Used VeriFone 9V (9.3 V) 4 A power supply. This time, model SM09003A.

Power supply from the same boxes from our local shop. Compact size—the case is exactly the same size as the one from the previous power supply (model AU1370933g (the same one).
The same 2-PIN connector for mains input (230V). The only difference is the window for the LED. The cable is the standard one for newer VeriFone units, with a 16AWG cross-section (excellent). Usually, power supplies like these come with a standard 18AWG cable. In AWG numbers, the lower the number, the better (thicker wire).
Continuing the cable discussion—in this case, VeriFone went with two improvements. They increased the wire gauge (from 18 to 16 AWG) and raised the output voltage. The old units were rated at 9V; the new power supplies are rated at 9.3V. This is because their output cables are very long—nearly 2 meters.
I make my own cables for power supplies like these (and adapters such as 12V/2A) from NV-5 wire (0.75mm2). I really like this wire; it comes in loads of different colors. Of course, it’s not like silicone wire, but it’s very flexible and pleasant to work with. Compare an ordinary cable from a cheap AliExpress power supply, say a 12V - 5/6A one (judging by the copper cross-section), with this NV-5 wire (0.75mm2), and it’s like a Chihuahua next to an elephant. The Chinese cables have just a few strands of copper and thick insulation. In most cases, their wire/cable thickness is just an illusion. I made several spare cables at home in different lengths, with different lengths for the little “nose” on the 5.5х2.1mm connector.
I put a green ring on the cable and wrap the cable around it a few times. There was one peculiarity with this NV-5 wire (from Belarus).
The coils of this wire at our shop came from Belarus. There was something odd about them: wires of every color had the same overall diameter. But for some reason, the orange wire (not red) was the odd one out. Its insulation was thicker, making the wire itself one size thicker. Looks like a manufacturing defect.
The labels on the factory packaging were identical, though, in both the wire markings and cross-section.


Soldering various connectors onto cables, DIY bench power supplies, and so on.
You often read about people using acid and active flux, saying ordinary flux won’t work. In 90% of cases, that’s not true. You just need to remove the oxide layer from the surface. I simply scratch the surface with a sharp object (like an awl) and solder it with ordinary liquid rosin. In this case, the 5.5х2.1mm power connectors for DIY cables solder perfectly using this method.


Where should you put a ferrite ring or ferrite bead on a cable?
Here’s a quote from Ksiman on Muska: “The filter on the cable goes on the noise-source side.” This ferrite bead or ring helps in cases where, for example, a laptop or tablet screen (touchscreen) glitches. Often, the culprit isn’t the conventional filter choke in the power supply. It’s often the lack of a Y capacitor and a ferrite filter on the cable.
In other words, if you see a little ring on the power supply output with two wires wound around it, put the ferrite filter on the cable near the connector. If there’s no ring, and no filter choke either, put the ferrite filter near the power supply.

A quick note on tinning wires—I was surprised to see experienced people on Muska writing that tinning isn’t reliable either.
A wire breaks quickly at the boundary between the tinned and untinned sections if you twist it. Here’s what to do: strip a generous length of insulation from the wire. Tin the wire. I use rosin in solid form specifically. Then comes the most important part: take heat-shrink tubing slightly thicker than the wire and make sure to slide it over the tinned section by a few millimeters. Then, of course, heat the tubing. This prevents the wire from breaking at the boundary between the tinned and untinned sections. When repairing a board, I simply solder the factory wires back in so they don’t get in the way or break off. After soldering the wires, I secure them at the solder joint with sealant. If it’s a quick repair (replacing capacitors), I coil up the cable and secure it with a rubber band (I use the ones that come wrapped around money). I got some from a girl I know at the grocery store next door.

Board sealants.
I currently use three neutral-cure sealants at home.
1) Kazan Avtosil, 180g, temperature range from -50 to 300°C. Gasket sealant 11225, RTV silicone, version 2.0. Pros: inexpensive, large tube. Con: no dispensing nozzle. But a nozzle from other sealants will fit.
2) KERRY gasket sealant—neutral gray RTV. Temperature range from -50 to 330°C. 85g. Comes with a long dispensing nozzle and a spatula. Con: noticeably more expensive than the Kazan one.
3) Kafuter K-704 adhesive sealant. Temperature range from -60 to 250°C. I have 45g tubes. White/black. The most expensive one. I buy it locally at Profi. The local price is lower than on AliExpress and Banggood. Plus, it comes with a small dispensing nozzle. It’s handy for reaching awkward spots and applying the sealant in small, neat amounts. And the tube itself is junk—the metal is very thin, so you have to work carefully.
I didn't find any problems with any of these sealants.

Now let's move on to the star of the review: a used VeriFone 9V (9.3V)-4A power supply, model SM09003A.
The specifications are shown in the photo.


I shouldn't have sealed the case with T-7000 adhesive sealant. The first time, the case opened easily and with practically no marks. But after using T-7000 (I opened it again to take photos), noticeable marks were left on the case. I didn't use gasoline/solvent a second time,
and simply forced it open. I somehow managed to find my photo of the board as it came from the factory. The other board photos were taken after a small modification.
The first thing that caught my eye on the board was a jumper in place of the thermistor. I was unpleasantly surprised. I have no explanation for this jumper. The only logical explanation is that the factory ran out of thermistors at some point during production of the power supply. A thermistor isn't the kind of component you'd leave out to save money, if that were the goal. It's quite possible that the same power supply, made at a different time, would have a thermistor installed.




After inspecting the board, I got the impression that the previous power supply (model AU1370933g) was an interim model, or perhaps they simply ran short of power supplies and made an additional batch at another factory. This one has a 2-pin connector for the power supply, permanently secured with clips. It uses the familiar (for VeriFone) 100uF/400V capacitor. An SMD capacitor has been installed in parallel with it. I also tried doing this with my home power supplies (using a 1206 SMD capacitor rated for 1 kV), but people told me that adding an SMD capacitor across the high-voltage capacitor is pointless. The SMD capacitor's capacitance is just too small.

The board also has the familiar LED on the output. It stays lit for several minutes after the power supply is switched off (an SMD capacitor was added to it).



The connector (5.5 × 2.1 mm) on the cable is the standard VeriFone type. As usual, it's right-angled.


Compared with the previous power supply (model AU1370933g), the power components have been beefed up. They've switched to MagnaChip. The switching transistor is a T60R380P (plastic package, now rated for 11A/650V, Rds 0.38 Ω). The synchronous rectifier is an MDP1922 (now rated for 100V/97A). For transistors, pay attention to the current rating at 100°C (the Drain Current TC=100°C parameter). At first I was surprised to see that the MDP1922 has a metal package but no insulating pad. However, its heatsink isn't connected to anything. I couldn't identify the PWM controller IC.


The synchronous rectifier controller is definitely an MP6902DS (the marking is readable). The bridge rectifier looks exactly like the ones used in common power supplies (3A, 700V).
The input filter is again standard for VeriFone: a square plastic fuse (3.15A), varistor, thermistor (jumper), small inductor, X capacitor, and large inductor. No complaints here; everything is excellent (apart from the factory thermistor jumper). There's a proper Y capacitor. Companies often put ferrite beads on the leads of this Y capacitor.



Here, the wires to the 230V input connector are a reasonable gauge (not very thick). I couldn't understand why the previous power supply (model AU1370933g) had such thick wires for 230V. The current there is very low, even in countries with a 110V mains supply. These wires are crimped to the connector, rather than simply soldered.
A couple of things on this board surprised me a little. One lead of the T60R380P switching transistor has a ferrite bead on it. That's fairly common, except this bead is on the third lead (the source). I often put beads on transistor leads, but I've always put them on the first lead (the gate). The previous power supply (model AU1370933g) was the same, by the way.
As for these beads—in principle, you shouldn't overdo it on switching transistors. But I use them quite often and haven't noticed any downsides so far. If I ever do, I'll stop.
Here's another quote from Ksiman on Muska (about beads on transistors): “Ferrite rings on transistor gate leads damp parasitic resonant voltage oscillations at the gate (due to internal and external drain-gate capacitances), and also reduce parasitic interaction between transistors connected in parallel (in a PFC circuit). Beads are also sometimes placed in the output diode circuit to reduce switching noise. Whether beads are needed is determined during the design and tuning of the power supply (end quote).”
Feel free to put ferrite beads on diode assemblies. I always put them on the two outer leads. But I've seen branded power supplies where the bead was only on the center pin (common cathode).
The next thing that surprised me a little was that there was only one small electrolytic capacitor on the board. And it has a very unusual capacitance of 4.7uF/50V. The previous power supply was the same. What's more, the factory-installed capacitor on the previous power supply was large for that capacitance and had a scored vent on top. I didn't even believe the marking and thought it was a mistake. But I checked, and the capacitance really is 4.7uF. The output power traces are very wide and have been reinforced with solder plating—it looks as if the board was designed for low voltage and high current.
The next thing was that the transformer looked small in terms of its dimensions. And finally, in my previous review I didn't understand why they had made room for resistors (during the design) next to the large common-mode choke.
Well, those spots are still there, and what's more, the resistors are now installed (R6 and R9). They're 51 kΩ, size 1206.
I have no questions about resistors R17 and R19. The standard pair of 155s (1.5 MΩ) to discharge the X capacitor. I used to do the same thing on small Chinese power supplies. On your high-power Chinese power supplies, you'll usually find a regular 0.5 W resistor, often rated at 680 kΩ, to discharge the X capacitor.


After looking at several used power supplies from VeriFone, I noticed one thing. They glue all the power components—transistors, diode assemblies, and bridge rectifiers—to the heatsinks with thermal adhesive. I wanted to replace the diode assemblies in a couple of power supplies with 60V ones (the originals are 100V), but I couldn't get them off. I gave up and left the originals in place rather than heating them or soaking them in gasoline/solvent. I only added ferrite beads to their outer pins (on the Au-7992n model).
My observations about the board (subjective): I've already mentioned the thermistor. With an input capacitor of 100uF, I think a thermistor is essential. Even though there are two input chokes whose resistance partly serves the same purpose as a thermistor. My next observation is about the heatsink and the high-voltage capacitor leads.
The high-voltage capacitor is right up against the heatsink, and its leads aren't insulated. The heatsink isn't insulated either, even from the capacitor's positive lead. In its defense, the outer side of the heatsink is insulated.
And then there's the lack of an output filter choke. There's no ring-core choke on the output either, unlike the previous power supply. But instead, there's a large, genuine ferrite filter on the cable (at the power supply end). It's the equivalent of the ring core on the previous power supply's board. Check the ferrite filter on the cable with a magnet (it should be magnetic; the Chinese often use a dummy).

What I did (and usually do) to the board.
I removed the jumper and installed a Chinese 5D11 thermistor from AliExpress (covered with heat-shrink tubing). As usual, I replaced the output capacitors with 1000uF/25V ones (2 pcs. Samwha RD, 10/20mm). I always replace the small capacitor(s). I have some good-quality small capacitors on hand—Samwha WL and JAMICON TL (there were some very old batches in a shop for 2–3 RUB). I also replaced the high-voltage 100uF/400V capacitor on this board. I removed the original one (it'll come in handy elsewhere) and installed a branded 82uF/420V capacitor (18/25mm). I covered its leads with pieces of heat-shrink tubing and secured them with sealant. I desoldered the heatsink together with the switching transistor and bridge rectifier, then insulated the heatsink with thermal tape where it faces the high-voltage capacitor.
On the back of the board, I added 1206 SMD ceramic capacitors to the output. I resoldered the larger, high-current components with regular POS-61 solder. When replacing capacitors, use the highest point as a guide for their height. That's usually the transformer or heatsink. If you install the capacitors any higher, the case probably won't close (check right away, after soldering in the first capacitor).



For the temperature test under load, I tested the bare board. Apart from coating the board with varnish, I ran it for 1 hour at 4.2A and 1 hour at 4.5A.
The result: I was wrong to worry about the transformer size. I don't know what frequency it runs at, but after two hours under load, the transformer was warm, with plenty of headroom (I could comfortably keep my finger on it). The other components were similar—their temperatures were well within limits (with the board uncovered).
One thing I forgot to do: I wanted to replace the small Y capacitor with a larger 400V one (222) with beads. I remembered only after varnishing the board and sealing the components.

What I usually do with boards like this (using this board as an example).
I connect a load and set the current to 3.5A for 20–30 minutes. Then I disconnect the load from the power supply and, while the board is still warm, coat the chokes with varnish and add varnish to the transformer windings. I also varnish the wire connections to the 230V connector. Just to be safe, so nobody at home gets a shock while you're having a smoke in the kitchen (while the board is under load). Then I turn the load and power supply back on for about 30 minutes.
Then I turn the board over (once the varnish has dried a little) and coat the entire solder side with varnish. I power up the board upside down and leave it under load until the varnish dries. Then be sure to air out the room.
I have both PLASTIK 71 varnish and KO-921 (100 ml, Rexant 09-3789) at home. KO-921 is thicker, so I use it on transformers more often. I was running out of PLASTIK 71 varnish (100ml), so I had to buy 500ml (because of the shipping cost). You can't find varnish anywhere in town; it's disappeared. I top up the old 100ml bottle. I use a brush from an old bottle of flux ФКСп (20мл). After all the checks, I secure the components to the board with sealant.


The bottom line on this power supply: the temperature is fine. A synchronous rectifier is genuinely very efficient. In the past, the diode assembly was the main source of heat in these power supplies (the ones with plastic cases). And, unfortunately, the capacitors were right next to the diode assembly (and heatsink), so they got very hot. That problem is completely gone now. The main downside of a synchronous rectifier is that it works efficiently (and is used) only in power supplies with relatively low output voltage.
A general note on output capacitors in flyback power supplies (especially those with diode assemblies): pay the most attention to the capacitors before the filter inductor. The inductor is installed in series with the positive trace. Those capacitors matter most, so use name-brand ones. The capacitor after the inductor has a much lighter load. Typically, for example, two 1000 µF/16 V capacitors are installed before the inductor, followed by a small 470–680 µF/16 V capacitor. If your power supply or small adapter won't start and there are no visibly bulging capacitors, first check (or replace) the small capacitor in the IC's power supply circuit.
This is usually a 10–22–33 µF/35–50 V capacitor. When checking small capacitors, bear in mind that an ESR of a couple of ohms is normal for them. The SMD resistors in this power supply are really tiny. I couldn't manage to work with resistors like these at home, for example, if I wanted to change the output voltage. You need good lighting, a fine tip, and a microscope to solder them (smaller than 0805).
I marked three resistors in the primary circuit on the board (size 1206) that control overcurrent protection. To change the overcurrent protection, replace these resistors. The general rule is: lower the resistance to increase the protection current, and vice versa.


As an addendum to the review of the power supply (model Au-7992n).
Four resistors are soldered in a row on the board, in parallel: SMD 1206. Three are rated 2R0 (2 Ω), and one is rated 2R2 (2.2 Ω). To set the power supply to 5 A at 12 V, simply replace the 2R2 (2.2 Ω) resistor with a 2R0 (2 Ω) one. In other words, all four resistors should be 2 Ω. Depending on resistor tolerances, overcurrent protection will trip at 5–5.2 A. I spent ages looking for these 2R0 SMD resistors at home. I knew I'd bought a set of low-resistance SMD resistors (1206) specifically for this a long time ago. Eventually, I managed to find the little bag of them.

Now, about the output voltage on the modified 12 V power supplies (model Au-7992n)—here's how I set it.
The 18 V Zener diode works only up to about 12.45 V. I always set the voltage on power supplies like these at least 0.5 V higher. Account for voltage drop across the wires and connectors under load. I consider 12.5–12.7 V normal (up to 12.8 V with no load). In that case, use a 20 V Zener diode. Solder a 7.5 kΩ resistor onto the original resistor in the TL431 circuit (marked in the old Au-7992n review) to get about 12.5 V, or a 6.8 kΩ resistor to get up to about 12.8 V. It depends on resistor tolerances. Use 0805 resistors in the TL431 circuit (on the Au-7992n).
But using an 18 V Zener diode and setting the power supply to 12.4 V (solder an 8.2 kΩ resistor onto the original R39 resistor in the TL431 circuit) has one big advantage: the power supply's overvoltage and undervoltage protection works very precisely.
Using a 20 V Zener diode makes that protection much less sensitive.


A note on solder from Russian stores.
I have several solder brands at home: ЗУБР ПОС-61 (1 mm, with rosin), Рязанский ПОС-61 (0.8 mm, with rosin), Rexant с флюсом (various diameters, from 0.6 mm), and MECHANIC HX-T100 (with flux, various diameters, from 0.6 mm). They're all excellent solder.
A long time ago, people on Muska tried to tell me that ЗУБР and Rexant were Chinese junk. I can't speak for all their products, but their solder is excellent. I buy ЗУБР and Rexant during sales (or pricing errors) at DNS. Every now and then, for a couple of days, a 250 g spool of Rexant sells for 450–550 RUB. ЗУБР ПОС-61 is actually proper Ryazan-made solder. They even use the same spools.
If your soldering iron tip doesn't respond well to rosin, use Rexant solder with flux or MECHANIC HX-T100. I buy all my solder locally; MECHANIC HX-T100 is cheaper than on AliExpress.
I had some doubts about Rexant solder a while back, but it turned out to be excellent—I have no complaints at all. The only downside is that the listed weight includes the spool. For ЗУБР ПОС-61 and Рязанский ПОС-61, the weight is for the solder itself (they're 100% identical and come from the same factory). As for flux, 559 works great for soldering. Look for it on AliExpress, with a plunger and needles included (it needs to be washed off). I also often use plain liquid rosin.

That's all from me. Good luck and stay healthy.