Transistor getting very hot - suggestions please for a alernative

Thread Starter

Ford Prefect

Joined Jun 14, 2010
245
Since R6 and R7 are only used as a reference voltage, you can set their values much higher to reduce drain on the battery. IE 600K each or more for 6 volts
Yes, true. But won't this alter the hysteresis to a wider amount?
Being that the LDR will be at about 60k when I want the striplight to turn on/off, I have chosen a 50k preset to be adjusted somewhere in the middle, therefore the voltage at the inverting input will be about 8 volts and the preset can adjust the voltage about 1 volt either way.
See also Dodgydave #30 post above
R6 and R7 values were chosen so that the voltage at the non-inverting input would be also be about 8 volts
 
Last edited:

cmartinez

Joined Jan 17, 2007
8,218
... never ever leave input pins unconnected
Yes... I had to learn that the hard way... by watching how my circuits did spooky, strange and unpredictable things, until someone with more experience than me (who was not hard to find, btw) explained it to me...
 

Thread Starter

Ford Prefect

Joined Jun 14, 2010
245
Built this circuit yesterday and tested with a 5k1 resistor at R9 based on the #19 post above
This appeared not quite correct and the circuit did not work very well because as the light levels decreased to almost darkness, the striplight was not coming on but was just faintly glowing.
It seems the R9 resistor needs reducing to a value of between 1k and 4k3 so I decided to try a 5k preset (highlighted green) so that the resistance could be adjusted to the desired light level.
This seems to work well and in darkness the striplight glows bright.
Dark Light Circuit-v3.jpg
 

AnalogKid

Joined Aug 1, 2013
10,986
R10 still might be too large. As the ambient light level slowly decreases, does the strip light snap on or gradually increase in brightness?

ak
 

Thread Starter

Ford Prefect

Joined Jun 14, 2010
245
R10 still might be too large. As the ambient light level slowly decreases, does the strip light snap on or gradually increase in brightness?

ak
I want to test the circuit this evening and watch the strip light as the light levels slowly decrease.
If the strip light flickers as opposed to snap on, I shall decrease R10 tomorrow.
I am also interested to see what happens. I will post a message and let you know.
 

ScottWang

Joined Aug 23, 2012
7,397
As the below circuit when the led is turn off then there is no current flows through Ic of Q1, when the led is turn on then the Ic current of Q1 around 0.25mA.

AutoLight_Nmosfet_ScottWang.gif
 

AlbertHall

Joined Jun 4, 2014
12,343
How much voltage would be dropped across the LED strip just from the current through the hysteresis resistor? Does anyone know? It may be sufficient to mess up the hysteresis. A resistor connected across the LEDs would eliminate the problem. Perhaps 4.7k would be good. What does anybody think?
 

AnalogKid

Joined Aug 1, 2013
10,986
While hysteresis usually involves a resistor both sourcing and sinking current in a node, that is not a requirement. A resistor that is open circuit in one case and moving current in the other still has the hysteretic effect, just less of it compared to the same circuit with a source/sink drive. This comes up when moving a comparator circuit between an LM38 and LM393 for whatever reason. The 393's open collector output still affects the non-inverting input in two ways, just not the same ways.

ak
 

Robin Mitchell

Joined Oct 25, 2009
819
One method for solving your problem would be to create a VCO with a triangular waveform. Then, feed the triangle wave into a comparator + input and the LDR potential divider output into the - input. This creates a PWM signal that depends on the light level and this PWM drives a transistor.


 

cmartinez

Joined Jan 17, 2007
8,218
One method for solving your problem would be to create a VCO with a triangular waveform. Then, feed the triangle wave into a comparator + input and the LDR potential divider output into the - input. This creates a PWM signal that depends on the light level and this PWM drives a transistor.


I've built and tested that circuit before... and it works beautifully because it's able to produce the entire 0 to 100% PWM spectrum. Unlike most circuits based on a 555
 

Thread Starter

Ford Prefect

Joined Jun 14, 2010
245
As the below circuit when the led is turn off then there is no current flows through Ic of Q1, when the led is turn on then the Ic current of Q1 around 0.25mA.

View attachment 124829
This appears to be the circuit I am looking for...thank you for posting.
I want the strip light to illuminate (glow bright) when the LDR (CDS1) using a GL5528 has the value of between about 60k and 100k ohms.
Can anyone work out if the other component values will do this or do any of the values/components need changing?
 

ScottWang

Joined Aug 23, 2012
7,397
This appears to be the circuit I am looking for...thank you for posting.
I want the strip light to illuminate (glow bright) when the LDR (CDS1) using a GL5528 has the value of between about 60k and 100k ohms.
Can anyone work out if the other component values will do this or do any of the values/components need changing?
I already tested the circuit and it works fine, but your condition of application maybe different, if you want to limited the values of resistor between about 60k and 100k ohms, you could adjust the R2 and VR1 to suit what you want, if adjust those two values still can't match what you need then you can also adjust the values of R3, when you do the test and see how's going.
 

Thread Starter

Ford Prefect

Joined Jun 14, 2010
245
I already tested the circuit and it works fine, but your condition of application maybe different, if you want to limited the values of resistor between about 60k and 100k ohms, you could adjust the R2 and VR1 to suit what you want, if adjust those two values still can't match what you need then you can also adjust the values of R3, when you do the test and see how's going.
Thanks for the reply. I shall give it a go. :)
 

Thread Starter

Ford Prefect

Joined Jun 14, 2010
245
R10 still might be too large. As the ambient light level slowly decreases, does the strip light snap on or gradually increase in brightness?

ak
I want to test the circuit this evening and watch the strip light as the light levels slowly decrease.
If the strip light flickers as opposed to snap on, I shall decrease R10 tomorrow.
I am also interested to see what happens. I will post a message and let you know.
Ref. post #43 above
I tested the circuit below and it seems that the value of 100k for R10 appears to be quite suitable. I have also replaced R6 with a 10k resistor and R9 with a 2.2k resistor.
The striplight appears to snap on at a certain light level (just before darkness/twilight).
Dark Light Circuit-3.jpg
However, the striplight (some of the LED's) continually glows very dimly in daylight and I have noticed that the voltage level at the junction of R10, Q2 and the striplight with respect to ground is 7.32v (at the red arrow).
I need some way to increase the voltage when the striplight is off (LDR in daylight) so that the striplight does not glow dimly.
One idea I thought was to add a low value resistor in parallel to the striplight.
I haven't tried it yet but would this work? What are your thoughts?
 

cmartinez

Joined Jan 17, 2007
8,218
Ref. post #43 above
I tested the circuit below and it seems that the value of 100k for R10 appears to be quite suitable. I have also replaced R6 with a 10k resistor and R9 with a 2.2k resistor.
The striplight appears to snap on at a certain light level (just before darkness/twilight).
View attachment 124975
However, the striplight (some of the LED's) continually glows very dimly in daylight and I have noticed that the voltage level at the junction of R10, Q2 and the striplight with respect to ground is 7.32v (at the red arrow).
I need some way to increase the voltage when the striplight is off (LDR in daylight) so that the striplight does not glow dimly.
One idea I thought was to add a low value resistor in parallel to the striplight.
I haven't tried it yet but would this work? What are your thoughts?
The way the circuit is drawn at this moment, allows for current to flow through the LED strip, through R0 and then through R9, even when it's supposed to be completely off. This also applies a small voltage to Q1's base, turning it slightly on. One way to avoid that would be to move the mosfet so that it switched the LED strip from its high side, instead of its low side as it is now. To do that more easily would require changing the mosfet to a p-fet, and make changes in the way that Q1 is triggering the mosfet's gate.
 
Top