hornby Three Rail conversion to DCC

[Yelrow]

Being the person, with the most to gain, as i run more 3 rail than anything else, thought it sensible to start a new Thread, enabling those who were bus ing, etc to continue, and allowing the purists their chance to learn about this new to me, groundbreaking

 

possibility. I am hopeful that this will reduce typing, whilst allowing the critical advice needed for an old git, to come ,in words of a few letters. I have servicing schedules, and am obtaining photos of insides of all locos, as intend to use elink, plus

 

hornby point motors etc, with their access decoder, together with hornby chips, unless, someone knows better. There are a few people on this forum, who are very knowledgaable on DCC electrics, and their may be those, readind but not joining in who run 3 rail

 

NOW IS THE TIME FOR ALL GOOD 3 RAIlERS, OR THOSE WITH INFORMATION TO IMPART, TO COME TO THE AID OF THE EXPAT. thanks, john

[Fishmanoz]

Well you need to start with the 3-rail DCC system as described in the DCC bus around the layout thread,then work from there to here, if typing is to be saved.

 

Now to continue, don't even think about using the 8249 chip on old locos as it will not

 

have enough power. Will need at least an 8245, and maybe more powerful, I'm not sure. To be certain, look up how to do a stall current test with that multimeter you are buying.

[Raven_Electric]

Well here we are on a 3 rail model railway thread. My goodness isn't life interesting.

 

To recap:

 

1. One set of leads come from the "Programming" output of your controller. ONLY EVER use these to connect a special programming track for one

 

decoder equiped item at a time. You CAN use a two pole three way switch so locos can run onto the track when it is in "Main" mode and tben be programmed when you throw the switch to "Programme" mode. Throw the switch back again and the loco can leave. There

 

are very good but long and tedious reasons why this is best practice.

2. The leads from the "Main" output of your controller can run one track system or be split to run more than one such aa a two rail layout and a 3 rail layout. There is no theoretical

 

or practical objection to this indeed block isolation is not exactly new in railways.

3. You really must obtain NEM 652 wire harnesses. Soldering wires from decoders without an expensive anti-static soldering setup is not recommended for your wallet.

4.

 

You need to strip, clean and service each loco. If the collector mount (possibly brittle with age) is screwed on then remove it and check the condition of the wire going up to the motor. DCC puts far more demands on insulation than dc operation. The 15vac

 

is constant whereas dc systems rarely get anywhere near 12 volts.

5. Because you will be cutting ends off the NEM 652 harnesses you will have lots of colour coded suitable wire to replace wire where a. the insultion is hard and brittle b. the termination

 

is loose or of poor quality c the colour coding is confusing. It is good pratice to use the NMRA DCC colour coding (below) thoughout your fleet.

6. You also need some 2mm od heat shrink to hand in case you need to join wires.

 

7. Solder the ORANGE

 

wire of your NEM connector to the motor terminal the collector was soldered to.

7. Solder the RED NEM connector to the collector wire or terminal.

8. Disconnct the other motor connector wire and solder the GREY NEM connector wire to the motor terminal.

 

 

9. Solder the Black NEM connector to the wheel pickup wire.

10. Plug in a Hornby R8249 decoder (Avoid static)

11. Ensure NO wires can rub or vibrate and damage their insulation.

12. Ensure the R8249 decoder cannot touch anything or short and

 

has free air all round it.

13. Box up and place on a 3 rail programming track and write CV1 address

14. Play trains

 

NEVER use heatshrink on decoders. It is hydrooarbon derived and has a high calorific capacity, very poor thermal tranfer and

 

seriously impairs air circulation round the the ICs and board. Power chips in decoders are highly stressed they MUST have adequeate cooling.

 

NEVER lift a DCC equiped loco on or off the track while it is selected (or already in running mode} by the

 

controller. contr

 

Be aware insulation on very old motor windings can be in a state of terminal breakdown. This CAN be tested but usually puts them beyond repair. In this state they may work extremely slowly in dc (for a while) because of the low voltage.

 

They are likley to fail in DCC or trip the power supply cutoff.

 

In any event once the loco is working both back EMF response nnd power curves need setting. This should be done with the age of the motor in mind.

 

The Hornby R8245 Sapphire chip

 

contains further diagnostic CVs and at least one should be obtained even if not need for functionality to be used to produce diagnostic reports on the state of the fleet's motors - this could prevent failure.

 

Details on an auto 2/3 rail switch track

 

and decoder installation for use with modern locos to run on three rail will follow later. You need to acquire some power collectors.

 

 

[Raven_Electric]

The R8249 can propel a current Hornby range R2339 Mallard hauling 10 Gresley R417x series 61'6" coaches at a scale 92 mph in level running and well exceed scale acceleration with train lighting on. I accept 10 tinplate carriages with bearings of unknown

 

resistance will likely offer a MUCH greater load. However, the R8249 is rated at 500Ma continuous with a 1Amp peak. From what I remember of those motors I would be very loathe indeed to put more than 15Watts through nearly eighty year old varnish insulation,

 

(at first anyway) once it is gone it is GONE. The R8245 offers 1.5Amp and unless a boiler is going in for train heating that would worry me at my age let alone nearly eighty!

In any event I suggest an R8245 for diagnostics and it will easy to swap the R8249

 

to newly prepared locos for testing because NEM sockest are being fitted and the choice of decoder for the whole fleet can be deternmined after tests.

 

[Yelrow]

Raven Electric, thanks, have just bought some nem harnesses, and will look at 8245 chip. Have already got points decoder. Looked at maplins, but as usual postage to france, NO. cAN BUY 1 FROM EBAY WITH BUZZER. uSB LEAD ALSO NO GO FROM MAPLINS, BUT ARE

 

LOOKING AT cONRAD FRANCE. What is a power collector ?, new to me, unless i call it by another name. You, i belive have not used elink, where writing to CV1, can be a bit of a lottery, due to my lack of knowledge. Apparently, much easier on elite. Still, we

 

are some time away from that. Fishy , had already spotted the more powerful chip, sorry about the pun. Wil remove bodies from locos one by one and photo, as just for a start, one can be chosen as easier. I have read about problems with ringfield motor, insulation,

 

and i have several of these. Most old locos, have round carbon brushes with holders each side of motor, eg dutchesses, 060. 462, and i have a lot ofthem, but take your point about age. Thanks again for your ideas, john

[Raven_Electric]

I call the power collector the two sprung spoons underneath that run along the third rail. If I remember it usually sat on a bit of brown circuit board but I'm blowed if I can remember if it was screwed or rivited on. Screws would be best for checking

 

the wiring etc.

 

No I don't have an eLink. I've got some expensive German and American stuff but I prefer the the Elite which of course can use Railmaster though I usually use either JMRI or software I wrote myself. The latter usually for reading/writing

 

CVs as it flows straight into/from my spreadsheet entries where I keep/tweak my various decoder CV settings and performance curves etc and I write/read ALL bits of CVs memory by way of error checking first (some very expensive decoders have had dodgy static

 

ram). My prog is compiled so runs much faster than JMRI which is written in interpreted Java but I still don't sit around waiting while it runs!

 

I have tried Railmaster for reading writing CVs without problem although of course it can't be automated

 

(as far as I've looked) so there will be no difficulty there when you get to that stage.

 

If yu've found an electronic store in France don't forget some 2mm and a bit bigger (works as cable ties too) heatshrink. At least a couple of metres of both. I

 

am sure Premier Farnell will deliver first thing next day anywhere in France and lots of the eBay traders will too.

 

Your other issue later will be beefing up continuity on your rails and points so get a little solder at different temperatures and a

 

pot of flux and some wire wool (Baker's Fluid was the stuff but Health and Safety saw that off) and play around a bit. If you've got loads of track sitting around soldering tails onto a few lengths would be handy. It is up to you how you deal with points.

 

DCC wants the whole thing live but you may wish to be able to continue to drop power to unused roads. You could solder on long enough wire to chop the wires later and rejoin or safely terminate in chocolate block I suppose. (Better get some 3mm and 4mm chocolate

 

block too)

 

If you are not up to much soldering now hold on doing the chip harnesses because you will be solderer of the year after a few rails.

 

 

[Yelrow]

now understand power collector, but chocollate block? will look at other things, pity you dont know where we are, as nearest shop is 35 miles, in any direction, but do have ebat, france, and germany. I have bakers fluid, kinda old but it works. Will continue

 

to accumilate. john

[Fishmanoz]

For chocolate blocks, take a look at http://www.maplin.co.uk/p/3a-terminal-blocks-3-pack-l96ar

 

I agree with Raven that you may end up with reliability issues on your older locos. But all of their motors are repairable including rewinding armatures

 

and remagnetising old magnets. Various places do such service including Norfolk Coast Models.

 

And I repeat that old motors in original condition are most unlikely to use less than the 500mA that an 8249 can deliver. Fully serviced with reconditioned

 

magnet and they may, but you need to do a stall current test to find out. Take a look at http://www.brian-lambert.co.uk/DCC.html#NORMAL to find out how this is done.

[Raven_Electric]

Hi Fishmanoz

I recall yelrow saying he had an R8249 which is why I suggested it as a handy, good value, reliable starting place. I had a pleasant day on the North Yorks Railway a while back being pulled by Sir Nigel at 25mph max. I'm not suggesting

 

that such a scale speed should be the limit of yelrow's ambitions but he is after all running a Heritage Railway and in my view some respect needs to be shown to these venerable old beasts! (I include myself on both counts).

 

I never had a Hornby Dublo

 

three rail layout so I don't really know why I have bits and bobs of stuff (which I am sure I do) but I do remember a couple of pal's stock flying round at about a scale 250mph with a few carriages barely hanging on the track behind them. After hours of this

 

they weren't glowing red and (unlike my own) didn't smell like spontaneous combustion was imminent so the motor/mechanism cannot be that inefficient. I would be very happy with a scale acceleration to a max of 70 mph for my mainline "Heritage" stock and in

 

OO I would consider that something was wrong if I could not get that from a digitally poked 10watts and R8249s deliver that for me without complaint. However, I defer to your greater knowledge and if yelrow needs more speed then the R8249 can be swapped out

 

for something more beefy before he buys more as he is wisely fitting sockets.

 

I have a little experience of very much larger and a great deal more than considerably more expensive motors than are found in model railways of any scale. Anything can be

 

re-wound or mended, even improved with better and closer tolerance bearings etc. The point is at what cost and would it be worthwhile? I don't consider putting a relatively high dc current through an artificially seized motor to be any kind of "Test" whatsoever

 

more of a high risk gamble with a possibly perfectly satisfactory winding to find out something I don't need to know. How fast yelrow's locos could go or what their maximum trainload may be is not important. What matters is that they go fast enough and pull

 

enough to deliver his timetable with a little bit more reliability than Sir Nigel's finest ever managed. If they do then Job Done as Sir Nigel never got to say.

 

Just one of the joys of DCC is that locos can be profiled not only to help protect the motor

 

but also to enable very accurate calculations of just what can be done and at what speed with a fair stab at quantifying the associated risks.

 

[Fishmanoz]

Hi Raven, yes, I agree that this is good pragmatic advice and may work out, then again maybe not. Now I said I have no experience with 3-rail, but I do have some with older loco experience, and I have done a deal of reading.

 

The first thing my

 

reading tells me, time and again including on these forums, is that older locos typically draw more than the 500mA that the 8249 can provide, and more so if the magnet is weak with age. My reading also tells me that a stall current test is a good idea, with

 

the Brian Lambert link I gave on how to do it being confirmation.

 

So with that background, I purchased a second hand Ivatt from the 70s with an X04 motor and proceeded to convert it to DCC using a decoder capable of 1 Amp continuous. One thing I failed

 

to notice with initial DC running before fitting the decoder was that the wheel quartering was out so was inclined to bind from time to time. Fast forward to DCC test, wheel quartering still out. Loco is proceeding slowly on sensible throttle but comes to

 

a halt on a point. Being on a point, I am thinking loss of contact to the rails but, as I soon found out, it was wheel binding and I was already giving the decoder a low throttle stall current test, which it was passing. However, what did I do next? I advanced

 

the throttle a little to see if that made a difference (loco was of course out of reach on other side of layout, as Mr Murphy will tell you will always be the case when this happens). And yes, it made a difference - connection to the loco was fine and 2 puffs

 

of smoke in quick succession and the bridge rectifier at the front end of the decoder was very much non-operational.

 

Conclusion - stall current on an X04 exceeds 1 Amp! And you wouldn't find me using an 8249 to check if an even older 3-rail loco's

 

motor draws much less. All very well to intend sensible scale speed operation with speed curve, back EMF adjustment etc to match, but you will be operating on the brink of disaster. And Murphy will find a way to prove that for you, sooner rather then later,

 

just as he did for me.

[Raven_Electric]

I am sorry, timed digital control of power supplies to ANY apparatus has absolutely nothing to do with a current being "drawn" by a potential difference across a circuit. Equally, because of designed in redundancy a locked stationary motor would be capable

 

of considerably more torque and thus current "draw" than would be needed to commence rotation in even peak load conditions so, without the design spec, or output performance testing to create it, the information obtained when stopped is of indeterminate accuracy

 

or value for anything. A more useful piece of information would be just how many milliseconds it would take to commence destroying windings insulation on those eighty years old motors by subjecting them to what is in effect an unlimited current system short

 

and this is NOT an either or situation, excessive heat for only a very short time can cause a progressive insulation breakdown to commence as air gets into cracks AND this, depending on insulation type, may continue after the excess heat is removed. I ask

 

myself, would I learn something of use to me about an 1100 volt traction motor by risking destroying it (and those around it) through bolting up the shaft and hooking it up to a power station? I think not. I can assure you, this is not how things are done

 

in all the best circles. Many great minds have, over many years, devoted a lot of ingenuity to devising some really interesting pieces of equipment (especially for coal mines) that exist only to prevent the very circumstances you seek to encourage! There is

 

a fine 190 year tradition that if the train doesn't go fast enough stuff something bigger in and that works equally well with boilers and decoders!

 

As far as operating on the brink of disaster goes I have been too close to too many big heavy things

 

going very fast for too many years to consider doing that. I threw the old envelopes out and built a computer for the calculations forty years ago for that very reason. I mentioned before the R8249 is in play because I recall yelrow saying he had one. There

 

are no profiling facilities but it is small so easier to install and more than compensates for its low spec with its size. low price AND very reliable current sensing (I have a few and I've tested them all a few times with 100% success). Your XO4 motor would

 

have been in safe hands running on the R8249 because the cut out is set at 1amp. I have suggested to yelrow that an R8245 be obtained for diagnostics. He may find that with profiles and other whizz bangs setup that, at least for his top link shed, he can spare

 

the extra cash to get hold of the extra fun! From my point of view and interest the R8245 is just dandy as Hornby's coup do grace has been to push the boat right out and support Lenz's RailCom so those diagnostics will flow straight off the tracks into my

 

spreadsheets - that is what I call testing..

 

Frankly other than simple guidance on which bit to solder to which bit and what to check for continuity and insulation I think yelrow's biggest problem on this project isn't technology it is sanity. There

 

must be a limit on just how much soldering a man stuck in darkest France can actually cope with before the men in white jackets arrive to cart him off. What is needed on this thread while the soldering continues is encouragement. Anecdotes on just how good

 

a loft full of die cast trains flying around really are as they decapitate all in sight when they come off on the bends! Memories of badly cut fingers from tin plate carriage ends and Dinky cars with no tyres welding themselves to the level crossings. We've

 

given all that up for "finescale" and sixteen engines all crashing into each other like rats in a sack.

 

I made it to the loft this evening and have returned with one of those beautiful Hornby die cast grey controllers with the lovely long chunky chrome

 

throttle lever. THAT IS the business! What railways are all about. I think I will have the guts out of a Select and put my top shed local controller where it should always have been in a beautifully finished, chrome topped die cast box.

 

 

 

[Tradman]

Not quite 3 rail, but I have managed to convert several HD 2 rail and 3 rail loco's to DCC. The first one took around 3 hours ( a Duchess) , and I reckon now it takes me about 2 hours.

The main thing I found is to replace the motor magnet with a new

 

neo magnet, on the duchess/city loco's I find this reduces the running current ( 8 standard hornby coaches) from an average of 560mA to below 300mA. The next thing is to isolate the brushes from the chassis, one is already isolated, the other I drill out teh

 

existing tube ( after removing armature) and replace using 3/16 brass tube insulated with heat shink tubing superglued in. I make my own NEM connectors, and using polymorphic plastic make my own mounts. The connectors are fairly standard, and it remains just

 

to plug in the decoder after checking the wiring. I have used several different decoders but the basic hornby one works fine.

 

As well as duchess/city class I have converted several Co-Bo's , 4MT's and deltics. If you need any more info let me know.

 

Mike

 

Mike

[Raven_Electric]

Hi Tradman

 

I was delighted to read your post because I encouraged yelrow to go for it on his Hornby Dublo 3 rail DCC project. It will be a lot of work for him and to justify that it needs to be a pretty impressive success.

 

If you read back

 

Fishmanoz identified power demands from these old locos as a potential problem and he specifically drew attention to the tired out old magnets so your experience has confirmed his fears. A 40%+ drop in power demand to 300mA is startling, it just goes to show

 

how important attention to detail is in these projects.

 

Keep your eye on this thead because I sure when he sees your post you will be hearing from yelrow. And, I expect the cry will go out how does one get magnets out to darkest France!

 

 

[Fishmanoz]

Certainly at 300mA there is no problem using an 8249.

 

Tradman, I'm not familiar with these locos so what style of motor are we talking? Is it like an X03/04 or a Ringfield or something different?

 

Your description sounds like the Ringfield

 

version that has the left hand brush shorted to chassis. This can be fixed with a plastic screw which sounds simpler than your drilling out method.

[The son of Triangman]

An X.04 on a loco in good coniditon and with good wheel quartering will be find on the Saphhire even at stall, I have converted loads of them. Condition of motor and loco before conversion is the key. The quartering and giving the loco more beans pushed

 

the motor current consumption above it's normal stall current Fishmanoz, that's what blew the decoder.

[The son of Triangman]

Some Hornby dublo three rail locos are a nightmare to convert to DCC operation, the brushes have to be insulated for instance.

[Yelrow]

Tradman, hi, good to know you have converted 3 rail locos, Son of T, thats also good, if you can remember which ones are a nightmare, that would help. Alternately, i can list mine, and forum can advise which to avoid. It was never, and still is not my

 

intention to convert all 3 rail, as my wife would prefer me at home, rather than in an institution, for demented souls. I would like to do one initially, just to prove my state of mind, and my ability to solder. Would like to know about these magnets etc.

 

Some of mine have ringfield motors, and what i DONT want to do, is to damage/devalue any of engines, as they are my childrens inheritance. So how do i proceed, answers on a postcard, john

[The son of Triangman]

Cities, Castles and Deltics are about the worst John to convert. Magnets are best left in place, removing an H/D magnet will usually mean having to get it remagnetised.

[Yelrow]

leaving aside my dutchesses, castles, and diesels, i am left with 6, no 062 tanks, all with upright armetures, and brushes either side, 2 no 264 black locos, 80054, construction, same as others, an 8F , and a silver king, 10 in all, need to pick a guinea

 

pig. Any views, john

[Tradman]

Note that I run my loco's, that is they are not for show, not for sale etc, so basically I will do what I have to, to convert them, note I was binned from one of my local clubs when I mentioned that I had changing the magnets on them and note had them

 

remagnetised. I found that the duchess's, city's, 4MT , class 20's, deltic's and Co Bo's are relatively easy to do, I can provide photo's of conversions. I do not have any castles so cannot comment. The deltics and co bo's only, have ring field motor bogies,

 

the others have the standard HD vertical motor (not an XO3/4).

Note that all the parts I use are easily obtainable by post from various supplies or if you use it, ebay .

 

As an aside there was an article in one of the modelling magazines some time

 

back showing how to convert HD loco's to DCC. Forget which one though.

 

Mike

 

Mike

[frederick-1232245]

a most interesting thread. It does not seem to have been active since 2014 so I wondered if it had been continued somewhere else

[RDS]

@frederick

Moderator Note:

Could you please select the 'View my Community’ in your Dashboard on the Forum and choose a nickname in the ‘My profile' tab, rather than your First name that has been automatically allocated by the system. Hopefully a Forum change is coming soon to prevent this automatic allocation but a manual change is required at the moment and there will be more than one member with your name.