Keep-alive in the current iteration of Atlas gold GP9

[bbussey]

All,

I finally cracked open one of my Atlas gold GP9 units from the last run to see what was involved in adding a TCS KA-N1 keep-alive component.  The model uses the Nano and has a chassis design similar to the latest RS3, so the integration procedure of the KA-N1 would be the same.  The frame however fills the entire shell and there appears to be no free cavity to park the KA-N1.  There also appears to be no section of frame that can be sacrificed for space.  Does anyone have any thoughts of where the keep-alive could be placed?

[peteski]

No idea where you can locate the KA unit, ut one of those light boards is on the way to me so I can do a write-up like I did for the other Atlas light board board. Agin, Atlas with their lack of foresigth did not include a GND pad off of the E24 connector. That is piss-poor design for adding any sort of KA circuit. Really disappointing. I suspect all the " new design E24 boards for various models were designed by the same person, probably at the same time when they remade the chassis for multiple locos. Too bad nobody at Atlas is a seasoned modeler familiar with DCC to advise them to include the GND pads on the new boards.

While it is too late for these boards, maybe @EmdFan could communicate this need to the appropriate Atlas employees who work directly with the Chinese engineers/designers, for GND (and U+ would also be very helpful) solder pads to be added.  With E24 connector this is a no-brainer and adding those pads to the PC board layout should not incur any additional costs.  We all know that with smaller scale models the wheel-to-track contact reliability is not ideal. We need keep-alives for smoothest slow speed performance.

[bbussey]

I fear the RS-11 will be similar, with no cavity for a keep-alive module. I just ordered a pair of Atlas ex-LifeLike/Walthers RS-2 units with DCC. I suspect the mechanism will be similar to the RS-3 and that I’ll have to stash the keep-alive in the upper cab after cutting away the body roof inside of the cab.

[jdcolombo]

I don't have one of these, but looking at the photos on Spookshow, would milling out the fuel tank area work?  You probably could run the wires around the speaker-side of the motor housing and up a small groove to the decoder.

John C.

[Ed Kapuscinski]

I continue to believe that selling N scale sound equipped models without keep-alives in 2025 should be a federal crime.

[jdcolombo]

It continues to amaze me how the manufacturers simply ignore the wealth of relevant information on the subject of sound installations that has been developed by modelers actually doing this stuff.  We all knew for years that to have decent sound, you needed to have a speaker in a separate sealed enclosure, but it took a decade for manufacturers to even start down that path, and some still haven't gotten it.  A bare speaker mounted on a frame is going to be awful, and the shell is not a proper enclosure.

Similarly, we've all known for years that "keep alive" (put in quotes as a recognition that Peteski divides the world into "real" keep alives that can keep an engine running for a second or more and "anti-sound-dropout" devices, which won't keep a motor running but will avoid sound dropouts and re-starts from momentary power loss) is required in N scale models to keep us all from going bonkers listening to sound decoders reset even on relatively clean track with relatively clean wheels.  Yet do we get it?  No.  Not even the ability to wire it up conveniently after we by hook or crook or both find some tiny space to stash a couple of 220uf chip caps wired in parallel.

This is not nuclear fusion research, folks.  The answers are on this site.  Just read the install threads.

John C.

[bbussey]

I don't have one of these, but looking at the photos on Spookshow, would milling out the fuel tank area work?  You probably could run the wires around the speaker-side of the motor housing and up a small groove to the decoder.

In the fuel tank is definitely an option, if the motor does not sit low.  That's were I looked to place it in the RS3 originally but there wasn't enough room.  Fortunately there is plenty of room in the cab if a hole is cut in the roof of the bodyshell that sits underneath the cab.

[John]

It continues to amaze me how the manufacturers simply ignore the wealth of relevant information on the subject of sound installations that has been developed by modelers actually doing this stuff.  We all knew for years that to have decent sound, you needed to have a speaker in a separate sealed enclosure, but it took a decade for manufacturers to even start down that path, and some still haven't gotten it.  A bare speaker mounted on a frame is going to be awful, and the shell is not a proper enclosure.
This is not nuclear fusion research, folks.  The answers are on this site.  Just read the install threads.

The designers know better than us ..     It seems that the decoder manufacturers should be able to add a couple of these capacitors onto the board, rather than as a separate assembly

These look promising .

https://tonystrains.com/product/tcs-2010-ka-n1-n-scale-keep-alive-capacitor

[bbussey]

The KA-N1 is what I installed in the RS3 units. I’ll pop the GP9 plastic fuel tank cover today to confirm there is only zinc in that void.

[Ed Kapuscinski]

The designers know better than us ..     It seems that the decoder manufacturers should be able to add a couple of these capacitors onto the board, rather than as a separate assembly

EXACTLY. Part of the problem with keep alive installations is that they ARE separate boards that you have to find space for. While having the smallest decoder board possible is really handy at times (cough, STRT 10), there are plenty of situations where the equivalent of the ESU "micro" fits just fine. If the same hardware fits in the footprint of the "nano" I don't know they don't offer a version of the micro board with caps in that extra space.

[peteski]

The designers know better than us ..     It seems that the decoder manufacturers should be able to add a couple of these capacitors onto the board, rather than as a separate assembly

These look promising .

John, sure those could be added to the decoder but micro-Farads require space. As you know, that would make decoders physically larger, and we already have problems finding empty space iinsode our tiny models.  Lats thing we want are larger size decoders.

And manufacturers do offer decoders with on-board keep-alive circuits, but they are on physically larger decoders designed to fit in larger scale models.

At this time the best N scale solution is to use off-decoder capacitors, so they can be squeezed somewhere in our tiny models.  For a while now I have been experimenting with multi-layer-ceramic-capacitors (MLCCs) instead of tantalums. They are non-polarized and do not literally burn up  when they fail.  While the highest MLCC capacitance is only 47µF  they are small and can be assembled into  custom-shaped banks of capacitors which can be squeezed into spaces where tantalums would not fit.

One of these days I'll do  a write-up about MLCCs.  But we still need easily accessible GND and U+  solder pads on the decoder or on the light board in order to connect them to the decoder.  I really hope model manufacturers listen to our pleas.

[jdcolombo]

Ed got it right.

Yes, adding caps will need more space.  But the LokSound nano is 20mm long, which is the length of the "shelf" on either end of the decoder chassis on most hood diesels.  So why can't you design a board decoder with the equivalent of the Nano at one end, a speaker in an enclosure at the other, and keep alive caps in the middle?  OK maybe thickness (height) is an issue; but the current Atlas GP9 board sort of has this construction, with most of the decoder bits at one end, the speaker at the other, and the middle with some things sticking up but vast ("vast" for N scale, anyway) tracts of open land in the middle (there are some tiny resistors there; move 'em). The height of the JST connectors (if that is what they are - the white plastic bits) isn't less than a typical 220uf tantalum chip cap, and as Peteski points out, there are even smaller options available today to fill in that empty space and end up with 400 - 600uf of capacitance to act as an electrical dropout buffer, if not a true keep alive.

It is certainly within the realm of the doable.  So just do it already.  Or as Peteski notes, AT LEAST give us pads to solder to!

John C.

[GM50 4164]

The KA-N1 is what I installed in the RS3 units. I’ll pop the GP9 plastic fuel tank cover today to confirm there is only zinc in that void.

Brian, I have a SD7 with the same board as your GP9. While I've already added a KA-N1 to one end as there is room. The GP9 is shorter. I did lay a couple of caps across the speaker area just past the plug on the board and the shell fits with no modification. Also, it doesn't impact the sound at all! How this helps some.

[bbussey]

Brian, I have a SD7 with the same board as your GP9. While I've already added a KA-N1 to one end as there is room. The GP9 is shorter. I did lay a couple of caps across the speaker area just past the plug on the board and the shell fits with no modification. Also, it doesn't impact the sound at all! How this helps some.

Yeah definitely not enough room on the end for a KA-N1.  Are those caps above the speaker on your SD7 in the photo?  If so, what is the manufacturer and part#, and how did you wire them in?  At the same U+ and ground points that you would a KA-N1?

One additional note – the SD9 photo has the stock open frame motor that is sitting above the sill, while the RS3 now uses a can motor that sits partially below the sill into the fuel tank area.  So if the GP9 has the same motor as the SD9, and I think it does, then there should be enough clearance to mill the frame inside the fuel tank to house the KA-N1.

[jdcolombo]

FWIW, photos on Spookshow's site show that the motor in the 2024 GP9 release is the open frame motor.

http://www.spookshow.net/loco/atlasgp79.html

John C.