DCC conversion of a 3 car unit

[37lover]

Can anyone please advise on the following.

In the past I have converted Bachmann Class 159 to dcc running. At both end cars I remove the lights and solder new LEDs into the old "bakalite" [sp?] templates, I insert a 10k resistor on the white circuit, nothing on the red, but then a 6.8k on the common blue return. These were connected to a separate decoder in each car and I was happy with the resulting led brightness [the white is slightly brighter than the red]. At one car I swap the white/yellow wires when connecting to the decoder so the lights are opposite to the other car.

This has worked fine but needs three decoders.

I thought I would try and wire everything back to a single decoder in the middle [power] car. I had already converted the lights and installed the resistors when I changed tack and thought about bringing the wires back to the centre. Unfortunately this means I have glued back the bulkhead [which has to be removed to gain access to the screw holding the lights in place!] so it's a quite a job to cut out again to replace the resistors, which is something that is looking like I might have to do.

The lights did not work although I've made a right pigs ear of the wiring with too many suspect 3 pin micro connectors. I am going to redo the wiring but this time I am going to run continuous wires from the lights all the way to a Y connection at the central decoder.

My question is by splitting the decoder output I am obviously putting the circuit through the white resistor but also two resistors in the blue returns. But I am struggling to understand if I've split effectively into two parallel circuits, rather than a series, would this have an effect on the brightness?

[Fishmanoz]

Should be easy to tell whether you have them in series or parallel. If in parallel, you should be able to trace from your blue wire (plus resistor) directly (not via any other LED) to each LED.  Then the other side of each LED will go directly to either the white or yellow on the decoder.   If any are in series, then the blue will only be connected to one LED and you will pass through at least another LED before reaching the white or yellow wire on the decoder. 

[37lover]

Fishmanoz,

Thanks for the very quick response, sorry I wasn't clear I have connected in parallel. What I was trying to ask was would this result in a reduction in brightness

[96RAF]

If any leds and/or resistors (including those in the blue lead) are in series then brightness may be effected depending upon the individual led characteristics (see spec sheet) but if they are in parallel then each led brightness will be affected only by its own resistors (includes any in the blue lead).

Rob

[Chrissaf]

Just be aware that different coloured LEDs that are in parallel that SHARE a common resistor and don't have an individual resistor each can result in one of the colour LEDs not lighting up at all.

.

For example, take a Red and White LED. White LEDs typically have a forward bias voltage of 3.2V, whilst a Red LED has a typical forward bias voltage of 1.75 volts. The forward bias voltage is the voltage required to turn the LED on.

.

Put a White LED in parallel with a Red LED and then feed them both through a common shared current limiting resistor and this is likely to happen. At switch on, the current will start to flow through the circuit. Once the voltage across the LEDs reaches the forward bias voltage of the Red LED, it will switch on and the voltage across the both LEDs will stabilise at the Red LED forward bias voltage. Thus the voltage rise will never reach the voltage needed to switch the White LED on.

.

This only happens when you mix certain LED colours in parallel. One needs to look at the LED technical specifications and refer to their forward bias voltage parameters. The whole issue can be eliminated if you don't use any shared resistors (except for LEDs of the same colour) and use individual resistors per LED that are chosen for that particular LED brightness.

.

This only relates to LEDs wired in parallel.

[37lover]

Guys,

Thank you. Unfortunately if I post a picture, or circuit diagram, it will not show until Monday at the earliest but I think I am a little clearer on the problem of shared resistors. I will do a quick refresher on how voltage & current split in parallel circuits. I now think I've simply got too high a resistance given that the circuit used to work fine before combining, notwithstanding that there may be a problem with the forward bias voltage stabilising.

I will rig up a trial circuit and run a few tests. I don't like admitting defeat but it will be a lot easier to revert back to using three decoders than ripping out my newly installed LED's to change their resistors, which I would have to do if I find a solution that works with the parallel circuit.

[37lover]

I haven't run a trial yet but have successfully removed a bulkhead from one car so refitting with a different circuit looks viable. I've prepared a simple sketch to show my intended wiring, I've only shown the circuit for say the white output, the yellow would be a mirror image but using the same common return. Apologies for the crude presentation.

Assuming 15V decoder output, Red LED Vf 2.5V the voltage reduction through the red resistor needs to be 10V. If I assume 30 mA current then the Resistor is 333 Ohms, but the power loss is 0.3W so I propose using two 200 Ohm in series as my resistors are only rated at 0.25W.

A similar calculation for the White Leds [Vf 3.1, current 25mA] gives R = 350 Ohm. This time the power loss is 0.22 so I should be okay with say a 470 Ohm at 0.25W rating.

In practice it seems Led's still work with much higher resistor values than the above theorectical calcs indicate, I guess the theory is for maximum brightness.

Could anyone please confirm that my suggested circuit should be okay?/media/tinymce_upload/c8c68bcca8f42f47a1e022d8846c76d1.png

[Chrissaf]

A very common error is to read too much into the LED specifications with regard currents. The 30mA figure is usually stated as the maximum operating current that the LEDs should be fed. It is not intended to be the figure that your design should aspire to giving.

.

A modern 30mA LED will still function quite adequately on currents less than 5mA. In my designs I tend to favour a resistor value that gives between 5mA and 10mA and still obtain no noticeable reduction in brightness. If I want a dimmer LED, then that can result with resistors being selected in the 5,000 to 10,000 ohm range.

.

A 10mA current will increase the life expectancy of the LED as well as reducing the need for higher wattage resistors. Not only that, but the lighting current load on the decoder function output is cut to a third of that used by your 300 to 400 ohm resistor values. I suggest using 1,000 ohm resistors.

.

Two LEDs in series sharing a common resistor is fine.

[37lover]

Chrissaf,

Thank you for the clarification regarding low currents. I knew from my previous experience that much higher values of resistor could be used but it's only just clicked how to work backwards from the low current to show the resistor value. There was a bit of delay before my Monday post was cleared for publishing and unfortunately I was back at work today and cannot log on whilst at work.

[Chrissaf]

37L, you might find this online LED resistor calculator interesting.

.

If you have two LEDs in series, just add their forward bias voltages together and input that value into the calculator. The calculator may protest if it doesn't recognise the value, but just ignore the error warning. It will still perform the calculation to give a resistor value.

[37lover]

Chrissaf,

Just had a quick look at your link and actually went on to the Led in series link. I've already set up a similar calculation on an Excel spreadsheet, it's just that I didn't appreciate that much lower currents can be used [in the calculations!]. I'll do a few simulations and get some resistor values then set up my trial. I've always thought the red lights should be dimmer than the white so I'll play with resistor values to check out the effects on relative brightness.

Thanks again, I am now much more confident of what I'm trying to do with just the single decoder.

[37lover]

I have managed to get the lights working although I've had to re-fit one pair of red led's, not sure what happened but probably a bad soldered joint. I ended up using a 3.3k resistor in line with the two whites, and a 2.2k with the reds. Now I'm happy with the brightness I can refit the light clusters into the car ends and reinstate the bulkheads. I'm going to extend the fine wires across into the middle car with black wires to hopefully make them less conspicuous then fit a three way set of mini plugs. Waiting for the plugs ex Ebay. The plugs can be withdrawn through the end of the middle car, but not the end cars, so I can store the set in the original boxes.

/media/tinymce_upload/9643af2d56aba6b413a65e8927afff5c.JPG

 

 

/media/tinymce_upload/f7920d7cc64a522e837015f251c51d77.JPG

 

The second photo shows one light cluster, with the white reinstated bulkhead, before I decided to link the lights to the centre car!