I'm building this circuit to work with a 12v 5w panel, ten 1.2v 4800mAh NiMH batteries, 200mA of LED light strip.
http://www.reuk.co.uk/Use-Solar-Panel-As-Darkness-Detector.htm

I've used a 1N4007, 10k half watt, and a BC328 for the PNP transistors. However the circuit isn't working. The LEDs aren't turning on when the panel is covered.

Is it the transistor I've selected, the resistor ohms or what's the best way to trouble shoot the circuit?

 

Measure voltage on the base of the transistor. Should be about 0.6V below battery voltage to turn on LED.

 

try putting a 10k resistor from the transistor base to batt neg, it maybe that the panel is still giving out a voltage to keep the transistor off.

 

To saturate BC328, base current should be around 20 mA, so change 10k to 620 ohms.
You need about 15 V to charge battery, so if open ckt. V of panel is only 17 V ?, then will charge weakley on best of solar conditions.

 

200mA of LED light strip

How many LEDs? Presumably white LEDs? How are they series/parallel connected?
The 10k resistor will allow a maximum of only ~1mA of base current. To drive a 200mA collector load the base current should be ~20mA to ensure the BC328 is turned on fully. If only partly turned on it will be dissipating a lot of power as heat.

 

The link is a poorly conceived circuit.

Most panels have the series diode D1 sealed inside the panel housing, and a connection to the anode is not possible without opening the weatherproof enclosure. This circuit solves that problem by adding a diode D2. Note the way that the base current for Q1 is supplied.

The simulation shows an entire day (1s=1m). As long as the panel is delivering more than about 1mA, then Q1 is off, and all the panel output goes to the battery through D2.

 

Hi, thanks for the help. I put a potentiometer on and found at 400ohm it lights 100% without the panel charging. The transistor does get hot to touch, my infrared thermometer days it's 50 degrees, but feels like more.

The lights are RGB LEDs running through a PWM to tune to a particular colour. I've set them at the brightness I want and it's drawing 200mA's including the PWM.

 

The panel puts out 20v open circuit. It's been raining for the last 4 days so I haven't been able to test it's charge voltage or current. It's rated at 5w. It also does not have a diode inbuilt.

So it's working but I'm worried about heat of the transistor.

 

Sorry, correction. Not all of the LEDs on the strip were at 100% for some reason. I tried again and the potentiometer at 100ohm lit them all with panel covered. But this is 120mA going to the base. Is that excessive? Is that current wasted into heat or does it go to the LEDs? Sorry, I don't have much of an idea how transistors work.

 

Are you using the post #1 circuit, or Mike's post #6 circuit?

 

Post #1 circuit.

So that #6 circuit joins base to panel side of the diode without a resistor and then adds a 47k resistor between positive and negative of the panel. Is the diode on the panel a must have? My panel doesn't have one. Should I add one?

 

So I built the circuit in post #6. But the LEDs are always on.

In the picture:
Red and black is panel
White is battery
Grey is LEDs. This is using 1N4007 and BC328. I assume they will work for the circuit.

 

Please ignore my last posts... My stupidity lead me to run the battery in the wrong polarity. I thought I'd wired the black strip as negative where it was actually positive.

The problem I have now is there is only 8v at the collector (on the LEDs)

 

How much current do the LEDs draw when the collector voltage is 8V?

 

In the simulation of post #6, I assumed, and the plot showed, a LED current of ~30mA, like a single LED. If your LED string requires much more, than the circuit must be adjusted. How much current does your LED string require?

 

I'll have to check today how much they are drawing at the 8v. But they need to draw between 150 to 300mA at 12. The PWM means they are adjustable.

 

Here is revised circuit which will switch ~300mA @12V. Note that I had to add a second PNP to get the required current gain. D2 is now a 1A Si diode. The circuit switches when the panel gets dark or at first light.



A simpler circuit if you can procure a PFET:

 

Thank you, I'll give it a go.

When I search for pfet Google corrects to mosfet. What is the difference or is there another name for this 'pfet'?

 

... is there another name for this 'pfet'?

Sure, there are lots of them if you can deal with surface mount. Very few through-hole parts left, though.

 

Ok so it's a P channel MOSFET? My local electrical parts place only had very low current such as 20mA or 7A + models. Will a high amperage model require more power to switch? Would the 33k resistor have to be changed to match the MOSFET?

Thank you for all this help.

As for the current draw at the 8v, it's about 50mA draw coming from the LEDs at 8.8v.