ESU new Next18 Nano decoder

[reinhardtjh]

I thought that functionally both LokSound 5 micros and LokSound 5 nanos were identical. They are both 5th generation LokSound. Are you saying that the nano so more feature-rich with more advanced sound capabilities?  I don't see any such differences in the LokSound manual.

As for the AUX outputs, since they use Next18 connector, both will have the same number of AUX outputs on the connector (plus whatever solder pads are on the PC board.

I'm with you. I thought it was the same hardware in different format.  But unless ESU has a typo...

LS 5 Micro DCC
https://www.esu.eu/en/products/loksound/loksound-5-micro-loksound-5-micro-dcc/

Sound

The LokSound 5 micro decoder can reproduce up to 10 channels simultaneously. Each channel offers up to 16 Bit / 31250 kHz and thus we have finally achieved hi-fi sound quality on your layout. For all practical purposes one cannot detect any difference to the original. A double class D audio power amplifier with up to 3W output power feeds the speakers with a permitted impedance of 4 – 32 Ohms. The huge 128 MBit sound memory assures sufficient storage capacity.

LS 5 Nano DCC:
https://www.esu.eu/en/products/loksound/loksound-5-nano-dcc/

Sound

The LokSound 5 Nano Decoder can play up to 12 channels simultaneously. Each channel can be resolved with up to 16 bits / 31250 kHz and offers hi-fi quality on your system. There is virtually no difference to the original audible anymore. A Class-D audio output with up to 3W output power controls the speakers, which may have between 4 ohms and 32 ohms impedance. A 128 Mbit sound memory provides enough capacity.All individual noises can be adjusted individually in the volume. The super-flexible sound engine without a rigid schedule allows the exemplary simulation of all conceivable rail tracks.

Checking the others...

12 Channels:
LokSound 5 Fx
LokSound 5 micro DCC Direct Atlas Legacy
LokSound 5 micro DCC Kato
LokSound 5 micro DCC Kato USA
LokSound 5 micro DCC Kato USA Widebody - Nano on a board
LokSound 5 micro DCC Direct Atlas Legacy - Old Atlas/IM

10 Channels:
LokSound 5 DCC Direct - I wonder if that's old copy because it's a Nano on a board so shouldn't it be the same?
LokSound 5 DCC Direct with integrated PowerPack - Same as above, Nano on a board
LokSound 5 / LokSound 5 DCC
LokSound 5 micro DCC Direct (Atlas/IM drop in)
LokSound 5 L / LokSound 5 L DCC
LokSound 5 XL


In general, it seems to be the more recent LS 5 boards that do 12 channels.  Maybe it's a factor of a part being replaced and it happens to be more capable.

[peteski]

Interesting.
Of course none of this is mentioned in their English version of the decoder manual.  Well, to be honest, they really don't seem to cover the specifics of the sound circuitry.

I also wonder if the "12 channels" is simply a typo.  I'll have to check what they used for microcontroller in each type of decoder.
I also don't know how existing projects will utilize the extra 2 channels.  Not that I'm an expert in creating sound projects.

[kiwi_bnsf]

Interesting.
Of course none of this is mentioned in their English version of the decoder manual.  Well, to be honest, they really don't seem to cover the specifics of the sound circuitry.

I also wonder if the "12 channels" is simply a typo.  I'll have to check what they used for microcontroller in each type of decoder.
I also don't know how existing projects will utilize the extra 2 channels.  Not that I'm an expert in creating sound projects.

12 Channels vs 10 Channels is definitely not a typo. UPDATED - SEE FURTHER POST BELOW WITH CORRECTED INFO

I ran into this when customising my Sound Projects to support the ProtoThrottle.

I developed a number of sound slots for logical functions: 5-step variable brake, custom Brake release, custom Reverser Center, custom Start Delay, custom consist-aware independent front and rear lighting (and others).

Combined with the fact that ESU hard-code sound slots 1-3 as reserved channels, I quickly ran out of the channels. This led to several weeks of refactoring my logic into fewer sound slots (and therefore fewer active channels). I had to get reasonably creative with sound priority, and using function maps to subdivide logical sound slots to only be active when absolutely necessary. Finally, I had to make all my low priority sounds random so that they automatically recover from being cancelled due to a channel max condition in a way that isn't too obvious to the ear.


Thankfully I used the ESU 58741 in most of my locos with 12 channels. It's only my ScaleTrains factory Loksound V5 Micro Next18 decoders that I will swap out (and ScaleTrains has already moved to Nano E24s on their new releases).

UPDATED: Further testing with current V5.2.6 LokProgrammer software and the later V5 firmware shows that the limit is indeed 12 channels on the ESU 58828 V5 Micro Next18 today. The same as the Nano.

[peteski]

So the extra channels do show up in LokProgrammer?  I have not ventured into sound project design/editing, but I'm curious.
But OTOH, what happens when you install that 12-channel sound project on a LokSound 5 decoder which only has 10-channels?
It will not install, or will just be missing some functionality?

[kiwi_bnsf]

So the extra channels do show up in LokProgrammer?  I have not ventured into sound project design/editing, but I'm curious.
But OTOH, what happens when you install that 12-channel sound project on a LokSound 5 decoder which only has 10-channels?
It will not install, or will just be missing some functionality?

Channels are not visible in a Loksound project.

The channel limit applies to how many sound slots can be active simultaneously (and sound slots can be audible or logical or both).

If you have a Loksound Micro V5 then you can have a budget of 12 sound slots that can be active at once. [CORRECTED]

If you have more than 12 slots go active at one time, the 13th sound slot is cancelled. Which slot is cancelled is based on its priority as defined within the sound slot (where 15 is the highest and 0 is the lowest). Sound slots 1,2,3 are reserved (typically used for Prime Mover, Horn, Bell) and therefore are considered always high priority, so you have an effective budget of only nine more sound slots that can be active simultaneously.


Most people would probably never run into this issue, but I learned it the hard way. My logical functions would intermittently stop working — like a brake lever not responding when another sound was playing. Or my lighting stopping responding.

The channel budget really forced me to think about how to use sound slot priorities, and how to architect sound slots not to be active except when a very specific function and logical conditions were active (e.g. the reverser sound slot does not need to be active when moving). It also forced me to consolidate several logical functions into single sound slots that contain multiple modes (activated by SHIFT conditions, or Share2.x interactions with other sound slots).

Finally, this issue made me realise that you have to really think about what a custom sound slot does when it gets cancelled — does it resume or abort, and does it flow back to the same logical state even if all the conditions that initially trigger it are no longer present?

I ended up moving a lot of low priority sounds to be random rather than being always-on while containing random delays. This means they consume a channel only while active, and when they get cancelled, they come back randomly a short time after channels become available again.

Hope this helps explain channels vs sound slots.

[milw156]

Channels are not visible in a Loksound project.

The channel limit applies to how many sound slots can be active simultaneously (and sound slots can be audible or logical or both).

If you have a Loksound Micro V5 Next18 then you can have a budget of 10 sound slots that can be active at once.

If you have more than 10 slots go active at one time, the 11th sound slot is cancelled. Which slot is cancelled is based on its priority as defined within the sound slot (where 15 is the highest and 0 is the lowest). Sound slots 1,2,3 are reserved (typically used for Prime Mover, Horn, Bell) and therefore are considered always active regardless of their priority, so you have an effective  budget of only seven more sound slots that can be active simultaneously.


Most people would probably never run into this issue, but I learned it the hard way. My logical functions would intermittently stop working — like a brake lever not responding when another sound was playing. Or my lighting stopping responding.

The channel budget really forced me to think about how to use sound slot priorities, and how to architect sound slots not to be active except when a very specific function and logical conditions were active (e.g. the reverser sound slot does not need to be active when moving). It also forced me to consolidate several logical functions into single sound slots that contain multiple modes (activated by SHIFT conditions, or Share2.x interactions with other sound slots).

Finally, this issue made me realise that you have to really think about what a custom sound slot does when it gets cancelled — does it resume or abort, and does it flow back to the same logical state even if all the conditions that initially trigger it are no longer present?

I ended up moving a lot of low priority sounds to be random rather than being always-on while containing random delays. This means they consume a c3rdhannel only while active, and when they get cancelled, they come back randomly a short time after channels become available again.

Hope this helps explain channels vs sound slots.

Canyon post a screenshot of your function mapping screen as an example?

[Steveruger45]

 Regarding 10 channels vs 12 channels.  Didn’t ESU change some components due to warranty issues?   Could that be the reason for the channel difference?

[peteski]

Hope this helps explain channels vs sound slots.

Yes, thank you!

That looks similar to what occurs in musical electronic keyboards. I has an inexpensive synthesizer which only had 4-channels.  If I simultaneously played 4 notes, it would sound ok, but if I played a 5th note on the keyboard it would not play until I released one of the other 4 notes I was playing.

[peteski]

Regarding 10 channels vs 12 channels.  Didn’t ESU change some components due to warranty issues?   Could that be the reason for the channel difference?

You might be thinking of a substandard voltage regulator chip on LokSound 5 decoder which was causing lots of failures of new decoders. That is not the same as the microcontoller chip (that's where the channels reside).

Out of curiousity I checked the microcontroller chips on both micro Next 18 and a nano E24.

The 58828 micro Next 18 decoder's microcontroller chip is marked:
ATMEL
ATSAMC21
J17A-U
2224 F TW
2244MR  ARM

The nano E24 decoder's microcontroller chip is marked:
ATMEL
ATSAMC21
J17A-U
2203 F TW
20314P  ARM

So as far as I can tell, the chips are identical on both decoders.  I believe the last 2 lines on each chip that are different denote date code and other manufacturing info.  I'm not sure how one can have 10 channels while the other one 12.

[nstars]

I think it’s rather surprising that soundslots are used by none sound features like lighting and consist. 10 or 12 soundslots do sound enough but this way i can understand running out of soundslots.

Marc

[kiwi_bnsf]

Out of curiousity I checked the microcontroller chips on both micro Next 18 and a nano E24.

The 58828 micro Next 18 decoder's microcontroller chip is marked:
ATMEL
ATSAMC21
J17A-U
2224 F TW
2244MR  ARM

The nano E24 decoder's microcontroller chip is marked:
ATMEL
ATSAMC21
J17A-U
2203 F TW
20314P  ARM

So as far as I can tell, the chips are identical on both decoders.  I believe the last 2 lines on each chip that are different denote date code and other manufacturing info.  I'm not sure how one can have 10 channels while the other one 12.

As is so often the case with threads on The Railwire, this one has caused me to question my own thinking!


When I first ran into the problems with channel budgeting on Loksound V5 decoders, my source of information was the excellent Loksound Group https://groups.io/g/Loksound/topics

This led to me understanding that all sound slots that are active use up a channel (not just ones that play a sound). This makes sense as even a logic-only sound slot can contain a lot of complexity and I suspect that the channel limit is a CPU and memory constraint.

I confirmed this at the time by creating a diagnostic sound project with twelve logic-only sound slots that simply enter a loop when the corresponding Function is active. I set these sound slots all to the highest priority of 15, and mapped them to F11-F22. Then I added an air compressor on F10 with the lowest priority of 0.

With this project loaded on the ESU 58741 V5 Kato decoder, I observed the compressor sound to cut out when the 12th logical function was activated. I tested it on a 58821 and have a recollection of it cutting out after the 10th logical function was activated — I wrote myself a note at the time to consider upgrading all my 58821s in the long-term.


@peteski's observation today that the Loksound V5 Micro and Nano use the exact same ARM controller got me thinking as to why would there be a difference in channels?

Well I just loaded my diagnostic channel test project back onto a Scale Trains Dash 9 equipped with an ESU 58828 V5 Micro Next18, and this now happily supports 12 channels.

I suspect the older Micros may have benefitted from firmware updates somewhere along the way. My first batch of ScaleTrains Dash 9s had 5.1.85 firmware from 2019, but that's now been updated automatically to 5.9.160 (with fancy tone control and many other new features). Either that or I just remember it wrong.

I've just attempted to do a regression test, but the oldest available LokProgrammer 5.1.5 still includes newer firmware and will not launch if I delete any of the firmware files (and you are forced to update decoder firmware when you write a sound project).


Overall this is very good news — I don't have to upgrade any of my Loksound Micros to Nanos to get 12 channels!   

I've just amended my original post on this thread so that no one is misled by the 10 channel limit for a V5 Micro. ESU could update their website, but not everyone has a LokProgrammer and can update firmware, so perhaps they are being conservative?


FYI, the Loksound V4 Micro is definitely limited to 8 channels — but that makes sense given that its an earlier generation of chip.

[kiwi_bnsf]

I think it’s rather surprising that soundslots are used by none sound features like lighting and consist. 10 or 12 soundslots do sound enough but this way i can understand running out of soundslots.

Marc

Just to clarify: a Function that is not mapped to a sound slot does not consume a channel when active. You can have as many Physical Functions active as you like (including all the cool logical modes that ESU offer in the function mapper, and in the physical output configurations).

It's only when you are mapping a Function to a sound slot (either to control a sound or provide logic) that it consumes a channel (and only while active).

For example, the default ESU Starting Delay sound slot on their V5 projects will consume one channel while the loco is stopped because it is active and looping at speed=0 waiting to apply the start delay logic. As soon as the loco moves and the delay timer expires, this sound slot is no longer in an active state, and the channel is released.

Hope that helps.

[peteski]

Ah, the TRW think-tank at work.  Excellent research Tim!  It does make sense that the changes were (features added) in later version of firmware.  The upgradable firmware is a nice perk of ESU decoders (like the recently added tone control), but it also can create confusion or misunderstandings. Especially when the documentation is not all that clear.