How can I enhance RLD circuits with an LED for period doubling?

[rmiller70015]

First a little background. In my lab physics class we performed a lab to determine the Feigenbaum constant. I decided to do my research project on period doubling in non-linear circuits. I started thinking about the RLD circuit and how I could make it more interesting by adding different non-linear components instead of a diode. I settled on a LED (because my department has a lot of monochromatic LEDs).

So I am wondering how similar are LED's to regular old biased diodes. Do they have the same type of pn-junction and can I model them the same as I would for a diode?

I was also wondering if anyone had a better way of determining at what voltage the bifurcations occur. Currently I have an old HP analog oscilloscope and a Tektroniks oscilloscope hooked up to it. The analog gives pretty pictures and a not so good estimate of the voltage, but the Tektronics oscilloscope gives a FFT of the signal which seemed to make finding the bifurcations easier (the number of frequencies used doubled and I believe that this is proportional to the number of periods.)

Finally, can someone point me towards some journal articles where people have investigated different aspects of this type of circuit (outside of the articles published by Feigenbaum, which I already have). The source or cost isn't really important my university has access to most journals and those it doesn't can usually be obtained through other libraries in the area.

Any help or advice you can give about this type of circuit would greatly be appreciated, I'm still in the literature review phase of the research and don't have anything concrete to work off of just yet.

 

[Baluncore]

An LED is very similar to a PN junction diode, but it has a couple of inconvenient differences.
1. The voltage needed to forward bias the LED is greater than the simple diode. To produce light, it needs to have a sufficient potential drop or energy step for the electron. E = h · v.
From which comes V_fwd = 1239.8 / λ_nm.

2. Reverse bias by a few volts of an LED will easily damage it. So you must put a diode in series with an LED to protect it from reverse current. A parallel LED also works to protect an LED, but may not work in your RLD circuit.

This means the RLD circuit will need to be biassed differently to a simple PN junction diode. The transition between conduction on and off will be of very similar character. All that will be achieved by using an LED is to move the switching voltage further away from zero.

 

[rbelli1]

Reverse bias by a few volts of an LED will easily damage it.

A few volts is usually >5V. They will also usually withstand much higher voltages for a while before being rendered inoperable.

Is your testing below 5V? Is it for a sufficiently short time if above 5V? I have seen LED's used in reverse bias in 9V circuits. I'm not saying I would design a permanent device with that condition. You should be able to find one that will survive at least that much long enough for your test.

An alternative is to find some very high voltage diodes.
http://www.digikey.com/product-detail/en/sanken/RP-1HV1/RP-1HV1CT-ND/4572586

Forward voltage even at 10mA it is about 4V. It quickly ramps up to 7 volts with extra current and frequency.

BoB

 

[rmiller70015]

A few volts is usually >5V. They will also usually withstand much higher voltages for a while before being rendered inoperable.

Is your testing below 5V? Is it for a sufficiently short time if above 5V? I have seen LED's used in reverse bias in 9V circuits. I'm not saying I would design a permanent device with that condition. You should be able to find one that will survive at least that much long enough for your test.

An alternative is to find some very high voltage diodes.
http://www.digikey.com/product-detail/en/sanken/RP-1HV1/RP-1HV1CT-ND/4572586

Forward voltage even at 10mA it is about 4V. It quickly ramps up to 7 volts with extra current and frequency.

BoB

I think the voltage will be between 1 V and 6 V. I am not sure how long I will need to run it for, so I should probably plan on ruining a couple while trying to get everything set up.

As a follow up question, is there an equation that describes the wavelength or frequency emitted by the LED or would I be able to find this information in an electronics textbook?

 

[rbelli1]

I should probably plan on ruining a couple while trying to get everything set up.

The good thing is that LED's can be had for a few cents each.

is there an equation that describes the wavelength or frequency emitted by the LED

Higher frequencies generally require higher voltages but temperature and composition also play a role. The farthest infrared often require less than a volt and the very far ultra violet can be 7 volts or higher. It's complicated.

Try this. Take a 20mA green LED and give it 100 to 200 mA and watch the color shift as it get near thermal destruction. You may want to do this outside as the smoke you let out will be rather unpleasant smelling. This may also work for other colors.

BoB

 

[Baluncore]

As a follow up question, is there an equation that describes the wavelength or frequency emitted by the LED or would I be able to find this information in an electronics textbook?

From which comes Vfwd = 1239.8 / λnm

To a first approximation, a LED appears as a voltage source in series with a junction diode. The colour or wavelength of the light is related to the voltage by the equation I gave earlier. Shorter wavelength light at the blue end of the spectrum has higher photon energy and so needs a greater voltage step in the diode.
The chemistry of the semiconductor alloy decides the voltage difference between two energy levels and hence the colour of the photons.

 

[Baluncore]

I did some LTspice simulations. The RL-D chaos circuit has an inductor in series with the diode. When there is sufficient AC voltage to turn on a LED, there is a high reverse current flow, with a negative voltage spike that exceeds –8V, (limited by the LED reverse breakdown voltage). That happens once each cycle, which is over 250kHz. Even with a standard diode there is a –3V transient on each cycle, but then there is no reverse current.