Ok, Used a very simple circuit. Two Leds in parallel Reverse polarity from each other with a 4.7K limiting resistor. I know if one led opens the reverse voltage will kill the other. Experience with Chinese eflea LEDS?

Original Post

Not sure what you're asking here. What voltage did you use? Are you trying AC or DC? Why reverse polarity?

I've gotten plenty of LEDs from China. Where do you think most American companies get theirs? They work fine. Maybe you need to be more specific.

()

Reversed polarity because you can drive with AC directly with limiting resistor. I believe gunrunner had some bad experiences with LED suppliers on the bay

Removed two grain of wheats for number boards on a legacy engine. I was able to solder two leds in the grain of wheat mounting areas on the pcb. Then add the limiting resistor into the connector line going to the board.

I just hooked one up fast and took a picture. Those are test leds..

#### Attachments

Photos (1)
()

Gunnerunner, I was thinking of putting a bipolar cap in series with the power connection. Think it’s a good idea?

()

No need for a series cap, a 4.7k resistor is plenty for track voltage.  In point of fact, you can probably go much lower if you want them brighter, say 750 ohms 1/4W.  If one of the LED's goes out, you'd be replacing them anyway, so I wouldn't worry about one failing.  As Consolidated Leo says, where do you think everyone gets their LED's?

()

If you have 2 LEDs in reversed in parallel, you actually can use just a capacitor instead of a resistor. You'd be surprised how many consumer products have the current limiting for LEDs implemented this way.  For example, a 750 ohm resistor (as mentioned above in GRJ's post) is "equivalent" to a 3.5uF capacitor.   If the voltage was 120 rms and the LEDs need 15 mA, then the capacitance value would be about .24uF. If the product has only one LED, a Si diode placed backwards across the LED can be used.

()

Since the resistor is a lot cheaper than a bi-polar 3.5uf capacitor, what would be the point?

()

The capacitor may alleviate voltage transients?

()

If you have 2 LEDs in reversed in parallel, you actually can use just a capacitor instead of a resistor. You'd be surprised how many consumer products have the current limiting for LEDs implemented this way.  For example, a 750 ohm resistor (as mentioned above in GRJ's post) is "equivalent" to a 3.5uF capacitor.   If the voltage was 120 rms and the LEDs need 15 mA, then the capacitance value would be about .24uF. If the product has only one LED, a Si diode placed backwards across the LED can be used.

.24 microfarid? Yep! That’s right - the rate the cap for the peak voltage plus? So, A 50 volt will do for the Lionel ac side.

Now, How will switch sine wave affect the whole ball of wax

()

Use the resistor, it's a tried and true method and has been used for tons of installations.

()

The ,24uF is for 120V rms - For track voltages something like 3.5 uF is necessary.

My back of the envelope calculation was as follows. To get the peak voltage from RMS multiply by 1.414. Subtract the voltage drop of the LEDs (about 2 volts) from this number. Assume that the desired current is no more than 15mA, Dividing these two gives the magnitude of the Capacitive reactance, Xc, in ohms.  The value of the capacitor is approximately 1/(377 * Xc). The voltage rating on the capacitor should be at last the peak voltage.

A quick web search shows nonpolar 100V  4uF capacitors sell for about a quarter each, less in quantities.

I agree that a switched sine wave will act slightly differently, but I doubt if anyone's eyes could really tell the difference.

()

My question is why bother with the cap when the resistor is dirt simple and also dirt cheap?  What are you gaining with the capacitor?

()

John,

Just really posting a question/ A the availability of AC LED's. Which have two led's back to back with the capacitor integral to the package. I believe I lost the LEDS on my test lighting board due to votlae spikes from derailment. I believe the series cap would help? Although, the frequency and the duration of the spike may not be subdued by the capacitor?

()
shawn posted:

John,

Just really posting a question/ A the availability of AC LED's. Which have two led's back to back with the capacitor integral to the package. I believe I lost the LEDS on my test lighting board due to votlae spikes from derailment. I believe the series cap would help? Although, the frequency and the duration of the spike may not be subdued by the capacitor?

I guess I could put  a transient protector across the ac input to the LEDS?

()
Consolidated Leo posted:

Not sure what you're asking here. What voltage did you use? Are you trying AC or DC? Why reverse polarity?

I've gotten plenty of LEDs from China. Where do you think most American companies get theirs? They work fine. Maybe you need to be more specific.

Aren't youy glad I asked this question...better then going to school with some of these responses

()
shawn posted:
Just really posting a question/ A the availability of AC LED's. Which have two led's back to back with the capacitor integral to the package. I believe I lost the LEDS on my test lighting board due to votlae spikes from derailment. I believe the series cap would help? Although, the frequency and the duration of the spike may not be subdued by the capacitor?

I can't imagine the cap being better than the resistor, and I don't see any reason to use a bulkier and more expensive component.  The one penny (or less) resistor is just fine for me.  In point of fact, the capacitor will actually pass a spike better than the resistor, so if anything, it would be worse!  The capacitive reactance is being calculated on 60hz, the spike is a much faster rise time, hence it will blow right through the capacitor at full amplitude.  The resistor, OTOH, always has the same resistance, no matter what the frequency.  New and different isn't always better!

()
gunrunnerjohn posted:
shawn posted:
Just really posting a question/ A the availability of AC LED's. Which have two led's back to back with the capacitor integral to the package. I believe I lost the LEDS on my test lighting board due to votlae spikes from derailment. I believe the series cap would help? Although, the frequency and the duration of the spike may not be subdued by the capacitor?

I can't imagine the cap being better than the resistor, and I don't see any reason to use a bulkier and more expensive component.  The one penny (or less) resistor is just fine for me.  In point of fact, the capacitor will actually pass a spike better than the resistor, so if anything, it would be worse!  The capacitive reactance is being calculated on 60hz, the spike is a much faster rise time, hence it will blow right through the capacitor at full amplitude.  The resistor, OTOH, always has the same resistance, no matter what the frequency.  New and different isn't always better!

Yep! That’s why I’m asking! Now, how to protect the lad from large transient spikes without the addition of regulators and capacitors

()

Lol, I’m getting old! I checked the tracing of both lands for the old grain of wheats back to the hot and chassis goring if the pc board. I checked this a couple of times.

I knew to reverse the polarity of the leds in relation to each other. So, guess what!  I had both leds in the same polarity.. not good for reverse voltage  wow,

()