduce the charging current to the level mentioned above and draw the freed-up power from the main output. But it turned out not to be that simple: I can reduce the charging current, but judging by the load characteristic, that won't affect the main output at all... Still thinking.

And of course, the usual question: what is the transfer time to battery power?
The answer is that this type of UPS has no such thing as a "transfer time". In practice, the battery is permanently connected to the output, while the charger adjusts the inverter's operating mode to charge the battery and provide the required output current. That's why the output voltage depends directly on the battery voltage.


That's all for today. If I manage to modify this UPS, I may add to this review or write about it separately.
I hope this was useful, and thanks again to Vladimir for letting me tinker with interesting devices.