Adding a TCS KA-N1 Keep Alive to an ESU Loksound 58721 decoder?

[vermonter_railfan]

I've been looking into adding some keep alive to my N scale locomotives and figured I would try with the recently released Atlas B40-8W. These units have ESU Loksound 58721 decoders installed from the factory. I purchased a TCS KA-N1 for the keep alive side of everything, but while I know how to add tons of LEDs to these decoders I don't understand what I would need to do to add the keep alive. Anyone have any suggestions?

[peteski]

There is no convenient way to add a keep-alive to those.  No dedicated pads.  See:
https://www.therailwire.net/forum/index.php?topic=54278.msg744107#msg744107

[vermonter_railfan]

There is no convenient way to add a keep-alive to those.  No dedicated pads.  See:
https://www.therailwire.net/forum/index.php?topic=54278.msg744107#msg744107

Thank you! So if I'm understanding this correctly, since the KA-N1 is a 1000uf design I need to attach them to the purple and green markings the image of yours above? Basically to the onboard capacitors? Do you know if it matters which wire goes to which?

[peteski]

Thank you! So if I'm understanding this correctly, since the KA-N1 is a 1000uf design I need to attach them to the purple and green markings the image of yours above? Basically to the onboard capacitors? Do you know if it matters which wire goes to which?

Yes that would be my recommendation.  The Keep-alive is polarized, so its negative side goes to the green marked areas, and positive to the purple marked areas.  Try to be quick when soldering because those MLCCs (Multi-layer ceramic capacitors) are sensitive to thermal shock.  Of course also keep in mind that doing this will void the decoder's warranty.

[vermonter_railfan]

Yes that would be my recommendation.  The Keep-alive is polarized, so its negative side goes to the green marked areas, and positive to the purple marked areas.  Try to be quick when soldering because those MLCCs (Multi-layer ceramic capacitors) are sensitive to thermal shock.  Of course also keep in mind that doing this will void the decoder's warranty.

Thanks for the direction! I will report back my findings once I get everything installed. I'm already well passed the warranty with the decoder so that won't be a problem.

[jdcolombo]

Be sure to take a VERY good look at Peteski's drawings on those decoders.  Note in particular that there are three different "stages" to feed in a keep alive.  If you want the keep alive to keep the motor running, rather than just keep the sound circuit going, you'll need to use the RED areas on Peteski's photos, not the purple areas.  The purple areas will keep the sound circuits alive, but not the motor.

The TCS KA-N1 has large enough capacitance that you may want to connect it to the red areas, instead of the purple areas, but that would take much more careful soldering . . . 

John C.

[peteski]

Yes John, I did explain the different power stages in the write-up I pointed to earlier, and also what will be the effect of connecting a keel-alive to different stages.

I don't think that a 940µF capacitor-based keep-alive will provide enough energy for the motor and LEDs plus the decoders electronics.That is why I recommended to connect it to the stage which keeps the electronics-only powered.  Hopefully the flywheel (even at slow speeds) will get the loco coasting past the power drop, while keeping the electronics powered.  And as a secondary reason, becuase it is much more difficult to solder the positive wire to one of those small areas indicated in red (the raw rectified voltage).

I'll also reply to the PM I received from vermonteer_railfan.

You understand my recommendations correctly.  Positive wire goes to any of the areas you circled in red and white, and negative to any areas circled orange.
The green dots basically indicate that the large copper areas of the decoder are all "ground" or common.  You would have to gently scrape the solder mask coating until the copper is exposed.  The problem with soldering to those areas is that the large copper planes act as heat sinks, so it will be difficult to heat them up to where the solder wets the copper (for a good solder joint).

If you want to follow John's recommendations, you would have to solder the keep-alives positive wire to to one of the two small areas indicated by red rectangles.  Negative connection would be the same as for the other scheme.

[jdcolombo]

Hi Peteski.

Yeah, you're probably right about not trying to power the motor with the KA-N1.  I had a brain freeze and was thinking of the TCS KA-1, which I regularly use in N-scale steam tenders to power both the electronics and motor with terrific results (I even use the KA-1 in HO "legacy" engines that aren't DCC ready).  But I think the "regular" KA-1 has something like 36,000uf (36 farads) vs. the 1 farad of the N1. 

On the other hand, before the N1 came along, a lot of us had been using 440uf or 660uf in N scale diesels via two or three paralleled tantalum polymer caps connected to the regular keep alive pads on a LokSound micro (which would route the power both to the electronics and motor circuits) with good results.  So maybe trying the N1 for both motor and power to see how it works would be the thing to do, IF it can be soldered to those red areas without too much risk.  I don't use any of the LokSound board decoders in my installs, so I don't know how hard it would be to accomplish the solder job.  Given the size of some of the components I see on LokSound micros, however, I'd be VERY careful about my soldering skills (and equipment) before going down that road; soldering to the bigger caps (the purple areas) would be much easier, I think.

John C.

[peteski]

I never looked inside KA2 (the larger unit), but I know that TCS KA2 (smaller unit) has five 1F 2.7V caps wired in series to make up a 0.2F (or 200,000µF) 13.5V capacitor. Yeah, that is a "real" keep alive which can power the entire N scale model for couple of seconds (especially slow moving model since the motor current draw at that time is fairly low).

[vermonter_railfan]

Alright the KA-N1 finally arrived today and I got it installed. I followed your guys recommendations and soldered the positive wire to the red mark and then I soldered the other to a green ground. The extra 1000uf really doesn't do much but it does make a difference too. Testing a modified decoder and non modified on the same track (both new run B40-8W with factory 58721's and no CV changes) the chassis with KA-N1 added has about a split second before the lights dim while the non-modded drops immediately plus when you reconnect track power the keep alive chassis takes right off while the non-modded one takes a second to get back up to speed.

There is also enough room that you could probably add a second or larger keep alive behind the decoder in these B40-8Ws so there might be some options for a stronger keep alive

[GaryHinshaw]

Here is a very rough rule of thumb for keep-alive capacity: the charge stored in a capacitor is Q = C*V where Q is the charge in Coulombs, C is the capacitance in Farads, and V is the voltage across the capacitor in Volts.  A KA-N1 has C = 1000 uF and it operates with a typical track voltage of ~10 V (in round numbers), so it can store a charge of:

C = 1000 x 10^{-6} F x 10 V = 10^{-2} Coulombs.

1 amp of electric current is 1 Coulomb per second, and a typical N scale loco draws up to ~100 mA = 0.1 Amps.  So if the power is interrupted, a KA-N1 would discharge in ~0.1 sec if the current draw is 100 mA, and proportionally longer (shorter) if the current draw is lower (higher).  This neglects details like the actual profile of the discharge current and such, but it gives a good rule of thumb.

My experience with non-sound decoders is that even a fraction of a second of back-up power can make a world of difference in loco performance.

[peteski]

You also have to remember that the useful energy stored in the capacitor is only down to about 6-8V (depending on the decoder).  So any energy still remaining below that voltage will not be useful.   So only the amount of time it takes for the cap to discharge from let's say 11V down to 7V is actually useful.

[GM50 4164]

Yes that would be my recommendation.  The Keep-alive is polarized, so its negative side goes to the green marked areas, and positive to the purple marked areas.  Try to be quick when soldering because those MLCCs (Multi-layer ceramic capacitors) are sensitive to thermal shock.  Of course also keep in mind that doing this will void the decoder's warranty.

Peteski, what is the reason for not wiring to the board's U+ pad under the board?

[peteski]

Peteski, what is the reason for not wiring to the board's U+ pad under the board?

Because on this decoder ESU decided to make V+ supply 5.3V (not the more typical U+ 12V).  Adding caps to that pad will not supply power to some parts of the decoders.   Keep alive needs to be connected to either the raw rectified voltage side A (on my diagram) or B (on my diagram).  On A it will power everything (motor, functions and all the electronics) or on B it will power all of the decoder's electronics sans motor.

[vermonter_railfan]

Because on this decoder ESU decided to make V+ supply 5.3V (not the more typical U+ 12V).  Adding caps to that pad will not supply power to some parts of the decoders.   Keep alive needs to be connected to either the raw rectified voltage side A (on my diagram) or B (on my diagram).  On A it will power everything (motor, functions and all the electronics) or on B it will power all of the decoder's electronics sans motor.

I'm guessing that is because of the built-in resistors for the AUX pads on these decoders? I was originally wondering the same thing but that would make sense now. Is that still true if you change the AUX function to the "keep alive setting" within the decoder programming?