Component identification question

[kingsolmn]

Greetings fellow enginerds! I am in need of your assistance in identifying a couple of components for a repair I am attempting. On my bench is the power supply out of a UniFi 500W PoE switch, it appears to have suffered from a power surge due to not being plugged into a proper UPS/surge protector. I have identified most of the failed components, but am having problems with a couple that I hope you all can help with. Being a UniFi product the schematics and service manuals are not available, at least that I've been able to find.

Of the components that I have yet to ID, the resistors I feel kind of confident that I can figure out the values of what appear to be standard SMD resistors. What I would like help with is ID of the MLCC adjacent to the blown SMPS control IC (is rather discolored), I'm not ever sure where to start wit this one. And what at first I thought was a gate transistor for a power MOSFET, but two of the leads are bonded together in a way that doesn't really make sense to me - even for some kind of dual diode.

Sorry if the pics are a bit blurry, my photography skills leave much to be desired. I'm going to be borrowing a better camera to get more (clearer) pics if needed, just let me know what/where more detail is needed.

I thank you all in advance for any help or insight you can provide! 

[golden_labels]

The only way to identify SMD capacitors is to measure them and guess under what conditions they operate. They do not have any markings or other indications of their parameters.

Your best bet is to look into datasheets for U5. In the majority of cases components around the component are copied directly from a circuit shown in the datasheet. If not, in many cases the change is caused by cutting costs or BOM optimization, and the circuit still works right with the default values.

The D9 component is, as the name indicates, almost surely a diode. Without seeing the markings it will be impossible to tell, what device this is. But I do not see, how connections make no sense. There could be two diodes between 1-3 and 2-3. Or a single diode between those pins, with the other pin unused.

The photo seems unreadable due to motion blur, not poor camera. I would say: use more light. Or you may put the PCB vertically against something and rest your camera on the edge of a table for stabilization.

[kingsolmn]

Awesome insight, thank you! As for the diode, that makes a lot of sense. I'm planning to see if that part of the circuit is repeated elsewhere and hopefully I can find ID markings for ordering a replacement. I had looked at the typical implementation in the datasheet to try and ID the component at D9 but didn't think to see if the capacitor is that, I'll look at that also when I get a chance. Thank you for giving me a new direction to look in! 

Also, great advice about getting better pictures. I guess it's time I start practicing for better images. Thanks for this also! 🤓

[kingsolmn]

An update on my journey of component ID, I've figured out that the dual diode in question are likely Schottky Barrier Diodes. A closer look at the board shows that the manufacturer did work to reduce BoM cost by reusing this diode package in like 7 or 8 places, all of which are marked with "A4" with a sideways "L" (If I can find a card reader that still works, I'll post a pic from my cheap scope giving a good view of one). The next issue I need to figure out is which of the possible configurations these diodes are using, the datasheet https://media.digikey.com/pdf/Data%20Sheets/Avago%20PDFs/HSMS-280x.pdf for the most likely that I've found has several possible configs. The biggest stumbling block that I'm coming across now is that for all the like components on the board they meter out the same across all the leads! I'm probably going to need to pull one off the board and meter it that way to figure out the internal configuration of the diodes. THT component ID and troubleshooting is so much easier! Haha 

[golden_labels]

The behavior may be checked with a meter: diode test is available on pretty much any DMM. Also derived from the circuit, as only a single type of operation will make sense.

Diodes in SOT-23 package and A4 markings seem to all share the same configuration: common cathode at pin 3, with anodes at pins 1 and 2.

[kingsolmn]

Yay me! \o/ Lol

I figured out that the dual diodes I needed here are the Avago HSMS-2804, which has been superseded by the ST BAS70-05FILM. This is based on the markings, usage, and metering of the pins after removing it from the board.

Since UniFi products are less than repair friendly, I have had to go on a bit of a journey to determine the values of some of the components I needed to replace. One resource for these values that actually surprised me was the official UniFi forums. Since it would not surprise me if UniFi removed the post from a generous user with some of the component values on the 500W power supply in the UniFi US-48-500W 48 port PoE switch, I am copying the post (https://community.ui.com/questions/Failed-US-24-500W-13-months-of-use-repair-options/47cc80ec-f0ff-4f24-9eee-45b897f81abc#answer/bba5f739-1310-4ab9-b887-151b8192f1cb) with the values below from user chadaaron (https://community.ui.com/user/chadaaron/37b4be3a-823b-4e74-8f64-dc868510c44e) in case someone else needs them.


I don't have a schematic but I do have a bunch of component values. I think the differences between the 750w and 500w are small enough that you can use most of these values for both. I pulled them from various sources throughout the interwebs and youtube. I'm guessing they are 90% correct. Most are from a 500w

R27=22 Ω

C50=270pF

C36=1uF

C33=100pF

C35=C103=100nF

C38=2.2nF

D9=BAV70

D4=S1M

R5=.2 Ω,2W (probably want to use a 3W or 5W)

R39=4.7 Ω

U5=EA18363

Q5=P10NK70ZFP (I think the 750w has the P10NK80ZFP)

R52=10k Ω

R47=470 Ω

Q22=MBR20200CTG

R33=22k Ω

[kingsolmn]

Finally got the picture feature of my cheap microscope figured out (cuz I've never used it  lol) AND I found an SD mini card that works. And wow does lighting makes a difference, I guess now I should upgrade my bench lighting! Lol

[fzabkar]

TEA18363T, NXP, GreenChip SMPS control IC, SO8:
https://www.nxp.com/docs/en/data-sheet/TEA18363T.pdf

Page 7 of the following document shows a similar drive circuit for the MOSFET chopper:
https://cdn14.21dianyuan.com/download.php?id=180494

The diode in that circuit is a BAS316. Your diode is probably a BAV70x (as you have already discovered):

https://assets.nexperia.com/documents/data-sheet/BAV70_SER.pdf

[kingsolmn]

Howdy again all! I finally have the time and components to finish up this repair and I'm down to needing to replace a 2W resistor but I'm stuck scratching my head on the value. This resistor was contained in a shrink tube so I missed the burnout at first, but could the banding color change as a result of the burnout? I've attached a picture of the resistor in question, and I'm hoping that someone will tell me how I am reading the marking wrong. As best I can tell the resistor is called out as either a 2W .2R or 200R depending on the source, but I haven't been able to tell if it's a typo or not tho, but I read the bands as a 2G resistor.

I wish I had at least a partial schematic to show you all, but this is the PSU from an UniFi PoE network switch so that's not happening. Thanks, UniFi! 

[kingsolmn]

Ok, I see what I did wrong reading the resistor color codes!  The third band is silver, not grey! 

Yeah, so it is a .2 ohm not a 2G ohm.

But, I don't have any 2W resistors, just 3W, and I ended up having to use a .33R as that was the closest value I have. But, it works now! Another bit of electronics staying out of the landfill! 🤓