Come close to it

Ken,

I've come close to doing what you want. In the three 2-6-0 steamers I modified to run on DCC voltage via radio control. I used a bridge rectifier and a voltage regulator to bring the voltage down to 5 volts for the radio receiver to work and power a 6 volt motor. The engines run very smoothly and at a lower speed than any dedicated DC power supply.

Now you can take that and perhaps make a circuit, but I think you'll need a transistor throttle to control the voltage. You wouldn't be able to use the DCC dedicated throttle. Best bet just for a test track would be totally separate units. In other words your DCC rig and a pure DC power supply rig. A switch to either the DCC power supply or the pure DC power supply. The only other way would be a selector switch to take DCC power supply run it through a bridge rectifier with some very large micro-farad capacitors to take out the ripple and then through a transistor throttle.

Questions?

Bernd

Possible but not simple or practical

It is possible but the design would not be "simple" requiring some logic or a microprocessor to detect the stretched "0", the polarity of the "0" and more. The availability of cheep ( http://3-12V, 2A Selectable Output Supply) and inexpensive ( http://0-30VDC, 0-3A Variable Benchtop Power Supply) regulated DC bench supplies makes the two source solution a better option. DC bench supplies usually have Volt and Amp meters which without additional engineering would be missing from the filter design. DPDT switches are cheep. Add another DPDT or use a double pole triple throw switch to the mix for switching to the programming output and you have complete dependable solution.

High amps decoder

Would it be possible to use a high amp decoder 2+, hook the truck side of the decoder to the DCC Power Bus and feed the rails from the motor side of the decoder?

Doug W

Absolutely

This is essentially what I do with AristoCraft Large Scale receivers ( CRE57002SD REVOLUTION RECEIVER  or CRE57005 BASE STATION / SUPER RECEIVER) to drive the loops on the garden layout. The output is PWM and not DC, but simple filters LC filters can be used to generate smoother DC. Although you could use a smaller decoder I would go with a decoder like the Digitrax DH465 which would give you about a 4 Amp capability or a Large Scale decoder like the DG583 which would give you about a 5 Amp capability. AristoCraft use to produce the CRE57090 Revolution Sound adapter boards PWC to Linear and CRE57091 Revolution PWC to Linear boards for track side receivers which may still be available through Crest ( http://shop.crest-electronics.net/Accessories_c7.htm). You will still need the switches to route the power.

Remember what address 0 really does . . .

It makes the pulses on one side of the DCC waveform longer, time wise, than those on the other.

It does NOT eliminate the opposite polarity.

Thus, If you have, say 15 volts in the DCC waveform, the resulting DC equivalent using address 0 would be only about 8 volts. Not what you want for testing.

The issue with a high amp decoder is that shorts frequently occur as wheels run over track. Decoders don't take well to this, whereas boosters are designed to self-protect.

An inexpensive power pack or power supply is a better answer in the long run, to me. I use a Molex connector between the booster and the track on simple layouts. I can unplug the booster and plug in the power pack.

Logical, not practical...

Dear Doug, Ken K,

If we accept that the typical DCC signal path is

Booster> buss> track> decoder> motor

it might seem logical that by reshuffling we can create a "DCC-controlled analog track" situation thusly:

Booster> Buss> Decoder> Track> motor

Honestly compels me to admit I've even been-there-done-that myself

One big, (and it is Big) "But"....

Unlike a booster output stage, or a typical analog throttle,

the Motor Output Stage of most decoders do not have adequate "short circuit protection" to handle a (apparently "normal" operating procedure, if typical forums posting is to be believed)

"loco-runs-thru-turnout-not-set-for-it" short-circuit condition,
let alone a "wheel back-to-back momentary short against switchrails" condition.

(Believe Loksound and some Euro-manufactured decoders can handle a dead-short accross the motor outputs for a short period of time, many US-manufactured decoders cannot. Ken K, have you confirmed what happens if you present a dead-short to your 5Amp garden decoder? My gut-instinct expects either "smoke" or "welding flash"... :-( ).
 

You'll only do this once, and having let the "magic smoke" out, you're unlikely to try again.

As far as Ken R's OP is concerned:

- Making some form of _active_ filter,
which _only_ messes-with/adapts "zero bit stretching" signals into variable amplitude/polarity "analog DC", would require some relatively complex DCC-packet analysis, processing, and output. $$$ and complexity for a convoluted solution. (Breaks Ken R's "second question") 

- Making a simple bi-polar DC smoothing/filtering capacitor+diode array would "fill in the zeros" in the DCC packet signal, and result in relatively smooth "analog DC" output. As noted, the resulting signal would be indecipherable/indeterminate/useless to a DCC loco, this is understood.

However, it would also be of fixed polarity
(no way to reverse an analog loco's direction from the DCC system supplying the filter),

and would be of fixed amplitude/analog "speed", (approx 8VDC half-wave rectified, or around 12-14VDC full-wave rectified via a bridge-rectifier), 
not _variable_ as is required for our loco-driving application.

In short Ken, nice idea, but realistically, just get/build a flatline DC throttle for the test bench
(Willy Occum was onto something)

Happy Modelling,
Aim to Improve,
Prof Klyzlr

light bulb

Prof Klyzlr

Would adding a tail light bulb between the decoder and the track, handle the short issue?  (aka like Joe did on his layout for short protection)

Second option maybe a resistor of the correct wattage to act as a fuse, to burn out if the current became to high, as in a dead short.

Doug W

Bulbs for "pseudo analog" ; not fast-acting enough

Dear Doug,

Bulbs sink (some of) the excess power, but they do not actually stop the fault condition from occuring,
(IE they do not cut the actual supply, or "remove the short condition" from being applied).

The typical 12v "tail light bulbs" are also relatively slow-acting power sinks, even when presented with a "constant high voltage" as in a DCC situation. Under analog the "voltage being short-circuited" could be anywhere between 0 and 12-14VDC,
with a related time-delay at the bulb measured in seconds, if not minutes.

Given that it only takes literally fractions of a second to kill the output stage of a typical 1Amp decoder under "short circuit" conditions, the "bulb technique" simply isn't fast-acting-enough under a "pseudo analog VS short circuit" risk condition.

As such, no, I would not trust such a solution in this situation.

Happy Modelling,
Aim to Improve,
Prof Klyzlr

resister as fuses

Prof Klyzlr

Back in my days working with Building Automated Systems, some of the circuits would have a resister of a designed wattage in the circuit that would burn out if the current in the circuit was to high.  Would this work?

Doug W

Using Decoders to Power Tracks

Prof Klyzlr,

I said it could be done not that it was a good idea for the average modeler for exactly the reason you noted: that there are risks due to short circuits. As with most decoders the smaller Aristo receivers are not well protected for shorts, but the larger unit are designed for "base" use. However being the cautious type I installed current limiters on all receiver outputs (The supplies in my six channel box are rated for 50 Amps at 24 Volts so a short could weld wheels to rail and generate a lot of smoke in the box).

If someone were to use a decoder to drive even a test track, I would definitely recommend some sort of over current / short circuit protection between the decoder and the track. But for DC testing, as I commented, a bench supply is a significantly better option.

Sacrifical resistors?

Dear Doug,

A "sacrificial resistor"? Theoretically maybe,
if it was part of a properly-designed and integrated zener-diode+resistor "clamping limit" circuit.

In practise,
for an average model RRer attempting to kitbash additional protection onto a commercially-produced decoder,
for an application which the decoder was not inherrently designed to accomodate?
(IE staying true to the spirit of Ken R's OP),

the answer has to be a qualified "no".

Happy Modelling,
Aim to Improve,
Prof Klyzlr

Sacrificial resistors

Thanks for the comments

I appreciate all of the advice and comments, guys.  I guess I'll just have two separate throttles. I was hoping that there would be a way to add a capacitor to filter out the AC, letting only the net DC voltage pass through, and that if the capacitor were not polarized that it should work with either polarity.  Guess it's not that simple - it rarely is!

Oh well.  One more idea down, many many many more to come!