[Resolved] Part Identification Help - J15 SMD Transistor

[Grytpype]

Hi and Happy New Year all!

I'm trying to replicate a circuit and have identified all the parts bar a couple of (what I assume are) transistors in the bottom-right of the image.

They have J15 on the silkscreen and seem to be the 4th component in the list here: https://www.electronica-pt.com/db/smd.php?ref=J15 (CHT857BPT-S)

From a really sparse datasheet I think it's a "PNP Muti-Chip General Purpose Amplifier" (?) but can't find out much more about them, or find them on ebay, mouser etc. They seem to be discontinued from what little info I can find so am wondering if anyone can shed any light on what they actually are and suggest an alternative.

Thanks all!

[BillyO]

I see nothing on the silk screen.  Do you mean "J15" is written directly on the devices?

If you reverse engineered it, can you post the schematic you came up with??

If it is a PNP GP amplifier a 2N3906 will probably suffice.

[Jwillis]

BC857   https://assets.nexperia.com/documents/data-sheet/BC856_BC857_BC858.pdf

[golden_labels]

That datasheet gives markings 3E to 3H, not 1J. In this circuit the bottom right element would have its rails directly shorted through BE, reverse polarized. See the attachment. So I find that unlikely.

Perhaps some N-MOSFET? Matches BSS138LT1.

[Jwillis]

The BC857 is the equivalent to the specs of the J15 in the link that the OP posted . 45V, 100mA, B=420..800, 100MHz and the type listed is CHT857BPT-S. J15 is a PNP

[gamalot]

The marking is J1 not J15, the character 5 is not part of device code.

[Grytpype]

Thanks for all the replies. I had to use a microscope to read it as it's soooo faint but yes there are two "J15"s.
Will try BC857s and let you know how I get on, and I'll put the schematic up once I've finished it.

Thanks again!

[gamalot]

Thanks for all the replies. I had to use a microscope to read it as it's soooo faint but yes there are two "J15"s.
Will try BC857s and let you know how I get on, and I'll put the schematic up once I've finished it.

Thanks again!

No, you are wrong.

They shouldn't be PNP transistors by any means, almost all SOT23 transistors (including BC857) have pin 2 as the emitter, and they shouldn't be connected to the blue one of the two wires that look like they come from a solar panel.

[Grytpype]

Ah okay. So wrong in that it says J15 / J1(5) or wrong in substituting the BC857 ...or both? What would you recommend?
J1's seem to mostly be voltage regulators from what I can see (?)
Ta!

[gamalot]

Ah okay. So wrong in that it says J15 / J1(5) or wrong in substituting the BC857 ...or both? What would you recommend?
J1's seem to mostly be voltage regulators from what I can see (?)
Ta!

I didn't repeat it because I saw that @golden_labels already pointed out that heit should be BSS138 and I totally agree with him.

[Grytpype]

Smashing - thanks!

[golden_labels]

As for reverse-engineering, a few hints:

If I recognize it correctly, the upper chip is LMC7660. The way pins 8 (V+), 1 (NC) and 3 (fat traces, probably GND) are used supports that guess. If yes, it’s very likely it is a copy of one of the circuits in the datasheet. That would solve the upper right quadrant. If I am seeing connections right on the photo, figure 18 seems promising.

The mystery J1(5) elements in the bottom seem to me like something that may be used to detect solar power being present, and switching fragments of the circuit on. If it is indeed that N-MOSFET, the left one would be enabling the entire LMC7660 network by binding its GND to the negative power rail.

I can't see markings on the SOT23 element between solar +/- connections, but I would start the search with diodes. In particular Schottky diodes, possibly dual (see BAT54 as an example).

[Grytpype]

..just as an update, I took a look at another later issue broken board and the '5' of the 'J15' is clearly separate from the J1 on both trannies.

Am flu'd up at the moment but will post a circuit diagram when better.

Ta!

[Grytpype]

Hi all,

I've finished the PCB layout if anyone's interested and added the parts list too.

I've added each component's complete silkscreen info, along with what I think is the corresponding part / code where applicable. If anyone has time to sanity-check it then that would be great!

Ta!

>>>

Q1, Q2: J1 5 - BSS138 n-channel mosfet
Q3, Q4, Q5: P3 L44 - BAT54S Schottky diode
D1, D2, D3: N7 A2 - 1N4148WS diode
R1, R4: 1004 - 1M ohm resistor
R2, R3: 000 - 0 ohm resistor
R5: 2702 - 27K ohm resistor
U1: 16 AJ LM 555 CM - 555 timer
U2: 21 AG LMC 7660 - switching voltage regulator
C1, C3, C4: 10nf - ceramic
C2, C5: 10uf 16v - electrolytic

Plus a small solar panel, LED, and 3.7v rechargeable battery.

NB Q3, Q4, Q5, and U2 have floating / unconnected pins.

[mariush]

Well, I don't know what others would say, but I think it's really bad layout.

why use bat54s when you use only one diode ... just use some other generic schottky diode with two pins / contacts instead.
.. ex mouser has 200k+ of the RB521S in stock : https://eu.mouser.com/ProductDetail/onsemi/RB521S30T1G?qs=xZq1yRCsb1eKFKS8TVH%2Fqg%3D%3D

try having ALL resistors and diodes horizontal instead of vertical (except maybe R2)

may make more sense to have the trace from pin 4 of U1 go to pin 8 under the chip and have R2 to the right of U1, instead of jumping over two traces.
You can use a jumper link or a 0 ohm resistor to jump from pin 2 over the trace on pin 3 and then connect to pin 6

Some quick and dirty messing around in paint :




absolutely no need to have traces at weird angles ... keep it 90 degrees, 45 degrees for better looks.

[Grytpype]

Nice one - thanks for the advice and 'dirty' paint

I was following the layout from the original PCB but thought it was very strange to leave things floating - maybe they just designed it with what they had lying around; or it worked out cheaper to do it that way in bulk.

BTW I didn't mention that this was from an old motorway cat's eye which I always assumed were just reflectors.

Ta!

[golden_labels]

Well, I don't know what others would say, but I think it's really bad layout.

It is not OP’s layout? This thread is about reverse-engineering an existing layout.

As for using BAT54: for example because of BOM? Because of the assembly line limitations? Because it was cheaper? Because it is already well known to the designer? Because it is widely used in other circuits on the market? Because PCB was designed for a different diode and later they realized it’s not meeting the requirements? Because time spent on even considering that would pointlessly increase costs? There are many reasons.

[golden_labels]

I checked the reverse-engineered layout. It seems to be in agreement with the photo. At least where I can see the details well enough. But it makes some sense.

The 555 part looks like a classic astable multivibrator, with diodes added to control duty cycle. But it seems to do nothing more than driving the LED. I wonder: am I getting something wrong or saving energy is so important in this device?

[Grytpype]

Thanks again for all the help - I made the circuit up and it didn't work; I've got +3.7v both sides of the unlit LED so something's rotten in Denmark.

I don't seem to have any shorts so think I've not mapped it properly. I've stripped the components off the dead board and will put it on a light box to work out where I've gone wrong.

Ta!

[golden_labels]

If I made no mistake, this is the corresponding schematic. I marked D1 as 1N4148, following your identification, but please double check it. I believe it’s meant to be a zener diode.

As for debugging, there seem to be three distinct sections (if I made mo mistake! ). It should be possible to test them independently.

[Grytpype]

Thanks for taking the time to do that! Brilliant - and it makes a lot more sense laid out like that

I'll go back over it at the weekend and let you know how I get on.

Ta!

[Grytpype]

Hi all - I got my PCB circuit at least partly working; I haven't tested the charging side completely but it looks promising - just need some decent sun! Thanks again golden_labels for the advice and diagram!

It seems to me that it was probably always working but I was testing it without the solar cell connected. Maybe this left something high / low that stopped the light part of the circuit becoming active - whilst still covered so it doesn't provide any power, disconnecting it stops the LED.

I also got one of the original broken circuits working and took measurements on all the parts. This is attached and has the readings next to the pins. Any pin without a reading had zero / negligible millivolts.

The oscilloscope gif is what happens to pin 3 when the solar cell gets light - just thought it looked pretty cool.

Ta!

>> Update: It charges too!