TL-SG108E switch voltage regulator

[sandros]

Hello,

I gave 9V to this 5V switch and the magic smoke came out.
I am basically in the same situation as https://www.reddit.com/r/ElectronicsRepair/comments/1fsl8w7/tplink_switch_tlsg108_repair/

I have measured 2 ICs near the power connector and found that at least one of them (U2 on the image in reddit thread) is shorted. (U2 beeps on 2 pairs, U1 does not on the same legs)

U1 is GF2D9R
U2 is JWM3J 3J2cG

I would like to replace both but none of them are available from the source I have access to in my location (https://www.hestore.hu/index.php?lang=en)

What would be alternatives to these?
I have no idea what am I looking for exactly that can replace these tbh.

Thank you!

[floobydust]

Welcome to the forums

SMD markiing JWM3_ is JW5250A. "JW" is JoulWatt, SOT23-5 package. It is a buck converter, datasheet

Not sure about the second IC.

[sandros]

Thank you for your reply.

I was able to find this as a potential replacement for U2: https://www.aliexpress.com/item/1005009341259202.html?mp=1
For U1 I am unable to find a replacement.

I also measured where U1 and U2 connect from their Pin 3. These are connecting to an RTL8370N on AVDDH and AVDDL with exceptions.
According to https://static.chipdip.ru/lib/594/DOC041594562.pdf AVDDH is 3.63V max., AVDDL is 1.21V max.

Not sure if the RTL chip has received higher voltage is if it's ok and just the U1/U2 is damaged.

I am looking for a replacement for U1 too. Any recommendations are welcome.
Thank you!

[GraemeG]

I don't know what part U1 is but the circuit layout and components around U1 are almost identical to those around U2. It's almost certainly another switch mode voltage regulator just with a different output voltage.

U2 is a buck converter so it steps the voltage down to something lower than the input. U1 could be the same or could even be a boost converter to increase the voltage. If you can map out the circuit around both devices you might be able to relate them to the datasheet floobydust provided. If they are both set up the same then I'd assume U1 is also a buck converter and, if you are desperate, you could try using the same part as U2. If the circuit layout is a bit different then it may be a boost regulator designed to output a higher voltage. I have no answer but it might narrow down your search of switch mode regulator chips.

[GraemeG]

So I thought I would try AI to see if it could help identify the chip. There was a fair bit of "discussion" and Google's AI assured me that it could be a JW5250A while Microsoft's Copilot was certain that it was NOT a JW5250A!

So I asked Copilot this. Everything below is from Copilot:
The code GF2D9R is on a SOT-23-5 package used in a circuit that appears to be a switch mode regulator. Manufacturer is unknown and it could be a boost or buck regulator. I need to identify the IC.

Here’s the good news: with the details you just added — GF2D9R, SOT‑23‑5, and used as a switching regulator — we can narrow this down much more effectively. And the even better news: GF2D9R is not a random code — it matches a known family of Chinese switch‑mode regulators.

Below is the clearest, most actionable identification path.

🎯 Most likely identity of GF2D9R
Based on the marking style, package, and function, GF2D9R is almost certainly a Chinese SOT‑23‑5 buck converter, typically a clone of one of these common devices:

1. MP2307 / MP1584 / SY7208‑style buck converter clones
These are extremely common in low‑cost electronics and often use 6‑character alphanumeric codes like:
GFxxxx
GExxxx
2Dxxxx
9Rxxxx

These codes vary by factory and batch, not by part number.

2. LN226x / XL1509 / XL1507 derivatives (SOT‑23‑5 versions)
Several Chinese manufacturers produce SOT‑23‑5 buck converters with markings in the GF2xxx pattern.

3. A 1–2 A synchronous buck converter
The SOT‑23‑5 footprint and typical application strongly suggest:

VIN: 3–6 V or 4–24 V
VOUT: 1–5 V
Switching frequency: 1–1.5 MHz
Internal MOSFET
EN pin
FB pin

This matches the behavior of the JW5250A, but the marking does not.

❌ Why it is NOT a JW5250 or JW5250A
JoulWatt uses long, two‑line markings beginning with JWM3.

They do not use short codes like GF2D9R.

No JoulWatt datasheet lists anything resembling GF2D9R.

So we can confidently rule that out.

🔍 How to identify it with high confidence
You can confirm the exact IC by checking pin function with a multimeter and the PCB traces.

SOT‑23‑5 buck converter pinouts are very distinctive:
Typical buck converter pinout (most Chinese clones):
1 – SW (switch node)
2 – GND
3 – FB (feedback)
4 – EN (enable)
5 – VIN
If your chip matches this pinout, then:

➜ It is a 1–1.5 MHz synchronous buck converter, not a boost regulator.
Boost regulators in SOT‑23‑5 have a different pinout:
1 – SW
2 – GND
3 – FB
4 – EN
5 – VIN
…but the PCB layout around them looks different (inductor on VIN side instead of VOUT side).

🧭 What I need from you to identify the exact part
If you can provide any of the following, I can narrow it down to the exact IC:

Photo or description of the PCB area 
(Where is the inductor? Where are the diodes? Is there a Schottky diode or is it synchronous?)

Pinout tracing

Which pin goes to the inductor?

Which pin goes to VIN?

Which pin goes to a resistor divider?

Input voltage and output voltage 
(This often uniquely identifies the regulator family.)

Any second‑line marking 
Many Chinese ICs have a second line with date/lot codes.

[GraemeG]

So I will leave it to you to take it further with Copilot if you want to. But I have had various AI engines give me totally wrong information so double check anything it gives you!!

My guess is that it MIGHT be a JW5250A the same as U2 but likely it is a similar but maybe different spec'ed chip. When I told Copilot that the inductor is connected to pin 3 it replied with:

About “pin 3 goes to the inductor”
If pin 3 really goes to the inductor, that pin is the SW node. That’s useful:
SW pin → inductor → output cap = classic buck topology
If another pin goes straight to input jack or bulk cap, that’s VIN
One pin to ground plane = GND
One pin to a resistor divider from VOUT = FB
Remaining pin often goes to VIN via a resistor or directly = EN
Once you confirm that pinout, you don’t actually need the exact part number to repair/replace it.

It would have helped identify a suitable replacement chip if I confirmed what was connected to each pin, input and output voltages. I'll leave that to you if you want to go down that path. Bottom line, I found it more useful than I expected and will probably use it again if I need to identify parts.

[gamalot]

GF2D9R https://www.lcsc.com/product-detail/C7425124.html

JWM3J https://www.lcsc.com/product-detail/C5331999.html

[GraemeG]

GF2D9R https://www.lcsc.com/product-detail/C7425124.html

Nice work tracking down the GF2 code gamalot! Can I ask how you found it? I don't do a lot of SMD work and my hit rate with component codes is maybe 80% but I'm always looking for better ways to find datasheets.

[gamalot]

GF2D9R https://www.lcsc.com/product-detail/C7425124.html

Nice work tracking down the GF2 code gamalot! Can I ask how you found it? I don't do a lot of SMD work and my hit rate with component codes is maybe 80% but I'm always looking for better ways to find datasheets.

My first choice is to search for part numbers by the markings on Taobao, which is basically the domestic version of Aliexpress.

[sandros]

Thank you everyone for the help.
I was able to resurrect the device by replacing the U2 only. The other one was not shorted so I gave it a try and it works.

U1 input is 5V, same as the other.

Off: I may remove the DC jack and add an 5V USB PD psu to avoid problems like this in the future.