Remote Control Audio for Rolling Stock

Just curious, why did you run the 70L05 off the output of the main 5V supply?  It would seem to make more sense running off the input, and it would be further isolated by the little switching module.  I can certainly try that, it's easy enough to do.

I have plenty of room for a regulator, I won't be needing the speaker filter.

As far as trace widths, I have a 1/2A max supply, and the module draws about 100ma blasting out audio.  For a 500 ma current and 1mil thickness, required trace width is 4.55mil.  Since I don't see anything close to those currents, I think the trace width shouldn't be a factor.

Here's the Trace Width Calculator I'm using.

Yup, schematic revised. I was so busy trying to edit a pixelated image rather than a schematic...

Well, there's theory and then there's practice.  It's not a question of how much current a given cross-section of copper can carry, or right or wrong.  It's my opinion that if you have the space to run fatter traces for power you should do so.  It makes for more robust traces for prototyping (less chance of nicking or cutting traces as you fidget with modifications), it's easier to mod boards with resistors/caps as there's more copper area to solder to, there's less inductance per inch which reduces glitches for fast current edges, etc. etc.  But to each his own...

I'm going to give the separate regulation a try, that sounds like it may be a big help.  I think I'll stick the 'scope on the power and see what it looks like as well.

I hear you on the traces, nothing wrong with your thinking.  On my little motion sensing board, I did widen the traces to the relay contacts since they are capable of 5A if you want to use them that way.  I'm a ways from actually going to boards on this one, I just do the layout to see how things fit.  I'll revisit the trace widths when I'm actually ready to make a board.

I was thinking about your suggestion to use a processor for the logic, it has some interesting benefits.  One thing it would do is allow me to have both the TMCC and RF version as a single board.  The TMCC serial data would be used for the TMCC version, that could come from a variety of TMCC products.  By plugging in the receiver, I could receive RF data and operate from RF inputs.  That allows more flexibility from a single board design, not to mention minimize the costs of having two different versions.

When you're right.... YOU'RE RIGHT!  That's it exactly.  Powering them with a separate supply and all is well!  I tried a choke and a 220uf cap with a .1uf cap to cover all bases, but with one power supply, they still have an issue.

I looked at the power on my 'scope, and it does dance around with the audio playing, even with a 220uf cap.  That does seem to be the issue.

Time to move on, I think the noise issue is behind me.

Updated for the uP option, looks like everything still fits on the board.

Why use a DIP uC pkg if you have an ICD interface?  Presumably you'll be (re)programming  using the ICD so should never have to remove the chip.

So then I'm thinking if you have more space using a TSSOP, you can easily fit a 20-pin version to get another 6 i/o pins for, say, 50 cents more.  What to do with 6 more pins?!

I still think you need a solution to multiple addresses so you can have more than 1 receiver running per layout.  I'm hoping you're thinking this can be the first revision toward something you can sell to bring this to the masses.  Perhaps using the flash/EE on the uC, you can come up with some kind of "learning" mode where the receiver learns its address and stores it in non-volatile memory...rather than having a bunch of jumpers of DIP-switches to set the address.

A downstream option is to accept the 99 cent receiver module's 3-signals: 5V, GND, DATA.  Not so much the cost, but the space would be greatly improved for fitting into rolling stock if plug-in rcvr board was half the size of the 4-channel rcvr.  Yes, would require a bunch of code to decode the bit-stream.

I'm imagining some output pins to drive other functions besides sound.  Whether it be some LEDs, or a coupler coil, or a PWM'd hobby servo, or fill-in-the-blank.  Perhaps a gross violation of KISS...

I'm imagining at least one input pin to allow something to trigger a sound locally.  In other words it can be playing sound #1, but when it gets a trigger it goes to sound #2...or something like that.  Kind of half-baked in my mind, but I'm thinking of the animal sounds that trigger when there's motion.  eBay has low-cost motion/vibration detection modules. 

But I do grasp the rapid progress being made!  As you well know in a for-profit corporate world this would take forever.

Funny you should mention that, I was just looking at the addressing and thinking how I might address it.  I agree that allowing more than one working at the same time is a very desirable option, so that's got to be incorporated.  Looks like I'll have to have some connections to the unused pins on the receiver over to my board.

As far as triggering sounds locally, if there were a way to mix audio channels, that would be a more attractive option.  That's getting way beyond what I am going to do here.

One reason I went to the DIP was easy of hand assembly, I'm getting tired of soldering tiny little leads.  I'd probably with a surface mount, but not the TSSOP as that's pretty small.  I just added the ICD interface after I put the chip in, I was initially thinking I'd pull them out and program them on my debug board.  Then I decided to stick the ICD in, so it probably makes sense to go with a smaller footprint SMD option, the SOIC is probably good and pretty easy to handle as well.

I did think about other functions as well, but I'm leaning to getting the audio function working as a first step.  I'm not ready to spend the time to decode the bit stream since someone already did the work.  One thing that the 20 pin chip would do is allow me to have the TMCC serial input on the chip as well, but then I have to open that can of serial decodes.

Truthfully, I'm not too worried about the size of the module with the receiver plugged in, it'll be less than 3" long and 1" wide.  Even a small boxcar will fit that easily with a speaker.

Understood.

I think it be sloppy to jumper over multiple "address" lines from the 4-channel rcvr module to your PCB.  My 4-channel rcvr uses the SC2272-M4 decoder chip; the SC2272 comes in a few variants for latching, momentary, etc. operation.  If you look at the 4-channel rcvr compared to the barebones rcvr, it's pretty much "just" the SC2272 that makes it different.  Hence, one option is to use the barebones 99 cent rcvr and put the SC2272 on your PCB.  I can't speak to the incremental PCB cost for the additional real-estate but the chips themselves are insanely priced on eBay for about 25 cents!

Then you plug in the 4-pin (3-signal, 5V, GND, DATA) barebones rcvr and feed the DATA line to the SC2272....AND have access to the addressing pins on the 2272 to connect to address switches, jumpers, or even spare uC output pins (if you choose to save the address in the uC EEROM).

Edited=no direct links to ebay!

Just a thought....

At this point, I'd like to get something going.  It only costs me $36 to get a few PCB samples of a given layout, so I want to arrive at a workable solution and pick up a set of boards and get some miles on it.

I could engineer this forever, but we used to joke about that when I was actually working for a living.  If I let the guys go, they'd "improve" on the design forever, but we wouldn't have a product to ship, and the costs would go through the roof!  

FWIW, I have a couple of the bare bones transmitter and receiver sets, did you ever determine if they had as good a receiver as the 4-channel remote?

One thing I'm waiting on is I ordered one of the superhet 4-channel sets, if that's vastly superior in range, I'd probably want to use that.  It happens to be pin compatible with the 4-channel ones I have now, so I could just drop it in.

John,

Have you looked at Adafruit:

https://learn.adafruit.com/ada...sound-board/overview

http://www.adafruit.com/search?q=pam8302a

I bought a solder sucker from them a few weeks ago and their site has some interesting stuff.  One of the posters on the Free Rails forum gave me these links.  I was looking for steam sound but he said the sound was not adjusted by speed of the engine.  Still they may have something useful.

Bob, I haven't seen that before, but the MP3 module I have I get for $7/ea, and I can play up to 5 sound files, each with a single button press.  I can also connect a serial stream and control many more sounds.  It includes a 3W mono speaker amp and dual stereo headphone or external audio outputs.  Since the long sound files I'm going to be using for rolling stock are in the 40-50 megabyte range, the limited capacity of this board is too restrictive for my use.  This one accepts SD cards to allow any capacity I desire.

I don't know of anyone short of the actual model train folks that have sound synchronized to the speed of engines or rolling stock.

No, I have not compared the bare bones vs. 4-channel super-regen rcvrs.

In looking at the super-het 4-ch rcvrs I noticed some have "learning" mode.  Looks like OGR is now enforcing no eBay links but easy to find.  Is this the one you got? 

I figure you press the button and then send a command within x seconds and stores away the other 8 address bits (if it was a 10-bit address)....no address pin/jumpers/switches required.  Looks like an LED next to the button which presumably flashes to provide feedback on learning mode.  Makes sense and would eliminate running extra address wires to your board.  Presumably this uses the 2262/2272 chipset serial protocol or obviously would need to get the mating 4-button remote.

Moving on then...here's my next issue.  Handling track power dropouts. 

With peak speaker currents of, say, 5V thru 8 ohms, that's over 1/2 Amp.  I was looking at your module pinout at it does not look like there are separate logic voltage and power voltage inputs.

Since I have a separate audio-amp module, I can split off (diode) the power to the audio-amp.  The MP3 electronics would have a cap with enough energy to ride-out a X second dropout without resetting or whatever.  Obviously the speaker output would dropout and there may be a pop or click on each transition.  Anyway that's my thinking right now.  I guess a big enough supercap is one approach but would need to run-the-numbers so to speak.  This dropout could obviously be from a direction change for conventional operation though I realize this could be a command-mode-only design.

What's your take on this?

I'm going to see what experience tells me, but track power interruption is certainly a potential issue.  I did look at how I could add a superCAP.  I'm thinking it could get a bit complicated to do this properly, probably a two-stage system.  Regulate track power down to 10 volts, put a couple of 5.2V 2.7F superCAPs in series to provide the power, then regulate down to 5V.  I don't think just putting the cap across the 5V output will cut it, though that might be a band-aid when I first try to run these.  I suspect it would provide a short time power bridge, but you need more voltage to use a decent amount of the stored energy.

I have no way to regulate the volume with the existing module, so I experimented with series resistors.  Right now, the volume with an 8 ohm speaker (ERR model) is pretty loud, maybe too loud for many applications.  I can say that volume won't be an issue!  I think the maximum power is used with a 4 ohm speaker, but I find the 8 ohm to be more than loud enough.  I'm adding a provision to use a resistor to control volume.  You could use a 50 ohm pot or just solder a selected 3W resistor directly to the board.  I found that 30 ohms dropped the volume by about 1/2 the apparent loudness, and was probably about as loud as I'd want it.

That looks like the superhet one I have coming, if it works out well, I'll consider that as the "standard".  I did observe that it has the learning provision.  One issue is how you do the transmitter, do you need a different transmitter for each of these receivers?  With the other one, you can change addresses at the transmitter side to address multiple units.

As far as addressing, if you look at the back of the super-regen receiver, you'll see they have already conveniently placed ground and 5V bus bare traces to easily jumper addresses.  For the first cut of this, that's going to be the addressing provision.

OGR is really starting to get a bit absurd, I think the commercialization is going a bit overboard.  The 'Bay links to stuff are a valuable resource.

Dear Stan2004 and JOhn

I have been following as best I can with my limited knowledge of electronics. I a, learning a great deal thank you to both of you. When did they start the no e-bay link, this is the first I have heard of this.

I have no idea.  All I know is a link was removed by someone from OGR yesterday.

As it says on their terms of service:

OGR Publishing, Inc. reserves the right to delete any message, at any time, for any reason.

"for any reason" is clear to me and the OGR person at least had the courtesy to say "no direct links to ebay".  So be it, and I figure we should just leave it or else this entire thread might get deleted!

Yikes.  Perish the thought of having to add a 3W resistor!  The MP3 module I'm messing with has a non-volatile volume-control-button setting so I figure you set it once and forget about it.  But in your case I suppose you could complicate matters and add an external audio amp WITH volume control? 

This would also then allow splitting the lower current "logic" supply from the higher current "power" supply.  So this could be driven by the DACR and DACL outputs. I understand KISS, but there is something un-American about just putting in a power resistor to simply convert excess volume into heat.  IMO that is.

I suppose one benefit is the spare audio channel is now buffered with higher current drive capability which could, for example, drive a motor or LEDs or who knows what. 

I was thinking more about your use of "just" a 7805.  So with 20V DC on the rectified bridge output, and (for the sake of argument) a 1 Watt audio load on the 5V regulator.  1 Watt = 5V x 200 mA average current...so the 7805 is burning 15V x 0.2A = 3 Watts.  . 

So further throwing KISS out the window, I suppose a 99 cent step-down regulator integrated with your PCB and maybe even a provision to charge a superCap as you suggested for the recent thread on the Dark Trolley station stop.

If I do anything regarding the volume, when I round up the serial control specifications for the module I have, I'll just control it that way.  I've written the vendor I bought them from to get the serial interface specifications. 

Truthfully, I have no interest in adding yet another amplifier, even though it might be a way to solve this particular issue. 

As far as "just a 7805", look closer.   I'm using the Recom R-78E05 there, it's a small switching supply that is a replacement for the 7805T three-terminal regulator.  It will handle 200ma without breaking a sweat, it barely has any temperature rise with that current.  It costs about $2.50 in quantity, and it's the same one I use for the Super-Chuffer, so I buy them in quantity.  I started using that when I couldn't get enough power out of the 7005T without a heatsink.  I also figured that the total module power would be a lot less, and it is.  I'm not reluctant to spend a couple of bucks on a good solution for power like the Recom switching module, it's been very reliable for me on the Super-Chuffer and I haven't had any power issues.

I'm going to build a "simple" RF controlled MP3 player first, it'll be my test bed to work out any kinks.  I want to have a couple running around and see if my radio choices are really practical in the real world.  If that all works out, then I'll consider the larger 20-pin PIC part and do the TMCC serial, the RF interface, and the serial interface to the MP3 module for an all-inclusive module.  At that point I might consider moving the SC2272 onto the board in one of the SMT versions and using the small bare-bones receiver.  That would probably be the one that I'd try to make into a product, with enough flexibility, I may be able to get enough quantity to get them built.  I have to get to around 100 in an order to get the price down to a reasonable level, smaller quantities kill you for parts, PCB blanks, and assembly services.  

As far as "just a 7805", look closer.   I'm using the Recom R-78E05 ... the same one I use for the Super-Chuffer

Roger that.  Here I am holding your Super Chuffer board in my grubby hands and whadda you know!

Sounds like a plan.  I believe the big opportunity is for general purpose (as opposed to just audio/sound) remote control of rolling stock.  So whatever you do I implore you to keep in mind all the animated/activated rolling-stock cars that , in today's world, only work when positioned on an activation track, or runs all the time, etc. etc.

I'll certainly keep the extra controls in mind, but it may not be as easy to do some of the stuff.  Lights or couplers are pretty easy, but stuff like animation would probably be all over the map.  I have used a couple of the model airplane servos for that task, but you have to tune the PWM to the specific application.  Unless I have a way to calibrate the output, that would be problematic.  Obviously, a simple digital on/off control is no problem, and perhaps several outputs that can control a variable speed motor would be useful.

John and Stan (and other contributors) - this - as usual - has been very educational and entertaining (and way beyond my electronics knowledge at this point).  But two (perhaps random) questions - at some level, it sounds like you two are reinventing DCC (a topic about which I know precious little at the detailed level).  I recall a few years ago seeing a video on the Yahoo S-gauge forum where the proud layout owner (and a nice layout it is/was) showed how he was able to operate different unloading cars essentially anywhere on his layout via a DCC system with the appropriate DCC hardware mounted in the various action cars.  The cars operated without need of the usual 'accessory shoe/rail'.

Now that I've said that - I think John said something earlier in the thread about using TMCC, which I guess would be similar.  Remind me - is the idea of using an RF control a more 'universal' solution - meaning those of us without TMCC or Legacy would be able to use it...??  Or is it more the problem of coming up with a robust TMCC kluge/hack?

A second question - a repeat of someone else's observation earlier in this thread - is it feasible to use a blue-tooth based system, with the advent of blue-tooth enabled speakers (or, presumably, rolling your own blue-tooth circuit board connected to a remote/mobile speaker)??  Might there be some advantage to having a 'fixed' audio source, with only the speaker end of the system mobile??

You are correct, since we're running command systems that are not compatible with DCC, so one solution is to use TMCC.  That would be compatible with TMCC/Legacy or DCS as they both run with straight AC track power.   The RF solution is the "universal" solution that can be used with DCS or TMCC, or indeed with conventional operation.  Currently, there is no DCS receiver capability to do a strictly DCS system, that's one significant hole in the MTH product line IMO.

I'm thinking in the near future that I'd probably merge the technologies and have RF and TMCC all in one package and be able to select between them.  I'm also waiting on several different RF packages with what may be higher quality radios for longer range, that will figure into what I end up building.

I saw the mentions of BlueTooth, but I want to have the ability to have perhaps five or ten pieces of rolling stock on one layout, all with unique audio files playing.  That seems like it would get a little tricky without writing a pretty involved BT application and having a computer to run it.  Of course, then you'd also have to have the BT host in the picture and be controlling the audio from that platform.  If I were only doing one channel, the BT might be a workable solution.  Even with multiples as I'm talking, I suspect it can be done with BT, but I know how to do it with the technology I'm using.

I've also seen other references to mesh networks like Zigbee, but having no idea of how complex that would get, nor the cost, I'm probably not going to invest time and energy in chasing that.

I'd say that hits the nail on the head. The hobby is great because it's a big tent. There are many ways to participate, enjoy, spend time (and money), etc.  I find the electronics aspect educational and entertaining, and no doubt there are countless times where I have "re-invented the wheel" so to speak. 

What's new are these insanely (low) priced electronic modules from Asia allowing  Lego-like construction of gadgets useful to our trains. It's not quite there yet requiring component-level "glue" to make them work together and that's how I see this particular project.  In some cases we need to change the glue, in other cases we need to change the pieces.

What's also new is forums like OGR to engage with like-minded folks. Not so long ago discussions like this were limited to the knowledge of the guys at the club, or to the Radio Shack clerk if you were lucky, or thru a snail-mail dialog with a magazine.  So keep your ideas (and questions) coming!

I now step down from the soap box.

GRJ, regarding your separate 5V supply, I notice the uC powered from U2 rather than the main 78E5.  Yes, the uC draws little current and should be relatively quiet on +5 but I'd think whatever RF module you end up with would be better off with a dedicated supply.  Stated differently, unless you're using some analog functions on the uC, I don't think it will be bothered by the 5V audio-side noise.  Just a thought...

The uP doesn't draw much, and I figured I'd rather not have it seeing the rather large swings I saw on the 5V supply with the MP3 player.  Note that I haven't powered the MP3 board with the switching regulator I'm using yet, I have to do that and see how well it regulates.  I think I have a better chance of it working well without the uP on the noisy MP3 supply.

I do note that with an 8 ohm speaker, the MP3 module is only drawing around 100ma on average, so I'm not getting nearly 2-3 watts to the speaker.  That's fine as it's already too loud using the little 1 1/2" ERR speaker and baffle.  Since I get one of those with each ERR RS-C I buy, and I don't often use them in the installations, they're piling up and will be just the right size to put into rolling stock.

I did go in and change the trace width from .006" to .01", that's twice the trace width "required" for half an amp of current, so I think it's conservative enough for what's going on here.  Let's be pessimistic and say we have 200ma, the computed trace width is 1.28mils, and I'm at 10mils.  Even allowing for parallel runs that increase the requirements, I think it's pretty conservative.  In addition, as I said, I really only see around 100ma to the board, so there is only about 1/2W of power available.  I suspect virtually all of it goes into the audio.

Regarding dropouts, if 200mA is indeed pessimistic then maybe a single 5V superCap might be OK (vs. stacking 2).  So at 0.2A, a 1F cap will discharge at "only" 0.2 Volts/sec - and a 2.7F less than 0.1 V/sec.  Your MP3 module says operation down to 3.6V and surely any modern uC will be fine at 3.6V.  So if you can get it charged up to anywhere near 5V, you have several seconds over which volume will fade slowly but I'd think imperceptible for short-time dropouts over dirty rail/switches etc. 

Your regulator datasheet specifies a max load capacitance of 220uF which I'm guessing is for load stability in its feedback loop.  So if I were designing in a superCap on the output, I'd want to know a bit more about why and hence the effect of isolating the capacitance with, say, a charging resistor which is what you'd probably use.  Also, I if you're isolating the superCap I figure the incremental cost of a Schottky diode may be be worth it to get another 0.4-0.5V of starting voltage.

Another idea is now that you have a uC in there, perhaps you could use its Brown-Out detect circuit at ~4 Volts or whatever and send a command to the MP3 module to mute.  I know you're not trying to squeeze every last penny out of the design but those superCaps are somewhat spendy relative to the other devices.  In which case a lower cost electrolytic maybe on the input side of the regulator could store enough energy to handle 0.2A of 5V audio for, say, 0.1 sec after which audio volume is cut reducing current to, say, 0.02A (10%) keeping the digital logic out-of-reset for a while longer in anticipation of voltage recovery so that the song doesn't reset/restart. 

Let's see. 1000uF cap charged to 20V = 0.2 Joules.  Let's say 60% of that is available after the step-down switcher so 0.12 Joules.  At "pessimistic" full-current load you need 0.2A @ 5V for 0.1 sec = 0.1 Joules.  So it's at least in the ballpark.

Are we having fun yet?

I did a brief measurement of the MP3 player at full volume on the bench, it varies between 100 and 110 ma on the 'scope.  That's playing an audio file with a rather high percentage of "noise", some of the quieter files actually run in the 70-80ma range, and I'm using an 8 ohm speaker.  When I insert some resistance (the only method without serial communication to the module) in the speaker leads, the current drops way down in the 50-60ma range.  I find a 30 ohm resistor still gives me probably as much volume as I'll want for most applications.

Given those power figures, let's do some calculating.  Figure the switching regulator at around 80% efficiency, which is below it's lowest specification.  We have at least 20 volts on the input with the rectifier and the cap.  So with the existing 330uf cap, we have about .066 Joules.  If I figure 100ma load for .1 second, that's .05 Joules.  So, the way I see it, I already have some protection against dropouts.  If I drop the volume as I suspect I'll have to do in most practical installations, I have even a bit more margin.
I know that the Super-Chuffer has only 1/2 wave rectification, and I can draw 100ma from that supply with a cab light, headlight, and the smoke motor, etc.  It has no problem with dropouts running normally, so I think I'm probably not going to have that bad a time with it.

In the future when I have serial communication to the MP3 module, I can do something a bit fancier.  I'm still looking for information on serial communication to the BY8001 module, that's a pacing item.

Yep, we're having fun!

If you don't get a response on the serial commands from the seller, I found this recent Arduino discussion which seemingly has what you need if you can read Chinese!

http://forum.arduino.cc/index.php?topic=306442.0

That resistor is looking better and better!

You said it!   I'm trying to get an English datasheet, but it's an uphill battle.

GRJ, what about going back to the most common denominator and thinking inside of the box.

With all of the effort expended to use a remote, why not use Rare Earth Magnets (REM,o)oriented for effect, strategically placed on the track. A reed switch with equal orientation, triggered by passing over the (REM,o)  could supplant a remote. After all, using a remote, there are just four functions, in your design, at this time. One is an OFF function that stops any other sounds. Three others are regional sounds' on the track, like Station sounds, Mountain sounds and use your imagination sounds. Others could take advantage of this angle by placing the REM,o magnets wherever they want and tailor their sound-scape to the present 4 channel application.

I sense that I am going to be fired from this forum for posting this. Since I will miss all of you guys more than you would miss me; and I will miss you all for sure, I will leave my account open in case John & Stan didn't completely finish me off.

Having been keeping up with this thread from the beginning, but not having anything to offer that wouldn't add complexity, haven't had anything to say.  This post by Up with UP got me thinking, however, that a little complexity might be fun.  

Rather than using the radios, or the simple magnetic switches, you could use a RFID reader and and place RFID tags where you want sounds to start and stop.  I thought of using this a while ago for other uses, but it had not occurred to me to use RFID tags to turn on and off sounds.  this would allow a practically limitless number of files to be used once you have the serial data figured out.  and furthermore to allow the sounds to be played at the appropriate location on the layout.  

As for radios, I'm partial to the very inexpensive 2.4GHz type, but these get a bit tricky in the coding end if you plan to use more than 6 devices.  

I'm sure the concept can be expended in the future, but I'm concentrating on the remote aspect right now. 

When I manage to get the serial data specifications for the MP3 module, then I could have practically limitless sound clips.  I actually have some RFID stuff that I was playing with here, that would allow you to trigger specific clips at a chosen place on the layout.

Truthfully, a combination of location triggered sounds and remote triggered sounds is the best of both worlds.  While we're at it, maybe an accelerometer to trigger curve specific sounds.

One of the hardest part is getting appropriate sound clips for all the things you might like to do with the sound modules.  I might have to have a website to start collecting all the various sounds.

Just found and read through this topic this morning. I have been considering a project like this for well over a year now, but have no where near the electrical experience or time to get deep enough knowledge to make it happen. The possibilities that these kinds of projects open in the world of O gauge are HUGE. Even larger if down the road you introduce logic circuits into a board like this. I believe there is a market for this product and I also believe that the core board is not the only marketable part of the product. I'm interested to see where this goes.

It'll be interesting to see where it goes.  I've been thinking about something like this for some time as well.  Recently, as the quality MP3 player modules started showing up, it seemed like a more practical project, so I decided to jump in and do some tinkering.

In the long run I see this having a core microprocessor and multiple activation methods.  I like the RFID idea, that allows position dependent triggering of the sounds.  The remote triggering gives you ad-hoc sounds wherever you are on the layout.  Finally, other sensors can be used, as I mentioned above, for specific sounds on curves or when running or stopped.  I'd like to have it compatible with TMCC control as well as the RF option.  That allows it to be used for TMCC/Legacy, DCS, or even conventional operation.

As you say, endless possibilities.

 This thread has been very interesting to me.  I have done 4 cars now with sounds, but I am limited to one sound per speaker - because that is the technical knowledge I can work with.  with some of the plans you have here my sound car fleet will get much more impressive

I have some prototype boards coming, so it'll be interesting to see how it shakes out.  

Guns,

   You get this worked out, it looks like a money maker to me, hope you make a fortune with it.

PCRR/Dave

Stan, where did you see this?  Can you provide the auction ID?

Not at all!  This is what this forum is all about. 

You could simply bypass the RF portion of circuit and trigger the MP3 module with your oriented reed switches.  So now you can start (up to) 4 unique sounds based on postion on the layout.  There would then be a 5th reed switch orientation that stops any sound.

As I see it, in the big scheme of things it's the recent availability of these DIY insanely (low) priced MP3 modules that is the windfall to O gauge trains.  The engine manufacturers have done a pretty good job with sound quality but the accessories and rolling stock have just not kept up.  And not just the sound quality but the length of sounds.  For years the "standard" was a set of IC chips with greeting-card voice-quality sound that came in increments of like 10 seconds.  A 2 minute chip was "big" and had a big price.  Now you get hours of music-quality sound for less $.

But beyond the triggering there are many other details that GRJ has worked out to install the sound module into a piece of rolling stock...such as managing the whole power supply issue and that pesky AC-voltage used in O gauge. And he openly shares how he is doing it so anyone can decide if it's something they want to undertake or use as the basis for their own project.  That's a real contribution to the hobby.

321685761503

I sure hope the sensitivity is -107 dBm (not +107 dBm)

Interesting, it's a four channel receiver, and a two channel transmitter.

FYI, here's the one I bought: 131512680390

I figured I'd have plenty of transmitters as it's the standard decoder on the receiver.