Mean Well PSC-160A-C: fixing shortcomings, modifications, and an addendum to the review
I’ll start today’s article by apologizing for the mistakes I made while preparing the previous review, which made the test results look worse than they really were. I’ll probably revise the previous review, but today I’d like to explain what happened and why.
Last time, I wrote and demonstrated that the power supply could deliver no more than 9 A from its main output. I was surprised that this current limit didn’t depend on whether a battery was charging at the time. That was genuinely strange, so I put off investigating it “until later.”

Imagine my surprise when, while preparing a video about this power supply, I had no trouble getting 15.6 A at the output during a load test.
How could that be? I’d tested this power supply, and it consistently delivered only 9 A, maybe a little more...

With the battery charging at the same time, the protection kicked in at a current 4 A lower. You can see it in the way the output voltage starts dropping sharply, then rises again; the power supply shuts down, switches to battery power, restarts, goes into protection, switches to battery power, and so on.

Something was clearly wrong. Then I remembered that just a few days earlier, I’d managed to burn out my electronic load by running it at maximum power and the maximum voltage while it was already very hot.
Of course, I repaired it right away, and had even planned to write about it. Now I definitely will.
After it shorted out, I opened it up and removed the power board. I left all the wires connected.

After unscrewing the board, I saw that there was clearly too little thermal paste under one of the transistors—and there wasn’t much overall, either.

But it turned out that the transistor that had failed was the one with more paste under it.

I remembered that I’d bought the right transistors for both modifications and new devices, so I grabbed one of the bags and picked out a pair with identical markings.

I cleared the holes in the board with a toothpick, installed the new transistors, and screwed the board back in place without thermal paste. Only then did I solder them in. The transistors are mounted rigidly, so they need to be positioned precisely, without any misalignment. That’s why they’re soldered in place after installation.
Then I removed the board and applied the thermal paste.

After all that, I installed the board properly and reassembled the load.

During the power supply test, everything seemed to be going well. The only odd thing was that, when I measured the ripple, some kind of stray mains-frequency interference appeared. I didn’t investigate it right away; I simply switched to another load and ran those tests with it instead. It didn’t affect the other test results, but it did make me suspicious.

The problem came to light completely by chance, and only indirectly. While testing batteries, I found that my load couldn’t draw more than 9.5–9.6 A from one. The fact that this current was close to the figure from the previous test was just a coincidence.

That’s when I realized what the problem was. I didn’t even need to take the load apart—everything suddenly fell into place, as they say.
Apparently, the new transistors need a higher voltage to turn fully on than the old ones did. That wouldn’t be a problem by itself, but the output op-amp that drives the transistors is probably powered by 5 V. So at low voltage and high current, they can’t turn on properly. Since even a small change in gate voltage in this mode can affect the on-resistance, interference from the power supply (mine is a regular transformer) creates small ripples at the mains frequency—or, more precisely, twice the mains frequency.
This barely affected operation with batteries, but apparently the Mean Well power supply was sensitive to this kind of problem. It may also have been causing the load to behave erratically with its own high-frequency ripple.
In short, the power supply’s protection was tripping earlier than it should have. The culprit was a combination of the load’s “quirks” and the power supply’s poor stability.
For the review, I eventually used my main electronic load, which I’d tested over the years, but it didn’t occur to me at the time to raise the current above 8–9 A.

Once I realized that the UPS could do more than I’d shown in the review—and that, on top of that, the current available at the main output depended normally on whether or not a battery was charging—I decided to “finish the job.”
At first, I was confused by the fact that the load current didn’t depend on whether the battery was charging, which seemed odd. Especially since the charging current was 4 A, more than half the UPS’s output current.
The second reason for modifying it may be useful not only for increasing the load current, but also for extending the service life of a lead-acid battery.
The thing is, while a 4 A charging current might even be considered low for modern LiFePO4 batteries, it’s very high for a lead-acid one.
For example, the manufacturer’s CSB HRL-1234W-F2-FR battery is rated at 9 Ah and 12 V. According to the manual, its maximum continuous discharge current is 66 A, and its short-term maximum is 130 A. Yet the same manual gives a maximum charging current of just 3.4 A—and that rating is printed right on the battery!
So 4 A is way too much for it.

So I decided to take care of several “birds” with one modification.
1. Reduce the charging current to 1–1.2 A
2. If I could reduce the charging current, I could increase the load current at the main output by the same amount.
3. Reduce the acoustic noise: the power supply makes an unpleasant squealing sound while charging the battery.
4. Reduce the parasitic voltage drop when running from the battery.
To do all this, I had to get into the circuit. It’s built around two chips: a TEA1761 synchronous rectifier controller and an AP4310 op-amp. Current is measured across a 20 mΩ shunt.

In fact, the TEA1761 handles feedback for both voltage and current. The datasheet even shows a typical application circuit where Rsense is the current-sense shunt.

Of course, the simplest solution would be to reduce the current-sense shunt’s resistance, say by replacing it with a 0.1 Ω resistor. The charging current would then drop proportionally to 800 mA.
But the problem is that current also flows through this shunt when the output is powered from the battery. Even in the original configuration, the input-to-output voltage difference is almost 0.4 V. At 10 A, that means 4 W of heat. I only noticed later that my load tests had caused the insulator under the board to melt a little.
The difference is smaller in charging mode, of course—0.15 V—but that wasn’t what I was interested in.
So the shunt needs to be reduced, not increased, but that means making more substantial changes to the circuit.

The only problem is that the current-feedback threshold voltage for the TEA1761 is 50 mV. What’s more, there’s a divider after the shunt that also smooths out short voltage spikes. So at 4 A, the shunt produces 80 mV, and the divider brings it down to the required 50 mV. There’s no easy way to lower this threshold.

To avoid cluttering the diagram with unnecessary components, I’ll show only what you need to know for the modification. The component and IC pin numbers match the actual ones.

And here they are on the board.

Since the goal was to amplify the signal from the shunt, I added an op-amp. Initially, I used 20 kΩ and 2 kΩ resistors for R2 and R3, giving a calculated gain of 11, which should have brought the current down to 360 mA. But for some reason, the current dropped much further, so I replaced R2 with a 10 kΩ resistor, giving a gain of 6. The op-amp is an LM358.

During the experiments, I also blew the fuse and slightly scorched the original shunt. I replaced the fuse with a small 20 A one, which reduced the total resistance by about 2 mΩ.

The charging current came out to about 460 mA, but it fluctuated quite a bit.

I soldered another 20 mΩ shunt in parallel with the original one, so... Now the shunt is 10 mΩ, but the current has only increased by a factor of 1.5.
Adding a capacitor to the op-amp power supply made a noticeable difference. A 10 µF electrolytic capacitor plus a 0.1 µF ceramic capacitor brought the current to about 0.97 A.
The electrolytic capacitor causes some problems because the gap between the circuit board and the chassis is quite small; even a small capacitor only just fits.

This is what the resulting setup looked like.

It seemed to work, but something was off. For example, the charging current depended on the load current: up to 1.5–2 A, it was the 0.96 A shown above, but above that it rose to 1.2 A and stayed there steadily.
But the problem with the maximum current being limited to 9–9.5 A also came back—only this time it happened with a load I had already verified. The power supply was clearly behaving strangely, and it was still accompanied by the same whine, whose pitch changed as the load current increased.

I experimented for a while and was already thinking of putting everything back the way it was when I decided to add a 0.1 µF capacitor to the gain-setting circuit and remove the resistor I had installed between the shunt and the op-amp input.
The resulting circuit now looked like this.

I also replaced the wire supplying power to the op-amp with a thicker one, mostly for reliability.
Earlier, I had also added ceramic capacitors in parallel with the filter capacitors and reinforced one of the traces to reduce the voltage drop across it.
Power is taken from the capacitor above the TEA1761. There is an option to take power directly from the output near the op-amp, but that makes the circuit unstable.

All the components can easily be mounted point-to-point. In the photo, the first pin is marked with an arrow; pin 4 is soldered to a tinned pad on the circuit board, which was already tinned. The other pins are bent slightly upward.
Pin 5 is connected to pin 4, and pins 6 and 7 are connected together; this is how I disabled the unused second op-amp.

In this configuration, the power supply worked almost perfectly. There was a little noise at load currents of 1.5–2 A, and another sound I couldn't explain at 3.9–6 A, which only appeared when the load current was reduced to that range.
The latter can be eliminated by adding a jumper between the negative terminal of the TEA1761 power supply capacitor and the capacitor in the op-amp output signal divider.
You can see it in the photo, too, but that's a minor detail. The noise in the 3.9–6 A range disappeared, while the noise in the 1.5–2 A range remained. In any case, it's barely audible, unlike the original whine, which was quite loud.

The charging current also now matches the calculated value: 4 A / 6 (op-amp gain) × 2 (halving the shunt resistance) = 1.33 A
At the same time, the parasitic voltage drop across the fuse, relay contacts, shunt, and circuit board trace fell by more than half, to just 227 mV, down from the original 390 mV.
Of course, you could say that the difference in voltage drop before and after the modification is negligible. Yes, in the context of the whole power supply, an extra 1.6 W at 10 A is minor—about 1%—but the difference between 2.27 W and 3.9 W being dissipated and heating up the circuit board is significant.

I tested the power supply in various operating modes, and everything works great. Now, even in the most demanding mode (for my use case), when battery charging has just started, the output current can reach 14.4 A, with up to 170 W at the main output. As the battery charges, the maximum power available can be even higher.

So, here's where things stand.
The power supply has been cleared of the charge of being unable to deliver its maximum load current: the problem was with the measuring equipment, not the unit under test.
The whine after a load change has been reduced somewhat in some operating modes, but it remains when charging the battery.
After the modifications:
1. The charging current was reduced from 4 A to 1.3 A. I could have reduced it further, but that would mean the battery voltage would rise less. Since I'll have a 12 V to 5 V DC-DC converter after the power supply, a higher voltage works better for me. I'm even considering setting the low-voltage cutoff higher than the original 10 V.
2. The noise has been almost completely eliminated. A faint sound remains over a narrow range of load currents. I could probably get rid of it too, but I'm happy with the result.
3. The shunt runs cooler because its resistance is now 10 mΩ instead of 20 mΩ.
4. With the lower shunt resistance, the power supply can now deliver 10.5–11 A continuously, or up to 14 A briefly or with additional cooling—that's almost twice the original current.
5. The overall parasitic voltage drop when running from the battery has been reduced.
6. The power supply is more stable overall, and the feedback loop is more robust. It's now less sensitive to the load.
7. Ceramic capacitors have been added in parallel with the filter capacitors, though that's more of a "cosmetic" change.
A note on the circuit design: I think the whole problem comes down to a poor choice of synchronous rectifier chip and feedback circuitry. Where temporary compensation in the feedback loop would have helped, it cannot be added. In addition, the current feedback trip threshold is too high, requiring a relatively high-value shunt. I even think the 4 A charging current was chosen for this very reason, which negatively affects performance: with a total power of 160 W, only 105 W could be drawn from the main output.
The total cost of the components for the modification is about 25 cents, even when bought at retail.
That's all from me. I hope this will be useful to someone :)
meanwell psc потребление холостого хода, psc-60a переделать, mean well psc 160a c make noise, psc-160a-c mean well, psc-60a-c mean well обзор, схема б.п. psc 160 a c, mean well как ограничить ток, https://www.kirich.blog/obzory/bloki-pitaniya/870-meanwell-psc-160a-c-ustranenie-nedostatkov-dorabotka-i-dopolnenie-k-obzoru, 1, tea1761t блок питания схема, tea1761 блок питания переделка, psc-60-r3, psc 160b c доработка, psc 160a c клеммы, psc 160a c mean well схема блока, схема включения qc160a, psс 160 sch, ac/dc psc 160a c, psc 160 r2 cn3, аккумуляторы какой емкости можно подключать к psc 160a, аккумуляторы какой емкости можно подключать к psc-160a, rps 160a r6 блок питания, блок питания с резервированием mean well psc 160a c, 13.8в 11.6а 13.8в 4а 160вт, psc 160a c схема включения, блок питания psс 160 принципиальная схема
Questions and answers
Why did the MeanWell PSC-160A-C appear to deliver only 9 A in the first review?
The electronic load skewed the results: after its transistors were replaced, it lacked the voltage needed to turn them fully on at high current. With a trusted load, the author measured 15.6 A.
How did the author modify the MeanWell PSC-160A-C?
He reduced the current-sense shunt resistance from 20 mΩ to 10 mΩ, lowering charging current from 4 A to 1.3 A. He also added ceramic capacitors in parallel with the filter capacitors.
What current can the modified MeanWell PSC-160A-C supply?
The author reports 10.5–11 A continuously and up to 14 A briefly, or with additional cooling. The noise during battery charging was also almost eliminated.
The figures from this review are in the measurement database: Power supplies
Discounted products
You may be interested in
Discounted products
All the content on the site and the advice are free, but we’d be grateful for your support of the project and the channel!
Comments20
Only registered readers can comment. It takes a minute to sign up, and you can log in with the same account you use on the forum.
Log inRegister




Comments are published in their original language.