Mysteries of the new LokSound in the Intermountain SD40-2

[peteski]

John,
The decoder photos you posted are much higher resolution, but still a bit fuzzy.  I'll try to post my thoughts about the circuit on Sunday night.

A quick observation of the bottom of the circuit seems to confirm 6 function outputs with current limiting resistors. I also see that the AUX pads are wired in before the resistors.  I haven't really had much experience with ESU (it has been a while).  It seems that they are confusing things by renaming function outputs (as defined by NMRA) into AUX outputs.  I have not looked their manuals for few years. Does this make sense?
NMRA    ESU
F0F         ?
F0R         ?
F1         AUX1
F2         AUX2
F3         AUX3
F4         AUX4

That makes it a 6-function sound decoder with motor control.

[AKNscale]

Yup, it does

[jdcolombo]

John,
The decoder photos you posted are much higher resolution, but still a bit fuzzy.  I'll try to post my thoughts about the circuit on Sunday night.

A quick observation of the bottom of the circuit seems to confirm 6 function outputs with current limiting resistors. I also see that the AUX pads are wired in before the resistors.  I haven't really had much experience with ESU (it has been a while).  It seems that they are confusing things by renaming function outputs (as defined by NMRA) into AUX outputs.  I have not looked their manuals for few years. Does this make sense?
NMRA    ESU
F0F         ?
F0R         ?
F1         AUX1
F2         AUX2
F3         AUX3
F4         AUX4

That makes it a 6-function sound decoder with motor control.

Hi Peteski.

ESU doesn't use the "F0 - F4" nomenclature to identify physical outputs on the board, because any physical output can be mapped to any function button.  The NMRA language is quite confusing when you do this - with NMRA, the ditch lights would be F1 and F2 mapped to F6.  With ESU, it's Aux1 and Aux2 mapped to F6.  I think it is much clearer to have different names for the physical outputs and the function buttons - things are less confusing when you say "Map Aux1 to F6" - with this, I know we're talking about mapping the physical output attached to Aux1 to button F6, rather than remapping button F1 to button F6.  The NMRA probably should have done this kind of thing at the beginning, but I don't think anyone realized where this would all end up 25 years later . . .

But be that as it may, yes, this is a six-function board.  Not exactly sure why Aux3 and Aux4 are duplicated, but they are.

As for the photos, this is probably as good as I can do with my 70D's 18-135mm.  What I really need for this is a macro lens, but I don't have one.  I might try with a tripod so I can stop down to F16 or more - that would help depth of field, which is probably why there is a little fuzz.  Taking photos of these multilayer boards is harder than I thought!

John

PS - I went ahead and disassembled mine today, cleaned out some of the excess grease, and then milled a slot in the back of the frame to stash two 220uf, 16v tantalum chip caps.  They fit perfectly there, and the decoder sits over them.  Already have the positive end of the caps wired to the U+ pad; if you can figure out where to solder the negative, I can have it ready to test in five minutes.

[nstars]

From Matt Herman on the Yahoo LokSound list:

The second part of the question about “startup delay” is available now in both the LokPilot V4 and the LokPilot Standard. It is a logic feature called “Virtual Drive Sound”. You must enable this logic feature in the function mapping on the same button as your sound locos. (Usually F8 in North America and Australia) From there you can adjust CV 252 to the amount of time needed to match the motion of the sound loco. If you are using the LokProgrammer, this is the last item in the “Driving Characteristics” menu if a LokPilot File is open.

John C.

That is exactly the way I did it and it works flawless. The only weird thing is that the loksound decoder has a similar setting as the CV 252 of the Lokpilot, but using the same value for the Lokpilot resulted in different delays (at least in my case). After some trial and error I got the same delay and now sound and non sound units start with the same delay. It is definitely worth to do it as it results in a very realistic start.

Marc

[RBrodzinsky]

Thanks Marc and John. I see the setting in both the LokProgrammer and in JMRI. Of course, reading the manual had me totally confused, nothing being said about it there.

[jdcolombo]

Thanks Marc and John. I see the setting in both the LokProgrammer and in JMRI. Of course, reading the manual had me totally confused, nothing being said about it there.

The manuals are out of date - the LokPilot manual dates from 2013; the LokSound from 2012.  Matt Hermann has been begging the folks in Germany to do an update, but it's not high on their list.  Sigh . . . but honestly, I guess if I had to choose between new products and a revised manual . . . I'd take the new products.

John C.

[nstars]

The manuals are out of date - the LokPilot manual dates from 2013; the LokSound from 2012.  Matt Hermann has been begging the folks in Germany to do an update, but it's not high on their list.  Sigh . . . but honestly, I guess if I had to choose between new products and a revised manual . . . I'd take the new products.

John C.

You can find it in the current manual, but I admit that the ESU manuals are not easy to read. Part of it is caused by the complexity of the decoders itself. They can be configured in so many ways.

Marc

[mu26aeh]

Not wanting to clog up the ditch light thread on the SD40-2, I'm wondering if I can use the Atlas ditch light board for the CW40-8's that I brought up earlier.

20170107_200244[1] by Adam Henry, on Flickr

[jagged ben]

Seems dubious.  Among other things, the new IM frame has a clip at the front of the frame, where that board would go, that seems fairly essential to holding the frame together.

[AKNscale]

What if he was able to fit it where the light pipes turn up to go to the board?

[mu26aeh]

What if he was able to fit it where the light pipes turn up to go to the board?

This is what I was thinking

[peteski]

Hi Peteski.

ESU doesn't use the "F0 - F4" nomenclature to identify physical outputs on the board, because any physical output can be mapped to any function button.  The NMRA language is quite confusing when you do this - with NMRA, the ditch lights would be F1 and F2 mapped to F6.  With ESU, it's Aux1 and Aux2 mapped to F6.  I think it is much clearer to have different names for the physical outputs and the function buttons - things are less confusing when you say "Map Aux1 to F6" - with this, I know we're talking about mapping the physical output attached to Aux1 to button F6, rather than remapping button F1 to button F6.  The NMRA probably should have done this kind of thing at the beginning, but I don't think anyone realized where this would all end up 25 years later . . .

But be that as it may, yes, this is a six-function board.  Not exactly sure why Aux3 and Aux4 are duplicated, but they are.

As for the photos, this is probably as good as I can do with my 70D's 18-135mm.  What I really need for this is a macro lens, but I don't have one.  I might try with a tripod so I can stop down to F16 or more - that would help depth of field, which is probably why there is a little fuzz.  Taking photos of these multilayer boards is harder than I thought!

John

I guess using the AUX-x nomenclature does make sense but it is not conventional.  Most (if not all) decoders out there have function remapping and I think it is even part of the NMRA specs.  But the NMRA-type remapping is not totally flexible (there are limits as to what function can be remapped to which function output).

Yes, the AUX3-4 appear to be duplicated but without resistors.

Here is my best guess as to the decoder layout.  Please verify it using ohmmeter and voltmeter before hooking up anything.

[ Guests cannot view attachments ]

I still think the components I drew the purple diode symbols on are the diodes which make up the bridge rectifier.  The "AC" (or track) inputs are the cyan color lines and the positive is red and negative dark-blue lines.  See if your ohmmeter/continuity tester confirms the connections I drew.

I'm not sure about the components circled in yellow and how they relate to the bridge rectifier but they seem to be part of the circuit. I would need to have the decoder in-hand to try to figure out what those are.

The components on which I drew the yellow dot seem to be strangely interconnected ceramic capacitors. Those seem unusual to me too.  Again, I would have to have the board in-hand to try to figure them out.

On the bottom side photo I circled what appears to be the AUX output circuitry.  The orange dots are on the resistors which are in the AUX outputs.  Since there are 5 LEDs on board then all but one AUX outputs are already utilized. Without seeing the board in-person I can't tell which one is the unused one.

[jdcolombo]

Hi Peteski.

I think you're right about the positive/negative.  I can certainly test continuity, but how to I check for the negative output (dark blue line) with a multimeter?  I assume I connect the positive lead of the meter to one of the track power tabs - easy.  Then what?  And what am I looking for?  A specific resistance? Or should I do this with power feeding and look for a specific DC voltage?

Thanks for your help - this is the last piece of the LokSound mystery!

John C.

[RAILCAT]

It would have been so easy to add to pads to the PCB and marked them appropriately
for connecting the keep alive. Is it sufficient to just wire in some tantalum caps without
any other additional components, a resistor and a diode.

[jdcolombo]

It would have been so easy to add to pads to the PCB and marked them appropriately
for connecting the keep alive. Is it sufficient to just wire in some tantalum caps without
any other additional components, a resistor and a diode.

Agree - it would have been pretty easy, particularly if Peteski is right about the negative, because that's a pretty wide trace coming off those 4 diodes (which I agree are very likely to be the rectifier diodes).

I have never used a resistor/diode with my keep alive caps, but I generally only install two, for 440uf of total capacitance.  The reason for the resistor/diode is to limit the in-rush current when the caps are first charged (e.g., when the track power is turned on, or the engine is put on the track) because it is possible that the current in-rush would look like a short to the command station.  With the small amounts of capacitance I use, however, I've never found this to be a problem, with one exception: I have my layout divided into power districts, with each power district fed by a booster (a Digitrax DB150).   On one of the sections, I've noticed that if I have more than 15 sound-equipped locos in that section and there is an actual short, the booster has problems resetting from that short.  It doesn't have any problems when I first power-up the layout, and I can solve the issue by simply turning the booster off and back on.  My other DB150's also don't have this problem - one booster feeds my engine terminal, and I have more than 15 sound-equipped locomotives in that section all the time (e.g., at least 8 steam locos in the roundhouse, and another 10 steam and diesel locos sitting on the ready tracks or in the WB yard).  I don't know if there is some point (e.g., 30 sound-equipped locos?) where the booster would refuse to reset, because it sees the in-rush current of all these combined locos as a short. 

If I had enough room, I'd probably do the resistor-diode thing, especially as I add more sound locos to my fleet; but I rarely have enough room (just finding room for the caps is hard enough).  However, it IS "best practice" to add the diode/resistor if you can do so.  If you're not planning on having more than a half-dozen or so keep-alive equipped locos fed by a single booster, however, I doubt you will have any problem, and you can always retrofit the diode/resistor to the blue wire if you start having problems.

John C.

PS - one of the ways to minimize the amount of room necessary for the resistor/diode is to use surface-mount components on a tiny piece of thin circuit board (single-sided, 1/32").  I've started doing this with my LED dropping resistors, and if you have the ability/patience to solder SMD components in this manner, then they can be stashed almost anywhere.