Best Of LokSound 5 Nano - Capacitor Connection?

[peteski]

@peteski I agree that 13.7v is too close for comfort on a 16v cap. It's easily within the 20% tolerance stated in the specs. The "second" batch is 220uF 25v - the ones I'm working with currently. I'll be out of town for a couple of days, but I'll connect the "conditioned" one (a 220uF 25v cap) when I get back and see how it goes. The 330uF and 470uF caps I have are 16v. 220uF was the highest 25v caps I could find.

@Steveruger45 Do you have a link for the tantalum-polymer caps you use? Also, do you know anything about the 220uF 16v caps sold by SBS4DCC? They're supposed to be MIL-spec, which I believe has a 5% or 10% tolerance. Are those worth a shot, or still too close for comfort re: voltage?

Again, I appreciate the help and suggestions, guys.

Chris

Chris, the tolerance specifications are for capacitance, not for max working voltage!!   That is an absolute value!  While others have used 16V in DCC installations I would never do that.  I also rather gang smaller capacitance caps together in parallel than use larger values caps.  Like 100 or 150 uF 25 V caps.

I have a feeling that the 25V caps will not burn up when you try them.

I have feeling

[Steveruger45]

Chris
I got my caps from digikey a couple of years ago
Here is an updated link to a bunch of suitable polytantalum caps rated at 20 and 25v.

https://www.digikey.com/en/products/filter/tantalum-polymer-capacitors/70?s=N4IgjCBcoEwAwBYCcVQGMoBcBOBXApgDQgD2UA2uHHAAQCtAYiMWAKy2PMgzz1PEBmARyYBdYgAdMUEAFUAdgEtMAeQBmAWXwBDAM65s+EAF9i8AOwpoIDJBwFiZSJV4A1LjFY134kFJkKyupaegZGpiAAtDCoNlh4RKQUIKwgosYZQA

I have used the 20v panasonic and avx caps with no issues but now use the 25v avx.
My power cab can pump out 13v so the 20v. Caps are within 80 percent of rated voltage guideline at 16v.
Ive had no failure’s of these.
I either use 4 x 100uF caps or 2x 220uF depending on the space i have to fit them in.
I never mix capacitors rated voltage in the same installation
Trust this helps.

[RBrodzinsky]

I've also made good use of the 330uF 20V PolyTa caps, DigiKey p/n  478-13505-1-ND (https://www.digikey.com/en/products/detail/kyocera-avx/TCN4337M020R0100E/10381493?s=N4IgTCBcDaIAQBYDsAOAtARgMwFYAMOmaAcgCIgC6AvkA)

A couple of these, with 100uH inductor (https://www.digikey.com/en/products/detail/taiyo-yuden/CBC3225T101MR/958010?s=N4IgTCBcDaIKwA4DsBaAjANjA9KByAIiALoC%2BQA) in series between cap and ground wire and never had a problem

[cbroughton67]

I'm back home and had some time this evening to tinker. The short answer is that the caps are the culprit. I started with a pair of the 220uF 25v caps with a 12v zener diode / 100ohm resistor "protection circuit" added. No issue. Tried it without the resistor - no issue. So, I removed the zener, so it was the pair of caps only, and after an hour of idling, there was no issue!

I then tried the same zener/resistor setup with the 16v 470uF caps. Boom. I followed that experiment up by testing the 470uF caps with the DC power pack on the variable DC side with a VOM connected. The cap failed before it got to 8v. So, they're junk.

For the record, I "smoke test" decoder installs without the shell on, exactly for this reason. In this situation, I've been using extra long wires to the caps, wrap the caps in kapton tape, and put them in a covered jar to protect the layout, etc. from "catastrophic failure." So, when I say it went "boom," everything was contained and nothing was damaged.

So, I guess I'll be ponying-up for the more expensive poly-tantalum caps you guys have recommended. $4-5 or so per cap is still cheaper than replacing a potentially torched locomotive shell.

Thanks gain for the help and education, gentlemen.

Chris

[peteski]

I'm back home and had some time this evening to tinker. The short answer is that the caps are the culprit. I started with a pair of the 220uF 25v caps with a 12v zener diode / 100ohm resistor "protection circuit" added. No issue. Tried it without the resistor - no issue. So, I removed the zener, so it was the pair of caps only, and after an hour of idling, there was no issue!

I then tried the same zener/resistor setup with the 16v 470uF caps. Boom. I followed that experiment up by testing the 470uF caps with the DC power pack on the variable DC side with a VOM connected. The cap failed before it got to 8v. So, they're junk.

For the record, I "smoke test" decoder installs without the shell on, exactly for this reason. In this situation, I've been using extra long wires to the caps, wrap the caps in kapton tape, and put them in a covered jar to protect the layout, etc. from "catastrophic failure." So, when I say it went "boom," everything was contained and nothing was damaged.

So, I guess I'll be ponying-up for the more expensive poly-tantalum caps you guys have recommended. $4-5 or so per cap is still cheaper than replacing a potentially torched locomotive shell.

Thanks gain for the help and education, gentlemen.

Chris

Good show Chris! 
Regardless of whether you go with the standard or polymer caps, with and without a Zener diode, I *HIGHLY* recommend going with ones rated for at least 20V.    Speaking of the Zener diode and a resistor circuit, if the decoder's internal DC voltage is higher than the Zener voltage, both the Zener and resistor will be conducting and "burning off" the extra voltage as heat.  That Zener protection diode circuit is really designed for occasional over-voltage rather than constant operation.  I prefer to minimize continuous heat generators inside the loco's shell.

Also, since you are now introduced to "burning in" caps before installation using external power source (DC throttle), I recommend using that instead of using the DCC system and the decoder's rectifier as the source of power for testing caps before installing them in your models. If the burning cap winds up with an internal short, you can fry the decoder's rectifier diodes.  It is safer using an old robust DC throttle.

I'm also still a fan of ganging smaller value caps to achieve the desired capacitance instead of using single large (like 470 or 220 uF) tantalum caps.  I know there is a trade off in size, but I think that is a safer combination.  I also don't think that you can get the larger capacitance devices in  voltages higher than 16V.

[cbroughton67]

Thanks for the cap recommendation @Steveruger45 and @RBrodzinsky . I went with the 330uF 20v caps, and they work and fit in the space available without issue. Thanks again for the help and guidance, guys!

Chris

[Ed Kapuscinski]

This thread deserves a Best Of. @GaryHinshaw , how do we do that again?

[GaryHinshaw]

We bug @tom mann and/or @John. 

[Ed Kapuscinski]

We bug @tom mann and/or @John. 

Dudes, make it so!

[tom mann]

Dudes, make it so!

Done buddy