Help identifying transistor

[ralun]

Hi! I have a broken led light I'm trying to fix, bought it second hand and it had gotten in to contact with saltwater.
The water caused some corrosion to a headerstrip, and a transistor which I think is part of the thermal protection is broken.

The problem is that I can't identify the transistor, it's marked with either B2H2 or 82H2, neither of which I'm able to find as a corresponding code.
I've attached two images of the transistor and the pcb it's on, if someone could help me out it would be great.

[bsfeechannel]

It can be a voltage regulator, not a transistor.

[magic]

U1 suggests it is indeed some IC rather than a transistor. A transistor would be Q-something.

What makes you think it's broken? If it's only a matter of corrosion on the upper solder joint, removing the old solder and applying new will get it to working condition unless the leg of the IC is completely rusted through.

[scatterandfocus]

Crazy times, when the components have gotten so small that an IC looks like a transistor, like a a resistor, like a capacitor, like an inductor.  In another 20 years, electronic components will have vaporized.   

[ralun]

U1 suggests it is indeed some IC rather than a transistor. A transistor would be Q-something.

What makes you think it's broken? If it's only a matter of corrosion on the upper solder joint, removing the old solder and applying new will get it to working condition unless the leg of the IC is completely rusted through.

Yes the U1 had me thinking it was something else than a transistor, but it tests out like a transistor. And then the fact that it seems to be part of the thermal protection makes me wonder if it isn't a bit of an overkill to use an IC rather than a transistor?

I should have been a bit more clear on the picture, the light I'm fixing has two identical led-PCBs, the picture is from the other pcb and it is still working.
When testing the broken transistor it reads 2 Ohms (would the take number with a tub of salt as I only have a cheap multimeter) between the upper leg and the left leg, in both directions.
The other leg and the other transistor reads ~500 Ohms in one direction only from the upper leg. The other one also reads ~1k Ohm from the left leg to the right leg.

I edited the picture of the pcb to make it easier to see some of the lanes.

[magic]

Resistance measurements on semiconductors are quite meaningless, use diode mode instead.

But if they are identical and one conducts in both directions while the other doesn't, then sure, one of them is toast.

[scatterandfocus]

Resistance measurements on semiconductors are quite meaningless, use diode mode instead.

But if they are identical and one conducts in both directions while the other doesn't, then sure, one of them is toast.

Can you elaborate on that a bit?  I just resistance tested some 2N3055's using a known good one as a reference, and it seems that resistance testing was useful.  Diode testing didn't tell me anything, but maybe I'm missing something.  Here is the thread where I talk about it:  https://www.eevblog.com/forum/repair/hpagilent-6227b-q1-q2-q8-q11-not-assigned-in-service-manual/

[ralun]

Resistance measurements on semiconductors are quite meaningless, use diode mode instead.

But if they are identical and one conducts in both directions while the other doesn't, then sure, one of them is toast.

The woes of a cheap multimeter unfortunetly, the diode mode is part of the Ohm scale and is separated from the "beep"-mode.

Anyone have a way of seeing if it's an PNP or NPN transistor?

Also do you think it would be possible to just skip the transistor all together, I mean if it's part of the thermal protection it shouldn't open until the thermistor gets hot enough right?

[Jwillis]

 SOT23 BH2B  B2H2 is identified as a TL431 adjustable shunt regulator. http://www.ti.com/lit/ds/symlink/tl431li.pdf

[amyk]

SOT23 BH2B  B2H2 is identified as a TL431 adjustable shunt regulator. http://www.ti.com/lit/ds/symlink/tl431li.pdf

Where do you see that?

From the famous Russian site, the following match marking and package:

Code: [Select]

B2xx 11AA020-I/TT 11AA020 SOT-23 2 Kbit, 1.8V UNI/O Serial EEPROM
B2xx R3111N122A R3111N SOT-23-3 1.2V Low Voltage Detector N-ch Open Drain Output
B2xx R3130N22EC R3130N SOT-23-3 2.2V Low Voltage Detector CMOS Output with Delay
It is unlikely to be an EEPROM, the other two make more sense but you can look at the pinout of how it's used in the circuit to make a better determination.

[Jwillis]

The package is a SOT23  and a B2H2 in that package is a tl431. On the B2XX the xx denotes 2 more numbers that would be the code that would be printed . and the only other one on that list that comes even slightly close is the B2HX but that is a SOT25.
upon further investigation the H2 could be a lot code so amyk could very well be correct.

[magic]

Can you elaborate on that a bit?  I just resistance tested some 2N3055's using a known good one as a reference, and it seems that resistance testing was useful.  Diode testing didn't tell me anything, but maybe I'm missing something.

You will get some result but it depends on which DMM and range you use. I'm also too lazy to try to guess how the resistance reading will depend on temperature and whether it's a big deal or not and how DMM measurement circuitry affects it, if at all.
Meanwhile diode mode is intended precisely for measuring forward voltage of PN junctions and it will show the forward voltage of the junction for some reasonable current such that any variation isn't going to affect the result by much. You can know that if it's something like 600-800mV then your transistor is probably okay but if it shows 400mV then it is likely busted. Or maybe it's a dual Schottky in 3-pin package or whatever.
Moreover, in diode mode you are guaranteed that the red lead is positive and black is negative. With resistance, some crappy DMMs reportedly use inverse polarity and you may end up thinking the junctions are backwards.

Anyone have a way of seeing if it's an PNP or NPN transistor?

The pinout of most SMD transistors is as follows:
left - base/gate
right - emitter/source
top - collector/drain

NPN conducts from base to emitter or collector, PNP conducts from emitter or collector to base.

I'm pretty sure it's an IC, though. On the good board, check which pin is connected to ground and what are the voltages on the other pins when it operates.

edit
LOL, an EEPROM in SOT23 and TO92 I think I have seen everything now

[LaserTazerPhaser]

Silk is larger than components in many cases. Soon there will likely be many boards without reference designators to increase part density.

[amyk]

The package is a SOT23  and a B2H2 in that package is a tl431. On the B2XX the xx denotes 2 more numbers that would be the code that would be printed . and the only other one on that list that comes even slightly close is the B2HX but that is a SOT25.
upon further investigation the H2 could be a lot code so amyk could very well be correct.

Where did you see that...? From the datasheet you linked, they all begin with "1".

[scatterandfocus]

Can you elaborate on that a bit?  I just resistance tested some 2N3055's using a known good one as a reference, and it seems that resistance testing was useful.  Diode testing didn't tell me anything, but maybe I'm missing something.

You will get some result but it depends on which DMM and range you use. I'm also too lazy to try to guess how the resistance reading will depend on temperature and whether it's a big deal or not and how DMM measurement circuitry affects it, if at all.
Meanwhile diode mode is intended precisely for measuring forward voltage of PN junctions and it will show the forward voltage of the junction for some reasonable current such that any variation isn't going to affect the result by much. You can know that if it's something like 600-800mV then your transistor is probably okay but if it shows 400mV then it is likely busted. Or maybe it's a dual Schottky in 3-pin package or whatever.

That would mean that all my 2N3055's are bad, as all of them are dropping 0.45-0.48 volts when testing with diode mode, which seems unlikely.  So I'm wondering what might be going on here.  I guess it is possible that I had 4 bad 2N3055's on the non-working side of my dual power supply and that I killed the reference 2N3055 from the working side of the power supply.  But it seems unlikely.  So I'm wondering what might be going on here.

Also, I ran across this:

https://www.androiderode.com/how-to-test-a-transistor-with-dmm/
Transistor BU 208 or 2N3055 Testing
Forward bias:   Connect DMM (+) Test Red lead  to Base
DMM (-) Test Black lead to Collector reading shows = 0.449v
DMM (-)  Test Lead to Emitter    = 0.502v
Verification: If the DMM reading is from 0.502v to 0.449v the condition is GOOD.
Reverse bias:

Connect DMM (-) Test lead to BASE
DMM (+ )Test lead  to Collector reading is = OL or 1 or open
DMM (+ )Test lead  to Emitter    = OL or 1 or open
DMM Reading shows = OL (over load) the condition is GOOD
Verification:  If you get reading in forward bias as 0000 or OL  or 1 or open and in reverse bias as 0000 (or) low values the the transistor can be FAULTY and needs replacement.

I would like to understand where these numbers are coming from.

[ralun]

From the famous Russian site, the following match marking and package:

Code: [Select]

B2xx 11AA020-I/TT 11AA020 SOT-23 2 Kbit, 1.8V UNI/O Serial EEPROM
B2xx R3111N122A R3111N SOT-23-3 1.2V Low Voltage Detector N-ch Open Drain Output
B2xx R3130N22EC R3130N SOT-23-3 2.2V Low Voltage Detector CMOS Output with Delay
It is unlikely to be an EEPROM, the other two make more sense but you can look at the pinout of how it's used in the circuit to make a better determination.

Thanks, I'll look in to these.

Anyone have a way of seeing if it's an PNP or NPN transistor?

The pinout of most SMD transistors is as follows:
left - base/gate
right - emitter/source
top - collector/drain

NPN conducts from base to emitter or collector, PNP conducts from emitter or collector to base.

I'm pretty sure it's an IC, though. On the good board, check which pin is connected to ground and what are the voltages on the other pins when it operates.

edit
LOL, an EEPROM in SOT23 and TO92 I think I have seen everything now

Thanks, and I think you might be right on the IC part seeing the above sudgested ICs. They are all connected to ground, which I think makes sense looking at this datasheet page 15?
I can't measure any voltages though as the lights failure proctection is stoping it from running at all, it only displays an error light.

I went ahead and split the broken part, I don't know if it's possible to read anything from that but I attached two images of it?

[ralun]

Ugh, tried to find a store that sells the R3111N122A-TR-FE.
Found two stores that stock it, one has a min. order of 893 units, the other one is waiting for a restock at the end of november. 

Took a walk and got an idea though, anyone think it would work to bypass it by wire like the attached picture?
Might need to skip the two blue wires though, don't know it would cause a short?

[magic]

It seems a few candidate ICs have been identified and other unknown ICs may exist with similar codes.

Before changing anything, be absolutely sure that you know which IC it is and that its connections to other parts match the datasheet.
Just looking at ground doesn't say much, there is 33% random chance that any 3-pin IC has ground on the same pin as yours

[ralun]

It seems a few candidate ICs have been identified and other unknown ICs may exist with similar codes.

Before changing anything, be absolutely sure that you know which IC it is and that its connections to other parts match the datasheet.
Just looking at ground doesn't say much, there is 33% random chance that any 3-pin IC has ground on the same pin as yours

Yeah that's true, got a bit excited that it might had been identified.
Guess to only real way of knowing is to measure voltages across it, but I need the broken pcb to send signal aswell for it to start.
Do you think my drawn jury rig above would work without shorting everything to hell?

[ralun]

It seems a few candidate ICs have been identified and other unknown ICs may exist with similar codes.

Before changing anything, be absolutely sure that you know which IC it is and that its connections to other parts match the datasheet.
Just looking at ground doesn't say much, there is 33% random chance that any 3-pin IC has ground on the same pin as yours

Yeah that's true, got a bit excited that it might had been identified.
Guess to only real way of knowing is to measure voltages across it, but I need the broken pcb to send signal aswell for it to start.
Do you think my drawn jury rig above would work without shorting everything to hell?

Turns out I'm more of an idiot than I thought, ofcourse I can measure voltages, how else would it know not to turn on.
In my defence my craptastic multimeter takes 2-3 seconds to measure voltage sometimes though, so I was a bit hasty when I tried it before.

I just can't get my head around this though, I get:
0,75 V from left to upper leg. (3,14? V on broken pcb)
3,14 V from right to upper (most likely across the C2 capacitor? Same reading on the broken pcb)
2,4 V from right to left leg. (0,11 V on broken pcb)

But I don't get any measurements on the lane going to/from the left leg.
I once again made a picture and added the values I got while measuring.