LED Circuit Design and Modern Transformers

Check the rated current usage of the led strip of lights.  The 3528 strip I have says 3 amps and the 5050 strip says 5 amps.  Both strips draw less than that when measured.  Both strips are designed to run off of a 12v dc power source.  

If your AC input is 18volt then your 16volt capacitor is way under rated and may be passing too much ripple current.   Increase the cap rating to 35 v rating and add some caps until the ripple decreases.  Your bridge may need to be heat sinked if you have substantial current going thru it.

Adding inductors would help as well.

The wall wart that came with my strip gets warm after a while.

You could always add a volt regulator ckt with cooling.

I am planning to use a computer power supply to power my strips.  It has readily available high current 12v sources.

If you're powering a full 5 meter strip, a 220uf capacitor isn't even close to enough to prevent flicker.  The capacitor is getting hot because it's seeing a LOT of ripple, like full voltage swings! 

I'd be considering a switching power supply, the above suggestion for a computer power supply is an excellent start if you have a number of them to light.

How about using one of these?

https://www.ebay.com/itm/120-2...rstClass!94062!US!-1

Hi MATTRAIN,

As AUSSTEVE and GUNRUNNERJOHN said, the 220uF capacitor is too small.

Here's a good on-line calculator that will help you figure out what you need.  You're minimum requirement would be at least 1000uF, but if you use the CW80's throttle to adjust the voltage down to 12 volts, the minimum isn't going to be even close.  That's because the CW80 uses a chopped sine wave throttle operating at 60HZ.  At 12 volts the throttle in the CW80 is cutting off power roughly 35% of the time.  That leaves a pretty big gap where the capacitor is trying to supply 5 amps to the lighting strip.

A safer and pretty economical way to go is a switching power supply.  As mentioned above, old PC power supplies are a good choice.  Another option that I use a lot on layouts are power supplies like these...

12VDC - 5amp $12.99

12VDC - 10amp $16.99

12VDC - 20 amp $21.99

Even after you add a 3-prong power cord, they're pretty cheap for the wattage, and you won't see any flicker in the LED strips.

I set up a similar circuit using an MTH Z1000 transformer which uses the same phase angle fired thyristor control as the CW 80.  I used a 5 amp bridge rectifier with A 470 UF 50 volt cap. I set the output to 12 volts with a 40 ohm load.  I let it run with the load for 10 minutes.  The bridge was warm which I would expect, but the cap was not.  The ripple voltage was .5 volts measured with a scope.  You can measure ripple voltage with a DVM set to measure A.C.  It seems that your circuit is drawing to much ripple current through the capacitor.  I would remove the load from the circuit and measure the A.C. with no load.  Also let the circuit run for 10 minutes or so no load and see if the capacitor gets warm.  The A.C. should be very low with no load.  If it is different, you may have a defective bridge rectifier.  With most of these devices, it helps to have a heat sink attached to reduce temperature.  Bridge rectifiers are available with holes for mounting a heat sink. 

mattrain posted:

Hi all,

With the wide availability and low cost of LED strip lights, I've started embracing them for lighting on and under the layout. For my first project, I've run a 16ft strip under my layout for illumination for working under the table.  Initially I was just using a 12V DC wall wart but given I have transformers already powering trains, lights, and accessories and I wanted to use my AIU to switch the lights on, I derived a simple circuit to wire it in to my accessories using a full bridge rectifier.

After hooking it up, it illuminated as expected, however I noticed quite a bit of flicker in the LEDs.  My suspicion is it is because of a squared/chopped sine wave being produced by the Lionel CW-80 transformer.  I hooked up my LED lights/rectifier to an old Lionel Type 1053, and sure enough the flicker is gone. I don't really want to switch to using the 1053 if I can help it as I like the integrated overload/breaker protection in the CW-80 (yes, I know I can install an in-line breaker or fuse, but I'd prefer to just use the integrated design of the CW-80). Rather, I'd like to try and address the flicker/chopped sine wave in my circuit design (if only for self-education in basic circuit design)

To get back to the original question, it is not unknown for the CW-80 output voltage to be a bit erratic, resulting in a random flicker. I stopped using mine for accessories for just that reason.

See if the CW-80 causes an incandescent lamp to flicker. If it does, you have your answer.

RoyBoy posted:

How about using one of these?

https://www.ebay.com/itm/120-2...rstClass!94062!US!-1

One problem.  No specifications.  It would be nice to know what the power handling capability is.

Phranzdan posted:

I set up a similar circuit using an MTH Z1000 transformer which uses the same phase angle fired thyristor control as the CW 80.  I used a 5 amp bridge rectifier with A 470 UF 50 volt cap. I set the output to 12 volts with a 40 ohm load.  I let it run with the load for 10 minutes.  The bridge was warm which I would expect, but the cap was not.  The ripple voltage was .5 volts measured with a scope.  You can measure ripple voltage with a DVM set to measure A.C.  It seems that your circuit is drawing to much ripple current through the capacitor.  I would remove the load from the circuit and measure the A.C. with no load.  Also let the circuit run for 10 minutes or so no load and see if the capacitor gets warm.  The A.C. should be very low with no load.  If it is different, you may have a defective bridge rectifier.  With most of these devices, it helps to have a heat sink attached to reduce temperature.  Bridge rectifiers are available with holes for mounting a heat sink. 

A 40 ohm load on 12V is less than 1/3 amp, a far cry from the several amps he was drawing with a full 5 meter strip.  Try repeating your experiment with a 6-8 ohm resistor and see what your results are.

2W10 bridge is rated 2 Amps.  If driving a 12V strip, that's "only" 24 Watts (= 12 V x 2 A).

My intent was to show that with a reasonable component selection, the ripple and hence the A.C. current through the cap should not cause heating.  One problem is we don't know  what the actual load on the power supply is.  If the current is in fact 5 amps then the 2 amp bridge he is using will not survive.  The bridge I used was a 5 amp and I wouldn't think of drawing 5 amps through it without a heat sink.  I looked on Amazon and found a 16 foot Led strip where the recommendation was for a 12 volt 2 amp supply.  It would help to know what is being powered up.  All of the modern transformers are using some type of Thyristor voltage control.  Depending on the setting and particularly at the lower voltage setting, when only a small part of the sine wave is being applied to the circuit, the waveshape looks similar to a spike.  The high frequency content will cause a high amount of current in the filter.  I will push the load up  and publish the details later. 

How about this: $4.49

12V DC switching power supply

Phranzdan posted:

All of the modern transformers are using some type of Thyristor voltage control.  Depending on the setting and particularly at the lower voltage setting, when only a small part of the sine wave is being applied to the circuit, the waveshape looks similar to a spike.

Au Contraire, the MTH Z-4000 uses FET outputs to create a pseudo-sine wave, and it doesn't generate a sawtooth waveform.