DCC conversion and Lighting update of the
Hornby Class 67 locomotive.
(The older pre-Oct 2012 type with a plastic chassis)
Class 67 in Royal Train colours passing through Ely Station July 2009
This page provides a summary of the process adopted to incorporate a DCC decoder and correctly operating external LED lighting in the EWS liveried Hornby Class 67 Locomotive unit (OO gauge).
Note that the smaller pictures can be enlarged by clicking on the image.
Class 67 at the edge of York Station in May 2008 (probably on Thunderbird duty)
|A few Initial issues with
the original Hornby Class 67 unit:
DC control: A bit jittery initially, but then settled down. Two rubber tyres are fitted (to one wheel on each bogie). I may replace these with non tyred wheels if I can find a source as they will be impossible to keep clean!
Lighting arrangements: The Hornby model includes directional headlight LEDs. These incorrectly illuminate both the headlights at the forward end, simultaneously. The marker lights and rear lights at each end of the locomotive are non-operational "painted-on" facsimiles.
DCC Conversion Approach:
The plan is to use a 6 function and back emf motor controller TCS T6X decoder sourced from Bromsgrove Models. (To provide correct operation of day and night running lights, 5 function outputs and some additional transistor circuitry are required.)
External Lighting Modifications:
It is planned to replace the existing Hornby headlight light-pipes and LEDs with new independent LEDs. New LEDs will also be added for the missing marker and rear lights. Daylight running will be via button 0 and night running via button 1. TCS Rule 17 dimming will be applied to the headlights only. Rear lamps will be enabled via button 9.
On the real loco, daylight running uses right hand forward headlights, while night running uses the left hand forward headlights. All forward top and side marker lights are active for both day and night running. Both aft end rear lights are active for night and day running if enabled. The loco rear lights are not used if the loco is hauling a train. (To see a way to produce a realistic flashing rear warning light for DCC freight trains: click here).
Two PNP transistors are used to switch the common positive supply (blue decoder wire) to the headlight LEDs under the control of the white and yellow decoder function output wires, which provide directional control, ensuring that only the forward end headlights can be activated. The negative ends of the headlight LEDs are switched via the green function wire for daylight operation (right hand side) and the purple function wire for night operation (left hand side). The marker lights and negative end of the rear lights, are controlled directly from the white and yellow control wires. The positive end of the rear lights is enabled or disabled by the third PNP transistor under the control of the brown function control wire. The resistors in series with each LED control the current flowing through that LED. This determines the intensity of the light generated by the LED. The above values were established by experiment, for the LED types employed. Some adjustment may be required if different LED types are used. The motor and track pick-up connections to the decoder will make use of the 8 pin NEM652 connector fitted to the model.
|Number 1 and number 2 Cab
The number 1 cab is adjacent to the large roof ventilator, while the number 2 end is adjacent to the roof mounted exhaust system.
Postscript: This definition came from the excellent 2008 1st edition of Traction Recognition by Colin J. Marsden...... However, readers should be aware that in his updated second edition of 2011, the end definition has been transposed.
Achieving Access to the loco:
First the upper body shell must be separated from the wheeled chassis unit. This is a tricky operation, as the upper moulding sides need to be gently levered outwards with respect to the black chassis moulding. I used a new Stanley knife blade (not fitted to the knife) to provide the initial movement and then slid in some thin cardboard pieces to separate the clips, first on one side, then the other.
Upper body shell and lower chassis, showing the clip locations
|Removing the Hornby Lighting PCB
The original headlights consist of a light-pipe and lens moulding, glued to the inside front of the upper body shell at each end. Each light-pipe is illuminated by a shrouded LED, mounted on a pcb, fitted to the chassis moulding. The light pipes need to be removed by breaking the glue joint and the chassis mounted LED PCBs are not required, so these (and their wires) can also be removed.
|Fitting the new LEDs:
The existing headlight holes need running through with a 2mm drill to take the new white tower LEDs. New 2mm diameter apertures are needed for the rear lights and upper marker lights. The rectangular apertures for the lower marker lights are also drilled and filed to shape. The black face of the main lighting assemblies are touched in with satin finish black paint, before the LEDs are mounted. The internal cab liners and side window mouldings are temporarily removed, to provide access for the installation of the top marker LEDs.
The painted tower LEDs are a tight fit in the 2mm holes, so do not need adhesive. Surface mount 0603 chip LEDs are used for the lower marker lights. These are first soldered to extremely thin 0.18mm enamelled copper wire and then held in place using super glue (cyanoacrylate adhesive). Click here for advice on soldering wires to chip LEDs.
|Eliminating "Light bleed"
between the new LEDs:
Light from the tower LEDs will couple into adjacent devices if fitted straight out of the bag, producing a very unrealistic effect. To prevent this, the LEDs are painted with gloss black modelling enamel, leaving just the outer face of the 2mm diameter tower exposed. I now also paint a white dot on the negative side of the LED body (shortest lead) before use, to avoid later wiring errors. Once the LED leads are trimmed, its easy to forget the polarity!
|Adding transistor circuit
and control resistors:
Many of the LED series resistors are soldered directly to the trimmed tower LED leads. The transistor circuitry (as shown in the circuit diagram above) is made up in the form of small circuit modules, which are fixed using double sided adhesive foam pads, to convenient locations within the upper body shell. Interconnections are made using thin 0.5mm flexible cable sourced from DCC Supplies.
|Connecting the T6X decoder:
The appropriate decoder wires (as shown in the circuit diagram) are connected to the LED and transistor circuits. The decoder orange, grey, red and black wires are then connected to the Hornby 8 pin socket via a suitable pcb plug. (The plug is the original Hornby link-plug, modified by the removal of the copper track links between pins.) The 8 pin plug and wire forms the complete electrical linkage between the upper body shell and chassis. The decoder is fitted to the underside of the roof, just clear of the Hornby PCB, using a double sided adhesive foam pad. The chassis unit and upper body shell then undergo programming and electrical testing, before being reunited.
|Programming the CVs:
Using the DCC controller (mine is a Bachmann Dynamis) the CV values were programmed into the decoder.
The "rule 17" dimming option provided in the TCS decoders has been used, so that when the train is stationary, the headlights are dimmed. As soon as the train is made to move, these lights come up to their normal intensity.
|Initial Testing of the Lights...
Basic testing demonstrated correct operation of the lights and enabled some optimisation of the series resistors... BUT it also revealed a problem:
The headlights have a very obvious yellow cast. I suspect golden white LEDs have been mixed up with the pure white type I had intended to use. Pure white types are now on order! (The good news is that the yellow cast on the top marker light looks in line with the real loco photos).
In fact, to be more accurate, a slight yellow cast to the lower white marker lights would be more correct, but I have found in the past that using "golden white" markers with the very low current needed for these fairly dim lights, makes them even more yellow than reality. So, I prefer to run with both pure white headlights and lower marker lights. The marker lights can be given a hint of yellow by the use of a thin coat of water colour paint. Photographs rarely show the real light colours as these bright points usually push the exposure at that point on the image, into saturation, making the colours look almost white.
Daylight running lights
(lights are saturated in picture)
Showing yellow tint on round lights
and blue tint to lower markers as the lights
come out of saturation with a shorter
photo exposure time
(out of saturation)
New pure white tower LEDs ordered (from Bromsgrove) received, painted and fitted.
|Testing the Lights:
Button 0 should activate the day running lights (right hand headlight with both marker lights in the forward end of the car and both rear lights in the back end, if enabled with button 9).
Button 1 should activate the night running lights (left hand headlight with both marker lights in the forward end of the car and both rear lights in the back end, if enabled with button 9).
|Fixing the problems:
Although the lighting operated as planned, the direction of movement of the loco did not correspond to that of the lights. Rather than making a cv change, the grey and orange decoder wires were simply transposed on the 8 pin plug. This set the default direction of movement at switch-on to number one end forward, as designed for the lighting.
The lower marker lights were too bright at the number one end (DCC sourced chip LEDs) so a 100k series resistor was fitted in place of the original 47k. (Number two end utilized chip LEDs sourced from Bromsgrove, for which 47k was about right.)
Parts of the lower chassis frame had to be trimmed to avoid clashes with the new circuitry. The raised ridge on the top of the chassis ends was filed flush with the remaining top surface, to clear the bodies of the tower LEDs and the upper sides of the chassis moulding were partially nibbled away with cutters in the cab area to avoid touching the fragile lower marker light wires and the headlight series resistors.
The loco occasionally stalled on the clean test track, due to loss of electrical signal from the track. I've since removed the lower bogie covers and increased the wheel contact spring pressure. This has definitely improved reliability, but I'm not happy with the traction tyres (fitted to one wheel on each bogie) as they must compromise electrical pick-up reliability, which to me is more important than pulling power. I may change the affected wheel sets if I can find non-tyred replacements. The drive mechanism also needs running in, so I'll see how the loco continues to behave during initial operations.......
Postscript: The tyred wheelsets have now been replaced with non-tyred equivalents and excess lubricating material has been cleaned off the contacts and wheels. Electrical power is now 100% reliable on clean track. CV3 has been increased to 20 and CV4 to 12. The loco now starts OK on the minimum speed setting.
The finished Class 67
|Supplier website links:
The photos of real class 67s were taken at York and Ely stations. The photos of the model were taken using a tripod on the kitchen worktop with fill-in flash.