Drivers from Discrete Components?

[huskyrunnr]

Can anyone point me to circuits of efficient switching drivers made of discrete components (rather than driver chips) that actually work for LED's? I have found and made a handful of 555 based circuits that work fine for rail guns and nixie tubes, but loads comparable to white LED's load these circuits terribly.

I just ran across this one and may try it. http://www.edn.com/contents/images/111402di.pdf Thing is, I have a million discrete components I got from an EE type who recently died, and I don't want to buy a chip if I don't have to.

 

[HarryN]

If you think about it, a driver circuit is just an RLC circuit that dumps the inductor through the LED. Any basic boost or buck RLC circuit will work.

If your millions of parts includes a resistor, then just size the resistor to the application. 2 ohms + a Li ion cell + Lux III or K2 = working driver circuit.

How fancy are you trying to get ?

 

[huskyrunnr]

Any 555-based boost or buck I've built gets loaded by an LED. Zout is too large. I have run across some discussion of this circuit loading at the link below, and the suggestions were to either use a self-boosted 555 or forgo the 555 for a real driver chip.

I am just wondering if anyone has pulled it off with a 555 or other oscillator, and if so, would they be willing to share the circuit?

http://groups.google.im/group/aus.e...ad/beadbb2f38b3d305/d06ba2b07a8cd4f5?lnk=raot

 

[Mr Happy]

Remember the 555 is a timer, not a driver. When significant loads are involved you must buffer the output with a transistor.

Efficient voltage boosting circuits use an FET to switch the current. Read up on the basic configuration and design of boost circuits, then try using the 555 to switch the FET on and off and let the main current go through the FET instead of the 555.

This is of course mainly of interest as a learning experience. I think you will always do better with a purpose designed chip than with home brew. For one thing, you will never make a discrete circuit as compact as the chip circuit.

 

[huskyrunnr]

Yes, I was using a power FET for the switch, a 555 as the oscillator driving the FET. I don't have a Schottky and didn't want to get into synchronous switching w/o a scope to see what's going on. What I've made works fine for it's intended load.

I think I'll chalk it up to an educational experience and move on to chips for the purpose.

Thanks to you both for the comments.

 

[HarryN]

I am still very much a beginner with electronics, but I have learned the hard way that without an O scope, it is really hard - nearly impossible - to know what is really happening in a circuit.

Some of the problems are that unlike water pipes and tanks which make perfect sense to a chem eng like me, electrical components have all sorts of secondary behaviors which just mucks stuff up. Also, a bad part (more common that we like to imagine) is hard to find.

In a water pipe, you can tell when things are plugged, but with an electrical circuit, you just can't see what the heck is going on. That is what the O Scope is for.

I guess in summary, if you don't have a o scope, I just cannot imagine trying to build a driver with ICs or discreets that is any better than a resistor or even close to a $ 20 driver from taskled.com. (cpf georges80)

 

[Mr Happy]

unlike water pipes and tanks which make perfect sense to a chem eng like me

Yo dude! Water pipes? You're only a plumber!

Now if anyone is curious about how boost circuits work, you can see a hydraulic analog using water and pipes here: http://www.youtube.com/watch?v=4mn58NGhQjA

The pump in the video is raising water from a height of 6 ft to 60 ft using just gravity with no external power. This is equivalent to raising a DC supply from 6 V to 60 V. The method of operation of the hydraulic ram pump that you see is almost exactly the same as a DC/DC switched boost converter. The valve that stops and starts the flow is the switch, and a long feed pipe takes the place of the inductor. The fat white upright pipe is equivalent to the capacitor and temporarily stores the high pressure water before feeding it into the delivery pipe.

 

[HarryN]

Mr. Happy - yes, I just used water for simplicity. Mostly I deal with UHP metalorganics.

The most difficult thing for someone trained in Chem Eng like me is to understand an inductor. The idea that it is really storing "energy" - not really volts or amp - and can magically make it into what ever voltage you need to push the current along took me years to accept. :shrug:

I chalked this up to the idea that while there is an analog to resistors, capacitors, voltage, and current with water, I could not think of an equivalent to an inductor. This seems like the physical equivalent of converting water to pressure or pressure to water.

It is interesting that you make the analogy of the long pipes and an inductor. I had always imagined that pump setup to be more like a step up charge capacitor arrangement with no inductance - I will have to give that one some more thought. Then again - maybe I shouldn't.

 

[Mr Happy]

The really important thing about an inductor is not that it can store energy, but that current flowing through it has inertia. It is really hard to start current flowing in an inductor, and once the current has started it is really hard to stop it again.

That is how inductors are used to make sparks. You get a big coil and start current flowing through it. When the current is flowing nicely, you disconnect one end of the coil. Now the current flowing through the coil really doesn't want to stop flowing, but you have taken the wire away and it has nowhere to go. So all the electrons pile up in a big heap at the point where the circuit is broken and then leap off into the air making a spark.

This is now your analogy with water and pipes and the hydraulic ram pump. The interrupted current in an inductor behaves like water hammer in a long pipe. If you try to stop the flow suddenly by closing a valve the water really doesn't want to stop and all the stored up energy rattles the pipe (or forces itself uphill in the case of the pump).

(By the way, hi! from another Chem Eng :wave

 

[huskyrunnr]

Well, given my stubbornness, I was able to tweak the first circuit I linked to push 6W through a 1.5 Ohm load. Maybe I'll try to blow up some LED's with it. I've read you can use rev. biased Zeners to simulate power LED's but I don't have any with the right breakdown voltage, so I'll have to live dangerously! I doubt it'll drive the LED's though.