I have a question 'cause I'm stupid

[carlso]

I have wired in some installs with the smd caps. I got them from Degi-Key and they are 100uf 20v items. I built a small circuit to place inside of a caboose and tested ,WOW!

Here is the circuit in order;

1. bridge rectifier
2. 5 caps wired all +'s together and all (-) together and wired properly to the dc side of rectifier
3. one 1.6K ohm resistor
4. 2 LEDs wired properly

I tested it on my work table and when I put the rail contacts on the bridge rectifier one of the caps exploded. There are 5 in a row and the very center one is the one that blew. Quite a quick fireworks show. The Power Cab measured 14.5v on the rail

Bad resistor ? ? ?

Carl

[Steveruger45]

Hi Carl,
I think it’s a bad cap.
I don’t see how a bad resistor could affect the caps. I’m assuming the resistor was/is there for the leds.
I’ve tested cheap Chinese unbranded 25v tantalum caps, and mentioned this in another thread a while ago, on an 18v power tool battery and had a good few blow up like you have experienced.   They appeared good on visual inspection before the test too, although I did find one of the few that blew with a damaged terminal.
Are your caps tantalum or poly-tantalum?
I ask as if the former you should only apply 50% of the rated voltage. For the poly tantalum caps you should only apply 80% of the rated voltage.
I always do a burn in test now on caps before installation.  I wire the ones I’m going to use plus a couple more in a ladder arrangement spaced about 8 inches apart and “proof or poof” them for an hour or so on a power tool battery.  I do this test on the garage concrete floor.
I haven’t had any failures on the quality caps from Digi-key.

[peteski]

Yes, like Steve said, it was a bad cap.  That is why some time ago I recommended a "burn in" for all the tantalum cap installs, before installing them in models.

Here is the post: https://www.therailwire.net/forum/index.php?topic=43415.msg553360#msg553360
Steve (few posts below mine) actually came up with using a power tool 18V battery pack for the burn-in. An excellent idea!

[MK]

I would think that the stuff from Digikey is of better quality.

[carlso]

You would think so, right. I am wondering. The caps I purchased were AVX Corp. units and were Tantalum caps 100uf, 20V  10%  2917, whatever all that ,means. I started using 16v a few years ago and then contributors on TRW sold me that 20v was much safer for DCC installs. So I bought these in April 2018 and just now tried to use them. Wow, what an awakening. I was not holding the damn unit but my hand was above it when the cap blew. Sort of fried the tip of one finger. That's my stupidity ! !

I think I will find another way (batts ? ?) or actually learn how to properly use electronic items.

Carl

[peteski]

Majority of electronic components are safe and very reliable.  But Tantalum caps are just one of those things that are a bit dangerous, due to their inherent design. Yes, it is strange to think that a widely used component can be dangerous (even if properly used),  but that is just the reality.  There could be many reasons it self-destructed.  Maybe even its epoxy case developed an hairline crack over time, and moisture got inside?  We'll never know, but this should be taken as a warning to be careful with these caps.  Tantalum Polymer caps are safer, but they are are usually larger physically (for the equivalent capacitance.voltage), and a bit more expensive.

If you have plenty of space inside the model (like a caboose), I recommend using plain old aluminum electrolytic caps.  The are much more robust, and less prone to catastrophic failures.

[Point353]

I have wired in some installs with the smd caps. I got them from Degi-Key and they are 100uf 20v items. I built a small circuit to place inside of a caboose and tested ,WOW!

Here is the circuit in order;

1. bridge rectifier
2. 5 caps wired all +'s together and all (-) together and wired properly to the dc side of rectifier
3. one 1.6K ohm resistor
4. 2 LEDs wired properly

I tested it on my work table and when I put the rail contacts on the bridge rectifier one of the caps exploded. There are 5 in a row and the very center one is the one that blew. Quite a quick fireworks show. The Power Cab measured 14.5v on the rail

Bad resistor ? ? ?

Doubtful that it's a bad resistor. Could be a random failure of the one capacitor, but more likely it could be a so-called surge failure.
If you plan on continuing to use the tantalum type capacitors, then you should add a resistor between the Power Cab output and the bridge rectifier input.
Try a value on the order of 470 ohms.

Also, what type of soldering iron did you use and how long did you apply heat to the capacitors?

https://escies.org/download/webDocumentFile?id=62199

[carlso]

I want to thank all of you for your input, very much appreciated.

First off, I had my head 'you know where" and did not have on my transparent belt, when I asked "Bad resistor?" I meant capacitor which is pretty much the consensus.

 Petseki - I have thought and may use the aluminum electrolytic. Should I use 16 or 20 volt cap?

Point 353 - I used a Weller iron and only applied heat a very short time. Thanks for the link.

Here are some images of my project:

Simple schematic


Bottom of ladder


Top of ladder


Bottom


I did not have the tweezers attached during power up, you can see the center cap was the bomb.
The schematic was drawn for using an aluminum electrolytic cap but I remembered having these old smd's, so used them.

[MK]

Carl,

Your copper "bus" seems to be at the bottom of each capacitors' contacts and touching the case.  As you are soldering the connections, you are also heating up the case of the capacitors.  In your second photograph there seems to be some melting of the cases.  I'm not saying this is the cause but I would solder the bus across the center of each pad instead of at the bottom.

Also I would use a thinner wire to join the caps.  No need for something so thick and thus requiring more heat for soldering.

Just an observation.

[Steveruger45]

Yes, what MK said.  I uses 30 gauge wire and solder on the sides of the caps or I place them in a row on a pcb with an insulating gap down the middle and surface mount/solder them to that after tinning everything first.
BTW the poly tantalum caps I use are Panasonic 100uF x 20v and/or AVX 220uF x 20v.  I’m applying a little less than 14v to them from my power cab so well within the 80% of rated voltage, my burn in test actually stresses them a bit with 18v.
The former are same footprint size but only 2mm thick so you can stack two of them in the same space.  The latter are the same size as the ones you have used.  I like the Panasonic’s a lot due to their size which gives quite a few more space options  to fit them in an n scale loco etc.

[Point353]

Here are some images of my project:

Simple schematic

The extra resistor, to limit the charging current to the capacitors when voltage is first applied, could alternatively be connected between the positive output of the bridge rectifier and the capacitors.
Also, if it's important for brightness of the LEDs to be closely matched, you may want to put a separate resistor in series with each individual LED rather than using a single resistor for all of the LEDs. Then, the resistor values can be adjusted, if necessary, to make the brightness of all the LEDs match.

[carlso]

 What a treasure trove of knowledge and experience here on TRW. Just ask and you shall receive.
      Very much appreciated!

Steve- When you have time, is the following Digi-Key part # the Panasonic that you use?
                P19854CT-ND

Point353- I shall add that to my next module, between the dc(+) and the + side of the caps.
                Not sure about several resistors. This was drawn up to add markers and interior
                 lighting. I may only use 2 for markers.

Thanks again to all, such great help,
Carl               

                         

[Steveruger45]

Hi Carl
Yes. Those are the ones.    Here is the link.
https://www.digikey.com/product-detail/en/panasonic-electronic-components/20TQC100MD2/P19854CT-ND/6202259

They are very handy. On some installs I’ve mounted them four in a line and placed them over the motor cradle, on others I’ve mounted them in L shaped blocks three long with one beneath the end one over the rear flywheel and others in just in a block two x two.

[peteski]

Carl, I recommend using 20V or higher rated caps (that goes for electrolytic, and especially tantalum).  Just to be on the safe side. At the last NTRAK show someone mucked around with the Digitrax booster and the track voltage was over 18V!  It was high enough do trigger a safety shutdown of a BLI(?) decoder. 

The initial current limiting resistor which was recommend is a good idea, but I feel it is not really needed here.  I feel that the resistance of the wiring and internal resistance of the bridge rectifier (plus finite amount of current the booster can supply) limit the initial current to a safe value for those caps.  The 470 ohm resistor also seems a bit on the high side.  47 ohm would be something I would recommend (if any resistor was to be used).

As for using a single resistor supplying current to multiple LEDs connected in parallel, that is not the best practice, for sure, but if the LEDs came from the same production batch, their electrical characteristics are likely similar enough for this type of hookup to work well.   Normally each LED shoudl have its own current limiting resistor.

[Point353]

The initial current limiting resistor which was recommend is a good idea, but I feel it is not really needed here.  I feel that the resistance of the wiring and internal resistance of the bridge rectifier (plus finite amount of current the booster can supply) limit the initial current to a safe value for those caps.  The 470 ohm resistor also seems a bit on the high side.  47 ohm would be something I would recommend (if any resistor was to be used).

You could certainly make the resistor value smaller than 470 ohm, but what is the power rating (and the surge withstanding capability) of that resistor.
Suppose that a capacitor fails in short-circuit mode.  Do you want the resistor to possibly fail, too?
Aside from a small space and cost penalty, is there any other downside to including a current limiting resistor between the bridge output and the caps?