Has DCC power damaged the lighting on my Pullman coaches?

[ColinRE]

Hoping someone may be able to throw some light on my experiences.

I swapped over to DCC about 6 months ago. I have an NCE Power Pro controller with a 5 Amp supply protected with an NCE EB1 circuit breaker. My layout includes 6 circuits and I operate up to 6 trains at once. I have about 25 loco mainly fitted with Hatton’s decoders. My rolling stock includes 16 Hornby Pullman cars of varying ages, all with lighting (which is permanently on with DCC). This has all worked amazingly well for more than 6 months.

A few weeks ago however, there was a short circuit and the EB1 cut the power as it is supposed to do. With lots of trains running and 40 points, short circuits are not uncommon. Usually they are momentary as a loco passed a point or a derailment occurs and the problems clear themselves. In this case it didn’t and the EB1 continued to operate switching the power on and off every few seconds. After lots of experimentation I narrowed the cause to 13 of the 16 illuminated coaches. Usually putting just 1 of these problematic coaches on the track and certainly 2 causes the circuit breaker to trip.

I cleaned the pickup wheels on each coach and established that there were no short circuits across wheels on the same axle and that there was continuity between the wheels on the power pick-up side of each bogie.  The power for the lighting is picked up from across both bogies (from diagonally opposite wheels) so you would expect to find continuity there – and I did. Inside the coaches the light to the window lamps is supplied via fibre optic leads from an LED at each end mounted on small circuit board which includes some resistors and at one end a capacitor – presumably to smooth any flickering.

I tried to measure the resistance between the power pick up wheels using a multi-meter, hoping to compare the difference between the working and problematic coaches, but I was unable to achieve a reading - perhaps because of the capacitor? I conducted a bench test putting 6 of the problematic coaches into a free standing length of track and connected up my old faithful Tri-ang DC controller from my childhood. With a few volts flowing it buzzed unhappily indicating a heavy load/short circuit. A second test was to disconnect the red feed wire between the power pick up and the circuit boards inside the coach. Result no lights (obviously) but also no short circuit identifiable on either DC or DCC. 3 coaches with disconnected lighting run happily on DCC.

My conclusion here, and I might be wrong, is that a power surge on DCC has damaged the circuit board in 13 of my 16 carriages – and some or all of the resistors are no longer effective.

Has anyone else encountered anything like this? I assumed 6 months ago from everything I read that Hornby’s Pullman coaches could make the transition without any special precautions. Have I missed something? Should I worry about the 3 remaining coaches that work? Are replacement lighting circuit boards available for the failed coaches?

Thanks in anticipation

Colin

[Chrissaf]

LEDs are electronic components that are very sensitive to over voltage induced damage and if the supply voltage exceeds their safe operating voltage they can be blown just like a bulb. But unlike a bulb that will go 'open circuit' when blown. An LED can just as easily and more commonly become 'short circuit' when electrically damaged.

.

In DCC, transient short circuits are capable of generating spike voltages up to three times the base track voltage.

.

This is the very first time I have heard of a fault of this nature. I think it must be very rare. Once you have got your coach lighting repaired you can fit a simple device to your DCC power distribution to restrict the magnitude of these short induced voltage spikes. The device consists of a 100 to 150 ohm resistor of 2 watts power rating in series with a 0.1uF 100 volt ceramic capacitor being placed across your DCC distribution wires. Not only would the addition of these components potentially protect your coach lighting circuits, they should also help protect your DCC decoders from a similar fate.

.

More information can be found here:

http://www.brian-lambert.co.uk/DCC.html#On

and here

https://sites.google.com/site/markgurries/home/dcc-general-best-practices/wiring-planing/snubbers-rc-filter

.

I tried to measure the resistance between the power pick up wheels using a multi-meter, hoping to compare the difference between the working and problematic coaches, but I was unable to achieve a reading

.

The reason that the multi-meter failed to get a reading across the wheels. Is because the first component the wheel pickups are connected to is a bridge rectifier. When you place an ohmmeter across the AC input of a bridge rectifier, a near infinite resistance value is perfectly normal. The bridge rectifier converts your DCC track voltage to pulsed DC (if the coach is placed on a DC track, then the DC current is just passed through the bridge rectifier). The Bridge rectifier introduces a 1.2 volt voltage drop to pass on to the next stage to light up the LEDs that are polarity sensitive (only light up when the voltage is the right way round). So in a DC layout environment the bridge rectifier is still necessary to support the track voltage being reversed when the train is travelling backwards.

.

My conclusion here, and I might be wrong, is that a power surge on DCC has damaged the circuit board in 13 of my 16 carriages

.

This would be my conclusion too.

.

Are replacement lighting circuit boards available for the failed coaches?

.

Best place to try would be 'Peter's Spares'. If parts not listed on the web site, contact by phone to confirm.

[96RAF]

My 3 Pullmans have been on the DCC layout fairly extensively over a period of about 7 years and the lamps work now just as they did when new.

 

These coaches are part of the VSOE boxed set R1038 and I haven't looked closely enough to see if the table lamps are leds or little bulbs.

[Chrissaf]

I have the R4427, R4428 and R4487 Pullmans with table lamps. I haven't taken them apart, but as far as I can tell the table lamps are lit via plastic light guides to a centralised LED light source.

[Wobblinwheel]

I've got about 12 of the various lighted Hornby Pullmans, and one of the worst DCC (was DC) track layouts ever conceived. Most wires are all the same color. Most are too small in gauge. Multiple switched track sections, no bus feeders anywhere, and a manually-controlled reversing loop. I'm also using a 6 amp power supply, fuse-protected at 3 amps. This thing is "short city"! My point is, I have never damaged any of my lighted Pullmans, even when I should have! Not throwing the polarity of the reverse loop at the right time is a prime example, using the coach as the source of the short circuit! However, I have had connectivity problems with some of the internal wiring/connections inside one or two of the coaches. Simple bending and cleaning of contacts does the trick. I would look really carefully for a connection issue, just to be sure. If I can't burn 'em up, NOBODY can....!

[brando]

If you can remove the circuit boards shine a strong light under them and this will show up any cracks in them.

[Fishmanoz]

I'm with Chris and agreeing with Colin that voltage spike, which occurs on recovery from a transient short, not when the short first occurs, is the reason for the failure.  And most of the components you have identified as resistors will in fact be diodes as Chris says and be in that bridge rectifier.  I think you can discount cracked PCBs, that would usually be from mechanical deformation of the board and would cause an open circuit, not a short.  The solution is most definitely fitting the terminators Chris describes.

 

This is the first time I can remember such a fault reported too, very unlucky as the lighting has been shown to be reliable I think.

 

Just because your wiring is a mess Wobbly doesn't necessarily mean you will be generating voltage spikes on recovery from shorts, although it probably is more likely.  It will depend to what extent your particular stray inductances and capacitances are reacting with your controller output on recovery.  Do you use terminators?  These are in fact low pass filters which short out the very high frequency components in the transient, so neutralising it, if that makes any sense.

[ColinRE]

Many thanks to Chissaf and Fishmanoz,

I feel confident you have clarified the problem and I will be putting the filter you suggest in place over the weekend. I did in fact consider this when I built the layout last year, but given my BUS is only about 7M long and well twisted I didn't think it to be necessary. Hindsight is a wonderful thing!

Colin

[Fishmanoz]

No problem Colin, all part of the service.  And if we are right then you are even more unlucky with that bus configuration.

[Chrissaf]

The other clue is that 13 out of 16 coaches have been affected. Logic dictates that the cause has to be an external global event such as a high voltage spike.

.

Loose connections within a coach might affect one or two and maybe many over time if an inherent design issue. But to have 13 coaches affected in bulk within one single time period by internal issues seems unlikely.

.

PS - I cannot find any of these boards on Peter's Spares site. If you were to take a close up photograph of the board in question and post here. I may be able to offer some advice on whether a DIY repair might be feasible and what components to acquire.

.

A further PS - Colin you say that your DCC Bus is "Well Twisted". It may be that you have too many twists. Having too many twists can be just as detrimental to DCC operation as having none. The number of twists per foot impact on the capacitance and inductance of the Bus. The recommendation is to have about 3 or 4 twists per foot, which is quite a light gentle twist. You might have inadvertently created an electrical environment that has the effect of amplifying the magnitude of voltage spikes when they occur.

 

[orcadian]

Your problem could be the Hattons decoders as the model rail press was awash with them causing all sort of problems not so long ago.

[Fishmanoz]

Welcome to the forums orcadian.  Did the press mention the decoders causing lighting failure?  It's hard to see this happening as there isn't a common connection between them apart from the DCC supply volts. 

[ColinRE]

Responding to Chrissaf’s kind offer to have a look at the coach PCBs, I took one of the problematic coaches apart (surprisingly easy) and pulled out the 2 PCBs for a photography session (see below). At one end there is just an LED and a 47 ohm resistor. At the other end an additional capacitor and, as Chissaf pointed out, 4 diodes forming a bridge rectifier.

Hoping I might see something obvious - like a fried component or some melted solder underneath, but it all looks pristine. I took the opportunity to test the components with a meter. The resistors are both fine with values correspond to the 47 ohm markings. The diodes all recoded 664 ohms in one director and infinity the other way; their marking aren’t fully visible (so cant confirm against design) but looks like N40 on the top line and M10 underneath. The capacitor showed it could charge up. I didn’t try to test the LEDs as I’m not sure how to this with meter, but I did connect up DC power to the coach wheels again and this shows they light up and that the capacitor does its job in smoothing flickers from inconsistent pickup 

So execpting confirmation of the diode design characteristics all the individual PCB components seem to be working? Or am I missing something obvious here? I also took the opportunity to check the internal red and black feed wire insulation, to make sure there is no continuity where there shouldn’t be….and there isn’t

The coach still trips the EB1 on DCC power and, as I said previously, 5 or 6 problematic coaches together on a free standing length of track cause my old faithful DC transformer to buzz unhappily.

In summary, I’m no longer sure I know what the root cause of the problem is but clearly there is a lack of resistance across the whole assemble.

Grateful for any suggestions on a way forward.

Point of detail and an aside from above, responding to Chrissaf’s question on Friday; the twisting on my BUS is about 3 per meter as I read Brian Lamberts guidance last year before building the layout.

/media/tinymce_upload/f8a9804dff68abf2589e7bd035716d1e.JPG/media/tinymce_upload/061859c1f7d4deb7c02d2460d8bc0908.JPG

[ColinRE]

Yesterday (Saturday) I uploaded a posting with photographs and measurements regarding the individual electrical components on the PCBs in a 'problematic' coach. Because of the photograph these havn't appeared yet as verification is necessary.

The results were not quite what I expected and I feel I'm still missing something. The only two components I haven't specifically measured with a meter are the capacitor and the LEDs themselves. As a sense check and to provide additional information I decided to have a go at measuring the resistance of  whole coaches by connecting up a single length of track to DC power, putting a coach on it, measuring both volts and amps and then calculating resistance from volts divided by amps.

The results were very clear, consistant and showed a big difference between 'good' and 'problematic' coaches

The 13 'problematic' coaches all have an overall resistance of 110 ohms plus or minus 10ohms

The 3 'good' coaches have an overall resistance of 365 ohms plus or minus 2ohms

I'll leave readers to combine the above information with that from yesterday's post (when it appears). I'm sort of concluding (based on rather weak electrical skills) that it may be down to the capacitors or LEDs themselves, although both seem to fullfil their basic functions at an 'observational' level.

All suggestions on a route forward greatly appreciated.

In the absence of better ideas, my next step would be to measure the individual components in a 'good' coach to see if anything is different.

Colin

[Chrissaf]

The diodes all recoded 664 ohms in one director and infinity the other way.

.

Sounds fairly normal for a silicon rectifier diode. So these can probably be discounted as being at fault.

.

I didn’t try to test the LEDs as I’m not sure how to this with meter.

.

It is near impossible to test an LED using an Ohmmeter. The resistance measures near infinity one way round and typically more than 200,000 ohms the other. But you say they light up on the DC track, albeit that the circuit draws a lot of current. The fact that they light up would infer they are not damaged.

.

The resistors are both fine with values correspond to the 47 ohm markings.

.

I concur that R2 is 47 ohms. I can't see the colours on R1 but if this is 47 ohms too. The values are much lower than I would have expected. The only conclusion I can draw is that the two LEDs and at least one of the resistors are wired in series. It is possible (can't see the PCB tracks to confirm), that one of the resistors is on the AC input side of the four diode bridge rectifier to provide an element of current limiting to protect the bridge. This actually makes sense, because when your power cuts out with the coaches on the track. The bridge rectifier is not being burn't out. If a 47 ohm resistor is on the input side of the bridge, the short circuit current would be limited to about 270 mA. The diodes are most likely rated at 1 amp. Alternatively the 47 ohm resistor may be on the DC side of the bridge prior to the capacitor connection. That configuration would have the same current limiting bridge protecting effect. The resistor being before the capacitor limits the surge current when power is initially applied  to a fully and deeply discharged capacitor.

.

Capacitor.

.

This only leaves the capacitor to discuss. I can't see the labelling on the capacitor, but it is a polarity sensitive electrolytic. It is possible for these to appear normal when tested with a meter or a low voltage power source, but break down internally as higher voltages are applied. Given all the previous testing evidence you have expertly provided (congratulations on the level of detail you have provided. It has been invaluable to me). The only further test I could suggest, is to unsolder the Capacitor from the PCB. And then try the coach on the track with the anti-flicker capacitor not in circuit. If that does indeed effect a cure. These can be easily replaced at about 20 pence each subject to their actual values that I cannot see. I would suspect possibly 220uF or 470uF (or maybe a lot less to save money) at 16 volts (25 volt replacements would be better).

.

Electrolytic capacitors are made up of strips of conductive foil separated by an insulated dielectric. It is entirely feasible for a high voltage spike to arc through the dielectric from one conducting foil to the other. This can create weak spots that will fail in the presence of higher voltages but seem perfectly OK at lower ones.

.

Note: If it proves that the capacitor is the source of the fault. Make a note of the + marking on the side and which PCB hole the capacitor lead on that side goes in. Any replacement will need to go in the same way round. If you fit one of these the wrong way round they will probably go with a bang the first time power is applied. A brand new electrolytic will have one lead longer than the other. The longer lead will be the + lead.

[Flashbang]

Hi

Personally, I would apply 12v DC to the red and black wires marked 'IN1' & 'IN2' observing polarity. If you don't have a readily accessible 12v DC power supply use a 9 volt PP3 battery instead.   The Two LEDs should light. If nothing happens removed power and then unsolder the wires that lead off to the sub board marked 'OP1' and 'OP2'  This will allow just the main board to be tested. Reapply 12V DC or the 9V battery to IN1 & IN2 connections. Does the main LED now light? If Yes then the problem is on the smaller board. If No then the problem is likely to be on the main board.  

All the components look OK in the photo, so the main suspect has to be the LEDs.  These could be replaced with any colour LED just to test. Observe LED polarity - Long lead on an LED is the Positive or Anode. The next suspect is the electrolytic capacitor which could also be removed to test, as without it all that will happen is there is no flicker free lighting. If the circuits then work with it removed replace it with an identical uF value and DC work voltage rated capacitor.

 

Edit to add.. Are there any other LEDs other than the two shown?  If so disconnect them when testing.

 

Chris... Snap! 

[Flashbang]

Chris...  Sorry I was typing my reply when you posted, hence my edit note of Snap!

Colin...  What is the voltage rating of the capacitor. It will something like 25v or 35v etc?  Also what is its uF value?

 

I am very surprised that the resistor values for the LEDs are so low?  Also there doesn't appear to be any anti surge resistors fitted. These help stop the huge inrush currents that occur due to the capacitors all charging when several lit coaches are powered up at the same time!

[Chrissaf]

FB, your suggestions are useful additional tests that can be done before going to the extreme of un-soldering the capacitor (probably making it unusable and needing replacement in the process). The more suggestions (intelligent ones like yours that is) the better.

.

Colin, Just one extra cautionary note if following FB's advice. If you apply 12 volts DC directly across the leads of the LED, you will blow it instantaneously. When applying your free floating 12 volt supply (or a 9 volt battery) DC test probes I would include a resistor (anything between 100 and 500 ohms in one leg of the power supply wires). The inclusion of the resistor will prevent any damage occurring but still allow the test to be valid.

.

FB wrote:

The Two LEDs should light. If nothing happens removed power and then unsolder the wires that lead off to the sub board marked 'OP1' and 'OP2'  This will allow just the main board to be tested.

.

I would do something slightly different here. The two boards might be wired in series. Thus disconnecting the wires at OP1 & OP2 would break the circuit and the main board would still not light up. I suggest instead, disconnecting the wires at the remote board marked J3. If the main board lights up fine, if it doesn't place a low value resistor >100 <500 across the two wires that go back to the OP1 & OP2 terminations. The resistor would simulate the presence of the remote board in a series circuit.

.

Note to FB. If both the resistors are 47 ohms as Colin has indicated and assuming they and the two LEDs are all wired in series (lowest current draw scenario). This still equates to a current in the region of 80mA. The Hornby design for this circuit is intriguing to say the least. There must be more going on circuit wise than we can see in the photographs.

.

All the components look OK in the photo, so the main suspect has to be the LEDs. 

.

Colin did say that on the test track (DC Power Supply). The LEDs did light up, albeit with the coach drawing a significantly greater amount of current than the working coaches.

.

PS - Whilst you were writing and editing your post I was extensively editing mine. Thus I too made reference in my revised edited post to in-rush surge currents and expressed surprise of the low stated resistor values. Not to worry, both singing from the same hymn sheet is what counts.

[ColinRE]

Hi Chris & FB,

Thanks for your thoughts. This is my lunch hour so a bit limited, but happen to have additional pictures I took yesterday on my work PC that clear up most, if not all your questions.

R1 is definately 47 ohms from markings and I checked it with the meter

The Capacitor looks to be 25V and 220 mico farads from the writing

/media/tinymce_upload/d0d0dcdb4d47c368cb168dbddab25fe0.JPG

/media/tinymce_upload/dd0ed4ce77a275dfb8d352f048f7de09.JPG

/media/tinymce_upload/68a77e05b3c0d70c32248fe9eba490a0.JPG

The pictures of the undersides of the PCBs are not the best, but perhaps sufficient to indicate the circuit design. I reiterate there is no melted solder underneath connecting what shouldn't be connected in spite of the reflected light hinting at this.

Working on the principle that the culprids are either the capcitor or the LEDs I will progress with your respective suggestions for both (capaciot first) and then FBs suggestions to separate the boards. 

I happen to have some spare (red) LEDs (good enough for a test) left over from putting the circuit breaker box together. Would I be right in assuming they are pretty standard and interchangeable? I know that one wire is longer than the other, BUT which way round do they go in the circuit board. From the photo the underside of the existing LED has been trimmed (so no clue there). I can of cause just try it and hope to be lucky first time, but would rather avoid  'fireworks'.

Suspect the next steps towards repairing the lights may not happen now till mid June as going on hols Friday night for couple of weeks, but I will update this post as soon as I do make progress. On the plus side I put the spike filter in place across the BUS last night.

Colin

[Chrissaf]

Colin,

Personally I would investigate the Capacitor first before trying to unsolder the LED. If you trace the PCB tracks that connect to the LED. The track that leads either to a red wire or the +ve side of the capacitor will be the LED long lead.

.

I will try and decode the additional photos, comparing them to the first batch and see If I can deduce an approximate circuit layout for the components.

.

EDIT1: The remote PCB board was easy to unravel. The resistor is in series with the LED and connected across the red & black wires connected to J3.

.

EDIT2: I have been able to back engineer the main board as well from the photos. I can confirm the following. The remote board LED and resistor is in parallel via the Red and Black wires with the LED and resistor on the main board. This can only have one possible conclusion. The LEDs used by Hornby are not standard 2 volt / 3 volt LEDs. They are made to a higher voltage, probably 12 volts. The 47 ohm resistors values have been chosen to provide a small amount of current limiting. A 12 volt LED on DCC voltages with a 47 ohm resistor would draw about 20mA this value is what would be expected for any LED of significant brightness. Just because they are rated at 12 volts doesn't stop them being lit at much lower voltages (for example on a DC analogue layout at slow to medium speed). 12 volt LEDs are available to buy, but not as common as the 2 and 3 volt variety.

.

Do not substitute the LED on the boards with a standard LED of 2 to 3 volt rating. With only a 47 ohm resistor a standard 2 to 3 volt LED will 'burn out' due to the current the LED will try to draw (approx 210mA).

.

If the LEDs prove to be damaged and need replacing with standard ones, then the 47 ohm resistors will need to be increased in value to compensate. I would suggest 470 ohms. Either that or source some 12 volt white LEDs as a direct 'one for one' replacement.

.

I will draw out the circuit and post it in the thread. Of course the moderation delay may mean the drawing will not be visible until sometime tomorrow.

.

Note for FB: There are no anti-surge resistors for the capacitors used in the Hornby design.

[Chrissaf]

Colin, as promised here is the circuit diagram I have deduced from analysing your provided close up photographs. See the two notes included on the diagram. The LEDs are in parallel so FB's suggestion of disconnecting the link wires to test the two boards separately will work. Now I can see the component connectivity, there is no need to include an additional protective resistor in the power leads that I previously suggested. However, I still suspect the capacitor as being the culprit.

.

Enjoy your Holiday.....

.

/media/tinymce_upload/285fba421b2c8a8017da733ab2997cd2.jpg

.

[Flashbang]

Isn't the resistor colours -  Yellow, Violet, Brown, Silver?   Hard to tell and be 100% sure?

Which if it is, according to my reckoning is 470R 10%  ??

 

If 470R then the LEDs vF is somewhere in the range of  2.2 to 4.0 volts giving a current drain of approx 0.2A or 20milliamp.

 

[Chrissaf]

Well spotted FB, your absolutely right.....Dohhh! must pay more attention.........I must admit I didn't look too closely at the resistor colours in the photos as Colin said he measured them with a meter. I saw the Yellow Voilet and recognised those colours as 4 and 7, then just assumed that Colin was correct as he said he had measured them. They're definitely Yellow Violet Brown = 470 ohms that makes much more sense. 47 Ohms would have been Yellow Violet Black.

.

I shall have to redraw and submit my diagram awaiting moderation and hope that Adam can delete the old one before publishing.

[Flashbang]

Having carried out a bit if research I found this...  http://www.rmweb.co.uk/community/index.php?/topic/38672-hornby-pullmans-power-consumption/ see item 19.

[Chrissaf]

Colin,

As promised, the circuit that I deduced from your posted photos. Please make note of the two notes included in the diagram. As FB kindly pointed out, the value of R1 and R2 resistors are 470 ohms not 47 ohms. This means that the LEDs will indeed be standard LEDs with voltages of about 2 to 3 volts. Not 12 volts as previously thought.

.

/media/tinymce_upload/cf66ab1005c6cb6460c8d0643508693a.jpg

.