sr.gif (1653 bytes)

Electrical Connectors for Model Railway Layouts

 

In August 1999, the Southern e-mail Group engaged in a vigorous exchange of opinion on the subject of inter-baseboard electrical connections. This had been prompted by Chris Gardner whose question at the outset was whether anyone in the group could suggest something more robust than the DIN connectors he was using for the hand held controller on his Cowes layout.

Graham Pettit was the first to enter the fray with a recommendation:-

have a look at the disco section of the Maplin catalogue. I used to do radio roadshows and discos, and the BIG disco connectors for speakers and amps really are robust - you can drive a lorry over them, and they will still plug in - lots of wiring options. Or go down your local musical instrument store. Have a look at the RS components catalogue www.rswww.com.

Dave Harris's reply came with the experience of many years exhibiting with the Basingstoke Club's Alresford and Bognor Regis layouts in 2mm scale:-

The 9 and 25 way computer D connectors, are cheap and very robust. We used them on our Alresford layout, which has been to 60+ shows over the years with hardly a failure. When we started out we also used DINs for the walkabout.. They did not last very long, and we replaced them soon after with 9way D's.

Problems with Ds are:

  • need care in assembly. (not a problem if using the IDC crimp type or surplus)
  • not suitable for higher currents
  • each pin has some contact resistance
  • won't take fat wires. (and thin wire has higher resistance)
  • the resistance problem means a noticeable voltage drop over distance for traction current. No real problem for point motors/relays and is not a problem on the 8ft max wiring length on our Alresford. (and locos do get slower at the fiddle yard ends - no bad thing!) For Bognor Regis(20ft max wiring), we use Ds for all control wiring, but we us audio 'Phono' connectors (2 wire coax style with extended center pin) for traction current jumpers between boards. These connectors are also cheap and robust and allow a thicker wire to be used and have a low contact resistance.


These can all be sourced from many places e.g. (UK) Greenweld, Tools-r-us, Maplin, Electrovalue, RS. We procured some 2nd hand 25way computer serial/printer cables (warning: not all connect all 25pins or the same 25 ways!) and these saved a lot of time, all we had to do was to fix a socket on the panel and fix a plug on the board. They range up to 25ft long.

Dave then went on with a fascinating explanation of how his club had tackled the voltage drop problem in the wiring on their Alresford and Bognor Regis (2mm scale) layouts:-

partly to combat the voltage drop problem, we do not switch track section currents in the control panel, instead the panel operates a relay physically next to the section to switch the track current. Each board has its own D cable from the control panel for this switching plus control of points/signals. The power controllers are in the middle of the layout. We have a 'traction current bus' going the length of the layout and each board takes traction current from this, the Phonos linking this bus. Where multiple phonos are used, they are colour coded.

This design makes sense if you realise that we do not us rail joiners or use the rail itself (2mm finescale/Code 40 NS is not a great conductor) as a traction current bus. Each individual rail has a dropper wire connecting to a 'section bus' on the underside of the baseboard. This bus is a copper wire shadow for each rail, the droppers soldered to it. Each section bus is energised from a point motor frog switch or a section relay or direct from the traction current bus, as required. Where a section bus crosses a board joint, we use phonos to jump across, unless at an end of sidings and that goes via Ds.

The exception to this scheme is where uncoupling electro magnets are involved. Our preference is to have push buttons mounted on each board and keep these high current devices away from the rest of the electrics. The operators normally want to look down on the shunting to eyeball the uncoupling.

This scheme, I believe, has much to offer.

  • With the traction current bus in place, modifications to track and sections become easier.
  • each board can be worked on and tested electrically as a separate unit
  • there are normally spare wires out of the 25 to each board to operate gadgets.


Earlier I mentioned 9way Ds for Walkabouts. These are our own club design and the box does not get hot nor the lead carry traction current. Basically if you put a standard controller in a panel and disconnect the speed pot, you wire the speed pot into the handheld box at the end of the lead and a direction switch that operates a relay in the panel that reverse the current. The only other bit needed is an LED/resistor wired to the output of the controller.

Roger Scott's contribution was a little more controversial. "Why use connectors at all?" he argued:-

Why do you actually need a connector? Answer, to connect two halves of layout together electrically. How often do you connect and disconnect? At the exhibition you connect it and leave it connected for the duration.... At home you connect it how many times during the year to check all is well and repair any faults that manifest themselves on the big day. Plus modifications undertaken in the light of experience working said model. Plus those occasions you entertain yourself and others "playing trains". Do you have room to leave it assembled? How much do these connectors cost? How long to wire up?

Even those among us here refer to connectors as "almost never letting you down". The connectors that I have seen referred to are of the crimp or telephone push fit type. Multi strand 16 20 32 wires. Turnouts require a minimum of three wires, plus signals, feeds and not forgetting buildings platforms etc. With all of those feeds channeled through one connector and anything electrical that can go wrong usually will at some time or another. Murphy's law states that it will the one that is hardest to get to the furthest away or the one "non-standard" connection on the whole layout that will go wrong first. A hard lesson I learnt a long time ago.

I therefore follow a simple rule. Everything electrical must be removable. So a turnout motor (I use PECO exclusively...I have purchased some SEEP but am now reluctant to use them) The connections to the turnout motor are soldered. Feed connections to the layout are soldered to the track. LOTS OF THEM! They all connect via block connectors to individual feeds for turnouts lights and other accessories but to a "ring main" for main power feeds. I pare back a small 5mm - 10mm of bare wire and "tin" it that is I solder it so that when placed into the block connector it is rigid and thus can be inserted and re inserted several times and maintain good contact. True, on the exhibition layout setting up will be longer but faced with the prospect of a malfunctioning feed within a 32 pin "D" connector or a row of block connectors that need individual attention I know which I would prefer. Would COWES really take that much longer to set up without a multi pin connector?

If the din plug and socket for a walkabout controller is unsatisfactory (I have no experience yet of such a thing but plan at least two "front of house" connectors) then a Jack or Phono type plug and socket may well be better. Our friends across the pond may have different names for them but basically one sort connects speakers to hi-fi's and the other connects headphones. The sort that connect headphones have a little "snap" fit to prevent headphones being pulled out.

The "no connectors" suggestion did not win much support though. Eric Penn had this to say

I note Rogers screw terminations but perhaps he has not had to set up a large layout? My club OO circular (oval) layout has 7x25 connections between the control panel and the layout, and these are then fed round the layout by multiple connectors between baseboards. The total number of connections must be in the region of 800. If we average at 15 seconds per connection (includes moving to the next set) it will take 3-4 hours to join. We would also have the problem of marking them up, and it would only need one error to upset the whole system. No thanks we will stick to connectors.

Mike Watts, too, an experienced exhibitor, wanted to assemble and dismantle his layout with utmost speed:-

As for experience, I have had many years of exhibiting, both here in Canada and the UK and have adopted a routine that I like best and has served me well. With the extremes here, of cold in the winter, and high, humid heat in the summer (3 weeks of humid 80s-90s recently), robustness is important.

Many years ago I settled on the 9-pin and 25-pin D plug. In recent years the basic insert-the-pin-and-solder kit has become available for both male and female, and much cheaper, of course. This suits me, as I am an avid soldering fan. Over the years I have learned through tens of thousands of soldering joins to make a solid, reliable join. I would guess the failure rate is in the one in thousands range, maybe higher. That is acceptable.

Also, contrary to all the experts' recommendations, I prefer solid wire, usually 20 s.w.g.. All my leads to connectors are rigorously bound, so effectively a cable run becomes multi-stranded, with sufficient flexibility. I've seen to much fraying of multi-strand wire, usually through incorrect stripping of the insulated cover. The experts will no doubt castigate me for this attitude towards multistrand wire. All I can say is years of experience in exhibiting, with a 99.9% reliability-rating in my electrics more or less proves to me that this issue is a matter of personal preference, not right or wrong.

I agree that the number of times in a year that you make or break a connector is quite small, although in my pre-exhibition testing of the WVR:


http://www3.sympatico.ca/lsw.lbsc/WVRWEB01.HTM

I assemble and disassemble the layout boards many times, and so subject the connectors to stress. There is a tendency to bend pins, if you are ham-fisted. But why should you be? I bet most of you treat your locos as though they were gold-plated, so why not treat the other parts of the layout in a similar fashion. Bent pins are now mostly a thing of the past for me. I do accept that each pin join offers resistance to current flow, and over a large layout, voltage drop may be a concern, but it is not on my 16'x12' WVR, which is about the maximum size of layout for one person to handle. In practice with the low current intakes of most motors these days, I would discount voltage drop on anything less than a 20ft run, when using these connectors. By the way, the unit cost for each unassembled 25-way D connector here in Canada (that is either male or female, not both) is around C$4.50, say just under two pounds.

Nowadays my track is cut to scale 60' lengths (we're talking C&L components here) and pre-made as panels where possible. Each rail length has two wire droppers soldered, and connected to the next rail length where appropriate. Each running track has a form of bus-bar for +ve and -ve, running the length of each base board. I tend to be a little more economical with sidings, and use the rails as a bus bar. So far in the 5 years of the WVR track, there has been 99% reliability in terms of flow of track power.

There is a problem with this multi-redundancy approach, though. As each section of rail has no less than 2 power feeds, and some have 4 feeds, then one or two breakages may go undetected. Eventually there will be breakages, unless the scrap-date of the WVR comes first!

Most of my wiring is done by following prepared tables and charts, with standardisation in circuit numbers and colours where possible. As an example, circuits 1 and 2 might always be 'Up Main +ve and -ve', and so on. My idea has always been that you wire up a layout, plug it in, and it works. With WVR I made only one wiring mistake. I do admit with such a rigid approach, solving this error took me several hours, but the time saved everywhere else made it all worthwhile. I am not in love with wiring diagrams; I much prefer the table approach.

In a similar vein, I do not like the concept of multiple feeds from the central control to each board. To me it is messy and unsightly. I am quite happy not to see any wires once the layout is set up. And the concept of running wires to a terminal block to be screwed in and tightened for each show I find, personally, completely unappealing to me. I would never entertain such activity on permanent basis; it has to all be plug-in for me. One day I will go DCC with radio control, i.e. only two power wires (well, in theory, at least:-))

While in theory I subscribe to easily removable electrical components, such as point motors, to actually replace one at the height of an exhibition, is too traumatic. Far better to rigorously test each item beforehand. I use H&M exclusively, having gathered over 70 over the years. I live with the 'clunk'. To replace all the 50 or so on both my layouts with Tortoises, would require an investment of $1000+, and even they are not faultless. My highly-tweaked H&M's are up in the 99% reliability range, and many are hid under sound deadening covers on the layout.

Set up time at exhibitions for me is critical. Local exhibitions here have got into the habit of setting up on Saturday mornings for 11am openings. As you can imagine, the halls are chaotic in the 8.30-10.00am time period. So my set-up time standard is 1 hour from the time the baseboards are unloaded. This is for 10 boards, each with fold-up legs, for one person. Consistently I beat this, having done it in 35 minutes. This is one area of operation where I won't tolerate slippage and constantly I look for ways of improving it.

I don't know if my requirement for setup time is unusual or not. I find other layouts at shows are still struggling after 2 hours to get trains running, or even getting boards aligned, whereas my objective is at the 30 minute mark to be patching up bits of scenery that may have got damaged in transit.

What is the experience of others in the group regarding set-up time?


Though an strong preference for "D" type computer connectors was now emerging, Eric Penn reported success with another type:-

Regarding connectors for controllers, I am using with success (except when someone trips over a cable) the Audio locking connectors in the Maplin catalogue (March - Aug 99) page 386, available in 2 to 8 way and with a locking ring and cable grip.

He also questioned Dave Harris's opinion that D connectors were only suitable for low voltages:-

I note that in a previous message... it was noted that D type connectors were not suitable for "Heavy Current" however the Maplin catalogue rates then at 7.5 amps, although how you get a wire in the socket to carry this is a problem. However is it not difficult if assembled with care to put 16/0.2 wire in them, rated at 3 amps. This should be enough for most applications although 'common return' may need two or three in parallel. If used between baseboards the connecting cables should be removable (i.e. plugs both ends) and of course spare cables kept in case of problems.

With at least two broadcasting professionals in the group, Chris Gardner asked what type of connectors would be used for outside broadcast. Richard Salmon was the first to reply:-

I suppose it's for me) to comment on Outside Broadcast (OB) connectors. The simplest would be the Audio XLR, three or five pin. Quite robust, but not enough ways. So we put several Audio circuits down a "Multi". These cables are about 3/4 inch diameter, and end either in "tails" with six or ten individual XLRs, or a multi-way connector. This latter is a somewhat bulky and expensive (but very robust - we call it OB proof) connector. Will stand up to being run over by a 15-ton truck... OTT for model layouts!

If using the "Power Bus" principle, as someone else expounded, I'd certainly recommend the Audio XLR for the power bus. It's way over the top for points/signals and relays though, due to small number of connections and bulk.

The D-type connectors are also widely used in broadcasting, and appear to be very reliable, even if not as robust as would be needed on an OB site. The better ones have a reasonable current capacity, but possibly not enough for traction power. The solution here, I'd suggest, is to double or triple up the pins on the circuits where more power is to be drawn. This is what I intend to do on my own layout (one day...). One can get very much better D-types than the computer-type ribbon-cable ones; we use them for digital TV equipment interconnections, where screened cables are an issue. Advantage is that a 25-way D is quite compact, and even allowing two or three pins per wire for higher current capacity, gives a good 12 or 8-way.

For cheap high power capacity connections, the 2mm or 4mm "Banana Plug" has a lot going for it. I use these for my portable chopper controllers.

For connections within a base-board I'd be perfectly happy with chocolate block. However, for exhibition use, solder or crimp tags on the end of a cable, joined on threaded studs with a suitable lock-nut will provide superior reliability, which is why this method is used for the wiring on real railway coaches (another tough environment).

Finally some important advice (which I assume most will already know): If using chocolate block, or any other screw-onto-wire connection, NEVER apply solder to the end of a multi-strand wire. The solder, being slightly soft, will creep under the pressure of the screw, and the connection will
come loose.

Support for D connectors continued to pour in. Ian Spalding commented that the well-known exhibition layout Walford used "D" connectors, without trouble, at 45 exhibition appearances.

Michael Gale Snr. endorsed Mike Watts' approach:

 

I can vouch for Mike W's quick setup and for the convenience of D-connectors. I intend to use them as well in much the same way as Chris G intended to use them--that is with sockets on either side of a baseboard join and a cable that links the two. I intend to screw in a printed circuit board (on standoffs) that has a PCB-mounted 25-pin D-connector and other connections to the rest of the baseboard to bus wires, point motors, signals, etc.

Chris G was also seeking advice on how to mount D sockets, as he had had difficulty sawing and drilling timber accurately enough to accommodate a D socket. Eric Penn was not the only one to come up with the solution:-

Use a piece of 1" x 1" aluminum angle (or steel), cut a section out the length of the socket (if the angle is thick it need not be a hole) and bolt the socket (or plug) in the space with two 3mm or 6ba nuts and bolts, the angle is then screwed underneath the baseboard structure.

I always prewire the fixed unit to a solder tag strip for convenience. I agree with Ian regarding D types, especially if you fit them up with retaining screws for show use, they won't fall out if touched accidentally.

He did sound a negative note in observing the limited life span of kit built D connectors and addressed his comments chiefly to Mike Watts:


I note your D types being cheaper in bits, the difference in the UK is so small (about 25p a connector) is just would not be worth the hassle of making them up. BUT the mating cycles for the kit ones is listed at 100 cycles, so perhaps you had better plan your exhibition appearances not to coincide with number 100!!!!!!!!!!

To which Mike replied

Actually I prefer soldering to kits ones, as I find soldering wire to a single pin easier. The cost difference here is a bit more. Without doing a current-day check, I think it is in the order of $2-$3, per connector. I never seen the 'life expectancy' of the kit version listed here, but the WVR will never see 100 exhibitions, so I guess I am OK.

Dave Harris, again, reflecting on experience with Alresford:-

I can not recall a single D connector failing, except forgetting to plug them in during setup and initial construction failures. We have never used the holding screws and I cant remember a plug falling out. Now I have seen people use D connectors without the wire restraining casing and that might well be a problem.

Eric Penn offered some more detailed advice on using D connectors:-:

Anyone using D types without covers and wire grips is asking for trouble, I also insist that at the club all plugs are fitted with screws and they are used at exhibitions. Have you never had a section or a signal fail only to find that the D type has been knocked and one end come loose, screws would have prevented it.

I always use gold flashed versions with solder buckets for terminations. The buckets are tinned before the wire is inserted. The connector is held in a small vice and a small piece of cored solder put in each bucket (about 3-4mm) and melted to tin the buckets. The wires (16/.02) are stripped about 12mm and the end tinned after making sure all the strands are grouped together, don't twist them or the bunch will be too big for the bucket. The wires are then cropped off about 3mm from the end of the insulation. Then place the wire on the top of the plug bucket, rest the iron on the bucket and gently push the wires into the bucket as the solder melts. Takes lots of practice to get it neat and tidy.

The wires are number sleeved for identification after the cover is fitted.

Chris Osment contributed to the subject of mounting the sockets:-

1. Make the hole in your baseboard frame as usual. Mount the connector in a small rectangle of metal sheet, veroboard or something flat and suitable, then screw this over the hole. Much easier to make a more accurate hole in the thin flat plate than the wood.

2. I don't like 90-degree PC-mounted connectors - subject to breakage of connections for all those regular insertion forces at right-angles with no support behind the connector body.

3. Always terminate baseboard wiring at female sockets and connect together using male-to-male jumper cables. Never try to save time/money by having a female on one board and a male socket on the end of a lead dangling from the other board - sooner or later it will get caught on something when moving the boards about and get damaged at a critical moment.

There was, however, support for "dangling leads" from Mike Watts on the basis that there are fewer "bits" to remember to pack, and Dave Harris who recommended they be held in place with Velcro when not in use.

Martin Truscott had an approach which reduced the labour of soldering the wires to the plugs:

My personal solution is to use cheap computer printer cables, 25 pin. Cutting them in half means that the two ends can be joined by plug and socket while the connections go via contact strip, solder or screw type to taste.

Colin Duff, the second broadcasting professional, cast doubt on the total reliability of any connector, before agreeing with the majority on D connectors being the best:-

After 24 years in the (audio) broadcasting work I can confidently state that there is no such thing as a totally reliable audio connector. BBC Radio Outside Broadcasts (and TV as well I think) used to use a multi way connector called the "DEF" series - multi connectors made to defence specifications. These things were so robust that you could probably run over them with a tank and they still not would break or distort, but that still does not mean that they were 100% reliable making connections.

Phono plugs and mono or stereo jacks - especially the domestic quarter inch, or even worse the 3.5mm and 2.5mm versions, are not reliable. The good old PO jacks are better but are still not totally reliable. One of the first jobs I was ever given in a radio studio was to clean all the PO jack plugs with Duraglit.....

Nowadays as Richard wrote we use a mixture of XLRs, PO jacks, (the dreadful) bantam PO jacks, and the current favoured multi connector is a varicon. Increasingly D plugs and phone type connectors are being used.

As a postscript, Roger Scott mailed the group some time later admitting to the error of his ways! This time it was a BT professional who swayed his opinion:-

Many thanks for your discussion on the recent subject. As you may know I have just completed a "Train Set" shunting exercise using "Chocolate Blocks". A friend who is an ex-BT was aghast at the site of all those "Chocolate Blocks" that he is going to ensure that never again do I use such things. He has taken it upon himself to "educate" me in the ways that these things should be done.....I did tell him that he wasn't the first to say so! Anyway I obviously stand corrected. Many thanks for the wealth of information you all provided and hopefully I shall soon now be very well connected!

So there we have it. The last word? Who knows? But at least it now seemed as though the Group was pretty much unanimous in its support for D type computer connectors for electrical connections on its layouts.



Edited by Chris Gardner
Alton, Hants
September 1999

 

station.gif (1056 bytes)