Transistor getting very hot - suggestions please for a alernative

Thread Starter

Ford Prefect

Joined Jun 14, 2010
245
Hi guys,

Heading should be:
Transistor getting very hot - suggestions please for a alternative.

I have made the circuit below (Circuit 1) that turns on the LED light when darkness falls. This is a circuit that seems to appear on different electrical sites all over the Internet. The output is 12.14v between the collector and emitter of the Q2 transistor which is a 2N2222 in a TO-92 case. The circuit works well with a single LED as in Circuit 1.
I would like to connect it to and I have tried with a 3528 SMD 12v LED striplight which is about 1.5 metres in length as in Circuit 2. But the Q2 transistor (highlighted in red) in Circuit 2 gets very hot, too hot to hold for any more than 5 or so seconds but has not yet blown! :) But I am waiting!! :D

The Q2 transistor in Circuit 2 needs replacing for a more suitable transistor or more powerful transistor. The amperage appears to fluctuate beween 270 and 300mA although I have noticed it drop to about 70mA and the voltage is 12.14v
I find reading transistor specifications quite confusing, I was think something an alternative like a BC639 but can anyone suggest a more suitable transistor that I can try preferably in a TO-92 or similar case? Or even a different circuit? But I would like just to replace the Q2 transistor without building a new circuit board.
The LDR is a GL5528 and the LDR when it toggles is about 60k ohms.

Dark Light Circuit.jpg
 
Last edited:

AlbertHall

Joined Jun 4, 2014
12,345
The first problem when driving the striplight is that R7 won't supply enough current to turn on Q2 fully so there will be more voltage across Q2 than there needs to be and so more power and heat than there need be. Change R7 to 330Ω.

The second problem is that as the light level falls slowly, Q2 will turn on slowly which again means that there will be both voltage and current and so heat in Q2. If you would be happy with the light being off and then switching fully on with no intervening fading then the circuit needs some hysteresis. This can be done easily by adding a resistor between the collector of Q2 and the base of Q1. Some experimentation may be needed to get the right value for this resistor but 47k would be a good starting point.

The 2N2222 will work just fine in this circuit providing it doesn't have to handle fading on and off. If you want it to fade then you will need a bigger transistor which can transfer heat to the air better, probably a TO220 case like TIP121. This is a darlington transistor which will also help.
 

Marley

Joined Apr 4, 2016
502
The main problem here is that the base drive current for the output transistor is supplied by R7. With 1k the maximum base current is therefore about 12mA. This will not be enough to fully saturate the transistor.

I recommend using a darlington transistor like the TIP121. This will give the gain required. However a darlington transistor will still have a voltage drop of about 0.7V when fully on so there will be some power loss. Might need a small heatsink.
 

AlbertHall

Joined Jun 4, 2014
12,345
The main problem here is that the base drive current for the output transistor is supplied by R7. With 1k the maximum base current is therefore about 12mA. This will not be enough to fully saturate the transistor.

I recommend using a darlington transistor like the TIP121. This will give the gain required. However a darlington transistor will still have a voltage drop of about 0.7V when fully on so there will be some power loss. Might need a small heatsink.
There's an echo in here...
 

Thread Starter

Ford Prefect

Joined Jun 14, 2010
245
Thanks everyone for your suggestions - very good of you to reply so quickly. :)

Change R7 to 330Ω.
If you would be happy with the light being off and then switching fully on with no intervening fading then the circuit needs some hysteresis. This can be done easily by adding a resistor between the collector of Q2 and the base of Q1. Some experimentation may be needed to get the right value for this resistor but 47k would be a good starting point.
Yes, thank you AlbertHall.
I did not have a 330Ω resistor so at first I tried a 150Ω resistor at R7 Well that got so hot that it took the skin off my thumb and forefinger when I felt it. :D It must have been about 150-200°C :eek: - My thumb and forefinger is a little scorched - but surprisingly the resistor is still ok and reading 150Ω
I then replaced this with a 560Ω at R7 (highlighted red) and that seems to be coping quite well.
I did not have a 47kΩ resistor but I tried a 1MΩ resistor at R10 (highlighted green).
This I thought would reduce the hysteresis to a very small amount and this also appears to be working ok.
But I may try a smaller value resistor if I am not happy with the 1MΩ
See picture below.

Bin Q2 for a mosfet, IRF510, IRF540....
Thank you Dodgydave. I was also thinking about this and think that a IRF510 would be more than suitable - I may try that later.

There maybe more posts later ;)

Dark Light Circuit-1.jpg
 
Last edited:

Thread Starter

Ford Prefect

Joined Jun 14, 2010
245
Now that the daylight has gone, the strip light comes on but the transistor Q2 is still getting quite hot.
What else can I try?
Can anyone suggest/draw a suitable circuit so the circuit is more stable?
I am worried that the transistor will eventually give up and stop working.
Thank you
 

ian field

Joined Oct 27, 2012
6,536
Hi guys,

Heading should be:
Transistor getting very hot - suggestions please for a alternative

I have made the circuit below (Circuit 1) that turns on the LED light when darkness falls. This is a circuit that seems to appear on different electrical sites all over the Internet. The output is 12.14v between the collector and emitter of the Q2 transistor which is a 2N2222 in a TO-92 case. The circuit works well with a single LED as in Circuit 1.
I would like to connect it to and I have tried with a 3528 SMD 12v LED striplight which is about 1 metre in length as in Circuit 2. But the Q2 transistor (highlighted in red) in Circuit 2 gets very hot, too hot to hold for any more than 5 or so seconds but has not yet blown! :) But I am waiting!! :D

The Q2 transistor in Circuit 2 needs replacing for a more suitable transistor or more powerful transistor. The amperage appears to fluctuate beween 270 and 300mA although I have noticed it drop to about 70mA and the voltage is 12.14v
I find reading transistor specifications quite confusing, I was think something an alternative like a BC639 but can anyone suggest a more suitable transistor that I can try preferably in a TO-92 or similar case? Or even a different circuit? But I would like just to replace the Q2 transistor without building a new circuit board.
The LDR is a GL5528 and the LDR when it toggles is about 60k ohms.

View attachment 123934
That much current is sailing close to the edge for a 2222 - highly likely you need a heatsink.

The circuit has insufficient Schmitt action, so there will be a region of light intensity that only turns the transistor half on. Fully saturated is maximum current but very small volt drop and very little dissipation. Fully cut off is all the voltage but only a tiny leakage current - also hardly any dissipation.
 

Thread Starter

Ford Prefect

Joined Jun 14, 2010
245
What transistor are you using for Q2,

Did you replace it with the Mosfet?
I am still using the 2N2222 transistors. I haven't been to the electrical component shop yet, maybe later this week. I think I shall have to scrap using the 2N2222's and get an IRFxxx MOSFET or something similar. I was even thinking of a 2N7000 but I think this is not powerful enough either.
Looks like I will have to design and build a new PCB after all. :(
Can anyone suggest a design a circuit with an IRFxxx MOSFET and with hysteresis?
Just looked...The component shop does not sell the IRF510 or IRF520 but they sell the IRF530 and IRF540 so I'll use the IRF530 MOSFET
 
Last edited:

ian field

Joined Oct 27, 2012
6,536
I am still using the 2N2222 transistors. I haven't been to the electrical component shop yet, maybe later this week. I think I shall have to scrap using the 2N2222's and get an IRFxxx MOSFET or something similar. I was even thinking of a 2N7000 but I think this is not powerful enough either.
Looks like I will have to design and make a new PCB after all. :(
Can anyone suggest a design a circuit with an IRFxxx MOSFET and with hysteresis?
Just looked...The component shop does not sell the IRF510 or IRF520 but they sell the IRF530 and IRF540 so I'll use the IRF530 MOSFET
MOSFETs have a fair bit of parameter spread on VGSthr, so could take more effort to get working.

With a MOSFET - my preferred route would be a dual op-amp. Its easy to wire one stage up as a Schmitt trigger and the other can be handy as a high input impedance voltage follower/buffer.
 

Thread Starter

Ford Prefect

Joined Jun 14, 2010
245
MOSFETs have a fair bit of parameter spread on VGSthr, so could take more effort to get working.
With a MOSFET - my preferred route would be a dual op-amp. Its easy to wire one stage up as a Schmitt trigger and the other can be handy as a high input impedance voltage follower/buffer.
Errr, yes if you say so, but you've lost me. I am but a novice!
 

MaxHeadRoom

Joined Jul 18, 2013
28,617
Sometimes in an application like this there is a need to avoid 'nuisance' switching if sudden dark time occurs or to ensure that it is not just a spurious event.
For this a time delay can be included such as a on-delay using a 555 etc.
The description in the JPG is not exactly this application but the principle is the same.
Max.
 

Attachments

AnalogKid

Joined Aug 1, 2013
10,986
My votes:
Reduce R10 to 100K to increase the hysteresis to 5%.
Increase R7 to 10K to reduce the standby current draw on the batteries by 20 mA.
Replace the right-side 2N2222 with any TO-220 n-channel power MOSFET.

ak
 

Thread Starter

Ford Prefect

Joined Jun 14, 2010
245
Dark Light Circuit-2.jpg
My votes:
Reduce R10 to 100K to increase the hysteresis to 5%.
Increase R7 to 10K to reduce the standby current draw on the batteries by 20 mA.
Replace the right-side 2N2222 with any TO-220 n-channel power MOSFET.
ak
Thank you AnalogKid - I know you always come up trumps. :)
I shall try your suggestions and let you know what happens.
Although I think I may try a 200K and maybe slightly higher but less than 1M for R10
 
Last edited:

AnalogKid

Joined Aug 1, 2013
10,986
The hysteresis calculation is complicated by the diode in parallel with R9, but as a starting point, ignore it. The Thevenin equivalent impedance at the top of R9 is R9 in parallel with the series combination of R6 and R8. For R8, use Ohm's Law to determine what its resistance is when the voltage at R8-R9 is 0.6 V. This is the voltage and resistance affected when the other end of R10 is either 0 V or open circuit. As R10 decreases, it has a greater effect on the R8-R9 node. Now you can recalculate the R8 value that holds 0.6 V at the node. This is what creates a gap between the light level to turn on and the one to turn off. The smaller R10, the bigger the gap. If the light from the LED strip might fall on R8, this gap keeps the circuit from oscillating.

ak
 
Top