PSMN020-150W and PSMN040-200W, MOSFETs for linear-mode operation, plus an "extra" in the form of an IRFP260M

Not long ago, I reviewed some completely “wild” and mysterious transistors marked IXUS, which I’d bought on Taobao for an electronic load. Some time has passed, and I decided to take another chance and order a few more types of transistors—ones designed to operate in DC mode, making them suitable for use in the power stage of electronic loads.


I won’t go into the details; I’ll just say that few MOSFETs are designed to operate properly in linear mode while dissipating a lot of power. In most cases, the limit is around 30–50 W, depending on the package, voltage, and other characteristics.
But some transistors are perfectly suited to this mode and can operate practically “at full power.”

Last time, the Chinese seller badly ripped me off, sending me who-knows-what instead of IXUS transistors. I’d even started thinking about ways to buy genuine ones, or just using a bunch of better-known IRFP250s, 460s, and so on. Then I suddenly remembered that my first electronic load came with an interesting transistor marked PSMN. I opened the datasheet again and confirmed that these transistors really can operate properly in this mode. What’s more, almost everyone has probably heard of IXUS, but I suspect fewer radio amateurs know what PSMN is, even though it’s also a well-known company: Philips Semiconductors.

In the end, I received three little piles of transistors, neatly packed in bags.


I ordered three types in total: 13, 11, and 7 pieces.


But I’ll start with the more familiar IRFP260M, not the PSMN, and answer the obvious question right away: Kirich, what the hell why did you buy these? Honestly, just “while I was at it,” and partly because of how buying from Taobao works: shipping within China usually has a fixed cost, so they ended up costing me 21 cents each, plus the forwarding service fee.

So, IRFP260M, price: $0.21, link.

The seller’s photo


Here’s what I received. The only real difference is that the ones in the seller’s photo are cleaner, but that’s normal: the transistors are used, and they cleaned them up for the photo.


You can download the full datasheet here; below, I’ll include only a few brief excerpts.

General description.


Note: these transistors are not intended for linear-mode operation, although they can be used that way if power dissipation is limited to around 40–50 W, since they’re designed for switching applications.


They look fine, and the leads are long, but it really looks like the leads have been welded on. I’ve come across similar comments in reviews several times. I wouldn’t use them in a commercial product, but they’re perfectly fine for my own projects.


A few tests.
The main parameters I usually measure on transistors and compare with the datasheet are gate capacitance and on-resistance. With counterfeits, these often differ noticeably from the specified values.


For most of them, the gate capacitance matched the datasheet at around 4100–4300 pF. One measured around 5500 pF, and I also found one oddball: at first I connected it and thought it was faulty because it looked like the gate was probably open-circuit. But after I wiggled the leads, it read 5300 pF. That last transistor probably has a bad connection somewhere, so I won’t use it.


I measured the resistance with gate voltages from 1 to 15 V in 0.5 V steps, but below I’ll give the readings for just four voltages: 3, 5, 10, and 15 V. I could plot resistance versus gate voltage, but I don’t see the point.
At 5 V, the resistance is 60 mΩ; at 10 and 15 V, it’s almost the same, around 34–35 mΩ, which is well within the claimed “less than 40.”


These transistors are much more interesting, so I ordered more of them.

PSMN020-150W, price: $0.35, link.
There’s something odd about the link. When I write reviews of products from Taobao, I take the links from the order page on the forwarding agent’s site; it’s more convenient. But in this case, the link took me to a different listing with a different photo and, accordingly, different transistors, even though the forwarding agent’s site showed everything correctly and I received exactly what I ordered. And it sent me to IXFB100N50P - link.

Photo from the product page showing the exact transistors I ordered.


As in the previous case, they look the same, except mine are dirtier.


I highlighted several groups here, each with a different part marking. There are four main groups. In my opinion, the fact that their markings differ suggests that these are indeed genuine parts.


Link to the full datasheet. There is also an abridged version, without graphs.


The SOA graph clearly shows the DC operating mode, and it covers power levels up to 300 W. But that doesn't mean you can dissipate the full 300 W with a transistor like this: you'd have to keep it at 25 degrees. According to another graph, at a case temperature of 100 degrees, you should limit it to 150 W (50%).


These transistors are clearly used: you can see traces of thermal paste and sealant, scuffs, and short leads.


Parameters from the datasheet.


The datasheet specifies a gate capacitance of 9537 pF. My measurements ranged from 9557 for one transistor to 11060 for two; the rest were around 10400 pF.


The resistance at 3, 5, 10, and 15 V, respectively, is 27 Ω and 17, 12.5, and 12.1 mΩ.


The third transistors are from the same series as the previous ones; the main difference is the voltage and, accordingly, the current/resistance.
PSMN040-200W, price $0.28, link.

By the way, the part number indicates the voltage and on-resistance: PSMN040-200W is 40 mΩ, 200 V, while PSMN020-150W is 20 mΩ and 150 V.

The product photo appeared only after I had bought and received the transistors. Initially, that spot was blank, so I was essentially buying a "pig in a poke". I noticed that all the photos were taken against the same background, which means they were most likely taken by the seller.


Photo of the item I received.


Several groups with similar part markings stand out here as well.


Link to the datasheet and a few screenshots from it.



The SOA graph is similar to that of the PSMN020-150W, but with a higher voltage rating. For example, the PSMN020-150W is rated for up to 150 V at 2 A; this one is rated for up to 200 V at 1.5 A. Personally, I think these transistors are better suited to electronic loads.


They look just like the previous ones, apart from the different markings, of course :)


And the electrical characteristics.


According to the datasheet, the gate capacitance is about the same as for the previous ones, around 9500 pF. My measurements were also similar to those for the PSMN020-150W, around 10500 pF, though one came in at 11330.


The on-resistance at 3, 5, 10, and 15 V, respectively, is 55 Ω and 28, 22, and 21.5 mΩ, i.e. about twice as much as for the PSMN020-150W, so everything checks out.


Someone once wrote in the comments that I was eating my sandwich measuring it wrong. Well, they were right: on-resistance is measured under load, so I put together this test setup.
Instead of using a power supply and loading its output through the open transistor, I used a battery, which I loaded with an electronic load through the transistor under test. The gate was driven at 10 V, as specified in the datasheet.


The battery was fully charged before the test.


The IRFP260M was tested last, but to keep things in order, I’ll show its results first.
And here I made a mistake. When checking the test current, I accidentally tapped 250 instead of 260 and got the data for the IRFP250M. Since I tested it last, I didn’t notice. The difference is that the IRFP250 specs are given at 18 A, while the IRFP260 specs are for 28 A.
Also, that value is specified for pulses shorter than 300–400 µs, but I have to test it under a sustained load.


Right after startup, the voltage was about 730 mV, but it started rising as the transistor warmed up, and by the time I took the photo it had reached 757 mV. As I mentioned above, I mistakenly set the current to 18 A, which gives us 730/18 = 40.5 mΩ. The datasheet specifies a maximum of 40 mΩ for pulses up to 400 µs, so I think that’s a pretty good result.


PSMN020-150W
The specs are given at the same gate voltage, but at 25 A. Because of the electronic load’s current limit, I set it to 20 A, but I don’t think that affected the result much.


The resistance rose noticeably more slowly because the transistor was dissipating considerably less power.
I got about 283 mV right after switching it on, which gives a resistance of 283/20 = 14.15 mΩ. The datasheet gives a typical value of 12 mΩ and a maximum of 20 mΩ, so I consider this test a success.


And the last transistor, the PSMN040-200W.
The stated test conditions are identical to those used for the previous transistor.


I got 513 mV, which works out to a resistance of 513/20 = 25.65 mΩ. On the one hand, that seems good; on the other, it’s suspicious, so it’s hard to say whether everything is fine or not. I clearly need another sample.



If you noticed, there’s an odd number of each transistor. That’s because I originally planned to open up one of each type, though after testing them I felt a little bad about destroying working components.

I started with the PSMN020-150W.
Unlike the IXTK, this one was much harder to open. I thought I’d break the vise while trying.
The base plate is 2 mm thick and looks like copper.


After a while, I finally managed to open it.


The die measures 7.2 × 6 mm, or 43 mm², which is considerably larger than the counterfeit IXTK dies, whose areas ranged from 25 to 36 mm², despite being advertised as more powerful.


I opened up the PSMN050-200W last, using a slightly different method. First, I heated it on a gas stove a couple of times, then cooled it quickly. Maybe that’s why part of the die remained intact.


It’s the same size as the one in the PSMN020-150W.



The third one I opened was the IRFP260M whose gate capacitance was “wandering.” It turned out that the gate lead had an intermittent connection: press it one way and the connection disappears; press it the other way and it comes back.

I thought the PSMN020-150W was hard to open. Nope—it was easy compared to this one, but I won in the end :)
The copper base is also 2 mm thick.


The die has a different shape and measures about 7.5 × 4.5 mm, with an area of 33.5 mm².


In this comparison photo, you can see that the paint on the PSMN040 has darkened. I pushed the temperature until the plastic package started to burn. That’s some good paint :)


I’d even say the PSMN transistors are genuine, at least based on the test results—if it weren’t for one thing...
I happened to come across a photo of one in a different package.


And indeed, according to the datasheet, the transistor has a package with semicircular side cutouts, not rectangular ones. So the question I can’t answer yet is: is this simply a good counterfeit, or does the manufacturer make transistors in this package? As far as I know, it is made by Philips and NXP, which essentially has its roots in Philips as well.


I even found a listing on 1688.com offering two versions.


I can’t say for certain whether these are 100% genuine or counterfeit. The Chinese are masters of tricks like this, but in my opinion, their characteristics are very close to those of the genuine parts. It would be good to know the die dimensions of 100% genuine PSMN020 and PSMN040 transistors; then we could be more certain. But the die in the IRFP260 does look more like the genuine one.

Anyway, I’m planning to order some more of these transistors. I just need to decide how many. Even if I don’t need them for this project, I’m sure I’ll find a use for them. By the way, I saw PSMN020-150W transistors with long leads in a local online store for about a dollar. I doubt they’re genuine. Another store had them for $3, which seems much closer to the price of genuine transistors—but that’s at least a fivefold price difference...

That’s all for now. I hope this information was useful. I’d be interested to hear what you think about whether they’re genuine.