Is there any way to fix transistor voltage drop?

[supak111]

Hey guys I have simple light activated circuit that when its ON is usually only ON for less then 1/4 of second. I usually power it from 12v or 24v. My problem is the voltage drop on the 0V side. Here is the circuit (mine doesn't have the LED):

Is there any way to fix the voltage drop? Maybe by adding a capacitor someplace or even a small solar cell? I just need to add about 1v to the circuit so that the whole circuit acts like a short for 1/4 of a second when its ON.

Total newbe at circuit design and I sure didn't go to school for this so any suggestions or help are much appreciated. Thanks

 

[KrisBlueNZ]

So the problem is that when light is present, the transistor turns on, but there's still some voltage between its collector and emitter, so the voltage across the load (in the collector circuit) isn't as much as it should be?

An N-channel MOSFET will give you a very low ON voltage, but needs to be driven from a signal with a wider voltage swing than you need for a transistor. You could try replacing the transistor with a BS170 (gate = base; drain = collector; source = emitter) and adjusting the potentiometer so it switches at the right brightness level. That may work better than the transistor.

What load are you driving in the collector circuit? How many milliamps does the transistor (or MOSFET) need to pass? If it's more than around 100 mA then you may need to use a bigger MOSFET, and there may be an issue with heat dissipation if the light level remains around the switching threshold, because the MOSFET (or the transistor) will be only partly conducting, which causes higher power dissipation than fully ON and fully OFF. In that case you should add a Schmitt trigger to provide a clean ON/OFF control signal to the device that drives the load.

 

[supak111]

Power coming into the transistor is plentiful. Load that the transistor is powering outside the circuit is usually less then 100ma sometimes as little 30ma. And this circuit is only being used for 1/4 of a second once or twice and then not used for hour so it has time to cool i think, well at least heat hasn't been an issue so far. Just need to figure out how to make this circuit act like a relay (full short) without making it much more complicated

 

[(*steve*)]

All transistors will have some sort of voltage drop, however this circuit will not turn the transistor on and off cleanly. You need a schmitt trigger for that. I suspect that this may also reduce the voltage across the transistor to a level that is acceptable to you.

 

[(*steve*)]

A circuit similar to the one on this page should be inserted before the base of the transistor you currently have.

 

[KrisBlueNZ]

Power coming into the transistor is plentiful.

No it isn't. Not "power" coming into the base, anyway.

Load that the transistor is powering outside the circuit is usually less then 100ma sometimes as little 30ma.

OK, that's fine.

And this circuit is only being used for 1/4 of a second once or twice and then not used for hour so it has time to cool i think, well at least heat hasn't been an issue so far.

It won't take long for a little MOSFET to lose its smoke.

It IS important for the MOSFET to switch cleanly, as Steve described.

Just need to figure out how to make this circuit act like a relay (full short) without making it much more complicated

None of us is going to advise you how to make a circuit that we know won't work reliably!

 

[hevans1944]

Just need to figure out how to make this circuit act like a relay (full short) without making it much more complicated

So why not, instead of a resistor in the collector of the transistor, use a relay coil, with a diode across the coil to suppress voltage spikes when the relay de-energizes? When closed, the normally-open contacts on an energized relay are pretty close to a dead short.

I would also add the Schmitt trigger circuit suggested by @(*steve*) because LDRs are notoriously slow in their response to changing light levels. The Schmitt trigger guarantees the circuit energizes at some light intensity and then de-energizes at a lower intensity because of hysteresis built into the Schmitt trigger.

 

[supak111]

Ok got ya. So then Schmitt Trigger to the base so that the transistor has a clean on/off. Do I still use my bc337 transistor or do I go with the mosfat bs170 you suggested above?

 

[Harald Kapp]

The Schmitt triggger in the base control circuit will not solve your "problem". As noted by Steve in post #4 any transistor will have a voltage drop, be it from collector to emitter for a BJT or from drain to source for a MOS).
Even a relay will have a small voltage drop due to the contact resistance.

 

[(*steve*)]

I think the first issue is to determine if the problem is a small voltage drop across a saturated transistor or a larger voltage across a transistor that is only partially turned on.

I suspect it is the latter. If so, the Schmitt trigger will be of great assistance.

 

[supak111]

Yea I'm not sure exactly what is causing the problem in my circuit to be honest as this circuit is turning ON many different devices. Works on most, but not on some.

Here is a little background. My circuit (call it LIGHTBOX) is being used and attached to other electronic DEVICES, so it basically acts as a light relay and turns stuff ON/OFF on the device you put it on. My circuit above works on most devices but NOT on some. What I have noticed is that on the DEVICES that it does not work on, if you just short circuit the V+ to V0 on the device itself the device does the function that my LIGHTBOX should be doing in the first place. After talking to you guys I now suspect that the problem might be that some Devices I'm trying to turn ON aren't doing so because the signal from my circuit is slow on and might not even get fully turned on. I think schmitt trigger might be the fix for this so I'll try that first.

Is there a small/cheap schmitt trigger ic that I can easily integrate into my circuit?

 

[Harald Kapp]

You can use e.g. LM239 or NE555 as a Schmitt-Trigger circuit. The datasheets show suitable application circuits. Or Google Schmitt-Trigger.

 

[supak111]

One other quick question, is there anything else that can just give a clean ON signal for the transistor? I don't care if its not clean OFF, or if my circuit only has an upper threshold and no lower one. I looked at varistors but I don't think that gives a clean ON signal right?

I also fail to maintain the light that triggers this circuit is a clean ON/OFF LED light. Wait could I use an LED instead of my photoresistor? And if I did would I have clean ON OFF then?

 

[(*steve*)]

Is the "LIGHTBOX" part of this circuit, or are you trying to detect it across some space (i.e between 2 different devices)?

If the "LIGHTBOX" has a LED *AND* it's part of the same circuit, then you might be able to use the signal which operates the LED to control your circuit.

BUT, we need to know more about the "LIGHTBOX" and the various "DEVICES". Even knowing what works and what doesn't may help us determine the problem and therefore a solution.

 

[supak111]

Sorry by me trying to make this less confusing I may have made it more confusing.

My circuit is the LIGHTBOX, and the device my circuit/lightbox attaches to is a same circuit because 2 wires that come out of my circuit/lightbox (12v and 0V wires in the diagram above) go to this device I'm triggering. DEVICE is the thing I'm triggering with my circuit/lightbox.

So it goes like this. LED light lights up (forget where the LED light is coming from, I can't tap into that circuit for sure), that light goes into my circuit aka LIGHTBOX, which then needs to give power to the end product aka DEVICE vie the 0V wire which now should now have 12V because my transistor is ON.

PS I will try the schmitt trigger, I hope it fixes my problem. Removing my whole circuit and just making a short from 12v to 0v does the job, but I want that job done by light. Relay works great too but I want it done without having an external power source.

 

[(*steve*)]

OK, so you're trying to turn on a device when an external LED turns on?

You really need a Schmitt trigger so that the device turns on and off cleanly, but also fully -- I'm glad you're going that way.

There are a lot of resistors in the Schmitt trigger, and you have to choose them carefully so that it turns on and off at the appropriate voltage levels. Those voltage levels are dependent on the LDR you're using, the brightness off the LED and the resistor used to bias the LED, and any amount of ambient light leakage you might have.

So the first thing you need to do is to separate the LDR from the transistor so all you have is this part:



Then (with the variable resistor in the middle of its travel) you need to connect a multimeter between 0V and the output and measure:

1. The voltage in darkness with the LED off
   
2. The voltage in bright light with the LED off
3. The voltage in darkness with the LED on
4. the voltage in bright light with the LED on
5. The voltage at V_IN (you may not need to measure this)

If all goes well, voltages 1 and 3 will be similar, voltages 2 and 4 will be similar, and both of these will be significantly different from each other.

Please tell us these voltages.

Then you might want to consider whether you wish to use the two transistor circuit shown to you, or one like this using an op-amp:


This uses an op-amp (or a comparator) in place of the two transistors. Whilst it looks simpler, it does introduce some complexities, especially if you have a low supply voltage. But the advantage is that there's a nice on-line calculator to determine the resistor values.

The calculations will need to be done manually for the circuit I gave earlier.

Let's get everything planned before you go and buy stuff.

Also you need to tell us the maximum current drawn by your "DEVICES". This is VERY important. Also if any of them are loads like lamps, motors, relays,... we need to know that too. If you can be explicit in telling us what those loads actually are, we can be far more confident in knowing that we are giving you good advice.

 

[supak111]

Do you mean #1&2 should be similar and 3&4 right?

Also I found this guy for cheap: Light switch module Couldn't I use Digital Output of this to switch ON my transistor? They say it works on 3-5v I don't know why when it uses a LM393 which works on up to 34v, maybe because of the 2 LEDs on it? I would remove those anyways

Here is the circuit of it:

 

[supak111]

I was just thinking, I'm not sure if this has been done or if its possible??

Can you use a zener diode and a cap on the base to turn a transistor all the way on really fast? Seem to me that this is a pretty simple bjt drive.

 

[(*steve*)]

No, 1 and 3 should be similar to each other, and 2 and 4 should be similar to each other. If they are not then ambient light is interfering. If that interference is significant enough it will swamp the changes caused by the LED.

 

[(*steve*)]

You could use that device.

1. It has no positive feedback, so it may not switch cleanly, although it will switch on and off quite hard.
2. The low output may be insufficient to turn off the transistor (and it is low when active). Can you use a PNP transistor between the +ve supply and the load instead of an NPN transistor in the ground lead of the load?
3. It is marked as 3 to 5 V because it is designed for use with microcontrollers that typically have power supplies between these voltages.
4. Yes, if you were to retain the LEDs you would need to change the series resistors to provide a lower current for the LEDs at higher voltages.