My enormously complicated, enormously fun DC throttle project

[peteski]

The club setting (multiple radio throttles) brings up a question of how will multiple units operate simultaneously?  How will be a frequency or channel (or throttle ID) selected?

[mmagliaro]

Here's how multiple units work (and bear in mind, I have not thoroughly tested this part yet!)
Each handheld has a 0-9 mini rotary on the end (it doesn't show in the video). You set it to a number,
and you set the throttle under the layout to same number and that is the only throttle that the hand-held will
talk to.
So you could have 10 DC throttles, and a hand-held for each, and they THEORETICALLY would stay out of each other's way.

How does that work?
All of them use 433 MHz.  I don't have anything fancier than that in the RF portion of this thing.
But each hand-held sends messages with their ID number in them, and the receivers will ignore anything that comes in with the wrong ID.
So in a room of 10 of these things, many of the receivers would be getting messages from the other hand-helds, but they would simply ignore them. 

But there is still the issue of RF interference from simultaneous transmissions on the same frequency.
Well...
Every message is sent out 5 times at randomly generated time spacings.  My hope is that this should assure that at least one
of them will go out when nobody else is transmitting and will get through.  In a really busy room with 10 throttles... that 5x might have to
be increased.

All the messages are sent via an Arduino library called "Virtual Wire", which puts a preamble and CRC check code in with every message.
So even if there is interference causing corrupted messages, it is very very unlikely that a corrupted message will be received and validated as a good message.  They will just get thrown out.

Like I said, testing this with many of them busily sending all at once is something I have not tested. I need to build a bunch and try that out.  Bear in mind, the transmission rate is very low on something like this.  The Hand-held only has to send out speed and direction changes, which are really not that frequent.  Even when you are moving the speed dial, it only sends < 10 messages in a second or so.

Transmission rate is 2048 bps, messages are 16 bits + the preamble + CRC, about 64 bits.  So it takes 30 mSec to send one.
That's another thing I can fiddle with (I can go to 4096 or 8192 bps and see how it does).

Range:
well, I've tested it out the train room, down the hall, into another room, so I've gone at least 30 feet, through two walls, and the reception is rock solid.  I spent a lot of time experimenting and studying antenna theory (oh joy), as well as different receiver types.
FUN FACT: If you play with these little inexpensive 433 MHz Tx/Rx things you can buy on eBay or Ali Express, do NOT use the
receiver with the little green block with the tuning screw in it.  Those are regenerative receivers and they are real c**p.
Look for receivers labeled "superheterodyne", such as the RXB8, or SRX882.

[peteski]

Thanks Max. Sounds like you already incorporated multi-channel capability. Sounds like it'll work well.  And the frequency you chose is far removed from the ever-present 2.4 or 5.8GHz WiFi frequencies.

My weather station remote sensors use those frequencies, and so do some land-line cordless phones (but those are disappearing like the Dodo birds).  The remote temp sensors only transmit once wvery 5 minutes, and garage door openers also operate very infrequently.

[Ed Kapuscinski]

As for "marketing these"  Ugh...  I really didn't expect that, because building one takes several days and the cost to cover EVERYTHING (a power transformer, the nice cases and cover plates, the PCB boards, all the components, plus my time to build it all) would be in the hundreds of dollars.  But if somebody wants to be a "guinea pig", message me and we will talk.

Give this a read: https://blog.kchung.co/taking-over-a-dead-iot-company/

Specifically.

The BOM is Too Damn High
A bill of materials or BOM is a list of the raw components that go into a product. Most electronics projects, especially serious ones, will have a detailed BOM that describes the item and price that it goes for.

Often the final cost of the BOM will be much lower than the retail price because of the cost of shipping, R&D, profit margin, etc. Small changes in BOM price can have a big impact on the final cost of an item. It’s not uncommon to switch vendors or parts to save just cents on the BOM.

One trick I use is that multiplying the BOM cost by 4 will often get you the retail price.

[Sokramiketes]

People (Like the guys that made them) thought the Proto-Throttle was too expensive and would never sell 100 units.

They are doing a drawing for the 1000th unit at Springfield.

Could you sell them, yes.  Could any idiot with $500 in his pocket ruin your hobby time for a week?  Also yes!

[mmagliaro]

Thanks Max. Sounds like you already incorporated multi-channel capability. Sounds like it'll work well.  And the frequency you chose is far removed from the ever-present 2.4 or 5.8GHz WiFi frequencies.

My weather station remote sensors use those frequencies, and so do some land-line cordless phones (but those are disappearing like the Dodo birds).  The remote temp sensors only transmit once wvery 5 minutes, and garage door openers also operate very infrequently.

My choice of 433 MHz is based on the fact that nearly all the inexpensive tx/rx modules out there are built for 433 or 315, because those are considered "unlicensed" little bands of the spectrum.  "Unlicensed" means you don't need a license to use them, BUT it doesn't mean there aren't FCC restrictions on how often you can transmit and how powerful a transmitter you use.  I could build a 1 watt transmitter and start blasting away, but the FCC would probably come knocking.   Garage door openers, home outdoor weather "stations", and handy home-owner remote control wireless devices use those two frequencies.  Car lock remotes use it too.  If there were 5 or 10 frequencies like that, and tx/rx modules available for them, I would have made throttle selection frequency-based.  But alas, there are not.  I think there is one in the 900 MHz vicinity.  But 3 would be about it and I wanted to have more than that. 

A nice thing about 433 MHz (vs the typical WiFi low Ghz bands) is that a lower frequency like 433MHz penetrates walls and other obstructions like people much more easily, so it can have good range and reliable transmission at low power.  I use a very specific coded message format, so I hightly doubt that any other noise coming in on that frequency would be misinterpreted by my receiver as a valid message, and vice versa for any other device out there that might happen to see one of mine.

When I get a chance, I will have an assistant walk around the house and yard with the throttle and see just what the maximum range on this thing really is.  I know it is more than 30 feet, but I suspect it's a lot more than that.

[mmagliaro]

Ed, Mike,
I take it from you pointing me to that sign article was to show me that people might pay a lot more for this thing than I figured.
Well, that's probably true.  But like you also point out, do I want to be stuck in my garage building these, even if I could sell them for $500 ?  Probably not.  No, I take that back.  DEFINITELY not.

Perhaps the way to go is something already suggested.   Created a detailed enough technical manual, and some assembly instructions, and then just sell boards and parts lists (or a parts package) as a kit --- effectively outsourcing the assembly and testing to the model railroaders who buy them.  The downside of that, of course, is that now I have to count on people having good assembly and soldering skills and not screwing them up and then hounding me for help.

[Sokramiketes]

The talk of unlicensed radio freq brings me memories of the TracTronics line.  Their CoolerCrawler throttles had the ability to take input from a certain garage door opener control.  I'll have to post a photo.

I have, in my collection of interesting bits of famous layouts, the TracTronic's founders home layout control board from the N Scale Pocohontas Div.  Think 3/4" plywood with 7 Cooler Crawlers, two transformers per crawler, garage door receiver boards, and uphill/downhill polarity switching.  Its a work of wall mounted art at this point.  But is another smart guy looking for the easiest solution to a radio throttle problem.  I like that. 

Rich Weyand designed electronics for pinball machines in a past life, which lead to the TracTronic's line.  That line also powers the Reid Bros layout to this day.  Now Rich is an accomplished SciFi author.

[jagged ben]

...
Perhaps the way to go is something already suggested.   Created a detailed enough technical manual, and some assembly instructions, and then just sell boards and parts lists (or a parts package) as a kit --- effectively outsourcing the assembly and testing to the model railroaders who buy them.  The downside of that, of course, is that now I have to count on people having good assembly and soldering skills and not screwing them up and then hounding me for help.

Thumbs up.  Don't undercharge, since you'll have to provide some support, of course.  But that should be less tedious. 

[greenwizard88]

Well, that's probably true.  But like you also point out, do I want to be stuck in my garage building these, even if I could sell them for $500 ?  Probably not.  No, I take that back.  DEFINITELY not.

You could look into quick turn pcb manufacturers. I've never used one, but my understanding is that you give them the PCB board and design, and the parts list, and they'll do the assembly via machine. They'll be more expensive, but you'll just need to do the finishing touches, connecting wires, or putting them into enclosures. At least that's my understanding from a co-worker of mine

[Tom L]

How about selling the design to a company already in the model RR electronics business or some such thing?  Not any of the big boys probably but there are a number of smaller types doing DCC, signaling, etc. 

Maybe the Proto Throttle guys would want to throw a bone to us dinosaurs from the DC age.

Tom L
Wellington CO

[John]

Imagine my surprise when this showed up at 4:30 AM in my youtube feed

/>
Why are you posting at 4 am Max ?   Why am I stalking youtube at 4:30?   Good job

[Bill H]

We always knew max was super human and probably never rests, but what was John doing up?

[mmagliaro]

I had no idea I did it that late.  Are you on the east coast?  Maybe it was 1:30 am my time.  That I could believe.  Sometimes, I'm up late working on something (but not 4:30 am... yikes!)

[Lemosteam]

Perhaps the way to go is something already suggested.   Created a detailed enough technical manual, and some assembly instructions, and then just sell boards and parts lists (or a parts package) as a kit --- effectively outsourcing the assembly and testing to the model railroaders who buy them.  The downside of that, of course, is that now I have to count on people having good assembly and soldering skills and not screwing them up and then hounding me for help.

This, with a robust disclaimer that they are assembling the unit at their own risk, and if they screw up the board, more are available for purchase.

As for the hounding bit, link them to the DCC electronics board here where they can ask and a multitude of folks can help them (and increasing TRW user traffic).

Max, the board is what intrigued me the most as I recall you were looking for a PCB manufacturer.  It appears you have found one and I think it looks fantastic!

Did you source them to design it for you or did you design it yourself in a specific software?