Capacitor Install ? ? ?

[carlso]

I need some guidance, please.

A simple install of a Digitrax  TL1, one function decoder to control on/off for caboose markers. It works great but I have some flicker, so I want to install a cap to control the flicker.

I have a 100uF, 20v smd that I want to install. I think one should be enough. What you think? It will be set in the cupola and there may be room for two. Also, do I wire it between the blue and white wire outputs from the decoder keeping polarity in mind. And should it be wired into the blue before or after the resistor?

I have looked at Digitrax website but can't seem to find answer to my questions.

Thanks in advance,
Carl

[Point353]

A simple install of a Digitrax  TL1, one function decoder to control on/off for caboose markers. It works great but I have some flicker, so I want to install a cap to control the flicker.
I have a 100uF, 20v smd that I want to install. I think one should be enough. What you think? It will be set in the cupola and there may be room for two. Also, do I wire it between the blue and white wire outputs from the decoder keeping polarity in mind. And should it be wired into the blue before or after the resistor?

My suggestion would be to connect the capacitor to the blue wire before (as in on the decoder output side of) the resistor.

Also, if you're going to use a capacitor with a value on the order of 100μF, I'd also install a second resistor - with a value of about 50-100Ω - between the blue wire decoder output and the capacitor to limit the current when the decoder output is first enabled.

However, such a large value of capacitor may not be necessary - you're not trying to keep a motor running.
I'd start with a capacitor value in the range of 0.1-1μF and see if that is sufficient to eliminate the flicker.

[carlso]

Point353,

Thanks for the reply, I appreciate the help. I guess I was on the right track in thinking of installing the cap between the decoder and the resistor, I know the 100uF is overkill but I have 10 of them on hand that I may never use, so save a few cents.

With the cap going on the output side of the decoder it will not charge up when the caboose is set on the track until I turn on the output and turn on the LED's. Is that correct? And would it only charge up to the voltage put out by the decoders one function? ?

Thanks for the help,
Carl

[Point353]

With the cap going on the output side of the decoder it will not charge up when the caboose is set on the track until I turn on the output and turn on the LED's. Is that correct? And would it only charge up to the voltage put out by the decoders one function? ?

Correct on both counts.
No voltage is applied to the capacitor until and unless the function output is turned on.
If the capacitor is only connected to the output of the decoder then it will only charge up to the voltage put out by the decoder.

[carlso]

Thank you !

[peteski]

Carl,
As mentioned in above post, a capacitor (with a series-connected ow resistance  current limiting resistor) hoked up between the function output and blue positive wire should keep the LED lit for a very brief time, but it will not keep the decoder circuitry (the microcrontroller) running during the  power dropout.  So, depending on the length of the power dropout, it can cause the LED to flicker (when the microcontroller reinitializes after the power is restored).  Also red LEDs are not as efficient as white LEDs, so a 100uF cap might not power the LED long enough to prevent flickering. But I suppose it won't hurt to try.

I would recommend adding a true keep Alive capacitor to that decoder. That means taking the shrink wrap off and finding the common (negative) circuit and hooking the cap (with a series resistor) between the common and blue lead.

[carlso]

Thanks Pete,

These are not red LED's but rather bright white's that I painted black and then scratched very tiny holes in the black paint. I then used Gallery Glass over the appropriate hole. They look pretty good for DIY. This application is in a very small KATO cupola caboose with not much available space.

Carl

[RBrodzinsky]

Please correct me if I am wrong, but wouldn’t putting a cap directly in line with the LED fail to prevent flickering? Since the LED is connected to the decoder’s common(+) and the light’s control wire(- if turned on, open if turned off), if the decoder loses power, there is no return path for the voltage from the capacitor, therefore the LED would still flicker until power is restored and a ground path available again.

[carlso]

Rick,

Thanks for the reply. I have no earthly idea, that is the reason for my post.

Actually I am more confused after Pete's reply and yours. Pete says to connect to the common (negative) and the blue wire. I thought the common was the blue wire and is positive. I may just pull off the shrink wrap to see the insides of the decoder.

Can a cap be placed between the red and black wires ? That would provide power to the decoder on dirty track spots to keep it "alive". Can a DCC layout be wired improperly to cause the positive rail to be the left hand rail ?

I am not in the know, just trying to learn,
Carl

[carlso]

I took the shrink wrap off and here are images of both sides of the TL1. I do not know but it appears to have a cap installed. I could not get it focused enough for a real close image.

That yellow item is lettered as follows:

       226J
     70F34

Is that a small cap?

Thanks for looking,
Carl

[jdcolombo]

Hi Carl.

For an effective keep alive, you need to connect the positive of the cap to the blue wire (the blue wire is the common positive connection), and the negative to the common ground for the decoder.  Finding the common ground is the problem.  Usually, you can wire the negative side of the cap to the negative side of one of the rectifier diodes.  In most decoders, the rectifier diodes are easy to find - there are 4 of them, same size, with a band at one end to signify the positive connection.  But this decoder doesn't have four rectifier diodes.  It DOES have a larger component with what I'm guessing is four legs on the upper LH side.  That four-legged creature looks suspiciously like a very small bridge rectifier, and one of those legs would be the common ground.  Do you happen to see +/- markings on that component?  Or on the two smaller components on the RH side of the photo at the top and bottom of the decoder?

Maybe Peteski knows.  It appears that the larger component has a number of 3321 on it.  Can you see what the numbers are on the two smaller components on the RH side at the top and bottom edges of the decoder?

John C.

[peteski]

Please correct me if I am wrong, but wouldn’t putting a cap directly in line with the LED fail to prevent flickering? Since the LED is connected to the decoder’s common(+) and the light’s control wire(- if turned on, open if turned off), if the decoder loses power, there is no return path for the voltage from the capacitor, therefore the LED would still flicker until power is restored and a ground path available again.

The ideal presented by others is to hook the capacitor in parallel with the LED with its series- connected current limiting cap.  Basically to use the cap as a very small capacity "battery" for the LED itself.  There is no return ground path needed in this circuit.

[C855B]

The ideal presented by others is to hook the capacitor in parallel with the LED with its series- connected current limiting cap.  Basically to use the cap as a very small capacity "battery" for the LED itself.  There is no return ground path needed in this circuit.

Bingo.

To John's idea, dollars to donuts the four-legged chip is a bridge rectifier. Its proximity to what appears to be a 22uF cap (yellow) is further confirmation, the combination comprising local power supply rectification from the DCC feed.

However, that said and as Pete confirmed, Carl just needs "keep alive" on the LED output, not the whole decoder. The 22uF cap would probably be enough to de-flicker if it wasn't for the decoder's power needs for logic. I would put Carl's suggested 100uF between the blue (positive) and white (switched negative), with resistors before and after the added cap, with provision to add another 100uF if it doesn't completely solve the problem.

[peteski]

Rick,

Thanks for the reply. I have no earthly idea, that is the reason for my post.

Actually I am more confused after Pete's reply and yours. Pete says to connect to the common (negative) and the blue wire. I thought the common was the blue wire and is positive. I may just pull off the shrink wrap to see the insides of the decoder.

Can a cap be placed between the red and black wires ? That would provide power to the decoder on dirty track spots to keep it "alive". Can a DCC layout be wired improperly to cause the positive rail to be the left hand rail ?

I am not in the know, just trying to learn,
Carl

Sorry that my explanation was confusing.

I would recommend adding a true keep Alive capacitor to that decoder. That means taking the shrink wrap off and finding the common (negative) circuit and hooking the cap (with a series resistor) between the common negative and blue lead.

There is no common negative wire or solder pad on this decoder. That is why Digitrax  does not offer a suggestion on how to add a Keep Alive capactor.

Was that clearer?

Here is a possible hookup diagram.



Looks like instead of a bridge rectifier or 4 discrete diodes they used 2 dual-diode components (black rectangles with 3 leads).  The yellow components  is (like you suspected) a 22 uF tantalum cap, but it is most likely used in the low power microcontroller power stage circuit.  You can't just add your cap in  parallel with it.  But its negative lead is pretty much guaranteed to be the decoder's common negative circuit.

However, to make certain that this is the case, before we continue, measure continuity between all the leads marked by red arrows.  That is the negative lead and the cathodes of the dual-diode components should all be connected (have continuity or zero ohm resistance between them).  If that is the case then you can go ahead and connect the Keep Alive circuit (shown in yellow) to the decoder.  That should eliminate any flickering or decoder dropouts.  If 100 uF is not enough, add more capacitance. 1000 uF was chosen arbitrarily (I don't expect you will need  more than few hundred micro Farads).

But that Keep Alive addition might make decoder programming problematic. If that is the case, disconnect one side of the Keep Alive for programming, or possible add a small coil in series with the Keep Alive (like it was discussed in one of the ESU decoder threads about adding Keep Alive).

[peteski]

Bingo.

To John's idea, dollars to donuts the four-legged chip is a bridge rectifier. Its proximity to what appears to be a 22uF cap (yellow) is further confirmation, the combination comprising local power supply rectification from the DCC feed.

Not quite - see above post.

However, that said and as Pete confirmed, Carl just needs "keep alive" on the LED output, not the whole decoder. The 22uF cap would probably be enough to de-flicker if it wasn't for the decoder's power needs for logic. I would put Carl's suggested 100uF between the blue (positive) and white (switched negative), with resistors before and after the added cap, with provision to add another 100uF if it doesn't completely solve the problem.

Also not  my goal.  I rather keep the entire decoder alive, than just the decoder.  I explained my reasoning earlier.