I have a no frost freezer which stopped freezing efficiently and after some research and trials I found the problem to be related to control board. And as of now I am pretty sure the problem is with this obviously burnt transistor that controls the compressor's relay.

The burnt transistor

Using a multimeter I figured it's a pnp type (Negative probe to base gave me a reading in diode mode) however the check was done on this other one that looks exactly the same and so I confirmed the type.

The other transistor

The burnt one did not give me overload when I touched either probes to base however it gave me a much lower reading for (probably B-C, I couldn't confirm C-E pins) than it read for (B-E). 24 , 420 respectively.

Now if my logic is correct, I want to know how can I figure it's precise model so I can change it? Marking reads A4t 95, as you can see, but a search on google was futile and I couldn't reach anything so figured asking here may return something.

Thanks in advance and sorry for my english.

  • 1
    \$\begingroup\$ It looks like the flyback diode thats used to suppress inductive kickback when the coil is switched. \$\endgroup\$ – Rendeverance Jul 24 '17 at 21:56

Googling "SOT-23 transistor A4t" yields http://www.s-manuals.com/smd/a4 which suggests it might be a BAV70 dual diode, and http://www.s-manuals.com/pdf/datasheet/b/a/bav70_series_nxp.pdf datasheet, see section 4 on page 3 : Marking which says "BAV70 - A4" and "t: made in Malaysia".

I was thinking it may be diode logic to power a relay under two different conditions, but it could indeed be just flyback diodes. As such, I believe it's probably adequate ( 150mA continuous, 4A for 1 us! ) but if it failed short-circuit (as diodes do tend to do) it would prevent the relay operating.

  • 4
    \$\begingroup\$ ... which might be kind of bad news, because if a diode is burnt, it's usually not the only thing that was damaged \$\endgroup\$ – Marcus Müller Jul 24 '17 at 20:58
  • \$\begingroup\$ I reached this but was not sure if it is the same. Thank you . But what about the 95 marking? \$\endgroup\$ – Wisam Naji Jul 24 '17 at 21:02
  • 1
    \$\begingroup\$ "95" is probably a date code. It may not be as simple as a year, it may mean 2009 month 5, for example. Searching "NXP SOT-23 date codes" may turn up an explanation. \$\endgroup\$ – Brian Drummond Jul 24 '17 at 21:07

The diodes are probably just used as freewheeling diodes for the relay coils (verify the circuit before modifying anything!) - and they look rather underengineered for that purpose, especially if something else might be wrong and actuating the relays more frequently than intended (which you'll still have to fix, along with verifying the relay is all right and not shorting mains into the control circuit. But constantly replacing diodes while investigating another problem can get tiring ;) ).

Not knowing the rest of the circuit, a safe replacement would likely be a series of two 1N4148 (or BAV70 indeed) paralleled by a beefier device like an 1N4003 - this would be as good at suppressing the back EMF quickly while unloading most of the current spike into the 1N4003...

  • \$\begingroup\$ Well the problem obviously is from our power source (here in Iraq we use generators along with power grid ) and the generators most of the time are unstable and under voltage. The relay use 24 vdc and if the voltage across the coil keeps fluctuating around 24 then maybe this what you mean by actuating the coil more frequently and could have burnt the diode. Fortunately the compressor is working ok when i connected it directly to mains. Now i have to test the relay which i think will work ok. \$\endgroup\$ – Wisam Naji Jul 25 '17 at 9:38
  • \$\begingroup\$ I would appreciate if you can give me a schematic for the replacement you suggested?.. I get it that i should connect the cathods of the two 1n4148 as one terminal. And the two anods will be the other pins. But how should i connect the 1n4007 which I have? \$\endgroup\$ – Wisam Naji Jul 25 '17 at 15:14
  • \$\begingroup\$ NO, that wasn't my intent! I suggesting putting the two smaller, faster diodes in series so they have a higher turn on voltage than the slower 1N4007 or 1N4003 parallel to them - so they will only take the back EMF until the 1N40xx turns on! \$\endgroup\$ – rackandboneman Jul 25 '17 at 16:04
  • \$\begingroup\$ Ok so in such respect this configuration will replace only a single diode of BAV70 and I will need the same configuration for the other diode of BAV70 which leaves me with 4 × 1n4148 + 2 × 1n4003 .. or I may have found another replacement which uses 2 × BAS16 which contains only 1 but exactly the same diode in BAV70.. Since i found alot of BAS16 in an old pc motherboard i have.. \$\endgroup\$ – Wisam Naji Jul 25 '17 at 19:30

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.