Dirt Cheap DCC controller

[rpierce000]

I have been in touch with a company in China that is willing to build very small DCC controllers. What are the MINIMUM functions you would accept on a controller? I am trying to get one that would fit in a generic N scale unit and sell for under $15, maybe even under $10 for bulk (say 10 units). I know directional lights and motor control are needed, can that be it? If that was it would you pay $15? $100 for 10? $120 for 10?

The goal is to create a decoder for the bottom of the market, the one you put in the "other" engines, not to compete with drop ins or "super decoders" with mars lights, ditch lights, cab lights and directional lights all on separate controls. I also want to make it small enough to fit in the really tiny spaces, but that may need to be a different unit, size is very dependent on wiring sizes.

This is FAR from a done deal, but I am tired of paying $30 for a decoder! (This would compete with the $20-$25 decoders from Digitrax, but not quite as fancy)

Please keep the replies in the thread, I want to have all the cross chatter I can get.

Many thanks!
Bob

[Martijn Meerts]

Hmm.. I'm actually happy with the current Lenz decoders. They might be a bit expensive, but they're pretty future proof and have everything I need now and in the (near) future.

What I would like to see though, is a simple 1-function decoder for controlling interior lighting in cars. Preferably with some circuitry to prevent heavy flickering and the likes. They shouldn't be expensive though, because your average 16-car shinkansen would require 12 or 13 of them, depending on how many motor cars there are. The thing is, turning the lights on and off isn't done often, which is why the decoders need to be cheap :)

[David]

It depends on who you want to sell them to. American trains are mostly served by drop in decoders. For Japanese trains (excluding Kato locomotives which follow American DCC ready design mostly) you've got a number of differences that certainly make the cost of the decoders high while negating much of their utility.

* The majority of Japanese model trains are EMU/DMUs, with 2 cab cars and a center motor car (at least 3 decoders, more if you want to control lighting in each car).

* With the exception of those Kato locomotives all are purely DC designs (multiple or complex lightboards/wiring dependent on DC polarity for function).

So what would I be looking for in a decoder for Japanese trains?

* It would be small, I would think the same size (or smaller if convient) then the DZ123/DZ125. Smaller is better for hiding it in an EMU with windows, but the DZ123 in particular seems to be the dimensions Kato used when it designed it's mid generation models which have a pre-milled underfloor spot in the motor car.

* It would have a total of 4 wires: Rail A and B and motor A and B (Red, Black, Orange, Grey).

* The motor wires would support at least 1amp(1.25 peak) like the Digitrax basic Z scale decoders, but preferably more (1.5, 2amp peak) to better support some Tomix models. This would cover all the bases in terms of Tomix motors and any bulb lights.

* The basic capability of the decoder of course would be to take a DCC signal in, decode it, and then create a PWM signal on the two motor wires. When the decoder detects a DC signal it should just pass it straight through (with voltage drop from the decoder)

* If possible it would support 'supersonic' PWM and BEMF to allow for smooth motor control.

* Should support basic CVs like direction of travel, 4 digit addressing, and if possible speed tables

Now where we start to diverge from American decoders a lot is the firmware - a CV is to configure how the motor wires are treated.

* Motor Mode: The decoder is just a 0-function motor decoder

* Directional Lighting Mode: Like motor mode, but full power even if the speed is zero. This mode would really make a difference in wiring cab cars, especially the ones that cram a lightboard somewhere inside the cab in the shell instead of the chasis. You simply have to wire the DC inputs directly to the decoder.

* Non-Directional Lighting Mode: Lighting is no longer directional. Polarity is now always the same.

* Bulb and LED Sub-Modes: Determines usage of PWM when in lighting mode. Can be used with a CV to set the brightness of the lights, and to create the fade in/fade out effect of bulbs with LEDs (so lights don't come on instantly).


I don't know if all this is possible for a $10 in bulk decoder. Given that a DZ123 can be bought for $15 the $10 bulk price would be a must to be competitive. To meet the price the first thing I would drop is the motor functionality - BEMF, supersonic, speed tables and peak amp capability. Someone could fill in how many amps you need peak for bulbs (the peak is rather high). So basically drop the motor mode completely (use existing decoders for the motor car and locomotives) and create a single function, dual polarity, light decoder. Effectively a cheap DC emulator for DCC supporting the basic CV features: direction, 2 and 4 digit address, and if possible have cheap PWM for LED brightness control in the firmware. If we went for the light only model it would really make a difference to have a capacitor of some type built in to remove flicker.

Ok, for the price lets focus on what is really needed for outfitting those Japanese EMUs and leave the motor decoder to someone else (this means that transponding/railcom can also be the users choice). How cheap can we go (assuming that the decoders only come in a bulk pack of 5 or something - no singles!)

Hardware:

* 4 wire miniature decoder. Enough peak amp draw to support a bulb turning on.
* Designed as a "dual polarity" 1 function light decoder. A 1 function decoder with the ability to switch polarity.
* Built in capacitor for flicker free lighting (cap size can be smaller, since we are mostly concerned with the LEDs which turn off instantly)
* Firmware based PWM for basic LED brightness control - smooth on/off, adjustable max brightness

CVs:

* Basic CVs required by standard (no effect for any motor related CVs)
* CV17 and CV18 for 4 digit address
* CV29 supports the flags for normal direction of travel, 4 digit address and DC operation

* CV x1: (lighting mode, default is all 0s)

Bit 0 changes to non-directional lighting (polarity is always "forwards"). CV29 will still swap polarity.
Bit 1 (not defined)
Bit 2 (not defined)
Bits 3-4 define how drawn out the PWM fade in/fade is a scale of 0 to 3. 0 means no fade and should be used for bulbs, while 3 draws out the  effect the most.
Bits 5-7 define the brightness reduction on a scale of 0 to 7. 0 is full brightness (no PWM) and should be used for bulbs, while 7 is the maximum PWM brightness reduction (very dim LEDs)

* CV x2 (lighting control, default is all 0s)

Bits 0-3 define which function key (F0, F1, F2, F3) is used to turn the lights on/off.
Bits 0-3 define which function key (F1, F2, F3) is used to manually toggle the direction of travel. 0 means the direction of travel is changed automatically based on DCC/DC direction of travel and CV29.


My only thoughts are that the function key binding might be changed a bit - if you have a sound decoder the F1-F3 are going to be taken since these have strongly defined sounds. Perhaps we could use those undefined bits as flags - ie flag will add +10 to the function binding except when 0.

[The_Ghan]

Bob,

I do pay only $15.  Well, $15.16, for the Digitrax DN-135D.  I also buy the DZ-123 for just a dollar more.  I think these are great value, especially with the Digitrax no-questions-asked replacement policy (which luckily, I've not had the misfortune to rely on as yet).

That said, I think there is room for a vanilla decoder or two, but I think they need to retail sub $10.  I think you'd be giving RailCom, Transponding, BEMF,  and supersonic a miss.  However, 4-digit addressing is a must. 

Here are my thoughts:

Standard decoders:

1. As a minimum, a zero function motor only decoder would be good for consists with 2 or more locos that are permanently together.

2. Single function version for internal lights and motor control.

Function only:

1. Single function as Martijn said.  Nothing else.  For interior lighting.  I like Martijn's idea of "flicker-free" circuitry. I calculated that I will need over 100 of these for my Kato fleet alone.  I started a thread on the subject around Christmas time.  Seriously, I won't be pursuing the idea with the Digitrax TL1, which is around $13 each.  But, if I could get 10 for, say, $59, I'd do it.  If I could get a flicker-free version for, say, $69 (10x), there would be no question.

2. As above but with headlight functions added working exactly the same as the headlight functions on a standard motor decoder.  Being for EMU cab cars I'd be selling them in pairs.  I personally wouldn't buy these because I want transponding in my cab cars.

In summary, vanilla decoders really need vanilla prices.  For me, I'll be using Digitrax products whenever I need transponding (locos, freight, and motor and cab cars on EMUs).  For those intermediate EMU cars I would definently go for the single function only decoder if you could do it for prices similar to that which I've suggested.  Much above that and I just wouldn't do it at all.

Cheers

The_Ghan

PS: As I clicked to post this Dave posted his comments.  They seem to be a more eliquent version of what I'm trying to say.

[Dani]

I fully agree with David:

* 4 wire miniature decoder. Enough peak amp draw to support a bulb turning on.
* Designed as a "dual polarity" 1 function light decoder. A 1 function decoder with the ability to switch polarity.
* Built in capacitor for flicker free lighting (cap size can be smaller, since we are mostly concerned with the LEDs which turn off instantly)
* Firmware based PWM for basic LED brightness control - smooth on/off, adjustable max brightness

One cheap decoder to use in all cab cars or illuminated cars. One function with "dual polarity" would be much better than two functions for non-dcc ready cars, easier to digitalize. In this way you don't need to modify the pcb of the cab cars (usually too small to work on it). And with possibilities to adjust brightness, this is a must!!! That's something I've been searching and never found it, one output able to switch polarity!!!!

I don't mind the motor car, I prefer to invest in a good decoder than buy a Chinese-cheap one. And if a want a cheap one I can also find it in the market.

And if it has and extra output for lighting, much better. One switching polarity output (cab lights), and other output for interior lighting. In this way I avoid to solder diode bridges before leds, and I would use this decoder for all cars (except motor car). But the main point is price....it's a lot of money digitalize all the cars and even more put lights in all them. Better the first "one switching output" option than this second if price is increased.

Cheers,
Dani

[CaptOblivious]

I would really like to see an inexpensive, small, motor-only decoder for the motorized carriages of MUs. No functions—takes up space and add to cost, and I won't use 'em. But the motor drive has to be solid: 1A or better, with smooth performance (esp at low end) to compensate for the lack of flywheels in many MU designs.

[keiman]

Like the good Capt, I to would like good basic motor only decoders, get fed up removing the other wires on mine.
As my trains/trams don't have lights, this is not a problem.Thinking fleet of B-Trains running DCC.

[KenS]

I'm not your target market.  I buy Digitrax/Kato because I want Transponding (and I like BEMF and ultrasonic too). That said, Dani's suggestion of a simple cab-car head/tail light decoder that does reversible polarity for compatibility with older lightboard designs is a very good one.  You don't need much to make that work, although being able to attenuate the output voltage would be handy, since older models designed for DC use don't expect 13-15 volts to the lights all the time, they'd be much happier with 7-9 I suspect.  I probably have some older trains I'd want something like that for eventually.

Quite honestly I don't think there's a need for a DCC decoder for interior lights. I'd never buy one.  A switch to kill power to the yard tracks is much cheaper to do. But if there is, make sure it can handle the amperage of a large bulb-based lighting system, because that's what people are most likely to want to turn off (I'm thinking 250mA minimum, for four bulbs, and maybe twice that). If they have LEDs, it's not likely that heat or lifespan is going to justify DCC decoders for all those cars.  Again, voltage attenuation could be a big feature here, although it probably needs to be jumper-selectable (or controlled by CV, but that might be a bit much for a simple decoder).

But I'll second what The_Ghan and Martijn said about flicker-suppression.  A circuit that does that on a very compact board isn't hard to design, and the components for a compact one cost $2 in small volume (not counting the board and assembly).  See my pages on such a circuit (and there are plenty of other examples on the net). 

It doesn't need to be DCC-controlled, but if I could get a board that did that for around $5-$7 I'd buy a hundred, maybe more (not all at once though  ).  Just make sure it's DCC-compatible, and that includes not tripping the breaker on the command station when you power up a 16-car train and charge all those capacitors.  That same warning applies to a DCC decoder with flicker suppression.  Sound boards with a single capacitor used to cause problems for some DCC command stations, and a train full of them is a bigger problem.  There's likely some work there to get something that plays nice designed.

Sooner or later I'll get around to making my own for $2 (component cost) and a heck of a lot of my labor.  It's worth a bit to avoid the labor.

[Webskipper]

I don't know about really cheap dcc.

I've dealt with really cheap 2.4 ghz in radio control and the reliability was, well really cheap. Distance was terrible.

[The_Ghan]

KenS,

FFB

I like your suggestion of a "flicker-free" board that is NOT DCC.  I plan to use transponding too.  I'll rely on transponding in cab and motor cars only.  I'd happily pay, say, $39 for 10x.  I think your design could retail for that price if made in volume.  We could replace your bridge rectifiers with Schottky's, but that might bulk up the capacitor size ... actually, it would probably work fine with the Panasonic alternative you suggested.

As an aside, I think there is a market for another DCC compatible part; the "Siding Power Switch".

SPS

I'm planning to use latching relays controlled by stationary decoders to turn sidings on and off.  I think latching relays are needed because stationary decoders I plan to use, such as the Digitrax DS64, provide momentary power.  I was going to wire the relays in parallel with the turnout motors.  Interested to know what others think of this idea.

Cheers

The_Ghan

[rpierce000]

So much feedback, thank you!

I had not thought of the "lights only" controller. Just in car lights and directionals. This would be a VERY easy and cheap unit. Countries with passenger service are the target markets, that leaves the US sadly behind.

If it did in car lighting and directional "end" lighting would it also need some kind of lighting consisting so that a 16 car train would be turned on and off as a whole? Is there another feature that could be added to substantially up the value without blowing the simple/rock bottom value of such a unit?

Thanks,
Bob

[cteno4]

bob,

yes i think the idea of a lights only (both car illumination and headlight/tail light) that would be really inexpensive would be the thing that might really fill a small void. right now its like $50 to just do a regular train with the motor car and the two end cabs for head/tail light alone.

i guess the same polarity with a settable voltage for the head tail light would be great. for lighting a rectifier circuit could also be used after the decoder to keep the lights on polarity independent. capacitors could be tough and make a small decoder, but a small inline capacitor block could also be something to look at.

i guess if the decoder output could be set for variable or fixed that would give you the lighting vs motor control.

personally i have always figured dcc can wait till i have more time to sit down and install in the number of trains i have. also figure prices for decoders will keep dropping and just get better on functionality.

cheers

jeff

[The_Ghan]

...

If it did in car lighting and directional "end" lighting would it also need some kind of lighting consisting so that a 16 car train would be turned on and off as a whole? ...

Thanks,
Bob

Bob,

This is achieved by setting all the decoders to the same address.

Cheers

The_Ghan

[KenS]

I'm planning to use latching relays controlled by stationary decoders to turn sidings on and off.  I think latching relays are needed because stationary decoders I plan to use, such as the Digitrax DS64, provide momentary power.  I was going to wire the relays in parallel with the turnout motors.  Interested to know what others think of this idea.

The DS64 comes factory-set for momentary (Digitrax calls it "pulse") output, but can be switched to continuous (Digitrax calls it "static") output to drive things like Tortoise switch machines.  See page 5 of the manual linked from Digitrax's DS64 page.  I think all you need is a relay and a resistor (to provide a limit on current).  Note that the whole DS64 must be set to "momentary" or "continuous" mode; you can't use just one output for a tortoise and have the others drive Kato's for example.

[The_Ghan]

Thanks KenS,

I was aware of that feature.  I also know that the 16s timeout for static mode can be disabled, presumably leaving the circuit ON.  I was just reluctant to keep current running through the circuit, but I guess it won't do any harm.  In that case, would something like  THIS work with the DS64?

Cheers

The_Ghan

[KenS]

I'm not actually very good at electronics.  I'm just good at picking battles (I mean projects  ) I can win.  My back of the envelope opinion is that "it should", based on the "2.5V - 20V High state (relay is ON)", but I could be wrong since I'm not familiar enough with transistors to really read the circuit diagram.  I stick with circuits that don't include semiconductor switching. 

The good thing is that I don't think it could hurt a DS64 due to the high resistance on IN1 (10K okms means a 12V output isn't going to run more than about 1 mA, which is what they list for the control amperage at 12V). The relay contact is rated for 10A at 30V DC or 250V AC, so it ought to be fine with any DC or DCC track power.

I may get one of those and play with it myself.  The seller has a US Ebay store with relays also.

[The_Ghan]

Hi KenS,

Yes, I'll probably play around with one too.  BTW, I've had my eye on this company for around 9 months.  I've previously emailed them asking about a latching relay version and they said they don't make it.  Still, it's a nice cheap way to turn sidings on and off when combined with a DS64.

Cheers

The_Ghan

[CaptOblivious]

I think these should be separate offerings. A two-function decoder with the option of bidirectional outputs (so it can be connected to an unmodified lightboard), and a one-function decoder for interior lights. On the hardware side, should be able to handle bulbs safely, with am adjustable voltage output for both models. Flickering headlights strikes me as far less of an issue than flickering interior lights.

Getting the basics right with robust, reliable hardware that is inexpensive is all the added value I need or want! Indeed, my problem is finding a decoder without all the bells and whistles....(sometimes literally!)

So much feedback, thank you!

I had not thought of the "lights only" controller. Just in car lights and directionals. This would be a VERY easy and cheap unit. Countries with passenger service are the target markets, that leaves the US sadly behind.

If it did in car lighting and directional "end" lighting would it also need some kind of lighting consisting so that a 16 car train would be turned on and off as a whole? Is there another feature that could be added to substantially up the value without blowing the simple/rock bottom value of such a unit?

Thanks,
Bob

[Dani]

I've previously emailed them asking about a latching relay version and they said they don't make it.  Still, it's a nice cheap way to turn sidings on and off when combined with a DS64.

I've been searching exactly the same, but finally I had to buy latching relays and build my own pcb.... It's a pity, for sure it will not be as reliable and professional than those from ebay.... :(
But if someone knows already done PCB with latching relays (I use them for loop management) please post!!!

[Dani]

Hello,

Continuing the discussion about a cheap function decoder with a "dual polarity" output, today I found this information in Lenz, sorry I don't know if it is really new or something that already everybody knows, but some months ago I searched a dual polarity decoder and all answers said "only Kato does it":

Many model railway enthusiasts miss a low-priced function decoder, for example for the direction-dependent change of light of control cars. Instead of producing a separate cost-intensive function decoder, we have opted for an intelligent software solution.

Starting with the next production series, the Digital plus locomotive decoders
STANDARD + and SILVER mini + will offer two operating modes:

Operating mode "Locomotive decoder": Motor outputs react to the speed step and the direction of travel.

Operating mode "Function decoder": Motor outputs do not react to the speed step (which would mean that the brightness of the light is dependent upon the speed step) but to the function key: If the function is enabled, full voltage is applied at all times and the polarity is dependent upon the direction of travel.

For all STANDARD + or SILVER mini + from version 7 onwards, we also offer an update free of charge. This update can be downloaded from our website and installed with the decoder programmer.

I think it's a very smart solution. With the same product you have a motor decoder with 2 outputs, or a function decoder with 3 outputs. All of them dimmable, light effects, function number assignment, master switch.... it's not the cheapest, but not far away from Digitrax DN135D (the only one with 3 functions to be equivalent). There are 5€ of difference here in Spain between both decoders, but having all decos the same you don't have to worry about distinct behaviours, configurations, CV numbers or functionalities.

What do you think about Lenz Mini? Someone has experience about them? I just have tried Uhlenbrock and Digitrax.

Regards,
Dani.

[Martijn Meerts]

The function decoder mode was announced quite a while ago, but I'm not sure if it's already implemented in the decoders. The manual doesn't seem to mention anything about it, but it hasn't been updated for a while.

Generally it's a good idea, but the lenz silver mini's aren't the cheapest decoders, especially not for a function decoder. I'll use them for cab cars that have dual polarity LEDs, but for cab cars with the more common 2 LEDs, I'll probably stick to cheaper function decoders.

As a motor decoder, the Lenz Silver+ mini is a great decoder, I pretty much use them exclusively in my trains.

[CaptOblivious]

Most TCS decoders will do this as well, and are considerably cheaper in North America. Indeed, I often use a TCS Z2 for Micro Ace cab cars for this very reason!

Hello,

Continuing the discussion about a cheap function decoder with a "dual polarity" output, today I found this information in Lenz, sorry I don't know if it is really new or something that already everybody knows, but some months ago I searched a dual polarity decoder and all answers said "only Kato does it":

Many model railway enthusiasts miss a low-priced function decoder, for example for the direction-dependent change of light of control cars. Instead of producing a separate cost-intensive function decoder, we have opted for an intelligent software solution.

Starting with the next production series, the Digital plus locomotive decoders
STANDARD + and SILVER mini + will offer two operating modes:

Operating mode "Locomotive decoder": Motor outputs react to the speed step and the direction of travel.

Operating mode "Function decoder": Motor outputs do not react to the speed step (which would mean that the brightness of the light is dependent upon the speed step) but to the function key: If the function is enabled, full voltage is applied at all times and the polarity is dependent upon the direction of travel.

For all STANDARD + or SILVER mini + from version 7 onwards, we also offer an update free of charge. This update can be downloaded from our website and installed with the decoder programmer.

I think it's a very smart solution. With the same product you have a motor decoder with 2 outputs, or a function decoder with 3 outputs. All of them dimmable, light effects, function number assignment, master switch.... it's not the cheapest, but not far away from Digitrax DN135D (the only one with 3 functions to be equivalent). There are 5€ of difference here in Spain between both decoders, but having all decos the same you don't have to worry about distinct behaviours, configurations, CV numbers or functionalities.

What do you think about Lenz Mini? Someone has experience about them? I just have tried Uhlenbrock and Digitrax.

Regards,
Dani.