DCC to DC voltage converter to power layout accessory lighting?

[ednadolski]

No doubt this has been asked and answered many times, but my search skills are eluding me today.   Would anyone know of a way to build or buy a voltage converter that can take a DCC track signal and convert it about say 5 VDC output?  I'm thinking primarily to power some structure LED lighting, so not a power-hungry application.

Possibly something like this:

http://dcyale.com/blog/?p=83

He references an Amazon link for a cheap converter that presumably could also work for this.  I have no idea how useful this could be, tho it seems cheap enough to try (as long as connecting it won't kill the command station):

https://www.amazon.com/HiLetgo-Super-Converter-Module-Adjustable/dp/B00LODGDYE/ref=sr_1_40?dchild=1&keywords=dc-dc+buck+converter&qid=1598994185&sr=8-40

Thoughts & comments gratefully requested.

Ed

[learmoia]

A Stationary decoder?.... Can you program output voltage?

Barring that... use a regular decoder and use the motor leads to power the items..
Set the decoder speed steps to be the voltage you want.. at all steps.

[ednadolski]

use a regular decoder and use the motor leads to power the items..
Set the decoder speed steps to be the voltage you want.. at all steps.

Actually, DCC decoders use a PWM signal to drive motors, as opposed to pure DC: https://dccwiki.com/Pulse_Width_Modulation

Ed

[nkalanaga]

PWM power should still work fine for structure lighting.  LEDs don't particularly care about the waveform, as long as voltages and currents are suitable.

But I'd suggest using a separate power supply, with it's own bus, for structure lighting.  No point getting fancy, or complicated.

[peteski]

I would not recommend using a DCC 9PWM) motor output from a decoder to drive LEDs designed for 5V.

Curious as to why tap of the DCC power for something that can easily use a dedicated DC power supply (inexpensive wall-wart)?

Also if the devices you need power for are simple LEDs, then just use different value current limiting resistor (or add one to limit the current above and beyond of what the light units have.

If you want to use the DCC track power, then to get DC voltage simply connect a bridge rectifier to the track, and on the output you end up with almost a perfect DC voltage, about 1V lower than the DCC pulses.  Probably around 12v.  Then use those additional resistors to limit the current (and by definition voltage) to the LEDs designed for 5V supply.  LEDs are current-based devices (but due to Ohm's Law, voltage also has to be considered).

If you decide to use a bridge rectifier and you have a multimeter, then I can help you in determining the resistor values.

[ednadolski]

Main idea is to avoid YAWW, while there is a perfectly usable power source available in the DCC waveform.

Bridge rectifier sounds like a good idea.  Might also want to include a conditioning cap and/or LC filter to take care of spikes/harmonics.  Let me see what parts I have on hand.

(Always risky tho to encourage a software guy to get back into HW circuits... )

Ed

[turbowhiz]

Bridge rectifier sounds like a good idea.  Might also want to include a conditioning cap and/or LC filter to take care of spikes/harmonics.  Let me see what parts I have on hand.

Ed,

The original blog post you linked went into this in good detail, talking of rectifying the “AC” DCC signal into DC. It notably mentions the problems with traditional integrated circuit bridge rectifiers not being fast enough leaking reverse voltage and his experiments using schottky diodes (just build you own bridge using 4 diodes) to remedy said issue.

The amazon buck convertors you link to unto themselves won’t work, you’ll need a (fast) rectifier circuit first. The buck convertor is useful to take the rectified DC voltage (i.e "track" voltage in this case) down to the target 5V output.

Andrew

[Sokramiketes]

Iowa Scaled has this available.  It's $10 and converts DCC to ~12v pure DC.

https://www.iascaled.com/store/ModelRailroad/Accessories/CKT-DCCPWR

[ednadolski]

Iowa Scaled has this available.  It's $10 and converts DCC to ~12v pure DC.

https://www.iascaled.com/store/ModelRailroad/Accessories/CKT-DCCPWR

Is the IAS using a bridge rectifier?

...  It notably mentions the problems with traditional integrated circuit bridge rectifiers not being fast enough leaking reverse voltage and his experiments using schottky diodes (just build you own bridge using 4 diodes) to remedy said issue.

Seems then 1A 40V schottkeys ought to suffice:

https://www.amazon.com/1N5819-Schottky-Barrier-Rectifier-DO-204AL/dp/B07PBVDCW2/ref=sr_1_3?crid=U60OT67YFLHM&keywords=diode%2Bschottky%2B1A%2B40V&qid=1698702461&sprefix=diode%2Bschottky%2B1a%2B40v%2Caps%2C156&sr=8-3&th=1

Ed

[peteski]

Yes plain old bridge rectifier using standard silicon diodes works well with DCC waveform. I use them all the time, and so do manufacturers of DCC powered car  interior lighting circuits.  Yes, Shottky rectifiers work too, but are unnecessary.

Decoder manufacturers do use 4 discrete Shottky diodes as the bridge rectifier, but I believe they choose them because they have lower forward voltage than standard silicon diodes (and not because they have faster switching time).  Lower forward voltage also mean less heat dissipated by the diodes, and that is an important factory for decoder design.

[ednadolski]

These look similar to the IAS:

https://www.amazon.com/gp/product/B0B535X3SH/ref=ppx_yo_dt_b_asin_title_o00_s00?ie=UTF8&psc=1

[peteski]

These look similar to the IAS:

https://www.amazon.com/gp/product/B0B535X3SH/ref=ppx_yo_dt_b_asin_title_o00_s00?ie=UTF8&psc=1

That should work, but the rather large value capacitor (4700uF) will have high initial (inrush)  current when the DCC booster comes up.  It will probably be OK though.

When using the power supply just for LEDs, there isn't much need to filter it. The square waveform of the DCC signal results in pretty steady rectified voltage. Yes, there will be some small gaps in the DC (during the DCC pulse transitions), but LEDs used for lighting structures don't care.

[ednadolski]

^ I wonder what cap is on the IAS?   I suppose one could always swap in a smaller one if it turned out to be a problem, but that sort of defeats the purpose of going with an OTS.

Ed

[John]

^ I wonder what cap is on the IAS?   I suppose one could always swap in a smaller one if it turned out to be a problem, but that sort of defeats the purpose of going with an OTS.

Ed

The schematic is here on their site

https://www.iascaled.com/blog/auxiliary-power-from-a-dcc-bus/

100uf 50v electrolytic

[mmagliaro]

The 4700 uF one is on there for the Amazon one because they rate that board at 6A.  If you draw current like that from a bridge rectifier without a huge filter cap, you will have a ton of ripple in the output DC.  Even with 4700uF, the ripple will be something like 6 volts of difference between the dips and peaks of the output.  It all depends on how much ripple you can tolerate.  Frankly, as Peteski said, ripple in a supply that's just for LEDs isn't going to matter.  You'll never see it visually with your eyes and it won't cause any damage to the device.  But for a board that's promising the be a "filter", I can see why they put that big cap on there.    I wouldn't worry about the startup inrush current for that big capacitor.   4700 is big, but not that big, really.

If you buy that Amazon one, I'd be curious to know if it really comes with a Nichicon cap on it.  That's one of the nicer premium brand electrolytics and usually cheap little boards through Amazon from China do not use components like that.

Nerd Alert
If anybody cares, the ripple voltage will be:

            I
--------------------- x 1,000,000                          I is current in amps, C is capacitance in uF
      120 x C

The 120 is the frequency you get from full-wave rectified 60 Hz AC.  The 1,000,000 is in there so you can
specify the capacitance in uF, which is more "friendly".

So for a 6A load with that 4700 uF cap, the voltage ripple will be  10.6 volts.  That's right.  It won't be very "smooth" DC
at all.  It will be rectified pulsy DC that swings between 2 and 12v, roughly (assuming the output is nominally 12v)